/ instituut voor de Nederlandse taal / Taalportaal
bibliography glossary about
  • Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
    11.3.3.Topicalization
    quickinfo

    This section discusses topicalization, the phenomenon that in main clauses virtually any clausal constituent (and sometimes also parts thereof) may precede the finite verb in second position, subsection I starts by showing that, as in the case of question formation, the moved constituent can have a wide range of syntactic functions and can be of any category, subsection II continues by comparing topicalization to question formation (as well as relativization) in order to motivate the claim that it is derived by wh-movement; we will see that, apart from the fact that topicalization is a root phenomenon, there are indeed compelling reasons to assume wh-movement to be involved in the derivation, subsection III repeats some arguments from Section 9.3 for rejecting the traditional view that subject-initial sentences are necessarily derived by topicalization; exclusion of such sentences from the set of topicalization constructions will lead to the conclusion that such constructions have two characteristic properties: they exhibit subject-verb inversion and have a non-neutral reading, subsection IV explores the latter issue, and will show that topicalized phrases often play a special role in discourse; they express a contrastive focus, act as a topic, or perform a special function in the organization of the discourse. Given this, we may expect for contrastively focused phrases and topics at least that wh-movement may pied-pipe a larger phrase if syntactic restrictions prohibits extraction and subsection V shows that this expectation is indeed borne out, subsection VI continues with a discussion of topicalization of clauses and smaller verbal projections: such cases are special because wh-movement of such constituents is not possible in the case of question formation and relativization, subsection VII concludes with a comparison of topicalization in Dutch and English, and will show that there are a number of conspicuous differences, which raises the question as to whether the two should be considered phenomena of the same kind.

    readmore
    [+]  I.  Syntactic function and categorial status of the topicalized element

    The traditional generative analysis holds that main clauses are derived by placing the finite verb in the second position of the clauses, the so-called C-position in (285), followed by topicalization of some constituent into the so-called clause-initial position, the specifier of CP; see Section 11.1 for details.

    285

    There seem to be virtually no restrictions on the syntactic function or the categorial status of the topicalized element. The examples in (286) start by showing this for nominal arguments: subjects, direct and indirect objects are all possible in sentence-initial position.

    286
    Nominal arguments
    a. Marie/Ze heeft haar broer/hem die baan aangeboden.
    subject
      Marie/she  has  her brother/him  that job  prt.-offered
      'Marie/She has offered her brother/him that job.'
    b. Die baan heeft ze her brother/him aangeboden.
    direct object
      that job  has  she  her brother/him  prt.-offered
      'That job, she has offered [to] her brother/him.'
    c. Haarbroer/Hem heeft ze die baan aangeboden.
    indirect object
      her brother/him  has  she  that job  prt.-offered
      'Her brother/Him, she has offered that job.'

    There are, however, two important differences between subject-initial sentences and sentences with an object in first position. First, clause-initial objects can be considered to be semantically marked in that they act as discourse topics or contrastive foci, or have some other special function in the organization of the discourse, while this does not necessarily hold for clause-initial subjects. Second, topicalized objects are often characterized by a special intonation pattern: the objects in (286b&c), but not the clause-initial subjects in (286a), must be accented, as is clear from the fact the latter but not the former can be a reduced pronoun. This suggests that subject-initial sentences may also be syntactically different from constructions with topicalized objects; we will return to this issue in Subsection III.
          Next, the examples in (287) show that it is also possible to topicalize prepositional objects: (287a) illustrates this for a prepositional indirect object and (287b) for the prepositional object of kijken (naar)'to look (at)'.

    287
    Prepositional arguments
    a. Aan haarbroer/Hem heeft ze die baan aangeboden.
    indirect object
      to her brother/him has  she  that job  prt.-offered
      'His her brother/him, she has offered that job to.'
    b. Naar dat huis staat Jan al een uur te kijken.
    prepositional object
      at that house  stands  Jan already  an hour to look
      'That house, Jan has been staring at for an hour.'

    Complementives can also be topicalized: we illustrate this in (288) by means of three examples with complementives of a different categorial status; they show that noun phrases, APs and PPs can all be topicalized.

    288
    Complementive
    a. Een liefhebber van Jazz ben ik niet echt.
    nominal
      a devotee of jazz  am  I  not  really
      'A devotee of jazz, I am not really.'
    b. Aardig is de nieuwe directeur beslist.
    adjectival
      nice  is the new director  definitely
      'Nice, the new director definitely is.'
    c. In de la heb ik de schaar gelegd.
    adpositional
      into the drawer  have  I  the scissors  put
      'In the drawer, I have put the scissors.'

    Adjuncts can also be topicalized. Example (289a) shows this for supplementives and examples (289b&c) for adverbial phrases. Observe that we did not mark the adverbial phrases for accent; assigning accent is possible but does not seem to be necessary. We will return to this issue in Subsection IV.

    289
    Adjuncts
    a. Kwaad liep hij weg.
    supplementive
      angry  walked  he  away
      'Angry, he walked away.'
    b. Op zolder slapen de kinderen.
    place adverbial
      on attic  sleep  the children
      'In the attic, the children sleep/are sleeping.'
    c. Na de vergadering vertrekken we.
    time adverbial
      after the meeting  leave  we
      'After the meeting, we will leave.'

