• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
Introduction
readmore
[+]  I.  Characterization and classification (Chapter 1)

Section 1.1 provides a brief survey of some conspicuous syntactic, morphological and semantic characteristics of verbs. Section 1.2 reviews a number of semantic and syntactic classifications of verbs and proposes a partly novel classification bringing together some of these proposals; this classification will be the starting point of the more extensive discussion of nominal complementation in Chapter 2. Section 1.3 discusses verbal inflection while Sections 1.4 and 1.5 discuss a number of semantic notions related to verbs: tense, mood/modality and aspect.

[+]  II.  Argument structures (Chapter 2)

Verbs can project in the sense that they take arguments (Chapter 2 to Chapter 5) and that the resulting projections can be modified by a large set of adverbial phrases (Chapter 8). We will begin the discussion of complementation by focusing on the adicity of verbs, that is, the number and type of arguments they can take. The traditional classification is normally based on the number of nominal arguments that verbs take, that is, whether a verb is intransitive, transitive or ditransitive.

1
a. Jan lacht.
intransitive
  Jan laughs
b. Jan leest een boek.
transitive
  Jan reads a book
c. Jan biedt Peter een baan aan.
ditransitive
  Jan offers  Peter a job  prt.

Chapter 2 provides evidence, however, that in order to arrive at a satisfactory classification not only the number but also the type of arguments should be taken into account: we have to distinguish between what have become known as unergative and unaccusative verbs, which exhibit systematic differences in syntactic behavior. Because the distinction is relatively new (it was first proposed in Perlmutter 1978, and has received wider recognition only after Burzio 1981/1986) but nevertheless plays an important role throughout this study, we will briefly introduce the distinction here.
      Unaccusative verbs never take an accusative object. The subjects of these verbs maintain a similar semantic relation with the unaccusative verb as direct objects with transitive verbs; they are both assigned the thematic role of theme. This is illustrated by the minimal pair in (2); the nominative noun phrase het glas'the glass' in the unaccusative construction (2b) maintains the same relation with the verb as the accusative noun phrase het glas in the transitive construction in (2a). It is therefore generally assumed that the subject in (2b) originates in the regular direct object position, but is not assigned accusative case by the verb, so that it must be promoted to subject, for which reason we will call the subject of an unaccusative verb a DO-subject. The fact that (2b) has a transitive alternant is an incidental property of the verb breken'to break'. Some verbs, such as arriveren'to arrive', only occur in an unaccusative frame.

2
a. Jan breekt het glas.
  Jan  breaks the glass
a'. * Jan arriveert het boek.
transitive
  Jan arrives  the book
b. Het glas breekt.
  the glass  breaks
b'. Het boek arriveert.
unaccusative
  the book  arrives

Hoekstra (1984a) has argued that regular intransitive verbs and unaccusative verbs have three distinguishing properties: (a) intransitives take the perfect auxiliary hebben'to have', whereas unaccusatives take the auxiliary zijn'to be'; (b) the past/passive participle of unaccusatives can be used attributively to modify a head noun that corresponds to the subject of the verbal construction, whereas this is not possible with intransitive verbs; (c) the impersonal passive is possible with intransitive verbs only. These properties are illustrated in (3) by means of the intransitive verb lachen'to laugh' and the unaccusative arriveren'to arrive'.

3
Intransitive
Unaccusative
a. Jan heeft/*is gelachen.
  Jan has/is laughed
b. Jan is/*heeft gearriveerd.
  Jan is/has arrived
a'. * de gelachen jongen
  the laughed boy
b'. de gearriveerde jongen
  the arrived boy
a''. Er werd gelachen.
  there was laughed
b''. * Er werd gearriveerd.
  there was arrived

Mulder & Wehrmann (1989), however, argued that only a subset of the unaccusative verbs exhibits all the properties in (3). Locational verbs like hangen in (4), for example, enter into a similar alternation as the verb breken in (2), but nevertheless the verb in (4b) does not fully exhibit the behavior of the verb arriveren, as is clear from the fact that it takes the auxiliary hebben in the perfect tense. It has been suggested that this might be due to the fact that there is an aspectual difference between the verbs arriveren and hangen: the former is telic whereas the latter is not.

