Taalportaal - the digital language portal

Dutch
Frisian
Saterfrisian
Afrikaans

Show all

11.3.1.1.Wh-movement in simplex clauses (short wh-movement)

quickinfo

This section discusses wh-questions derived by short wh-movement, that is, cases in which a wh-phrase is moved into the initial position of its own clause; cases of long wh-movement, in which a wh-phrase is extracted from its own clause and moved into the initial position of some matrix clause, are postponed until Section 11.3.1.2. The discussion is organized as follows, subsection I starts by showing that wh-movement is near-obligatory in the sense that one wh-phrase must be moved into clause-initial position, subsection II briefly discusses a hypothesis that aims at deriving this obligatoriness of movement from stating that wh-movement creates an operator-variable chain in the sense of predicate calculus (although some languages may also use alternative means like scope markers; see, e.g., Cheng (1991/1997), Bayer 2006, and also Section 11.3.1.2, sub V). An example like (96a) can be translated more or less directly into the semantic formula in (96b): if we ignore the feature -animate for the moment, the wh-phrase wat in clause-initial position corresponds to the question operator ?x, while the trace of the wh-phrase corresponds to the variable x. For completeness' sake, note that in formal semantics the question operator is normally expressed by the lambda operator: λx read(Peter, x). We will use more informal representations such as (96b).

a.	Wati	leest	Peter ti?
	what	reads	Peter
	'What is Peter reading?'

?x (Peter is reading x)

Subsection III restricts the discussion to wh-phrases consisting of a single wh-word like wie'who', wat'what' and hoe'how'; the aim of this subsection is to show that there are no restrictions on wh-movement related to the category or the syntactic function of the moved element, subsection IV briefly shows that the acceptability of embedded wh-questions depends on semantic properties of the matrix verb, subsection V discusses movement of larger wh-phrases, that is, phrases containing non-interrogative material besides the wh-element such as wiens boek'whose book' in (97a). According to the hypothesis to be discussed in Subsection II that wh-movement creates an operator-variable chain, it should suffice to simply move the wh-element, as the question is only concerned with the identity of the owner/writer of the book, as in the logical formula ?x (Peter is reading x's book); however, example (97b) shows that it is impossible to move the possessive wh-pronoun only. The fact that wh-movement may (or must) move a larger phrase than is needed for semantic reasons has become known as pied piping. We will say that in examples such as (97a) the wh-element wiens obligatorily pied-pipes the non-interrogative part boek of the direct object; example (97b) shows that stranding of this part is excluded.

a.	[Wiens boek]i	leest ti	Peter?
	whose book	read	Peter
	'Whose book is Peter reading?'

b.	*	Wiensi	leest	Peter [ti	boek]?
		whose	reads	Peter	book

Subsection V will show that pied piping can be forced by the fact that in some cases syntax simply does not allow wh-extraction. In other cases, however, stranding is possible or even required. There is, for instance, a contrast between pre- and postpositional phrases: while prepositions are normally pied-piped under wh-movement, postpositions are normally stranded, subsection VI will discuss a number of cases of wh-extraction.

a.	Jan is in die boom	geklommen?	preposition
	Jan is in that tree	climbed
	'Jan has climbed into that tree.'

a'.	In welke boom	is Jan geklommen?	pied piping
	in which tree	is Jan climbed
	'Into which tree has Jan climbed?'

b.	Jan is	die boom	in	geklommen.	postposition
	Jan is	that tree	into	climbed
	'Jan has climbed into that tree.'

b'.	Welke boom	<*in>	is Jan <in>	geklommen?	stranding
	which tree	into	is Jan	climbed
	'Into which tree has Jan climbed?'

The strongest hypothesis concerning pied piping and stranding would be that the two phenomena are in complementary distribution. We will formalize this by assuming a general constraint "avoid pied piping", which prohibits pied piping in constructions that allow stranding, subsection VI will show, however, that there are a number of potential problems with this constraint: there are cases in which pied piping and stranding are both excluded as well as cases in which they are both possible. For this reason we will briefly discuss the status of the constraint "avoid pied piping" in Subsection VII.

I. Wh-movement is near-obligatory
II. A functional motivation for wh-movement?
III. Categorial status and syntactic function of the wh-phrase
IV. Wh-movement in embedded clauses
V. Pied piping
VI. Stranding
VII. A note on the avoidance of pied piping

[+] I. Wh-movement is near-obligatory

The examples in (99) show that wh-movement is sometimes optional in interrogative main clauses; the wh-phrase normally occurs in clause-initial position but may also occur in clause-internal position in colloquial speech, provided that it is assigned a high tone, which we indicated by italics; cf. Zwart (2011:22).

a.	Wat	ga	je	doen?	regular form
	what	go	you	do
	'What are you going to do?'

a'.	Je	gaat	wat	doen?	colloquial speech
	you	go	what	do
	'What are you going to do?'

b.	Wanneer	ga	je	naar Utrecht?	regular form
	when	go	you	to Utrecht
	'When will you go to Utrecht?'

b'.	Je	gaat	wanneer	naar Utrecht?	colloquial speech
	you	go	when	to Utrecht
	'When will you go to Utrecht?'

The prosodically marked questions in the primed examples are normally ignored in syntactic descriptions of Standard Dutch, which may be due to the fact that they do not occur in written texts and formal speech. Unfortunately, we will not have much to say about these wh-constructions either, for want of sufficient in-depth research, although it is worth mentioning that leaving the wh-phrase in situ is a typical root phenomenon; Subsection IV will show that it does not occur in embedded wh-questions. Note further that the linear strings in the primed examples in (99) are also acceptable if they are construed as echo-questions: this reading requires the wh-element to be assigned emphatic accent. Echo-questions can be used if the hearer has the impression that he did not properly understand the speaker or if he wants to express surprise, disbelief, anger, etc.: echo-question (100a) could be used if B knows that A normally does not bother helping with domestic tasks, and echo-question (100b) could be used to express indignation or anger if A had promised B to spend the day together. We will not discuss echo-questions in what follows here.

100

a.	A:	Ik ga de afwas doen. B:	Je	gaat	wat	doen?
	A:	I go the dishes do	you	go	what	do
	'A: Iʼm going to do the dishes. B: You are going to do what?'

b.	A.	Ik	ga	vandaag	naar Utrecht. B:	Je	gaat	wanneer	naar Utrecht?
	A.	I	go	today	to Utrecht	you	go	when	to Utrecht
	'A: Iʼm going to Utrecht today. B. You are going to Utrecht when?'

The discussion of the examples in (99) has shown that wh-movement is more or less obligatory: it is the normal means to form a wh-question, although occasionally in colloquial speech it is not found in main clauses with a specific intonation pattern. The proper interpretation of the notion near-obligatoriness of wh-movement needs some special attention, though, as it pertains to the interrogative clause as a whole and not to individual wh-phrases. That wh-questions normally require the clause-initial position to be filled by some wh-phrase was already pointed out above. The so-called multiple wh-questions in (101) show, however, that it is possible for a wh-phrase to stay in its base position, provided the clause-initial position is filled by some other wh-phrase; it is in fact impossible to move both wh-phrases into clause-initial position, which can be attributed to the restriction that the clause-initial position can be filled by at most one constituent in Dutch; we will return to multiple wh-questions in Section 11.3.1.4.

101

a.	Wie	heeft	wat	gezegd?
	who	has	what	said
	'Who said what?'

a'.

Wie wat heeft gezegd?

b.	Wat	heeft	hij	aan wie	gegeven?
	what	has	he	to who	given
	'What has he given to whom?'

b'.

Wat aan wie heeft hij gegeven?

This subsection has shown that wh-movement is near-obligatory in the sense that the initial position of a wh-clauses must be filled by some wh-phrase; it is, however, possible for wh-phrases to remain in their original position if certain conditions are met, e.g., if the clause-initial position is already filled by some other wh-phrase.

[+] II. A functional motivation for wh-movement?

The near-obligatory nature of wh-movement in wh-questions can be attributed to the fact that this movement is needed to create an operator-variable relation in the sense of predicate calculus; see, e.g., Chomsky (1991) and Dayal (2006: Section 1.1.1). The syntactic representations in the primeless examples in (102), for instance, can be translated more or less directly into the (slightly informal) semantic representations in the primed examples. The preposed wh-phrases wat'what' and welk verhaal'which story' correspond to the question operator ?x plus a restrictor on the variable x (here: thing/story), while the trace of the wh-phrase corresponds to the variable x.

102

a.	Wati	heeft	Peter ti	gelezen?
	what	has	Peter	read
	'What has Peter read?'

a'.

?x [x: thing] (Peter has read x)

b.	[Welk verhaal]i	heeft	Peter ti	gelezen?
	which story	has	Peter	read
	'Which story has Peter read?'

b'.

?x [x: story] (Peter has read x)

Attractive as this may seem, it cannot be the whole story because it is not possible to translate the more complex wh-constructions in the primeless examples in (103) directly into the semantic representations given in the primed examples, as only a subpart of the wh-moved phrase corresponds to the question operator plus restrictor: the possessive pronoun wiens'whose' translates into ?x [x: person].

103

a.	[Wiens boek]i	heeft	Peter ti	gelezen?
	whose book	has	Peter	read
	'Whose book has Peter read?'

a'.

?x [x: person] (Peter has read x's book)

b.	[Wiens vaders boek]i	heeft	Peter ti	gelezen?
	whose fatherʼs book	has	Peter	read
	'Whose fatherʼs book has Peter read?'

b'.

?x [x: person] (Peter has read x's father's book)

The phenomenon of pied piping thus makes it impossible to propose a one-to-one relationship between syntactic structure and semantic representation: pied piping makes it impossible to state in simple direct terms that wh-movement creates an operator-variable chain. This problem is normally solved by assuming some form of reconstruction of the non-interrogative part of the wh-phrase in its original position. That such a mechanism is needed is clear from examples such as (104); since the anaphor zichzelf must have a c-commanding antecedent, the sentence is interpreted as if at least the non-wh-part gerucht over zichzelf'rumor about himself' still occupies the original position of the wh-moved phrase indicated by the trace. We will return to pied piping in Subsection V and to reconstruction in Section 11.3.6.

104

[Welk gerucht over zichzelfi]j	heeft	Peteriti	ontkent?
which rumor about himself	has	Peter	denied
'Which rumor about himself has Peter denied?'

Another problem we need to mention here involves multiple wh-questions such as (105a). Again, the syntactic structure does not directly correspond with the desirable semantic representation in (105b): because there is only one wh-phrase in clause-initial position, we would expect only one operator-variable chain in the corresponding semantic representation, while we seem to need two operator-variable chains to capture the interpretation of (105a). Section 11.3.1.4 will solve this problem by showing that the semantic representation in (105b) is actually not a proper semantic representation of (105a); multiple wh-questions do not quantify over entities but over ordered pairs of entities <x,y>, as indicated in the semantic representation in (105b').