    The discussion above has shown that topicalization is like wh-question formation in that constituents with various syntactic functions (argument, complementive and adjunct) and of various different forms (noun phrase, AP and PP) can be moved into sentential-initial position. Topicalization differs from wh-movement, however, in that it also allows preposing of clauses; this is illustrated in (290) for a finite clause. We return to topicalization of clauses in Subsection VI. Accent can be assigned at various places within the preposed clause.

    290
    Clauses
    a. Ik verwacht niet [dat hij dat boek wil hebben].
      I  expect  not   that  he  that book  wants  have
      'I don't expect that he wants to have that book.'
    b. [Dat hij dat boek wil hebben] verwacht ik niet.

    The examples in (291) show that it is also possible to topicalize the complement of perfect and passive auxiliaries, a phenomenon known as VP-topicalization. The (a)-examples show that topicalization of the participle is possible both with and without the direct object; the (b)-examples show that subjects are normally not affected. VP-topicalization will also be discussed in Subsection VI. Accent will normally be assigned to the object if it is pied piped by VP-topicalization.

    291
    VP-topicalization
    a. Ze hebben mijn huis nog niet geschilderd.
    perfect
      they  have  my house  yet  not  painted
      'They haven't painted my house yet.'
    a'. [<Mijn huis> geschilderd] hebben ze <mijn huis> nog niet.
    b. Mijn huis wordt volgend jaar geschilderd.
    passive
      my house  be  next year  painted
      'My house will be painted next year.'
    b'. Geschilderd wordt mijn huis volgend jaar.
    [+]  II.  Topicalization is a subcase of wh-movement

    Topicalization involves movement of some constituent into the initial position of the main clause. It resembles the formation of wh-questions in that the movement targets the position immediately preceding the finite verb; this is illustrated again in the (b)-examples in (292). This observation is not trivial; this does not hold for a language like English. We return to this in Subsection VII.

    292
    a. Jan heeft gisteren dat boek gelezen.
      Jan  has  yesterday  that book  read
      'Jan read that book yesterday.'
    b. Welk boeki heeft Jan gisteren ti gelezen?
    wh-question
      which book  has  Jan yesterday  read
      'Which book did Jan read yesterday?'
    b'. Dat boeki heeft Jan gisteren ti gelezen.
    topicalization
      that book  has  Jan yesterday  read
      'That book, Jan read yesterday.'

          The (b)-examples in (293) show that topicalization differs from question formation (and relativization) in that it is a root phenomenon. It cannot apply in embedded clauses.

    293
    a. Marie zei [dat Jan dat boek gelezen heeft].
      Marie said   that  Jan  that book  read  has
      'Marie said that Jan has read that book.'
    b. Marie vroeg [welk boeki Jan ti gelezen heeft].
    wh-question
      Marie asked  which book  Jan  read  has
      'Marie asked which book Jan has read.'
    b'. * Marie zei [dat boeki Jan ti gelezen heeft].
    topicalization
      Marie said   that book  Jan  read  has

    There is no way in which embedded topicalization in examples such as (293b') can be improved. The examples in (294), for instance, show that Dutch does not have the option found in German to have topicalization in embedded clauses with verb-second, as embedded verb-second is categorically prohibited in Dutch. We refer the reader to Haider (1985/2010) and Barbiers (2005: Section 1.3.1.8) for a discussion of embedded verb-second in, respectively, German and a number of non-standard varieties of Dutch; the German example in (294a) is taken from Müller (1998:42) in a slightly adapted from.

    294
    a. Marie sagte [dieses Buchi habeconjunctive sie ti bereits gelesen].
    German
      Marie  said   this book  has  she  already  read
      'Marie said that this book, she had already read.'
    b. * Marie zei [dit boeki had ze ti al gelezen].
    Dutch
      Marie  said   this book  had  she  already  read

    The examples in (294) also show that embedded topicalization cannot occur with a phonetically expressed complementizer, unlike what is the case in English examples such as (295a); cf., e.g., Chomsky (1977), Baltin (1982) and Lasnik & Saito (1992). Since there is no a priori reason to think that Dutch topicalization targets a different position than English topicalization, we have added example (295b'), in which the complementizer dat'that' precedes the topicalized phrase.

    295
    a. Marie thinks [that this booki you should read ti ].
    English
    b. * Marie denkt [dit boeki dat je zou ti moeten lezen].
    Dutch
      Marie thinks   this book  that  you  would  must  read
    b'. * Marie denkt [dat dit boeki je ti zou moeten lezen].
    Dutch
      Marie thinks   that  this book  you  would  must  read

          Examples (296a&b) show that topicalization is like question formation in that it allows long wh-movement if a bridge verb such as denken'to think' is present. It should be noted, however, that long topicalization is like relativization in that it is possible with a wider range of verbs than question formation; cf. Schippers (2012:105). For instance, the factive verb weten'to know' permits long topicalization (and long relativization), but not long wh-movement. It should further be noted that some speakers prefer the resumptive prolepsis construction in (296c) to the somewhat marked long topicalization construction in (296b).

    296
    a. Welk boeki denk/*weet je [dat Jan ti gekocht heeft]?
    wh-question
      which book  think/know  you  that  Jan  bought  has
      'Which book do you think that Jan has bought?'
    b. (?) Dit boeki denk/weet ik [dat Jan ti gekocht heeft].
    topicalization
      this book  think.know  I   that  Jan  bought  has
      'This book I think/know that Jan has bought.'
    c. Van dit boeki denk/weet ik [dat Jan heti gekocht heeft].
    prolepsis
      of this book  think/know  I  that  Jan it  bought  has
      'As for this book, I think/know that Jan has bought it.'