4
a. Jan hangt de jas in de kast.
transitive
  Jan hangs  the coat  into the wardrobe
b. De jas hangt in de kast.
intransitive
  the coat  hangs  in the wardrobe

The examples in (5) show that we can make a similar distinction for the dyadic verbs. A verb like bevallen'to please' in the (b)-examples behaves like an unaccusative verb in the sense that it selects the auxiliary zijn and cannot be passivized. Since the object would appear with dative case in languages with morphological case (cf. the German verb gefallen'to please'), such verbs have become known as nominative-dative (nom-dat) verbs. A verb like onderzoeken'to examine' in the (a)-examples behaves like a traditional transitive verb in that it selects the auxiliary hebben and can be passivized while in a language with morphological case the object would be assigned accusative case (cf. the German verb besuchen'to visit').

5
a. De dokter heeft/*is Marie gisteren onderzocht.
  the physician  has/*is  Marie yesterday  examined
a'. Marie is gisteren (door de dokter) onderzocht.
  Marie has.been  yesterday   by the physician  examined
b. De nieuwe voorzitter is/*heeft mij goed bevallen.
  the new chairman  is/has  me  well  pleased
b'. * Ik ben goed bevallen (door de nieuwe voorzitter).
  have.been  well  pleased   by the new chairman

Given that unaccusative verbs have a DO-subject, that is, a subject that occupies an underlying object position, we correctly predict that unaccusative triadic verbs do not exist. Consequently, if the distinction between what is nowadays known as unergative (verbs that in principle can assign accusative case) and unaccusative verbs is indeed on the right track, we have to extend the traditional classification of verbs at least as in Figure 1. Sections 1.2 and 2.1 will argue that there are reasons to extend the classification in Figure 1 even further, but we will not digress on this here.

Figure 1: Classification of verbs taking nominal arguments

      Section 2.2 discusses verbs taking various types of predicative complements. Examples are the copulas, the verb vinden'to consider' and a large set of verbs that may combine with a resultative phrase.

6
a. Jan is aardig.
copular construction
  Jan is nice
b. Ik vind Jan aardig.
vinden-construction
  consider  Jan nice
c. Jan slaat Peter dood.
resultative construction
  Jan hits  Peter dead

We will also show that verbs entering the resultative construction may shift from one verb class to another by (apparently) changing their adicity, as illustrated in the (a)-examples in (7), or their selectional properties, as in the (b)-examples.

7
a. Jan loopt (*het gras).
adicity
  Jan walks     the grass
a'. Jan loopt *(het gras) plat.
  Jan walks     the grass  flat
b. Jan veegt de vloer/$bezem.
selection
  Jan brushes  the floor/broom
b'. Jan veegt de bezem/$vloer kapot.
  Jan brushes  the broom/floor  broken

      Sections 2.3 and 2.4 discuss verbs taking PP-complements, like wachten'to wait' in (8a). and the somewhat more special cases such as wegen'to weigh' in (8b) that take an obligatory adjectival phrase. The discussion of complements in the form of a clause will be postponed to Chapter 5.

8
a. Jan wacht op vader.
PP-complements
  Jan waits for father
b. Jan weegt veel te zwaar.
AP-complements
  Jan weighs  much too heavy

Section 2.5 concludes by discussing another number of more special verb types like inherently reflexive verbs and so-called object experiencer verbs.

9
a. Jan vergist zich.
inherently reflexive verb
  Jan be.mistaken  refl
b. Die opmerking irriteert Jan/hem.
object experiencer verb
  that remark  annoys  Jan/him
[+]  III.  Verb frame alternations (Chapter 3)

The previous subsection has already shown that it is not always possible to say that a specific verb categorically belongs to a single class: examples (2) and (4), for instance, demonstrate that the verbs breken'to break' and hangen'to hang' can be used both as a transitive and as an unaccusative verb. And the examples in (7) show that the class of the verb may apparently also depend on other elements in the clause. This phenomenon that verbs may be the head of more of one type of syntactic frame is known as verb frame alternation will be discussed in Chapter 3. Another familiar type of alternation, known as dative shift, is illustrated in (10).