105

a.	Wie	heeft	wat	gelezen?
	who	has	what	read
	'Who has read what?'

?x ?y (x has read y)

b'.

<x,y> (x has read y)

Observe that we omitted the restrictors from our semantic representations in (105). For the sake of simplicity, we will follow this convention from now on whenever the restrictors are not immediately relevant for our discussion.
This subsection discussed the hypothesis that there is a direct link between the obligatory nature of wh-movement and the semantic interpretation of wh-questions, in the sense that wh-movement is instrumental in creating operator-variable chains. Although we have seen that there are a number of potential problems with this hypothesis, to which we will return in Sections 11.3.1.4 and 11.3.6, we will adopt this hypothesis as a leading idea in the following discussion.

[+] III. Categorial status and syntactic function of the wh-phrase

There seem to be few restrictions on the categorial status of moved wh-elements; the only requirement seems to be that an interrogative pro-form be available. We illustrate this here for clausal constituents. The examples in (106) start by showing that all nominal arguments can be questioned.

106

a.	Jan/Hij	heeft	Marie/haar	die baan	aangeboden.
	Jan/he	has	Marie/her	that job	prt.-offered
	'Jan/He has offered Marie/her that job.'

b.	Wie	heeft	Marie/haar	die baan	aangeboden?	subject
	who	has	Marie/her	that job	prt.-offered
	'Who has offered Marie/her that job?'

c.	Wat	heeft	Jan/hij	Marie/haar	aangeboden?	direct object
	what	has	Jan/he	Marie/her	prt.-offered
	'What has Jan/he offered [to] Marie/her?'

d.	Wie	heeft	?Jan/hij	die baan	aangeboden?	indirect object
	who	has	Jan/he	that job	prt.-offered
	'Who has Jan/he offered that job [to]?'

Note that the question mark on Jan in (106d) is not intended to suggest that there is a syntactic impediment on wh-movement of the indirect object if the subject is non-pronominal. The contrast between (106b) and (106d) suggests that there is indeed a tendency to interpret an animate wh-phrase in clause-initial position as the subject of the clause, but the fact that the use of a subject pronoun gives rise to a fully felicitous result in both examples shows that this tendency is not syntactic in nature. The examples in (107) show that we find the same tendency in the case of subjects and direct objects. The fact that we do not find a similar tendency in German or English suggests that Dutch clearly has a computational disadvantage compared to these languages, in which the intended reading is clear from morphological case marking and word order, respectively.

107

a.	Wie heeft	Jan/hem	gezien?
	who has	Jan/him	seen
	'Who has seen Jan/him?'

b.	Wie	heeft	?Jan/hij	gezien?
	who	has	Jan/he	seen
	'Who has Jan/he seen?'

PP-arguments like indirect and prepositional objects cannot be replaced by a simple interrogative pro-form. This does not mean that such arguments cannot be wh-moved, but that this is only possible if the wh-phrase pied-pipes the preposition, as shown in (108). Such examples will be discussed in Subsection V.

108

a.	<Aan>	wie	heeft	Jan die baan <*aan>	aangeboden?	indirect object
	to	who	has	Jan that job	prt.-offered
	'To whom has Jan offered that book?'

b.	<Naar>	wie	staat	Jan <*naar>	te kijken?	prepositional object
	to	who	stands	Jan	to wait
	'Who is Jan looking at?'

The examples in (109) show that complementives can easily be questioned: we illustrate this by means of three examples of complementives with a different categorial status.

109

a.	Wie	ben	jij	eigenlijk?	Een vriend van Jan.	nominal
	who	are	you	prt	a friend of Jan
	'Who are you? Iʼm a friend of Janʼs.'

b.	Hoe	is de nieuwe directeur?	Aardig.	adjectival
	how	is the new director	nice
	'How is the new director? Heʼs nice.'

c.	Waar	heb	je	de schaar	gelegd?	In de la.	adpositional
	where	have	you	the scissors	put	in the drawer
	'Where have you put the scissors? In the drawer.'

Example (110) shows that supplementives can be questioned as well. Note that hoe'how' can also be used as a wh-adverb so that the interpretation of the question Hoe vertrok hij?'How has he left' depends on the context.

110

a.	Hoe vertrok	hij?	Kwaad.	supplementive
	how left	he	angry
	'How did he leave? He was angry.'

b.	Hoe vertrok	hij?	Met de auto.	adverbial
	how left	he	with the car
	'How did he leave? By car.'

Finally, the examples in (111) show that adverbial phrases with various functions can also be questioned when a wh-proform is available. Typical simplex adverbial wh-phrases are: hoe'how', hoezo'why/in what way', waarom'why', wanneer'when', and waar'where'.

111

a.	Waar	slaap	ik	vanavond?	In Peters kamer.	place adverbial
	where	sleep	I	tonight	in Peterʼs room
	'Where will I sleep tonight? In Peterʼs room.'

b.	Wanneer	vertrekken	we?	Na de vergadering.	time adverbial
	when	leave	we	after the meeting
	'When shall we leave? After the meeting.'

c.	Hoe	heb	je	het	gelezen?	Oppervlakkig.	manner adverbial
	how	have	you	it	read	superficially
	'How have you read it? Superficially.'

The examples above have amply demonstrated that there are few syntactic restrictions on question formation: clausal constituents with virtually any syntactic function and of any categorial type can be wh-questioned. The main restriction is lexical in nature in that there must be a wh-word available that can be used to question the intended phrase. This accounts for the fact that non-gradable clausal adverbs such as misschien'maybe' cannot be questioned: cf. *zeer/hoe misschien'very/how maybe'.

[+] IV. Wh-movement in embedded clauses

The discussion in the previous subsections was confined to wh-movement in main clauses. The primeless examples in (112) show that wh-movement is also possible in embedded clauses, and the primed examples show that wh-movement is obligatory: the embedded clauses cannot be interpreted as wh-questions if the wh-phrase stays in situ. The number sign in (112a') indicates that the embedded clause is acceptable as a yes/no-question if wat is interpreted as an existentially quantified personal pronoun ("something"), but this is of course not relevant here.

112

a.	dat	Jan wil	weten	[wat	(of)	je	gaat	doen].
	that	Jan wants	know	what	comp	you	go	do
	'that Jan wants to know what youʼre going to do.'

a'.

datJan wil weten [of je wat gaat doen].

b.	dat	Jan wil	weten	[wanneer	(of)	je	naar Utrecht	gaat].
	that	Jan wants	know	when	comp	you	to Utrecht	go
	'that Jan wants to know when you go to Utrecht.'

b'.

dat Jan wil weten [of je wanneer naar Utrecht gaat].

The examples in (113) show, however, that embedded wh-questions have a limited distribution. The question as to whether they are acceptable depends on the matrix verb; while (112) has shown that weten'to know' can license a wh-question, the verb ontkennen'to deny' cannot.

113

a.	*	dat	Jan ontkent	[wat	(of)	je	gaat	doen].
		that	Jan denies	what	comp	you	go	do

b.	*	dat	Jan ontkent	[wanneer	(of)	je	naar Utrecht	gaat].
		that	Jan denies	when	comp	you	to Utrecht	go

A warning flag is in order here given that free relatives (relative clauses without a phonetically expressed antecedent) have the appearance of interrogative clauses and can therefore easily be confused with them. They can however be recognized by the fact that they may occur in argument positions, that is, in the subject/object position preceding the clause-final verbs, as shown in (114a). Caution is only needed when they are extraposed (which is possible with all relative clauses modifying an object) or when there is no verb in clause-final position: cf. Jan ontkent wat je zegt'Jan denies what youʼre saying'.

114

a.	dat	Jan	[wat	je	zegt]	heeft	ontkend.
	that	Jan	what	you	say	has	denied
	'that Jan has denied what youʼre saying.'

b.	dat	Jan heeft	ontkend	[wat je zegt].
	that	Jan has	denied	what you say
	'that Jan has denied what youʼre saying.'

For more discussion of the semantic selection restrictions on embedded clauses, we refer the reader to Section 4.1. The reader is also referred to Section 4.2 for a discussion that embedded infinitival wh-questions are mainly found in formal language; in colloquial speech they mainly occur in formulaic expressions such as Ik weet niet wat te doen/zeggen'I don't know what to do/say. Note in passing that such infinitival clauses are also frequently used as independent expressions (e.g. in instructions or as rhetorical questions): cf. Wat te doen in het geval van brand'What to do in case of fire'. More examples of this type can be found in Vos (1994:148).

[+] V. Pied piping

Subsection III dealt with wh-moved phrases consisting of a single word such as wie'who', wat'what', hoe'how' and waar'where'. This subsection will show that wh-movement may also affect larger phrases. This is illustrated in (115b-d) for nominal arguments with an interrogative demonstrative pronoun as determiner.

115

a.	Jan/Hij	heeft	Marie/haar	die baan	aangeboden.
	Jan/he	has	Marie/her	that job	prt.-offered
	'Jan/He has offered Marie/her that job.'

b.	Welke functionaris	heeft	Marie/haar	die baan	aangeboden?	subject
	which official	has	Marie/her	that job	prt.-offered
	'Which official offered Marie/her that job?'

c.	Welke baan	heeft	Jan/hij	Marie/haar	aangeboden?	direct object
	which job	has	Jan/he	Marie/her	prt.-offered
	'Which job has Jan/he offered [to] Marie/her?'

d.	Welke sollicitant	heeft	?Jan/hij	die baan	aangeboden?	indirect object
	which applicant	has	Jan/he	that job	prt.-offered
	'Which applicant has Jan/he offered that book?'

Wh-movement of larger phrases has become known as pied piping: the interrogative demonstrative welke'which' is said to pied-pipe the non-interrogative part of the noun phrase into clause-initial position. The reasons for using this notion will be made clear in Subsection A, subsections B to D continue with a detailed discussion of the restrictions on pied piping of, respectively, NPs, PPs and APs, subsection E concludes by showing that pied piping of (extended) verbal projections is not possible. We aim at keeping the discussion relatively brief, given that some of the issues are discussed more extensively elsewhere; more detailed discussion on the NP data in Subsection B and the AP data in subsection D can be found in N2.2.1, sub V and A3.1.2, sub IV, respectively.

[+] A. Pied piping as a repair strategy

The fact that wh-moved phrases consisting of a single word such as wat'what' in (116a) move into clause-initial position is expected on the hypothesis discussed in Subsection II that wh-movement derives an operator-variable chain in the sense of predicate calculus. This does not hold, however, for the fact that there are also cases of wh-movement in which wh-movement applies to phrases including non-interrogative material, like welke auto'which car', wiens auto'whose car', and wiens vaders auto'whose father's car' in (116b-d); the non-interrogative parts of the wh-phrases are in italics.