          That topicalization involves wh-movement is also suggested by the fact that it is island-sensitive, just like question formation and relativization. We illustrate this in (297b) by means of an embedded polar question. For completeness' sake, we have added (297b') to show that the intended meaning can be expressed by means of a resumptive prolepsis construction.

    297
    a. Ik vraag me af [of Jan dat boek gekocht heeft]?
      I  wonder  refl  prt.   if  Jan that book  bought  has
      'I wonder whether Jan has bought that book.'
    b. * Dat boeki vraag ik me af [of Jan ti gekocht heeft]?
      that book  wonder  I  refl  prt.   if  Jan  bought  has
    b'. Van dat boeki vraag ik me af [of Jan heti gekocht heeft]?
      of that book  wonder  I  refl  prt.   if  Jan it  bought  has
      'As for this book, I am wondering whether Jan has bought it.'

    Example (298b) illustrates the island-sensitivity of topicalization by means of an adjunct island. In this case, the resumptive prolepsis construction is not available as an alternative because the verb huilen'to cry' does not license a resumptive van-PP.

    298
    a. Jan huilt [omdat Marie dat boek gestolen heeft].
      Jan cries  because  Marie that book  stolen  has
      'Jan is crying because Marie has stolen that book.'
    b. * Dat boeki huilt Jan [omdat Marie ti gestolen heeft].
      that book  cries Jan because  Marie  stolen  has

    This subsection has shown that topicalization exhibits various hallmarks of wh-movement: it targets the clause-initial position, it can be extracted from clauses selected by bridge verbs and it is island-sensitive. What sets it apart from wh-movement and relativization is that it is a root phenomenon; it cannot target the initial position of embedded clauses. We refer to Hoekstra & Zwart (1994), Sturm (1996) and Zwart & Hoekstra (1997) for a discussion of the question as to whether this shows that topicalization targets a different position than wh-movement, as in fact would be claimed in the cartographic approach initiated by Rizzi (1997).

    [+]  III.  Subject-initial clauses versus topicalization constructions

    The standard view in generative grammar is that topicalization is responsible for verb second in declarative main clauses in Dutch. The verb is first moved into the C-position immediately preceding the canonical subject position, after which the specifier position of CP is filled by some topicalized phrase. This implies that subject-initial main clauses such as (299a) must be derived by topicalization, as indicated in the representation in (299b).

    299
    a. Mijn zuster/Zij/Ze heeft dit boek gelezen.
    subject
      my sister/she/she  has  this book  read
      'My sister/she has read this book.'
    b.

    If the derivation in (299) is correct, we would expect the placement of subjects to be subject to similar restrictions as other cases of topicalization, like in the examples in (300). We seen in Subsection I, however, that subjects crucially differ from objects in that they need not be accented. The effect is even more conspicuous with weak (phonetically reduced) pronouns; while (299a) shows that the weak subject pronoun ze'she' is fully acceptable in sentence-initial position, weak object pronouns like 'r'her' in (300a&b) are not because they cannot be accented; see, e.g., Bouma (2008:34) for more discussion. Adverbial PPs with a weak pronominal complement can be topicalized if the preposition can be assigned accent; see Salverda (2000).

    300
    a. Mijn zuster/Haar/*'r heb ik nog niet gezien.
    direct object
      my sister/her/her  have  I  yet not  seen
      'My sister/her I haven't seen yet.'
    b. Op mijn zuster/haar/*'r wil ik niet wachten.
    PP-object
      for my sister/her/her  want I not wait
      'My sister/Her I don't want to wait for.'
    c. Naast 'r zat een aardige heer.
      next.to her  sat  a kind gentleman
      'Next to her sat a kind gentleman.'

    The same contrast is found with the weak R-word er: the examples in (301) show that expletive er, which is normally assumed to occupy the regular subject position, can easily occur in sentence-initial position, but that this is excluded for er functioning as a locative pro-form or the pronominal part of a PP; topicalization is only possible with strong forms like daar'there' and hier'here'; see, e.g., Bouma (2008:29-30). We will ignore here that things are slightly complicated by the fact that (sentence-initial) er may sometimes have more than one function; we refer the reader to Section P5.5.3 for discussion and examples.

    301
    a. Er spelen veel kinderen op straat.
    expletive er
      there  play  many children  on street
      'There are many children playing in the street.'
    b. Daar/*Er spelen de kinderen graag.
    locative er
      there/there  play  the children  gladly
      'The children like to play there.'
    c. Daari/*Eri wacht ik niet [ ti op].
    pronominal part of PP
      there/there  wait  I  not  for
      'That I won't wait for.'

    That this contrast should have an impact on our syntactic analysis is clear from the fact illustrated in (302) that subject pronouns do exhibit a similar behavior as object pronouns if they are extracted from an embedded clause: whereas noun phrases like mijn zuster'my sister' and strong (phonetically non-reduced) subject pronouns such as zij give rise to a reasonably acceptable result, topicalization is excluded if the subject pronoun is weak.

    302
    a. (?) Mijn zusteri/Ziji zei Jan [dat ti dit boek gelezen had].
      my sister/she  said  Jan  comp  this book  read  had
      'My sister/she, Jan said had read the book.'
    b. * Zei zei Jan [dat ti dit boek gelezen had].
      she  said  Jan comp  this book  read  had

    Section 9.3 concluded from this that regular subject-initial constructions do not involve topicalization but are derived by simply placing the subject in the regular subject position, the specifier of the T(ense) head. This resulted in the following derivations of subject-initial clauses and topicalization constructions; cf. Travis (1984) and Zwart (1992/1997). Note that these analyses suggest that subject-verb inversion is a hallmark of topicalization constructions; cf. Salverda (1982/2000).