10
a. Marie geeft het boek aan Peter.
dative shift
  Marie gives  the book  to Peter
b. Marie geeft Peter het boek.
  Marie gives  Peter  the book

We will take a broad view of the term verb frame alternation and include voice alternations such as the alternation between active and passive clauses, illustrated in the (a)-examples in (11), as well as alternations that are the result of derivational morphology, such as the so-called locative alternation in the (b)-examples in (11), which is triggered by the affixation by the prefix be-.

11
a. Jan leest het boek.
passivization
  Jan reads  the book
a'. Het boek wordt door Jan gelezen.
  the book  is  by Jan  read
b. Jan plakt een foto op zijn computer.
locative alternation
  Jan pastes  a picture  on his computer
b'. Jan beplakt zijn computer met foto's.
  Jan be-pastes  his computer  with pictures
[+]  IV.  Clausal/verbal complements (Chapter 4 to Chapter 7)

These chapters in a sense continue the discussion in Chapter 2 on argument structure by discussing cases in which verbs take a verbal dependent, that is, a clause or a smaller (extended) projection of some other verb. The reason not to discuss this type of complementation in Chapter 2 is that it does not essentially alter the syntactic verb classification developed there: for example, many of the verbs taking an internal argument have the option of choosing between a nominal and a clausal complement. The reason for devoting a separate chapter to clausal/verbal arguments is that such arguments exhibit many special properties and introduce a number of complicating factors that have been investigated extensively in the literature. Even a brief discussion of these special properties and complicating factors would have seriously hampered the main line of argumentation in Chapter 2, and it is therefore better to discuss these properties in their own right.

[+]  A.  Selection of clauses and verb phrases (Chapter 4)

We start our discussion of clausal/verbal complements by reviewing a number of central issues pertaining to the types of verbal dependents that can be distinguished and thus provides the necessary background for the more detailed discussions in Chapter 5 to Chapter 7.

[+]  B.  Argument and complementive clauses (Chapter 5)

Chapter 5 provides an exhaustive discussion of dependent clauses functioning as arguments or complementives. Section 5.1 starts with finite argument clauses; we will discuss subject, direct object, and prepositional clauses. This section also includes a discussion of fragment clauses and wh-extraction.

12
a. dat duidelijk is [dat Marie de nieuwe voorzitter wordt].
subject
  that  clear  is   that  Marie the new chairman  becomes
  'that it is clear that Marie will be the new Chair.'
b. dat Jan niet gemeld heeft [dat hij weg zou zijn].
direct object
  that  Jan not  reported  has   that  he  away  would  be
  'that Jan hasnʼt reported that heʼd be away.'
c. dat Peter erover klaagt [dat het regent].
prepositional object
  that  Peter about.it  complains   that it rains
  'that Jan is complaining about it that it is raining.'

A typical example of fragment clauses is given in (13b); constructions like these are arguably derived by a partial deletion of the phonetic contents of a finite clause, which is indicated here by means of strikethrough.

13
a. Jan heeft gisteren iemand bezocht.
speaker A
  Jan has  yesterday  someone  visited
  'Jan visited someone yesterday.'
b. Kan je me ook zeggen wie Jan gisteren bezocht heeft?
speaker B
  can you  me also  tell  who  Jan yesterday  visited  has
  'Can you tell me who (Jan visited yesterday)?'

Wh-extraction is illustrated in (14b) by means of wh-movement of the direct object of the complement clause. In constructions like these the wh-phrase arguably originates in the same position as the direct object dit boek in (14a), that is, the embedded clause in (14b) contains an interpretative gap, which we have indicated by means of a horizontal line.

14
a. Ik denk [Clause dat Marie dit boek morgen zal kopen].
  think  that  Marie  this book  tomorrow  will  buy
b. Wat denk je [Clause dat Marie __ morgen zal kopen]?
  what  think  you  that  Marie  tomorrow  will  buy
  'What do you think that Marie will buy tomorrow?'

      Section 5.2 discusses three types of formally different types of infinitival clauses: Om + te-infinitivals, te-infinitivals and bare infinitivals. The examples in (15) are control constructions, which are characterized by the fact that they typically have an implicit (phonetically empty) subject pronoun, which is normally represented as PRO. It seems that the construal of PRO, which is normally referred to as control, is subject to a set of context-sensitive conditions. In certain specific environments PRO is obligatorily controlled in the sense that it has an (i) overt, (ii) unique, (iii) local and (iv) c-commanding antecedent, whereas in other environments it need not satisfy these four criteria.