116

a.	Wat	is de snelste auto?
	what	is the fastest car
	'What is the fastest car?'

b.	Welke auto	is de snelste?
	which car	is the fastest
	'Which car is fastest?'

c.	Wiens auto	is de snelste?
	whose car	is fastest
	'Whose car is fastest?'

d.	Wiens vaders auto	is de snelste?
	whose fatherʼs car	is the fastest
	'Whose fatherʼs car is the fastest?'

The hypothesis that wh-movement derives an operator-variable chain requires movement of the interrogative pronouns only; movement of the non-interrogative material in these examples is therefore superfluous from a semantic point of view. Consequently, there must be some other reason for the fact that wh-movement of the interrogative demonstrative and possessive pronouns in (116b-d) pied-pipes the non-interrogative parts of these noun phrases. This reason is syntactic in nature: it is is simply impossible in Dutch to extract determiners from noun phrases. The examples in (117) show that while it is possible to wh-move a full direct object, it is impossible to extract an interrogative demonstrative pronoun from it.

117

a.	[Welk boek]i	heeft	Marie ti	gelezen?
	which book	has	Marie	read
	'Which book has Marie read?'

b.	*	Welki	heeft	Marie [ti	boek]	gelezen?
		which	has	Marie	book	read

The examples in (118) show essentially the same for possessive pronouns: while it is possible to wh-move a full direct object, it is impossible to extract (a subpart of) a possessive determiner from it. The (a)-examples provide cases with the formal, genitive form wiens, while the (b)-examples provide cases with the more colloquial sequence wie z'n; in both cases pied piping is obligatory.

118

a.	[Wiens boek]i	heeft	Marie ti	gelezen?
	whose book	has	Marie	read
	'Whose book has Marie read?'

a'.	*	Wiensi	heeft	Marie [ti	boek]	gelezen?
		whose	has	Marie	book	read

b.	[Wie	zʼn	boek]i	heeft	Marie ti	gelezen?
	who	his	book	has	Marie	read
	'Whose book has Marie read?'

b'.	*	Wiei	heeft	Marie [ti	zʼn	boek]	gelezen?
		who	has	Marie	his	book	read

The examples in (119) are added in order to show that the wh-element need not be the determiner of the wh-moved noun phrase itself but can also be more deeply embedded: the wh-element wiens is the determiner of the noun phrase wiens vader, which in turn is the determiner of the wh-moved noun phrase wiens vaders boek.

119

a.	[Wiens vaders boek]i	heeft	Marie ti	gelezen?
	whose fatherʼs book	has	Marie	read
	'Whose fatherʼs book has Marie read?'

b.	*	[Wiens vaders]i	heeft	Marie [ti	boek]	gelezen?
		whose fatherʼs	has	Marie	book	read

c.	*	Wiensi	heeft	Marie [ti	vaders boek]	gelezen?
		whose	has	Marie	fatherʼs book	read

We conclude from the discussion above that pied piping is a repair strategy that is put to use if wh-movement of the wh-element itself is blocked for syntactic reasons. Since we will confine ourselves in the next subsections to providing an empirical description of the contexts that disfavor wh-extraction and thus favor pied piping, we refer the reader to Corver (1990:ch.7-9) for a more theoretical discussion of the syntactic restrictions on wh-extraction (as well as a cross-linguistic examination of the relevant data).

[+] B. Noun phrases

Example (120a) shows again that pied piping of noun phrases can be triggered by interrogative demonstrative and possessive determiners like welke'which' and wiens'whose'; Subsection A has already shown that this may be related to the fact that it is not possible to extract determiners from noun phrases.

120

a.	Welk	<boek>	heeft	Marie <*boek>	gelezen?	demonstrative pronoun
	which	book	has	Marie	read
	'Which book has Marie read?'

b.	Wiens	<boek>	heeft	Marie <*boek>	geleend?	possessive pronoun
	whose	book	has	Marie	borrowed
	'Whose book has Marie borrowed?'

Interrogative determiners are not only able to pied-pipe head nouns but also various other NP-internal constituents. This is especially conspicuous in the case of postnominal modifiers: while the primeless examples in (121) show that such modifiers can occur in extraposed position, the primed examples show that they must be pied-piped under wh-movement; cf. Guéron (1980). Note that the questions in the primed examples are special in that the use of the modifiers presupposes that the speaker has information enabling him to narrow down the set of potential answers: since we may assume that the addressee has greater knowledge of the situation than the speaker and consequently also has this information, explicit mention of the modifier may feel slightly forced. A more extensive discussion of the extraposition and pied-piping behavior of relative clauses can be found in Section N3.3.2.3.

121

a.	Jan heeft	[een boek	<met plaatjes>]	gekocht <met plaatjes>.
	Jan has	a book	with pictures	bought
	'Jan has bought a book with pictures.'

a'.	[Welk boek	<met plaatjes>]	heeft	Jan gekocht <*?met plaatjes>?
	which book	with pictures	has	Jan bought
	'Which book with pictures has Jan bought?'

b.	Jan heeft	[het boek	<dat hij gekocht had>]	gelezen <dat hij gekocht had>.
	Jan has	the book	that he bought had	read
	'Jan has read the book that he had bought.'

b'.	[Welk boek	<dat hij gekocht had>]	heeft	Jan gelezen <*dat hij gekocht had>?
	which book	that he bought has	has	Jan read
	'Which book that he has bought has Jan read?'

Subsection A has already shown that the wh-element does not have to be an immediate constituent of the pied-piped noun phrase, but can also be more deeply embedded: example (122a) illustrates this again for a possessive pronoun embedded in the determiner of a pied-piped noun phrase. Note in passing that constructions such as (122b) are sometimes judged as less felicitous because the noun phrases tend to become difficult to compute; this also holds for non-interrogative noun phrases like [[Peters moeders] auto]'Peterʼs motherʼs car'.

122

a.	We	mogen	[[haar moeders]	auto]	gebruiken.
	we	are.allowed	her motherʼs	car	use
	'We may use her motherʼs car.'

b.	[[Wiens moeders]	auto]	mogen	we gebruiken?
	whose motherʼs	car	are.allowed	we use
	'Whose motherʼs car can we use?'

It also seems possible to embed the wh-element hoe'how' in a quantifier phrase such as veel'much/many', although this fact may be obscured by the orthographic convention to write the formation hoe + veel as a single word. The fact that hoeveel in (123b) corresponds to heel veel'very many' in (123a) strongly suggests, however, that we are actually dealing with two separate words.

123

a.	Marie heeft	[[heel	veel]	boeken]	gelezen.
	Marie has	very	many	books	read
	'Marie has read a great many books.'

b.	Hoeveel	<boeken>	heeft	Marie <*boeken>	gelezen?	Heel veel!
	how.many	books	has	Marie	read	very many
	'How many books has Marie read? Very many!'

Example (123b) also shows that the interrogative quantifier hoeveel cannot be extracted from its noun phrase, and thus provides support for the hypothesis put forward in Subsection A that pied piping can be regarded as a repair strategy. Note that extraction of hoe is also excluded: *Hoei heeft Marie [ti veel boeken] gelezen?
Speakers occasionally seem to have varying judgments on pied piping triggered by a more deeply embedded wh-element. This can be illustrated quite nicely by examples such as (124b), in which the wh-element hoe'how' corresponds to the degree adverb erg'very' in (124a). In the earlier volume on adjectives (Section A5.2, sub I) in this series Broekhuis quoted similar examples as fully acceptable, while Corver (2003:292) has claimed that such examples allow an interpretation as echo-question only. Yet another verdict is levelled by Vos (1994:130), who assigns examples like (124b) a question mark. Clearly, it is difficult to decide whether the markedness of (124b) results from some syntactic constraint or from the computational complexity of the structure, which speakers can easily avoid by using the more or less synonymous but computationally simpler question Hoe groot is Els haar auto'How big is Elsʼ car?'.

124

a.	Els heeft	[een	[erg grote]	auto].
	Els has	a	very big	car
	'Els has a very big car.'

b.	%	[Een	[hoe grote]	auto]	heeft	Els?
		a	how big	car	has	Els
		'How big a car does Els have?'

For completeness' sake, note that it is also possible to find examples such as (125) on the internet. Although Vos (1994) claims that such examples are acceptable in colloquial speech, we doubt that the construction should be considered as part of standard language as many speakers simply reject it; see Corver (2003) for a more extensive discussion of this construction.

125

%	[Hoe	een	grote auto]	heeft	Els?	cf. zo'n grote auto 'such a big car'
	how	a	big car	has	Els
	'How big a car does Els have?'

The examples discussed so far show that wh-elements situated to the left of a nominal head may pied-pipe the non-interrogative part of the noun phrase. Pied piping seems to be more difficult, however, if the wh-element is situated to the right of the nominal head. This contrast becomes immediately clear when we compare the constructions with a prenominal genitive possessor in the (a)-examples of (126) to those with a postnominal van-phrase in the (b)-examples; cf. Vos (1994:130).

126

a.	Marie heeft	[Petersposs auto]	geleend.
	Marie has	Peterʼs car	borrowed
	'Marie has borrowed Peterʼs car.'

a'.	[Wiensposs auto]	heeft Marie	geleend?
	whose car	has Marie	borrowed
	'Whose car has Marie borrowed?'

b.	Marie heeft	[de auto [van Peterposs]]	geleend.
	Marie has	the car of Peter	borrowed
	'Marie has borrowed Peterʼs car.'

b'.	*	[De auto	[van wieposs]]	heeft	Marie	geleend?
		the car	of who	has	Marie	borrowed

At first sight, example (127a) seems to show that the contrast between the two primed examples in (126) can be derived from the hypothesis that pied piping is a repair strategy: the acceptability of example (127a) suggests that the postnominal van-PP can be extracted from the noun phrase. Closer scrutiny reveals, however, that the van-PP need not be construed as the possessor of the direct object. First, (127b) shows that the interrogative van-PP can also be used if the possessor is expressed by a possessive pronoun, which makes it very unlikely that the van-PP also functions as a possessor: cf. *[zijn auto van Peter]'lit.: his car of Peter'. Second, (127c) shows that the direct object can be pronominalized without affecting the van-PP, whereas pronominalization normally affects the full noun phrase. The examples in (127b&c) therefore show that the van-PP must have some other syntactic function; it probably is an indirect object with the thematic role of source (cf. Section 3.3.1.3).

127

a.		Van wie	heeft	Marie de auto	geleend?
		of who	has	Marie the car	borrowed

b.	Van wie	heeft	Marie zijnposs	auto	geleend?
	of who	has	Marie his	car	borrowed
	'From whom did Marie borrow his car?'

c.	Van wie	heeft	Marie	hem	geleend?
	of who	has	Marie	him	borrowed
	'From who did Marie borrow it?.'