    303
    a. Subject-initial sentences
    b. Topicalization constructions

    Observe that we are not claiming here that subjects cannot be topicalized, but only that they are not topicalized if they occur in a neutrally pronounced sentence. Examples like (304a) with contrastive accent on the subject may involve topicalization. That they do so is strongly suggested by expletive constructions like (304b); since it is normally assumed that the expletive er'there' occupies the regular subject position, the subject niemand can only occur in sentence-initial position as a result of topicalization. We added the locational adverbial phrase op de vergadering to example (304b) to block a locative interpretation of er'there' in order to ensure that er indeed functions as an expletive.

    304
    a. Mijn zuster heeft dit boek gelezen.
      my sister  has  this book  read
      'My sister/she has read this book.'
    b. Niemand was er op de vergadering.
      nobody  was there  at the meeting
      'Nobody was there at the meeting.'

    The analyses suggested in (303) are interesting in view of the fact that subject-initial clauses are the most neutral form of an utterance from a semantic view point: while topicalized phrases are special in that they play a specific role in structuring the discourse, sentence-initial subjects are often neutral in this respect. The representations in (303) thus enable us to express formally this by postulating that like question formation and relativization, topicalization is semantically motivated; see Dik (1978: Section 8.3.3), Haegeman (1995), Rizzi (1997), and many others. This will be the main topic of Subsection IV.

    [+]  IV.  Information structure: focus and topic

    The information structure of a clause is closely related to its intonation pattern. In utterances like the (b)-examples in (305), which present new information only if intended as an answer to the question in (305a), the main accent is located at the end of the clause, normally on the constituent preceding the clause-final verbs; see Section 13.1, sub III, for more detailed discussion. We will refer to utterances with this intonation pattern as neutral clauses (in order to not complicate things we will discuss main clauses only).

    305
    a. Wat is er gebeurd?
      what  is there  happened
      'What has happened?'
    b. Jan heeft Marie een brief gestuurd.
      Jan has  Marie  a letter  sent
      'Jan has sent Marie a letter.'
    b'. Jan heeft een brief naar Marie gestuurd.
      Jan has  a letter  to Marie  sent
      'Jan has sent a letter to Marie.'

    The intonation pattern of utterances can be affected by the information structure of the clause. In the primed examples in (306), which contain both presupposed and new information if used as answers to the questions in the primeless examples, the main accent must be located in the new information of the clause (henceforth: the new-information focus); in the cases at hand, this results in the placement of the main accent in a more leftward position. For more information about assignment of main accent in clauses we refer the reader to Booij (1995).

    306
    a. Wie heeft Jan een brief gestuurd?
    question
      who has  Jan  a letter  sent
      'Who has Jan sent a letter?'
    a'. Hij heeft Marie een brief gestuurd.
    answer
      Jan has  Marie  a letter  sent
      'He has sent Marie a letter.'
    b. Wat heeft Jan naar Marie gestuurd?
    question
      what  has  Jan  to Marie  sent
      'What has Jan sent to Marie?'
    b'. Hij heeft een brief naar Marie gestuurd.
    answer
      Jan has  a letter  to Marie  sent
      'Jan has sent a letter to Marie.'

    The following subsections will show that topicalization may also affect the intonation pattern of utterances; we will see that the way in which the intonation pattern is affected depends on the impact topicalization has on the information structure of the clause. There are also a number of cases in which topicalization does not seem to have such a great impact on the intonation of the clause; we will discuss some of the prototypical cases. Before we start, we want to note that the literature exhibits a great deal of variation when it comes to information-structural notions like focus and topic; cf. Erteschik-Shir (2007) for an extensive review. We aim at staying close to the use of these notions in É. Kiss' (2002:ch.1-6) description of the Hungarian clause, in which these notions play a prominent role.

    [+]  A.  Contrastive/restrictive focus

    The new-information focus can also be placed in sentence-initial position as a result of topicalization. So, next to the answers in the primed examples in (306), we also find utterances like (307a&b). The parentheses indicate that the presuppositional part of such answers is normally omitted.

    307
    a. Marie (heeft hij een brief gestuurd).
    answer to (306a)
      Marie   has  he  a letter  sent
      'Marie, he has sent a letter.'
    b. Een brief (heeft hij naar Marie gestuurd).
    answer to (306b)
      a letter   has  he  to Marie  sent
      'A letter, he has sent to Marie.'

    Jansen (1981: Section 4.2.1) claims that focus topicalization of the type in (307) is not very frequent (in non-interrogative contexts), which raises the question as to whether we are simply dealing with new-information focus or whether utterances such as (307) have some additional property. We tend to think that the accents in these topicalization constructions are stronger than those in the primed examples in (306), which may suggest that topicalization constructions express contrastive or restrictive focus in the sense that the proposition holds for the focussed phrases, to the exclusion of any other referent; see Section 13.3.2 for more discussion.
          This would be in line with the fact that utterance (307a) expresses that in the relevant domain of discourse only Marie was sent a book by Jan: if it were to turn out that Jan also sent a letter to Peter and that the speaker uttering (307a) was aware of that, he could be accused of not being fully informative by withholding information. The same would hold for utterance (307b) if it turned out that Jan also sent cocaine to Marie.
          That we are dealing with restrictive focus is also supported by the fact that it is often impossible to topicalize non-specific indefinite noun phrases, as these are typically used for introducing new information but cannot easily be used in a contrastive or a restrictive fashion. Example (308a') shows, for example, that topicalization of the existential pronoun iemand gives rise to a highly marked result, and (308b') shows that topicalization of an indefinite noun phrase such as een pianist is restricted to cases in which the speaker contradicts a certain presupposition on the part of the addressee: it would be acceptable as a reaction to the following question: Hoe was je ontmoeting met die cellist gisteren?'How was your meeting with that cellist yesterday?'.