15
a. Jan beloofde [om PRO het boek naar Els te sturen].
om + te-infinitival
  Jan promised  comp  the book to Els  to send
  'Jan promised to send the book to Els.'
b. Jan beweerde [PRO het boek naar Els te sturen].
te-infinitival
  Jan claimed  the book  to Els  to send
  'Jan claimed to send the book to Els.'
c. Jan wilde [PRO het boek naar Els sturen].
bare infinitival
  Jan wanted  the book  to Els  send
  'Jan wanted to send the book to Els.'

In addition to the control infinitivals in (15) there are also subject raising and accusativus-cum-infinitivo infinitivals. An example of the first type is given in (16b). The fact that the matrix verb schijnen in (16a) is unable to take a referential subject such as Jan suggests that the same holds for the verb schijnen in (16b). This has led to the hypothesis that the noun phrase Jan in (16b) is base-generated as the subject of the infinitival clause and subsequently raised to the subject position of the matrix clause, in a similar way as the underlying object of a passive clause is promoted to subject, subject raising is restricted to te-infinitivals and bare infinitivals and we will show that this can be accounted for by appealing to a generally assumed locality restriction on this type of passive-like movement.

16
a. Het schijnt [dat Jan een nieuwe auto koopt].
  it  seems   that  Jan a new car buys
  'It seems that Jan is buying a new car.'
b. Jani schijnt [ti een nieuwe auto te kopen].
  Jan  seems  a new car  to buy
  'Jan seems to be buying a new car.'

Accusativus-cum-infinitivo (lit.: accusative with infinitive) constructions are characterized by the fact that the subject of infinitival clause is phonetically expressed by an accusative noun phrase. In Dutch, this construction occurs with bare infinitivals headed by a causative or a perception verb only; cf. example (17).

17
a. Marie liet [hemacc dansen].
  Marie  make/let   him  dance
  'Marie made him dance.'
b. Els hoorde [henacc een liedje zingen].
  Els heard  them  a song  sing
  'Els heard them sing a song.'

      Section 5.3 concludes with a discussion of complementives, that is, clauses that function as secondary predicates; examples of cases that are sometimes analyzed as complementives are the copular constructions in (18).

18
a. Een feit is [dat hij te lui is].
  a fact  is   that  he  too lazy  is
  'A fact is that he is too lazy.'
b. dat boek is moeilijk [(om) te lezen].
  that book  is hard/not   comp  to read
  'that book is hard to read.'

Because the complementive use of clauses is extremely rare, it seems advisable to not immediately commit ourselves to the suggested complementive analysis. Closer scrutiny will in fact reveal that at least in some cases there is reason for doubting this analysis: it seems plausible, for instance, that example (18b) should be analyzed as a construction with a complementive AP modified by an infinitival clause.

[+]  C.  Complements of non-main verbs (Chapter 6)

Non-main verbs differ from main verbs in that they do not denote states of affairs, but express additional (e.g., aspectual) information about the state of affairs denoted by the main verb. This implies that non-main verbs do not have an argument structure and are thus not able to semantically select a clausal/verbal complement. Nevertheless, the use of the term selection is also apt in this case since non-main verbs impose selection restrictions on the verb they are accompanied by: the examples in (19) show that perfect auxiliaries like hebben'to have' select past participles, semi-aspectual verbs like zitten'to sit' select te-infinitives, and aspectual verbs like gaan'to go' select bare infinitives. Chapter 6 will review a number of characteristic properties of non-main verbs and will discuss the three subtypes illustrated in (19).