The discussion above shows that we should take care not to jump to the conclusion that wh-extraction of PPs from noun phrases is possible, but that one should investigate each case in its own right. For example, it is remarkable that most examples that have been analyzed in the literature as cases of wh-extraction of PPs from noun phrases involve PPs headed by van or voor. However, Section N2.2.1, sub VC, has shown that van- and voor-PPs can also be used as restrictive adverbial phrases. It is therefore imperative to investigate cases with other prepositions in order to establish conclusively that wh-extraction of postnominal PPs is possible, but the primed examples in (128) show that such cases normally do not allow wh-extraction; cf. Vos (1994:139-40) and Broekhuis (2014).

128

a.	Els zal	morgen	[haar klacht	[tegen Peter]]	intrekken.
	Els will	tomorrow	her complaint	against Peter	withdraw
	'Els will withdraw her complaint against Peter tomorrow.'

a'.	*	[Tegen wie]i	zal	Els morgen	[haar klacht ti]	intrekken?
		against who	will	Els tomorrow	her complaint	withdraw

b.	Het leger	heeft	[een aanval	[op de president]]	verijdeld.
	the army	has	an attack	on the president	prevented
	'The army has prevented an attack on the president.'

b'.	??	[Op wie]	heeft	het leger	[een aanval ti ]	verijdeld?
		on who	has	the army	an attack	prevented

Let us now return to the hypothesis that pied piping is a repair strategy. Given that stranding is excluded or aleast quite marginal in the primed examples in (128), this hypothesis predicts that the pied piping examples in (129) are acceptable, but this is not borne out: these examples are impossible as wh-questions.

129

a.	*	[Haar klacht	[tegen wie]]i	zal	Els morgen ti	intrekken?
		her complaint	against who	will	Els tomorrow	withdraw

b.	*	[Een aanval	[op wie]]i	heeft	het leger ti	verijdeld?
		an attack	on who	has	the army	prevented

From this, it follows that the hypothesis that pied piping is a repair strategy should not be interpreted in such a way that pied piping of the remainder of the noun phrase can be used to form the desired question whenever wh-extraction is excluded, that is, it may be the case that certain semantically plausible questions simply cannot be formulated for reasons yet to be determined; see De Vries (2002:section 8.5) for a specific proposal.

[+] C. PPs

The examples in (130) show that prepositional phrases with an interrogative pronominal complement require pied piping; this is illustrated for such PPs in various syntactic functions. Since stranding of the preposition would invariably lead to unacceptability, these examples are in full accordance with the hypothesis put forth in Subsection A that pied piping can be regarded as a repair strategy. We will ignore the stranding data in this subsection, but return to them in Subsection VI, where the stranding behavior of post and circumpositional phrase will be discussed. Note in passing that all examples in (130) involve the +animate pronoun wie'who'; we will see in Subsection VI that the -animate pronoun wat'what' is not possible in this context, but we will also ignore this for the moment.

130

a.	Op wie	wacht	je?	prepositional object
	for whom	wait	you
	'Who are you waiting for?'

b.	Aan wie	heb	je	dat boek	gegeven?	indirect object
	to whom	have	you	that book	given
	'To whom have you given that book?'

c.	Naast wie	zullen	we Peter	zetten?	complementive
	next.to whom	will	we Peter	put
	'Next to whom shall we place Peter?'

d.	Na wie	word	jij	geholpen?	adverbial
	after who	are	you	helped
	'After who will you be served?'

The wh-element triggering pied piping need not be the complement of the pied-piped PP but can also be more deeply embedded. The examples in (131) illustrate this for a prepositional object, a complementive and an adverbial phrase, in which the wh-element functions as the determiner of a nominal complement of the pied-piped PP.

131

a.	[Op	[welk/wiens boek]]	zitten	we	nog	te wachten?	prepositional object
	for	which/whose book	sit	we	still	to wait
	'Which/Whose book are we still waiting for?'

b.	[Op	[welk/wiens bureau]]	heeft	Marie het dossier	gelegd?	complementive
	on	which/whose desk	has	Marie the file	put
	'On which/whose desk has Marie put the file?'

c.	[In	[welke/wiens kamer]]	zullen	we vergaderen?	adverbial
	in	which/whose room	will	we meet
	'In which/whose room shall we have our meeting?'

The examples in (132) show that the degree of embedding can be even greater. Example (132a) shows that the degree modifier hoe'how' of a quantifier of a nominal complement of a PP will ultimately pied-pipe the complete PP. And (132b) shows that the degree modifier hoe'how' of an attributive modifier of a nominal complement of a PP will ultimately pied-pipe the complete PP; a Google search (5/14/14) on the string [met een hoe grote] has shown that such examples can easily be found on the internet both in main and in embedded clauses, despite the fact that they are more complex than the disputed example %[Een [hoe grote] auto] heeft Els?'How big a car does Els have' in example (124b) from Subsection B.

132

a.	[Met	[[hoeveel] mensen]]	gaan	jullie	naar Japan?
	with	how.many people	go	you	to Japan
	'With how many people are you going to Japan?'

b.	[Met	[een	[hoe grote]	groep]]	zijn	jullie	in Japan?
	wit	a	how big	group	are	you	in Japan
	'With how big a group are you in Japan.'

In the examples discussed so far the wh-element is located in prenominal position of the nominal complement of the PP. At first sight, it seems that the result is infelicitous if the wh-element is situated in postnominal position, as shown by the primed examples in (133). The percentage signs indicate that examples like these are often claimed to allow an echo-interpretation only (cf. Vos 1994:127), but that there are also speakers who allow them as regular wh-questions and attribute their markedness to computational complexity (cf. De Vries 2002:section 8.5).

133

a.	Marie wacht	[op	[de broer [van Els]]]
	Marie wait	for	the brother of Els
	'Marie is waiting for Elsʼ brother.'

a'.	%	[Op	[de broer [van wie]]]	wacht	Marie?
		for	the brother of who	waits	Marie

b.	Marie loopt	[achter [de broer [van Els]]].
	Marie walks	behind the brother of Els
	'Marie is walking behind Elsʼ brother.'

b'.	%	[Achter	[de broer [van wie]]]	loopt Marie?
		behind	the brother of who	walks Marie

Vos (1994) has also shown that pied piping is fully acceptable in wh-questions like (134a). She suggests that this is only possible if the postnominal PP functions as a modifier, but this would wrongly predict that examples such as (134b) would be unacceptable, given that the relational noun centrum clearly selects the PP van welke stad (see N1.2.3); to our ears this example has more or less the same status as (134a).

134

a.	[Op	[een taxi	[van welk bedrijf]]]	wacht	u?
	for	a cab	of which company	wait	you
	'For a cab of which company are you waiting?'

b.	[In	[het centrum [van welke stad]]]	zou	je	wel	willen	wonen?
	in	the center of which city	would	you	prt	want	live
	'In the center of which city would you like to live?'

An alternative explanation for the acceptability contrast between the wh-examples in (133) and (134) might be that the complex noun phrases in the primed examples in (133) alternate with the structures with a possessive pronoun ([op/achter [wiens broer]]), while such alternants are not available for the noun phrases in (134). The markedness of the primed examples in (133) can then be attributed to syntactic blocking, in the sense that the structures with a possessive pronoun are simply favored; this would be in line with De Vries' suggestion that the markedness of the primed examples in (133) is due to their computational complexity.
There may be other factors affecting the acceptability judgments. Vos (1994) claims that an example such as (135a) is unacceptable despite the fact that it seems to involve the same degree of syntactic complexity as the examples in (134). De Vries (2002) considers similar cases acceptable, but difficult to comprehend, which suggests that the infelicitousness of this example may again be non-syntactic in nature. What we would like to suggest here is that the cause of the awkwardness is located in the nature of the nouns involved: the nouns broer'brother' and vriend'friend' in (135) are both relational nouns and can therefore only receive a proper interpretation if the relational argument is known to the addressee. Out-of-context this condition is not fulfilled in (135a) for the noun vriend and it may be that this causes the degraded status of this example. That this line of thinking may be on the right track is supported by the fact that example (135a) does improve if we replace the relational noun vriend by a non-relational noun such as meisje'girl', as in (135b), which is as acceptable as example (134b).

135

a.	*?	[Met	[de broer [van welke vriend]]]	heb	je	gedanst?
		with	the brother of which friend	have	you	danced
		'With the brother of which friend have you danced?'

b.	[Met	[de broer [van welk meisje]]]	heb	je	gedanst?
	with	the brother of which girl	have	you	danced
	'With the brother of which girl have you danced?'

The discussion above suggests that wh-elements in postnominal position are able to trigger pied piping of PPs, unless they are part of a postnominal van-PP that has an alternative expression as a prenominal possessive pronoun. The discussion of the examples in (135) has further shown that in some cases there may be non-syntactic factors at play that obscure the proper syntactic generalization; since these factors have not been investigated in full yet, we have to leave them to future research.

[+] D. APs

Pied piping of APs is quite restricted and normally involves the interrogative degree adverb hoe'how'. This is illustrated in (136) for a complementive and an adverbially used AP. The fact that the adjectives cannot be stranded shows that cases like these are in accordance with the hypothesis put forward in Subsection A that pied piping is a repair strategy.

136

a.	Hoe	<oud>	ben	jij <*oud>?	complementive
	how	old	are	you
	'How old are you?'

b.	Hoe <zorgvuldig>	heb	je	dat <*zorgvuldig>	gelezen?	adjunct
	how carefully	have	you	that	read
	'How carefully have you read that?'

The degree adverb hoe can also be more deeply embedded as part of a gradable degree modifier such as druk'busy' in (137). The (b)-examples show that in such cases pied piping sometimes gives rise to a marked result; the preferred option seems to be wh-extraction of the complete adverbial modifier although preferences seem to differ from case to case and speaker to speaker, for which reason Corver (1990:ch.8) marked both (b)-examples as grammatical.

137

a.	Jan is	[[erg	druk]	bezig].
	Jan is	very	lively	busy
	'Jan is very busy.'

b.		[Hoe druk]i	is Jan [ti	bezig]?	stranding
		how lively	is Jan	busy

b'.	?	[[Hoe druk]	bezig]i	is Jan ti?	pied piping
		how lively	busy	is Jan

The contrast with respect to the stranding behavior of the simple degree modifier hoe and the complex modifier hoe A is illustrated again in the examples in (138): while the complex wh-phrase hoe goed in (138a) must be construed as a degree modifier of bereikbaar, the simplex wh-phrase hoe in (138b) cannot; it can only be construed as a manner adverbial.