    308
    a. Ik heb gisteren iemand ontmoet.
      I  have  yesterday  someone  met
      'I met someone yesterday.'
    a'. ?? Iemand heb ik gisteren ontmoet.
    b. Ik heb gisteren een pianist ontmoet.
      I  have  yesterday  a pianist  met
      'I met a pianist yesterday.'
    b'. # Een pianist heb ik gisteren ontmoet.

    The negative pronoun niemand'nobody', on the other hand, can be topicalized in constructions such as (309a) if the speaker wants to express that he did expect to see in Amsterdam at least one person from the given domain of discourse. Similarly, example (309b) expresses that the speaker did not expect to be able to meet in Amsterdam all individuals in the given domain of discourse.

    309
    a. Niemand heb ik in Amsterdam gezien (zelfs Jan niet).
      nobody  have  I  in Amsterdam seen   even  Jan  not
      'Nobody, I have seen in Amsterdam (not even Jan).'
    b. Iedereen heb ik in Amsterdam kunnen ontmoeten (zelfs marie).
      everybody  have  I  in Amsterdam  can  meet  even  Marie
      'Everyone, I have been able to meet in Amsterdam (even Marie).'

    Another indication that we are not dealing with mere new-information focus is that the topicalized phrase may be preceded by an (emphatic) focus particle like zelfs'even', alleen'solely', slechts/maar'only': cf. Barbiers (1995:ch.3).

    310
    a. Zelfs Marie heeft hij een brief gestuurd.
      even Marie  has  he  a letter  sent
      'He has even sent Marie a letter.'
    b. Alleen Marie heeft hij een brief gestuurd.
      only Marie  has  he  a letter  sent
      'Only Marie he has sent a letter.'
    c. Slechts twee studenten haalden het examen.
      only two students  passed  the exam
      'Only two students passed the exam.'

    For want of more detailed information on the question as to whether topicalized focus phrases indeed necessarily express more than merely new information, we have to leave our suggestions above to future research.

    [+]  B.  Aboutness topic

    The sentence-initial position is typically occupied by an aboutness topic, a phrase referring to an entity about which the sentence as a whole provides more information. Although the three examples in (311) express the same propositions, they provide additional information about completely different topics: in (311a) the topic is the subject Jan, in (311b) the topic is the direct object de brief'the letter', and in (311c) the topic is embedded in the complementive naar-PP. Observe that the comments in (311) typically contain new information and thus also contain sentence accent (which is again placed on the constituent preceding the clause-final verbs if the full comment consists of new information).

    311
    a. [topic Jan] [comment heeft de brief naar Marie gestuurd].
      Jan  has  the letter  to Marie  sent
      'Jan has sent the letter to Marie.'
    b. [topic De brief] [comment heeft Jan naar Marie gestuurd].
      the letter  has  Jan to Marie  sent
      'The letter, Jan has sent to Marie.'
    c. [topic Naar Marie] [comment heeft Jan de brief gestuurd].
      to Marie  has Jan the letter sent
      'To Marie, Jan has sent the letter.'

    The new information in (311) is provided by an argument, but the examples in (312) show that this can also be an adverbial element that can be used contrastively, such as the negative adverb niet, which can be contrasted with the affirmative marker wel, or adverbs such as morgen'tomorrow', which can be contrasted with adverbs like vandaag'today' or nu'now'. For more examples, see Salverda (2000:100-1).

    312
    a. Peter heb ik nog niet gezien.
      Peter  have  I  not yet  seen
      'Peter, I haven't seen yet.'
    b. Het boek moet je morgen maar lezen.
      the book  must  you  tomorrow  prt  read
      'The book, you should read tomorrow.'

    The aboutness topic is always part of the domain of discourse, which means that it must satisfy certain criteria: (i) it must be referential in the sense that it refers to an entity or set of entities and (ii) it must be specific, that is, the entity or set of entities must be identifiable in the domain of discourse. This implies that the aboutness topic is prototypically a proper noun, a referential personal pronoun, a definite noun phrase, a specific indefinite noun phrase, or a PP containing such a noun phrase; see É. Kiss (2002: chapter 2).

    [+]  C.  Contrastive topics

    Contrastive topics differ from aboutness topics in that they need not be referential or specific; the examples in (313) show that they can be non-individual-denoting elements like bare plurals, indefinite noun phrases, adverbial phrases and verbal particles; examples such as (313a&b) are of course also possible with definite noun phrases ( de zwaan/zwanen'the swan/swans') but this is not illustrated here. Contrastive topics are accented and followed by a brief fall in intonation on the following comment, which gives rise to a typical "hat" contour marked by the symbols "/" and "\". Contrastive topic constructions convey that there is an alternative topic for which an alternative comment holds (cf. É. Kiss 2002: Section 2.7); we made this explicit in the examples in (313) by adding the part within parentheses.