19
a. Jan heeft dat boek gelezen.
perfect auxiliary
  Jan has  that book  read
  'Jan has read that book.'
b. Jan zit dat boek te lezen.
semi-aspectual verb
  Jan sits  that book  to read
  'Jan is reading that book.'
c. Jan gaat dat boek kopen.
aspectual verb
  Jan goes  that book  buy
  'Jan is going to buy that book.'
[+]  D.  Verb clustering (Chapter 7)

Verb clustering is probably one of the most discussed issues in the syntactic literature on Dutch and German, and the topic is certainly complex enough to devote a separate chapter to it. Verb clustering refers to the phenomenon that verbs that are in a selection relation tend to group together in the right periphery of the clause (with the exception of finite verbs in main clauses, which must occur in second position). This phenomenon is illustrated in (20) by the embedded counterparts of the main clauses in (19).

20
a. dat Jan dat boek heeft gelezen.
perfect auxiliary
  that  Jan that book  has  read
  'that Jan has read that book.'
b. dat Jan dat boek zit te lezen.
semi-aspectual verb
  that  Jan that book  sits  to read
  'that Jan is reading that book.'
c. dat Jan dat boek gaatkopen.
aspectual verb
  that  Jan that book  goes buy
  'that Jan is going to buy that book.'

The examples in (20) show that verb clusters may arise if a non-main verb selects a past/passive participle, a te-infinitive, or a bare infinitive as its complement. Verb clusters may actually consist of more than two verbs as is shown in (21) by means of the perfect-tense counterparts of (20b&c).

21
a. dat Jan dat boek heeft zitten te lezen.
  that  Jan that book  has  sit  to read
  'that Jan has been reading that book.'
b. dat Jan dat boek is gaan kopen.
  that  Jan that book  is  go  buy
  'that Jan has gone to buy that book.'

Furthermore, verb clustering is not restricted to non-main verbs: it is also possible with main verbs selecting a te-infinitival or a bare infinitival (but not with main verbs selecting an om + te-infinitival). Example (22) provides some examples on the basis of the (b)-examples in (16) and (17), repeated here in a slightly different form for convenience.

22
a. Jan schijnt een nieuwe auto te kopen.
  Jan  seems  a new car  to buy
  'Jan seems to be buying a new car.'
a'. dat Jan een nieuwe auto schijnt te kopen.
  that  Jan  a new car  seems  to buy
b. Els hoorde hen een liedje zingen.
  Els heard  them  a song  sing
  'Els heard them sing a song.'
b'. dat Els hen een liedje hoorde zingen.
  that  Els them  a song  heard  sing

In the examples in (20) and (22) verb clustering is obligatory but this does not hold true across-the-board. In some examples, verb clustering is (or seems) optional and in other cases it is forbidden:

23
a. dat Jan <dat boek> probeerde <dat boek> te lezen.
  that  Jan    that book  tried  to read
  'that Jan tried to read that book.'
b. dat Jan Marie <??dat boek> aanbood <dat boek> te lezen.
  that  Jan Marie      that book  prt.-offered  to read
  'that Jan offered to Marie to read that book.'

Some descriptions of verb clustering take it more or less for granted that any string of verbs (or rather: verb-like elements) in clause-final position can be analyzed as a verb cluster. Section 5.2.2 and Chapter 6 show that many of such cases should in fact receive a different analysis: we may be dealing with, e.g., deverbal adjectives or nominalizations. These findings are important since this will enable us to present a much simpler description of verb clustering than is found in more descriptive grammars such as Haeseryn et al. (1997). Section 7.1 will therefore start by providing some diagnostics that may help us to identify genuine verb clusters. Sections 7.2 and 7.3 discuss the intricate relation between the hierarchical and the linear order of verb clusters. Section 7.4 concludes with a discussion of the permeation of verb clusters by clausal constituents, a phenomenon that is especially pervasive in the variety of Standard Dutch spoken in Flanders.

[+]  V.  Modification (Chapter 8)

This chapter will discuss adverbial modification of the clause/verbal projection. Section 8.1 will discuss the various semantic types of adverbial clause: the basic distinction is the one between adverbial phrases modifying the VP, like manner and certain spatio-temporal modifiers, and adverbial phrases modifying some larger part of the clause, like negation and modal modifiers. Section 8.2 will discuss the categorial status of adverbial phrase and show that there are often various options. temporal modifier, for example, can be APs (vroeg'early'), PPs (na de wedstrijd'after the game', NPs (de hele wedstrijd'during the whole game') and clauses (nadat Ajax verloren h