138

a.	Hoe goed	is dat dorp	bereikbaar?
	how well	is that village	accessible
	'How (easily) accessible is that village?'

b.	Hoe	is dat dorp	bereikbaar?
	how	is that dorp	approachable
	'How can that village be reached?'

The preference for stranding (if real) only holds for cases such as (137), in which the AP is a complementive. In other functions such as supplementive, pied piping is the only option. This contrast is illustrated in (139).

139

a.	Hoe goed	<?verzekerd>	is uw huis <verzekerd>?	complementive
	how well	insured	is your house
	'How well is your home insured?'

b.	Hoe goed	<verzekerd>	ging Jan <*verzekerd>	op vakantie?	supplementive
	how well	insured	went Jan	on holiday
	'How well insured did Jan go on holiday?'

For completeness' sake, the examples in (140) show that adjectives cannot be pied piped by their complement: (140a) shows that wh-movement of the PP op wie must strand the adjective boos and (140b) shows that wh-movement of the noun phrase welke opera must strand the adjective zat.

140

a.	<Boos>	op wie	is Peter <boos>?
	angry	at who	is Peter
	'Who is Peter angry with?'

b.	Welke opera	<*?zat>	is Jan <zat>?
	which opera	fed.up	is Jan
	'Which opera is Jan fed up with?'

We refer the reader to Section A3.1.2, sub IV for more data and a more extensive discussion. Pied piping of APs by a wh-element to the right of the adjectival head does not seem to occur; Subsection VI will show that such wh-elements normally strand (part of) the AP.

[+] E. Verbal (extended) projections

Wh-movement does not pied-pipe verbal projections. Example (141), for instance, shows that wh-movement of a direct object cannot pied-pipe the VP it is part of, but must be extracted from it. Note in this connection that Section 11.3.3, sub VIC, will show that VP-topicalization is possible: [Een boek gelezen]i heeft Jan ti.

141

a.	Wati	heeft	Jan [VPti	gelezen]?
	what	has	Jan	read
	'What has Jan read?'

b.	*	[VP	Wat	gelezen]i	heeft	Jan ti?
	*	*[VP	what	read	has	Jan

Similarly, a wh-phrase that is part of an embedded object clause cannot pied-pipe the clause (despite the fact that topicalization of clauses is possible), but must be extracted from it.

142

a.	Wati	zegt	Jan	[dat	hij ti	gelezen	heeft]?	wh-extraction
	what	says	Jan	that	he	read	has
	'What does Jan say that he has read?'

b.	*	[dat	hij	wat	gelezen	heeft]i	zegt Jan ti?	pied piping
		that	he	what	read	has	says Jan

The patterns in (141) and (142) are of course expected on the basis of the hypothesis put forward in Subsection A that pied piping is a repair strategy. It should be noted, however, that pied piping of an embedded clause is also impossible if wh-extraction is blocked, for example, if the wh-element is part of an adverbial clause. The examples in (143b-c) show that pied piping and stranding both lead to an unacceptable result and example (143d) shows that refraining from wh-movement is not an option either. As a consequence, it is simply impossible to phrase the desired question. Note that the linear strings in (143c&d) are acceptable as declaratives if wat is interpreted as the existential quantifier "something", but this is of course not relevant to our present discussion.

143

a.	Jan vertrok	[nadat	hij	het boek	gekocht	had].
	Jan left	after	he	the book	bought	had
	'Jan left after he had bought the book.'

Wati vertrok Jan [nadat hij ti gekocht had]?

wh-extraction

[Nadat hij wat gekocht had] vertrok Jan?

pied piping

Jan vertrok [nadat hij wat gekocht had]?

no wh-movement

The impossibility of formulating certain questions is not as exceptional as it may seem at first sight, as this is generally the case if a wh-element occurs in a so-called syntactic island; we will return to this issue in Section 11.3.1.3.

[+] F. Conclusion

This subsection has shown that pied piping is possible if the wh-element is embedded in a noun phrase, an AP or a PP, but impossible if it is embedded in an (extended) verbal projection. For the cases discussed here it seems observationally adequate to say that pied piping is possible whenever stranding (wh-extraction) is excluded, subsection VI on stranding will discuss more cases that are also covered by this generalization, while Subsection VII will show that there are also cases that run afoul of it. The discussion in this subsection was somewhat complicated by the fact that the judgments in the literature are sometimes contradictory; we argued that this may be due to the interference of a number of non-syntactic factors, which should be further investigated in the future. More extensive data sets on pied piping are given in Corver (1990:ch.7-10), Vos (1994), and De Vries (2002:section 8.5).

[+] VI. Stranding

Subsection V discussed cases in which wh-movement pied-pipes a clausal constituent. There are, however, also cases of wh-movement that partially strand clausal constituents: these will be discussed in this subsection. For reasons of presentation we start with wh-extraction from PPs, after which we will discuss cases involving noun phrases and APs. Wh-extraction from clauses is not discussed here; some core data were already presented in subsection V and a more detailed treatment will be given in Section 11.3.1.2.

[+] A. Wh-extraction from PP

Subsection VC has shown that wh-movement of the nominal complement of a prepositional phrase normally pied-pipes the full PP. This subsection will show, however, that there are also cases in which wh-movement of the nominal complement strands the adposition; this holds for pronominal, postpositional and circumpositional PPs. Our discussion of these cases is followed by an attempt at an analysis. We conclude with a discussion of stranding by wh-movement of the modifier of a PP.

[+] 1. Complements of pronominal PPs (R-extraction)

The primeless examples in (144) show again that pied piping of prepositional phrases gives rise to an acceptable result regardless of the syntactic function of the PP, while the primed examples show that stranding of the preposition is impossible. In (144) we are dealing with a pronominal complement, wie'who'; we refer the reader to Subsection VC for examples that show that judgments do not change if the wh-element is embedded in the complement of the preposition, such as the demonstrative pronoun welke'which' or the possessive pronoun wiens'whose'.

144

a.	Naar wie	kijk	je?	prepositional object
	at who	look	you
	'Who are you looking at?'

a'.	*	Wiei	kijk	je	[naar ti ]?
		who	look	you	at

b.	Naast wie	zullen	we Peter	zetten?	complementive
	next.to whom	will	we Peter	put
	'Next to whom shall we put Peter?'

b'.	*	Wie	zullen	we Peter	[naast ti ]	zetten?
		who	will	we Peter	next.to	put

c.	Na wie	word	jij	geholpen?	adverbial
	after who	are	you	helped
	'After who will you be helped?'

c'.	*	Wie	word	jij	[na ti ]	geholpen?
		who	are	you	after	helped

The results change drastically if the interrogative complement is inanimate. The inanimate pronoun wat'what' normally cannot occur as the complement of a preposition but triggers R-pronominalization; it surfaces as the R-wordwaar, which precedes the preposition. The examples in (145) show that wh-movement of waar may strand the preposition (which we refer to as R-extraction) if the PP is a complement of the verb or a complementive, but not if it is an adverbial phrase of time or place.

145

a.	Waar	<?naar>	kijk	je <naar>?	prepositional object
	where	to	look	you
	'What are you looking at?'

b.	Waar	<?in>	zullen	we deze ring <in>	stoppen?	complementive
	where	into	will	we this ring	put
	'What will we put this ring into?'

c.	Waar	<?na>	moest	hij	nu <*na>	weg:	het 1e of het 2e bedrijf?	adverbial
	where	after	must	he	prt	away	the 1st or the 2nd act
	'After what did he have to leave: the first or the second act?'

The question marks in (145a&b) are used to express that for many speakers R-extraction is the preferred option in colloquial speech; pied piping is, however, fully acceptable in formal speech and written language. The question mark in example (145c) is used to indicate that the use of the pronominal PP waarna'after what' is slightly clumsy and that speakers would normally use the simplex adverb wanneer'when'; nevertheless, the acceptability contrast between pied piping and stranding is clear. The main conclusion that we can draw from the examples in (145) is that stranding is readily possible in at least certain syntactic configurations. Note that the syntactic restriction is not simply that the PP cannot be an adjunct: some adverbial phrases such as the instrumental PP in (146) do allow wh-extraction. We will not discuss the syntactic restrictions on R-extraction here, but refer the reader to the extensive discussion of this in Section P5.3.

146

a.	Jan heeft	de wijnfles	met een schroevendraaier	geopend.
	Jan has	the wine.bottle	with a screw.driver	opened
	'Jan has opened the wine bottle with a screw driver.'

b.	Waar	heeft	Jan	de wijnfles	mee	geopend?
	where	has	Jan	the wine.bottle	with	opened
	'What has Jan opened the wine bottle with?'

[+] 2. Complements of postpositional phrases

Wh-extraction is also possible with complements of postpositional phrases, which have a restricted syntactic use as clausal constituent: they occur as complementives only. The examples in (147b&c) show that the wh-element can be the complement of the PP itself or be embedded in the complement of the PP. The use of the dollar sign indicates that example (147b) does not feel fully natural as a wh-question, but that the markedness is not syntactic in nature; the reason for assuming the latter is that stranding of the postposition is fully acceptable in (147c).

147

a.	De angstige kat	is	[die boom	in]	gevlucht.
	the frightened cat	is	that tree	into	fled
	'The frightened cat has fled into that tree.'

b.	Wat	<*in>	is	de kat	<$in>	gevlucht?
	what	in	is	the cat	into	fled
	'What has the cat fled into?'

c.	Welke boom	<*in>	is	de kat <in>	gevlucht?
	which tree	into	is	the cat	fled
	'Which tree did the cat flee into?'

It is not immediately clear why stranding the postposition in (147b) gives rise to a marked result. It may be due to the fact that the postposition in a priori restricts the set of possible answers to entities with an interior, which suggests that the speaker has specific prior knowledge, which may favor an echo-reading of this example. The echo-reading does not arise in (147c) because the speaker's prior knowledge has been made explicit in the non-interrogative part of the noun phrase; the cat has fled into some tree and the speaker simply wants to know which one.

[+] 3. Complements of circumpositional phrases

Circumpositional phrases are like postpositional phrases in that they are not used as prepositional objects or adverbial phrases, but occur as complementives only. The examples in (148b&c) show, however, that they also behave like prepositional phrases in that the interrogative pronoun wie cannot be extracted by wh-movement and that wh-movement of the interrogative R-word waar strands the remainder of the circumpositional phrase. This is illustrated in (148b&c).

148

a.	Jan sprong	[over Peter/het paaltje	heen]?
	Jan jumped	over Peter/the pole	heen
	'Jan jumped over Peter/the pole.'

b.	*	Wiei	sprong	Jan	[over ti	heen]?
		who	jumped	Jan	over	heen
		'Who did Jan jump over?'

c.	Waari	sprong	Jan	[over ti	heen]?
	where	jumped	Jan	over	heen
	'What did Jan jump over?'