    313
    a. [topic /Zwanen] [comment \heb ik niet gezien] (maar ganzen wel).
      swans  have  I  not  seen   but  geese  aff
      'I haven't seen swans, but I did see geese.'
    b. [topic / Een zwaan][comment \heb ik niet gezien] (maar wel een gans).
        a swan have  I  not  seen   but  aff  a goose
      'I haven't seen a swan, but I did see a goose.'
    c. [topic /Omhoog] [comment \ga ik met de lift] (maar omlaag via de trap).
      up    go I  by the elevator   but  down  via the stairs
      'Up I will use the elevator, but down I will take the stairs.'
    d. [topic /Tegen] [comment \stemden de socialisten] (voor de liberalen).
      against   voted  the socialists  for  the liberals.
      'The conservatives voted against (the bill), the liberals for.'

    The intonation pattern found in utterances like (313) is also possible with individual-denoting elements like the topics in (311). Applying the "hat" contour to these examples will result in similar contrastive readings as those in (313). For completeness' sake, note that examples such (313d) refute the persistent claim that verbal particles cannot be topicalized (cf., e.g., Zwart 2011:72); this is possible provided that they stand in opposition to another verbal particle (cf. Hoeksema 1991a) and thus allow a contrastive interpretation. We refer the reader to Section 13.3.2, sub II, for a more detailed discussion of contrastive topics.

    [+]  D.  Topic shift

    The distal demonstrative pronouns die'that' and dat'that' are very common in sentence-initial position. These pronouns are used to refer to some referent in the immediately preceding context, as in example (314). We added indices in order to unambiguously indicate the intended interpretation of the pronoun. Topicalized demonstratives differ from the topicalized phrases discussed so far in that they need not have contrastive accent; see, e.g., Salverda (1982/2000) and Bouma (2008:45).

    314
    a. Heb je Jani gezien? Nee, diei is ziek.
      have  you  Jan  seen  no  dem  is  ill
      'Did you see Jan? No, he is ill.'

    The demonstrative can be accented, in which case it receives a contrastive/restrictive focus interpretation. If it remains unstressed, it typically indicates topic shift, that is, a change of aboutness topic. In this respect distal demonstratives differ crucially from referential personal pronouns like hij'he' or zij'she', which typically refer to continuous topics. This is illustrated by means of the examples in (315); that the distal demonstrative brings about topic shift is clear from the fact that it cannot refer to the subject (the default topic) of the preceding sentence; referential pronouns are not subject to this restriction. We will not digress on topic shift here but refer the reader to Section N5.2.3.2, sub IIA1, for a more extensive discussion.

    315
    a. [Jani ontmoette Elsj ] en [hiji/*diei vertelde haarj dat ... ]
      Jan  met  Els  and   he/dem  told  her  that
    b. [Jani ontmoette Elsj ] en [zej/diej vertelde hemi dat ... ]
      Jan  met  Els  and   she/ dem  told  him  that

    Note further that distal demonstrative pro-forms like die'that' and dat'that' in sentence-initial position are often omitted in speech; we refer the reader to Section 11.2.2 for discussion of this.

    [+]  E.  Connectives

    The previous subsection has shown that unstressed demonstratives can be used to indicate a topic shift and are thus quite important for a smooth continuation of the discourse. Other topicalized elements with a similar function are connectives like daarom / dus'therefore', and desondanks'nevertheless', which are neither topical nor focal in nature but are simply used to indicate the relation between two successive sentences; cf. Salverda (1982).

    316
    a. [Marie is ziek] en [daarom kan ik niet komen].
      Marie is ill  and  therefore  can  I  not  come
      'Marie is ill and therefore I cannot come.'
    b. [Marie is ziek] maar [desondanks zal ik komen].
      Marie is ill  but  nevertheless  will  I  come
      'Marie is ill but nevertheless I will come.'
    [+]  F.  Formal movement: movement without semantic effect

    The cases of topicalization discussed in the previous subsections are all functionally motivated by information-structural considerations or considerations related to the organization of discourse. There are, however, many cases in which it is not so clear what the functional motivation of topicalization would be. Consider the examples in (317): it has been claimed that the locational PP in (317a) must be interpreted contrastively and thus be assigned accent, whereas the locational PP in (317b) can be interpreted neutrally and thus be pronounced without any phonetic prominence.

    317
    a. In Utrecht heeft Marie haar broer bezocht.
      in Utrecht  has  Marie her brother  visited
      'In Utrecht Marie has visited het brother.'
    b. In Utrecht is Els erg populair.
      in Utrecht  is Els very popular
      'In Utrecht, Els is still very popular.'

    This contrast between the two examples has been related to the semantic contribution of the PPs. The PP in (317a) is event-related in the sense that it is part of what is asserted: Marie has met Jan & this eventuality took place in Utrecht. This reading has the property that omission of the locational PP is possible without affecting the truth value of the assertion. The PP in (317b), on the other hand, is not event-related but is used to restrict the speaker's claim; this reading has the property that omission of the locational PP may affect the truth value of the assertion: from the fact that Els is popular in Utrecht we cannot infer that she is popular elsewhere. The contrast between the two examples in (318) shows that the difference between the two readings is associated with a difference in location of the PP in the middle field of the clause: while the PP can easily precede the subject in (318b), this gives rise to a marked result in (318a) (although the latter example improves if the subject is assigned contrastive accent). We refer to Maienborn (2001) for more detailed discussion.