Circumpositional phrases cannot readily be pied-piped as a whole in colloquial speech: the next subsection will show that the unacceptability of (149a) may be related to the fact that circumpositional phrases allow wh-movement to pied-pipe the first member of the circumposition, as shown in (149b); cf. P1.2.5, sub III for detailed discussion.

149

a.	*?	[Over	wie	heen]i	sprong	Jan ti ?
		over	who	heen	jumped	Jan

b.	[Over wie]i	sprong	Jan [ti	heen]?
	over who	jumped	Jan	heen
	'Who did Jan jump over?'

It should be noted that the same reasoning cannot be extended to account for the markedness of (150a), as (150b) is also degraded. The contrast between (149b) and (150b) may however be related to the contrast between the two examples in (148b&c): example (150b) may be syntactically blocked by example (148c), in which even less material has been wh-moved; example (149b) is not syntactically blocked because example (148b) is not acceptable.

150

a.	*?	[Waar	over	heen]	sprong	Jan ti?
		where	over	heen	jumped	Jan

b.	??	[Waar	over]i	sprong	Jan [ti	heen]?
		where	over	jumped	Jan	heen

[+] 4. An attempt at analysis

It looks as if a relatively simple explanation can be formulated for the data found in (144)-(150), but it will require a number of brief digressions. First, the fact illustrated in (149b) that circumpositional phrases can be split suggests that the first and second member of the circumposition do not constitute a single lexical unit; Section P1.2.6 concluded from this that circumpositional phrases should actually be analyzed as complex structures in which the second member (here: heen) is a postpositional-like element selecting a PP-complement. For our limited descriptive purpose here we will assume the structures in (151), but we refer the reader to Section P1.2.6 for arguments showing that these structures may actually be more complex in the sense that post and circumpositional phrases involve PP-internal movement.

151

Prepositional phrase: [PP P NP]

Postpositional phrase: [PP NP P]

Circumposition phrase: [PP [PP P NP] P]

Second, Koster (1987: Section 4.5) argued on the basis of examples like (144), (145) and (147) that the choice between pied piping and stranding depends on two syntactic factors, which we give here in an informal form as the descriptive generalizations in (152); see also Van Riemsdijk (1978). Since prepositions precede their complement, clause (152a) accounts for the unacceptability of stranding in (144). The two clauses in (152) are both satisfied in the case of the pronominal PPs in (145a&b) and the postnominal PPs in (147b&c), which are therefore correctly predicted to allow stranding. Since adverbial clauses are not selected by the verb, clause (152b) is not satisfied in (145c), which is therefore correctly predicted to be unacceptable.

152

Wh-movement of a complement may strand the head of a PP if and only if:

the adpositional head follows its complement; this holds for postpositions and prepositions that head a pronominal PP;

the adpositional phrase is selected by the main verb, the head of a dependent of the main verb, the head of a dependent of a dependent of the main verb, etc.the adpositional phrase is selected by the main verb, the head of a dependent of the main verb, the head of a dependent of a dependent of the main verb, etc.

It should be noted, however, that there is a problem with instrumental PPs in examples such as (146) because clause (152b) wrongly predicts wh-extraction to be unacceptable in these examples (unless we assume that certain types of adverbial phrases are in some sense dependent on the verb); we leave this problem for future research.
Third, the fact that stranding and pied piping are (normally) in complementary distribution requires us to assume that one of the two is the preferred option. This can be formulated as the constraint in (153), which can be seen as a slightly more precise version of the hypothesis put forward in Subsection VA that pied piping should be regarded as a repair strategy. This fully accounts for the acceptability judgments on stranding and pied piping in the examples in (144), (145) and (147).

153

Avoid pied piping: strand as much material as possible.

The set of claims in (151)-(153) also provides an account for the acceptability judgments on the circumpositional cases in (148)-(150). Consider again the examples in (148b&c), repeated as (154) in a form consistent with the hypothesis in (151c), according to which PP2 is the complement of PP1.

154

a.	*	Wiei	sprong	Jan [PP1 [PP2	over ti ]	heen]?
		who	jumped	Jan	over	heen
		'Who did Jan jump over?'

b.	Waari	sprong	Jan [PP1 [PP2ti	over]	heen]?
	where	jumped	Jan	over	heen
	'What did Jan jump over?'

That wh-movement of the pronoun wie cannot strand the circumposition in (154a) follows immediately from clause (152a): the preposition over precedes the pronoun and can therefore not be stranded. Clause (152a) does not prohibit R-extraction, as the preposition over follows its complement in pronominal PPs. R-extraction is also allowed by clause (152b): PP1 is selected by the head of PP2, which in its turn is selected by the main verb.
Now, consider again the examples in (149), repeated here in a slightly different form as (155). Example (155b) is predicted to be acceptable because wh-movement of PP2 is in accordance with both clauses in (152): PP1 is selected by the main verb and the head of PP1, heen, follows its complement, PP2. Since we have already seen that the head of PP2, over, cannot be stranded, pied piping is allowed by the constraint "avoid pied piping" in (153). Example (155a), on the other hand, is blocked by this constraint, as (155b) pied-pipes less material.

155

a.	*?	[PP1 [PP2	Over wie]	heen]i	sprong	Jan ti ?
	*?	*?[PP1 [PP2	over who	heen	jumped	Jan

b.	[PP2	Over wie]i	sprong	Jan [PP1ti	heen]?
	[PP2	over who	jumped	Jan	heen
	'Who did Jan jump over?'

Now, consider again the examples in (150), repeated here in a slightly different form as (156a&b). The descriptive generalization in (152) allows the structure in (156b) for the same reason as it allows the structure in (155b). The unacceptability of this structure must therefore be due to the constraint "avoid pied piping" in (153). And this is indeed the case: example (154b), repeated here as (156c), is the preferred structure, as it involves less pied-piped material.

156

a.	*?	[PP1 [PP2	Waar over]	heen]]i	sprong	Jan ti ?
	*?	*?[PP1 [PP2	where over	heen	jumped	Jan

b.	*?	[PP2	Waar	over]i	sprong	Jan [PP1ti	heen]?
	*?	*?[PP2	where	over	jumped	Jan	heen

c.		Waari	sprong	Jan [PP1 [PP2ti	over]	heen]?
		where	jumped	Jan	over	heen

This account of the surprising acceptability contrast between (155b) and (156b) completes our description of the acceptability judgment on pied piping/stranding in examples like (144)-(150), in which the wh-element is the complement of an adpositional phrase. The next subsection continues with a discussion of PPs stranded by wh-movement of their modifier.

[+] 5. Modifier of PP

Modification of PPs is normally restricted to spatial en temporal PPs. The following discussion of the movement behavior of these modifiers under wh-movement will be relatively brief because a more extensive discussion can be found in Sections P3.1 and P3.2. Here we will show that stranding/pied piping is sensitive to the syntactic function of the PPs: while the heads of complementive PPs are normally stranded, the heads of adverbial PPs are pied-piped. Prepositional objects like op vader in Jan wacht op vader'Jan is waiting for father' are not relevant, as these do not allow modification.
Section P3.1.2 has shown that modifiers of spatial PPs are normally of two kinds: modifiers of orientation like recht'straight' in (157a) and modifiers of distance like the adjectival phase diep

157

a.	Jan staat [PP	recht	voor de camera].
	Jan stands	straight	in.front.of the camera
	'Jan is standing straight in front of the camera.'

b.	De olie	zit [PP	diep in de grond].
	the oil	sits	deep in the ground
	'The oil is deep in the ground.'

The two types of modifier exhibit different behavior when it comes to modification: modifiers of orientation are modified by approximative modifiers like zowat'approximately/more or less' and by precies'exactly', while adjectival modifiers of distance are modified by degree modifiers like erg/heel'very'.

158

a.	Jan staat [PP	zowat/precies	recht	voor de camera].
	Jan stands	approximately/exactly	straight	in.front.of the camera
	'Jan is standing more or less/straight in front of the camera.'

b.	De olie	zit [PP	erg/heel	diep in de grond].
	the oil	sits	very/very	deep in the ground
	'The oil is very deep in the ground.'

It seems that approximative modifiers such as recht'straight' do not have an interrogative counterpart. The string Hoe recht staat Jan voor de camera? is fully acceptable but does not have the intended interpretation: the phrase wh-phrase hoe recht does not pertain to the location of Jan with respect to the camera, but to his posture; cf. Section P3.1.2. This means that the structure in (159a) is unacceptable. Degree modifiers such as diep'deep', on the other hand, do have an interrogative counterpart; the (b)-examples are acceptable with the intended interpretation.

159

a.	*	Hoe rechti	staat	Jan [PPti	voor de camera]?
		how straight	stands	Jan	in.front.of the camera

b.	Hoe diepi	zit	de olie [PPti	in de grond]?
	how deep	sits	the oil	in the ground
	'How deep is the oil in the ground?'

b'.	[PP	Hoe diep in de grond]i	zit	de olie ti?
	[PP	how deep in the ground	sits	the oil
	'How deep is the oil in the ground?'

Given the option of stranding in (159b), the constraint "avoid pied piping" in (153) predicts example (159b') to be ungrammatical, but nevertheless most of our informants do accept examples of this type. It seems, however, that actual usage is more in line with "avoid pied piping". A Google search (6/26/2014) on the string [Hoe diep in de grond zit] resulted in no more than one relevant hit, whereas [Hoe diep zit * in de grond] resulted in 13 relevant hits; in our search we excluded examples containing the string [tot hoe] and checked the remaining results manually.
Example (160a) shows that spatial PPs can also be modified by nominal measure phrases such as 2 kilometer. Such nominal measure phrases can also be interrogative and again it seems that stranding and pied piping both give rise to acceptable results; cf. Corver (1990:ch.9). Since it is not readily possible by means of a simple Google search to investigate whether actual usage is more in line with the constraint "avoid pied piping", we leave this issue to future research.

160

a.	De olie	zit [PP	2 kilometer	onder de grond].
	the oil	sits	2 kilometer	under the ground
	'The oil is located 2 kilometers under the surface.'

b.	Hoeveel kilometeri	zit	de olie [PPti	onder de grond]?
	how many kilometers	sits	the oil	under the ground
	'How many kilometers is the oil under the surface?'

b'.	[PP	Hoeveel kilometer	onder de grond]i	zit	de olieti?
	[PP	how many kilometers	under the ground	sits	the oil
	'How many kilometers is the oil under the surface?'

In all examples above the spatial PPs function as complementives. If the spatial PP functions as an adverbial phrase, pied piping is obligatory. This is illustrated in the examples in (161) and (162) for adjectival degree modifiers and nominal measure phrases, respectively.

161

a.	De speleoloog	verongelukte [PP	diep onder de grond].
	the speleologist	was.killed	deep under the ground
	'The speleologist had a fatal accident deep underground.'

b.	*	Hoe diepi	verongelukte	de speleoloog [PPti	onder de grond]?
		how deep	was.killed	the speleologist	under the ground

b'.	[PP	Hoe diep	onder de grond]i	verongelukte	de speleoloog?
	[PP	how deep	under the ground	was.killed	the speleologist
	'How deep underground did the speleologist have a fatal accident?'

162

a.	De speleoloog	verongelukte [PP	80 meter onder de grond].
	the speleologist	was.killed	80 meter under the ground
	'The speleologist had a fatal accident 80 meters underground.'

b.	*	Hoeveel meteri	verongelukte	de speleoloog [PPti	onder de grond]?
		how many meter	was.killed	the speleologist	under the ground

b'.	[PP	Hoeveel meter	onder de grond]i	verongelukte	de speleoloog?
	[PP	how many meter	under the ground	was.killed	the speleologist
	'How many meters underground did the speleologist have a fatal accident?'

Temporal PPs are normally used as adverbial phrases and the (b)-examples in (163) show that in such cases wh-movement triggers pied piping. This finding was confirmed by our Google searches (7/2/2014) on the search strings [hoe lang na] and [hoe lang * na]: the first search string resulted in nearly 200 hits, most of which instantiated the relevant construction, whereas a cursory look at the first 100 results for the second search string showed that hoe lang and the na-PP must be construed as independent adverbial phrases when they are not adjacent.

163

a.	De speleoloog	overleed [PP	kort na het ongeval].
	the speleologist	died	shortly after the accident
	'The speleologist died shortly after the accident.'

b.	*	Hoe langi	overleed	de speleoloog [PPti	na het ongeval]?
		how long	died	the speleologist	after the accident

b'.	[PP	Hoe lang	na het ongeval]i	overleed	de speleoloog?
	[PP	how long	after the accident	died	the speleologist
	'How long after the accident did the speleologist die?'

[hide extra information]

The discussion above has shown that wh-movement of a modifier may strand a PP used as a complement but not as an adverbial, which is in line with our discussion in Section 11.3.1.3 that adverbial phrases are normally islands for extraction. We further found that the (b)-examples in (159) and (160) constitute potential problems for the constraint "avoid pied piping" in (153), although the results of a Google search suggests that actual usage may be more in line with this constraint. For more discussion of (wh-movement of) adjectival and nominal modifiers of PPs, we refer the reader to Chapter P3.

[+] B. Wh-extraction from noun phrases

This subsection can be brief because there is little to add to what has been said in Subsection VB; we will confine ourselves to repeating some of the main findings. First, we saw that pied piping is obligatory if the wh-phrase is prenominal such as a demonstrative or a possessive pronoun. One example is repeated here as (164).

164

Welk/Wiens	<boek>	heeft	Marie <*boek>	gelezen?
which/whose	book	has	Marie	read
'Which/Whose book has Marie read?'

This leaves us with postnominal PPs like the possessive PP van Peter in (165a). Example (165b) shows that such examples are different from examples like (164) in that pied piping is excluded, and example (165b') furthermore suggests that, in accordance with the constraint "avoid pied piping" in (153), stranding is possible.

165

a.	Marie heeft	[de auto [van Peterposs]]	geleend.
	Marie has	the car of Peter	borrowed
	'Marie has borrowed up Peterʼs car.'

b.	*	[De auto	[van wieposs]]	heeft	Marie	geleend?
		the car	of who	has	Marie	borrowed

b'.		Van wie	heeft Marie	de auto	geleend?
		of who	has Marie	the car	borrowed

Things are, however, more complicated than this. The examples in (166) show that the van-PP in (165b') need not be construed as the possessor of the noun phrase, but may also be analyzed as an indirect object (source). First, (166a) shows that the interrogative van-PP can also be used if the possessor is expressed by a possessive pronoun, which makes it very unlikely that the van-PP functions as a possessor: cf. *[zijn auto van Peter]'lit.: his car of Peter'. Second, (166b) shows that the direct object can be pronominalized without affecting the van-PP, whereas pronominalization normally affects all noun phrase internal elements.

166

a.	Van wie	heeft Marie	zijnposs	auto	geleend?
	of who	has Marie	his	car	borrowed
	'From whom did Marie borrow his car?'

b.	Van wie	heeft	Marie	hem	geleend?
	of who	has	Marie	him	borrowed
	'From whom did Marie borrow it?'

Subsection VB has shown further that it is very hard (if not impossible) to construct cases that do not allow some alternative analysis; wh-moved van- and voor-PPs, for example, can in many cases plausibly be analyzed as restrictive adverbial phrases; see also N2.2.1, sub VC. Finally, it was shown that in many cases postnominal PPs cannot be extracted; one case illustrating this ban on stranding of the noun phrase is repeated in (167b). Example (167b') is added to show that pied piping is likewise excluded, which means that the intended question can simply not be phrased.

167

a.	Els zal	morgen	[haar klacht [tegen Peter]]	intrekken.
	Els will	tomorrow	her complaint against Peter	withdraw
	'Els will withdraw her complaint against Peter tomorrow.'

b.	*	[Tegen wie]i	zal	Els	[haar klacht ti]	morgen	intrekken?
		against who	will	Els	her complaint	tomorrow	withdraw

b'.	*	[Haar klacht	[tegen wie]]i	zal	Els morgen ti	intrekken?
		her complaint	against who	will	Els tomorrow	withdraw

The above suggests that noun phrases are absolute islands for wh-extraction, although more research is needed to establish this firmly; a similar claim was made earlier by Horn (1974), Bach & Horn (1976), Koster (1978:81) and, at least for definite noun phrases, by Fiengo & Higginbotham (1981), subsection VII will return to this issue and discuss one possible counterexample, the so-called wat voor split.

[+] C. Wh-extraction from APs

This subsection is again relatively brief given that much of what will be said here is discussed more extensively in Sections A2.3 and A3.1.4. We start by showing that wh-movement of a prepositional/nominal complement of an AP normally does not trigger pied piping. The result of wh-movement of the modifier of an AP depends on the nature of the modifier: some trigger pied piping whereas others are compatible with stranding.

[+] 1. PP-complements

Section A2.1 has shown that adjectives typically select a PP as their complement. Although such complements can normally either precede or follow the adjective, their base-position is the one following the adjective. There are at least three arguments in favor of the claim that the pre-adjectival position of prepositional complements is normally derived by leftward movement. We will illustrate this here by means of the examples in (168), in which the adjective boos'angry' selects an over-PP as its complement. First, (168a) shows that the over-PP cannot be placed between the modifier erg'very' and the adjective boos; given that complements are normally generated closer to the selecting head than modifiers, this would be unexpected if the voor-PP were base-generated in pre-adjectival position. Second, the freezing principle requires that stranded prepositions occupy their base-position; the fact that the stranded preposition over cannot precede the adjective in (168b) therefore shows that the PP originates in post-adjectival position. Finally, the (c)-examples show that topicalization of the full AP is not possible if the PP-complement precedes the adjective; this strongly suggests that the PP is external to the AP if it is in pre-adjectival position.

168

a.	Jan is <over die opmerking>	erg boos <over die opmerking >.
	Jan is about that remark	very angry
	'Jan is angry about that remark.'

b.	Jan is er	nog	<*over>	erg boos <over>.
	Jan is there	still	about	very angry
	'Jan is still angry about it.'

c.		Erg boos over die opmerking	is Jan niet.
		very angry about that remark	is Jan not

c'.	*	Over die opmerking	erg boos	is Jan niet.
		about that remark	very angry	is Jan not

Example (168a) suggests that the over-PP can be moved out of the AP into a landing site in the middle field of the clause, which is supported by the fact that the PP can be separated from the AP by a clausal adverb: cf. Jan is over die opmerking waarschijnlijk erg boos'Jan is probably very angry about that remark'. It therefore need not surprise us that the PP can also be wh-moved in isolation; cf. (169a). Example (169b) in fact shows that pied piping of the full AP leads to a degraded result, which is, of course, predicted by the constraint "avoid pied piping" in (153).

169

a.	Over welke opmerking	is Jan	[boos ti]?
	about which remark	is Jan	angry
	'About which remark is Jan angry?'

b.	??	[Boos	over welke opmerking]i	is Jan ti?
		angry	about which remark	is Jan

For more extensive discussion of leftward movement of prepositional complements of adjectives, we refer the reader to Section A2.3.1, which also discusses a number of potentially problematic cases for the brief sketch given here.

[+] 2. Nominal complements

Section A2.2 has shown that certain adjectives are able to take a nominal argument; cf. Van Riemsdijk (1983). Two examples are given in (170). We added German examples in order to show that the case of the nominal argument depends on the adjective not on the copular verb; zat/überdrüssig'fed up' select genitive, while vertrouwd/geläufig select dative case. Case assignment thus shows that the nominal object is an argument of the adjective (and not of the verb).

170

a.		Peter is	deze opera	zat.	Dutch
		Peter is	this opera	fed.up

a'.	Peter ist	dieser Opergenitive	überdrüssig.	German
	Peter is	this opera	fed.up
	'Peter is fed up with this opera.'

b.		Deze omgeving	is	hem	erg vertrouwd.	Dutch
		this area	is	him	very familiar

b'.	Diese Umgebung	ist	ihmdative	sehr geläufig.	German
	this area	is	him	very familiar
	'This area is very familiar to him.'

A potential problem with these cases is that the regular constituency tests do not show that the adjective and the genitive/dative noun phrase form a constituent; cf. Section A2.3.2. It is for instance awkward to place them into sentence-initial position together; judgments differ from case to case and from speaker to speaker, but examples like (171a&b) are generally considered degraded. The primed examples show that the noun phrase and the adjective can both be topicalized in isolation.

171

a.	%	[Deze opera zat]i	is	Peter nog	niet ti.
		this opera fed.up	is	Peter yet	not
		'Peter is not yet fed up with this opera.'

a'.

Deze opera

Peter nog niet zat.

a''.

Zat is Peter deze opera nog niet.

b.	%	[Hem vertrouwd]i	is deze omgeving	nog niet ti.
		him familiar	is this area	still not
		'This area is not yet familiar to him.'

b'.

Hem is deze omgeving nog niet vertrouwd.

b''.

Deze omgeving is hem nog niet vertrouwd.

The questionable acceptability of the primeless examples suggests that, for some unknown reason, the nominal argument must be moved leftward into some AP-external position. This is in fact also suggested by the fact that the nominal complement of the adjective must precede the modifier erg'very' in the examples in (172) and can even be separated from the AP by a clausal adverb such as waarschijnlijk'probably': cf. Cinque (1993:252).