    318
    a. dat <??in Utrecht> Marie <in Utrecht> haar broer bezocht heeft.
      that       in Utrecht  Marie  her brother  visited  has
      'that Marie has visited her brother in Utrecht.'
    b. dat <in Utrecht> Els <in Utrecht> erg populair is.
      that    in Utrecht  Els  very popular  is
      'that in Utrecht Els is still very popular.'

    There is a wide range of (especially) adverbial phrases that are not directly event-related, and which may occur in sentence-initial positions with no or little emphasis; see Kooij (1978), Salverda (1982/2000) and Florijn (1992). These include at least certain restrictive temporal, modal, and speaker-related adverbials.

    319
    a. In de middeleeuwen waren heksen heel gewoon.
    restrictive temporal
      in the middle ages  were  witches  very common
      'In the Middle Ages, witches were very common.'
    b. Misschien komt Peter straks nog.
    modal
      maybe  comes  Peter later  prt
      'Maybe Peter will come later.'
    c. Helaas kan Peter niet komen.
    speaker-related
      unfortunately  can  Peter not  come
      'Unfortunately, Peter cannot come.'

    Examples of the type in (317b) and (319) are sometimes accounted for by introducing special mechanisms. Odijk (1995:section 2.1), for instance, proposes that adverbials like misschien'maybe' and helaas'unfortunately' can be base-generated in sentence-initial position. Alternatively, Frey (2006) claims in his discussion of similar German examples that all elements that may (optionally) precede the subject can be moved into the sentence-initial position simply in order to satisfy the V2-requirement; topicalization of such elements is thus predicted not to have any effect on the information structure of the clause. Frey claims that this is confirmed by the fact that dative objects can be topicalized without any special effect in passive and unaccusative constructions; the topicalized phrase in the primed examples in (320) should be able to receive a neutral interpretation in terms of information structure and should not require any special phonetic prominence.

    320
    a. dat Peter/hem/'m gisteren een gratis maaltijd werd aangeboden.
      that  Peter/him/'m  yesterday  a free meal  was  prt-offered
      'that a free meal was offered to Peter/him yesterday.'
    a'. Peter/Hem/*'m werd gisteren een gratis maaltijd aangeboden.
      Peter/him/'m  was  yesterday  a free meal  prt.-offered
      'A free meal was offered to Peter/him yesterday.'
    b. dat Peter/hem/'m die voorstelling goed bevallen is.
      that  Peter/him/him  that show  well  pleased  is
      'that that show has pleased Peter/him a lot.'
    b'. Peter/Hem/*'m is die voorstelling goed bevallen.
      Peter/him/him  is that show  well  pleased
      'That show has pleased Peter/him a lot.'

    Although it does seem to be the case that the topicalized dative objects do not need any special emphasis, the primed examples nevertheless show that they differ from sentence-initial subjects in that they are not able to take the form of the weak pronoun 'm'him' (see also Bouma 2008:26); this may be incompatible with Frey's claim. Because the judgments on the contrast between the two examples in (317) are subtle anyway, we have to leave it to future research to further investigate whether formal movement in the sense of Frey really exists; it might be interesting, for example, to see whether Frey's claim that the presumed cases of formal movement do not involve any form of prosodic prominence can be confirmed by an in-depth phonetic investigation.

    [+]  V.  Pied piping and stranding

    Subsection IV has shown that topicalization is often semantically motivated. If we restrict ourselves to those forms of topicalization related to information-structure, we can say that topicalization may be used to create a focus-background, a topic-comment, or a topic-focus structure. As in the case of wh-question, we would expect that it would suffice to topicalize the focus/topic element, and this raises the question as to whether topicalization may trigger pied piping. It seems that we have to answer this question in the affirmative. Consider the question answer-pair in (321). We have seen that questions like (321a) involve pied piping: while movement of the interrogative pronoun wiens'whose' would in principle suffice to form the desired operator-variable chain, syntactic restrictions force movement of the complete noun phrase wiens boek'whose book'. Since the focus in the answer in (321b) corresponds to the wh-pronoun wiens we can immediately conclude that topicalization of a focus may trigger pied piping.

    321
    a. [Wiens boek]i heb je ti gekocht?
      whose book  have  you  bought
      'Whose book have you bought?'
    b. [Jans boek]i heb ik ti gekocht
      Jan's book  have  I  bought
      'Jan's book, I have bought.'

    The same can be illustrated by means of the question-answer pair in (322): while wh-movement of the nominal complement of the preposition op suffices in principle to create the desired operator-variable chain in (322a), the restrictions on preposition stranding in Dutch force movement of the complete PP op wie'for who'. As the focus in answer (322b) corresponds to the wh-phrase wie, this example again shows that topicalization of a focused phrase may trigger pied piping.

    322
    a. [Op wie]i wacht je ti?
      for who  wait  you
      'Who are you waiting for?'
    b. [Op Jan]i wacht ik ti.
      for Jan  wait  I
      'Jan, I am waiting for.'

    That pied piping depends on independent syntactic constraints can be seen once again by considering the question-answer pair in (323); the question in (323a) shows that stranding of prepositions is possible if the complement is an R-word like waar. The fact that the focused constituent de post'the post' must pied-pipe the preposition op shows that pied piping cannot be semantically motivated.

    323
    a. Waari wacht je [ti op]?
      where  wait  you
      'What are you waiting for?'
    b. [Op de post]i wacht ik ti.
      for the post  wait  I
      'The mail, I am waiting for.'