172

a.	Peter is	<deze opera>	erg <*deze opera>	zat.
	Peter is	this opera	very	fed.up
	'Peter is very fed up with of this opera.'

a'.	Peter is deze opera	waarschijnlijk	zat.
	Peter is this opera	probably	fed.up
	'He is probably fed up with this opera.'

b.	Deze omgeving	is	<hem>	erg <*hem>	vertrouwd.
	this area	is	him	very	familiar
	'This area is very familiar to him.'

b'.	Deze omgeving	is hem	waarschijnlijk	vertrouwd.
	this area	is him	probably	familiar
	'This area is probably familiar to him.'

Given the discussion above it will not come as a surprise that wh-movement of the nominal argument cannot pied-pipe the adjective.

173

a.	Welke opera	ben	je	zat?	stranding
	which opera	are	you	fed.up
	'Which opera are you fed up with?'

a'.

Welke opera zat ben je?

pied piping

b.	Wie	is deze omgeving	nog niet vertrouwd?	stranding
	who	is this area	yet not familiar
	'To whom is this area not yet familiar?'

b'.

Wie vertrouwd is deze omgeving nog niet?

pied piping

For completeness' sake, we add the examples in (174) to show that wh-movement of the (modified) adjectives hoe zat/bekend themselves strands the nominal argument.

174

a.	Hoe zat	ben	je	deze opera?	stranding
	how fed.up	are	you	this opera
	'How fed up are you with this opera?'

a'.

Deze opera hoe zat ben je?

pied piping

b.	Hoe vertrouwd	is deze omgeving	jou?	stranding
	how familiar	is the area	you
	'How familiar is this area to you?'

b'.

Jou hoe vertrouwd is deze omgeving?

pied piping

[+] 3. Modifiers

The (a)-examples show that wh-movement of the interrogative degree modifier hoe'how' obligatorily pied-pipes the AP; stranding of the adjectival head leads to a severely degraded result.

175

a.		Jan is erg verslaafd.
		Jan is very addicted

b.		[Hoe verslaafd]i	is Jan ti?	pied piping
		how addicted	is Jan

b'.	*	Hoei	is Jan [ti	verslaafd]?	stranding
		how	is Jan	addicted

Things are different, however, if the adjective is modified by a gradable degree adverb. The interrogative counterpart of (176) is compatible both with pied piping and stranding although the latter seems to be somewhat preferred (but judgments seem to differ from case to case and from speaker to speaker); cf. Section A3.1.2, sub IV.

176

a.		Jan is zwaar verslaafd.
		Jan is heavily addicted
		'Jan is severely addicted.'

b.	?	[Hoe zwaar	verslaafd]i	is Jan ti?	pied piping
		how heavily	addicted	is Jan

b'.	Hoe zwaari	is Jan [ti	verslaafd]?	stranding
	how heavily	is Jan	addicted
	'How severely addicted is Jan?'

A similar contrast can be found in the case of nominal modifiers, although there is a slight complication in this case. First, consider the examples in (177), which show that pied piping is obligatory if the measure adjective lang'long' is modified by the interrogative degree modifier hoe'how'.

177

a.		Het zwembad	is erg lang.
		the pool	is very long

b.		[Hoe lang]i	is het zwembad ti?	pied piping
		how long	is the pool

b'.	*	Hoe	is het zwembad [ti	lang]?	stranding
		how	is the pool	long

Example (178a) shows that measure adjectives like lang can also be modified by a noun phrase. The (b)-examples show that in this case stranding gives rise to a marked but acceptable result; judgments again seem to differ from case to case and fromspeaker to speaker.

178

a.	Het zwembad	is [100 meter lang].
	the pool	is 100 meter long
	'The pool is 100 meters long.'

b.	??	[Hoeveel meter	lang]i	is het zwembad ti?	pied piping
		how.many meters	long	is the pool

b'.	?	Hoeveel meter	is het zwembad [ti	lang]?	stranding
		how.many meter	is the pool	long

The markedness of (178) is probably of a non-syntactic nature; it may be an instance of blocking, due to the fact that the intended question can be more economically expressed by means of example (177b). That we are not dealing with a syntactic restriction is clear from the fact that nominal modifiers of the type in (178) can also be used in examples like (179), where the degree modifier te'too' blocks the use of the interrogative degree modifier hoe'how'. This means that syntactic blocking does not apply in this case and the result in (179b') is indeed fully acceptable. For more discussion of the behavior of modifiers of measure adjectives like lang'long' in (178) and (179), we refer the reader to Section A3.1.4, sub II.

179

a.	Het zwembad	is [5 centimeter	te lang].
	the pool	is 5 centimeter	too long
	'The pool is 5 centimeters too long.'

b.	??	[Hoeveel centimeter	te lang]i	is het zwembad ti?	pied piping
		how.many centimeter	too long	is the pool

b'.		Hoeveel centimeter	is het zwembad [ti	te lang]?	stranding
		how.many centimeter	is the pool	too long

The examples above have shown that wh-movement of simplex modifiers like hoe'how' obligatorily pied-pipe the full AP. Wh-movement of more complex modifiers like hoe zwaar'how heavily' in (176), hoe lang'how long' in (177) and hoeveel (centi)meter'how many centimeters' do allow stranding. The fact that pied piping is allowed as a marked option alongside stranding is again a potential problem for the constraint "avoid pied piping" in (153).

[+] VII. A note on the avoidance of pied piping

Subsection VI has shown that, depending on various factors, wh-movement of a subpart of a clausal constituent may involve pied piping of the full clausal constituent or stranding of its non-interrogative part. The two options are normally in complementary distribution, which was formally accounted for by means of the constraint "avoid pied piping" in (153). We have seen, however, that there are also a number of potential problems for this constraint. We suggested that in at least some of these problematic cases, pied piping is a marked/disfavored option, in accordance with what one might expect on the basis of the "avoid pied piping" constraint, although it should be added that it is still an open question whether this claim will stand further scrutiny. This subsection adds one problem for the "avoid pied piping" constraint that seems uncontroversial: the pied piping/stranding behavior of so-called wat voor-phrases. Since the relevant data are extensively discussed in Section N4.2.2.3, we will illustrate the problem by means of direct objects only. The examples in (180) show that wat voor-phrases freely allow both options.

180

a.	[Wat voor een boeken]i	heeft	Peter ti	gekocht?	pied piping
	what for a books	has	Peter	bought
	'What kind of books has Peter bought?'

b.	Wati	heeft	Peter [ti	voor een boeken]	gekocht?	stranding/wat voor split
	what	has	Peter	for a books	bought
	'What kind of books has Peter bought?'

One way out would be to assume that the two options express different meanings or obey different conditions on their actual use, in which case one might assume that the constraint "avoid pied piping" can be overridden by certain considerations of meaning/actual usage. However, to our knowledge this has never been claimed to be the case. This suggests that "avoid pied piping" is not a hard and fast rule; future research should investigate what other factors may affect its application.

References:

Bach, Emmon & Horn, George M1976Remarks on "Conditions on Transformations"Linguistic Inquiry7265-299
Barbiers, Sjef2002Remnant stranding and the theory of movementAlexiadou, Artemis, Anagnostopoulou, Elena, Barbiers, Sjef & Gärtner, Hans-Martin (eds.)Dimensions of movement. From features to remnantsAmsterdam/PhiladelphiaJohn Benjamins47-67
Bayer, Josef2006<i>Wh</i>-in-situEveraert, Martin & Riemsdijk, Henk van (eds.)The Blackwell companion to syntax5Malden, MA/OxfordBlackwell Publishing
Broekhuis, Hans2015Feature inheritance versus extended projections
Cheng, Lisa1991On the typology of wh-questionsMITThesis
Cheng, Lisa1997On the typology of wh-questionsNew York/LondonGarland Publishing
Chomsky, Noam1991Some notes on economy of derivation and representationFreidin, Robert (ed.)Principles and parameters in comparative syntaxCambridge, MAMIT Press417-454
Corver, Norbert1990The syntax of left branch extractionTilburgUniversity of TilburgThesis
Corver, Norbert1990The syntax of left branch extractionTilburgUniversity of TilburgThesis
Corver, Norbert1990The syntax of left branch extractionTilburgUniversity of TilburgThesis
Corver, Norbert1990The syntax of left branch extractionTilburgUniversity of TilburgThesis
Corver, Norbert2003Hoe leefbaar is het NederlandsNederlandse Taalkunde8285-300
Corver, Norbert2003Hoe leefbaar is het NederlandsNederlandse Taalkunde8285-300
Dayal, Veneeta2006Multiple-<i>wh</i>-questionsEveraert, Martin & Riemdijk, Henk van (eds.)The Blackwell companion to syntax3Malden, MA/OxfordBlackwell Publishing275-326
Fiengo, Robert & Higginbotham, James1981Opacity in NPLinguistic Analysis7395-421
Guéron, Jacqueline1980On the syntax and semantics of PP extrapositionLinguistic Inquiry11637-678
Horn, George M1974The noun phrase constraintUniversity of MassachusettsThesis
Koster, Jan1978Locality principles in syntaxDordrechtForis Publications
Koster, Jan1987Domains and dynasties. The radical autonomy of syntaxDordrecht/ProvidenceForis Publications
Riemsdijk, Henk C. van1978A case study in syntactic markedness: the binding nature of prepositional phrasesPeter de Ridder Press
Riemsdijk, Henk van1983The case of the German adjectivesHeny, Frank & Richards, B. (eds.)Linguistic Categories: Auxiliaries and related Puzzles1DordrechtReidel223-252
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vos, Riet1994Eurotra grammar. Wh-questions: a comparative study of Dutch, English, Portuguese and SpanishAlessandriaEdizioni del'orso
Vries, Mark de2002The syntax of relativizationAmsterdamUniversity of AmsterdamThesis
Vries, Mark de2002The syntax of relativizationAmsterdamUniversity of AmsterdamThesis
Vries, Mark de2002The syntax of relativizationAmsterdamUniversity of AmsterdamThesis
Vries, Mark de2002The syntax of relativizationAmsterdamUniversity of AmsterdamThesis
Zwart, Jan-Wouter2011The syntax of DutchCambridgeCambridge University Press

⇑ Back to top

About this article

Author(s): Hans Broekhuis, Norbert Corver

Category: Dutch Syntax

Keywords: AccentAdjective PhraseAdpositional PhraseClauseContrastive AccentFree Relative ClauseIntonationOperatorNoun PhrasePied PipingQuestionR-pronominalizationReconstructionRelative ClauseStrandingWh-questionWh-extraction

Version: 1.0, January 2016

Version history:

version	editor(s)	date	remarks
1.0	Hans Broekhuis, Norbert Corver	January 2016

report errorprintcite

Jump to:

introduction
references
related articles