    The examples in (324) illustrate that topicalization of contrastively accented phrases may also trigger pied piping.

    324
    a. [[Jans boek]i zal ik ti kopen] (maar Els' boek niet).
      Jan's book  will  I  buy   but  Els' book  not
      'Jan's book I will buy, but Els' book I won't.'
    b. [[Op Jan]i zal ik ti wachten] (maar op Els niet).
      for Jan  will  I  wait   but  for Els not
      'Jan I will wait for, but I won't for Els.'
    c. Mijn moeder is 115 jaar, maar [zo oudi [word ik echt niet ti]].
      my mother is 115 year but  that old  become  I  really  not
      'My mother is 115 years old, but that old I really won't become.'

    Although it is known that stranding and pied piping are relevant notions in the domain of topicalization (cf. Van Riemsdijk 1978), the literature normally focuses on wh-movement and relativization, because these allow us to investigate these phenomena without having to appeal to discourse; to our knowledge there is no detailed investigation of pied piping in topicalization contexts that takes information-structural considerations into account. We tend to think that there are not a great many differences vis-à vis question formation and relativization but this should be confirmed by a more careful investigation than we are able to conduct here.

    [+]  VI.  Topicalization of verbal projections

    Topicalization differs from question formation and relativization in that it allows wh-movement of certain types of clauses and other verbal projections. This difference is due to the fact that question formation and relativization normally affect some pronoun or other pro-form while topicalization affects full focus/topic phrases. This means that in the case of question formation and relativization the only way to get a clause in clause-initial position would be by pied piping, but this is prohibited across-the-board: wh-movement of a (part of a) clausal constituent is not able to pied-pipe the containing clause.

    325
    a. Wat zei hij? Dat hij Peter niet gelooft.
    question formation
      what said he that he Peter not believes
      'What did he say? That he doesn't believe Peter.'
    b. De opmerking [die me hindert] is dat hij Peter niet gelooft.
    relativization
      the remark   that  me  bothers  is  that  he Peter not  believes
      'The remark that bothers me is that he doesn't believe Peter.'
    c. [Focus/Topic Dat hij Peter niet gelooft] hindert me.
    topicalization
      that  he  Peter  not  believes  annoys  me
      'That he doesn't believe Peter annoys me.'

    It is often claimed that constructions with a topicalized verbal projection (and argument clauses in particular) should be analyzed as left-dislocation constructions with a deleted (phonetically empty) resumptive pronoun; see Koster (1978) and Odijk (1998) for, respectively, a fairly early and a fairly recent discussion of this issue. This subsection will also consider whether the topicalization constructions discussed in this subsection have a corresponding left-dislocation construction in order to see whether this claim can be maintained, subsection A starts by discussing topicalization of (finite and infinitival) argument clauses, which is followed in Subsection B by a discussion of topicalization of adverbial clauses, subsection C addresses VP-topicalization, that is, topicalization of verbal complements of non-main verbs, subsection D summarizes some of the main finding and draws some general conclusions.

    [+]  A.  Argument clauses

    Chapter 5 has shown that there are various syntactic types of argument clauses. The main division is that between finite and non-finite clauses, and the latter can be subdivided further into om + te-infinitival, te-infinitival and bare infinitival clauses. We discuss these (sub)types in the following subsections.

    [+]  1.  Finite clauses

    The singly-primed examples in (326) show that finite subject and direct object clauses can readily be topicalized, and the doubly-primed examples show that such clauses may also appear in left-dislocated position, followed by the resumptive pronoun dat'that' in clause-initial position. These examples thus seem to support the hypothesis that topicalization constructions are left-dislocation constructions with a phonetically empty resumptive element. An additional argument in favor of this hypothesis is that the anticipatory pronoun het'it' in the primeless examples cannot be used in the singly-primed topicalization constructions. This would follow immediately if these constructions indeed contained a phonetically empty resumptive subject/object pronoun: the anticipatory pronoun het could then simply not appear for the same reason that it cannot appear in the doubly-primed examples—it cannot be assigned an independent syntactic function.

    326
    a. Het hindert me [dat hij Peter niet gelooft].
    subject
      it  annoys  me   that  he  Peter  not  believes
      'It annoys me that he doesn't believe Peter.'
    a'. [Dat hij Peter niet gelooft] hindert (*het) me.
    a''. [Dat hij Peter niet gelooft], dat hindert me.
    b. Hij betwistte (het) [dat hij te laat was].
    direct object
      he  disputed   it  that he too late was
      'He disputed (it) that he was late.'
    b'. [Dat hij te laat was] betwistte hij (*het).
    b''. [Dat hij te laat was], dat betwistte hij.

    Things are different in the case of verbs selecting a prepositional object. Even verbs that do not require an anticipatory pronominal PP to be present do not allow topicalization of the clause. Left dislocation, on the other hand, is fully acceptable.

    327
    a. Jan twijfelde (erover) [of hij het boek zou kopen].
    PP-complement
      Jan doubted  about.it   if  he  the book would  buy
      'Jan doubted (about it) whether he would buy the book.'
    b. * [Of hij het boek zou kopen] twijfelde Jan (erover).
    c. [Of hij het boek zou kopen], daar twijfelde Jan over.

    Example (328) shows that omission of the pronominal part of the discontinuous PP daar ... over in example (327b) also gives rise to an unacceptable result for most speakers (although some speakers seem to accept it at a pinch). The impossibility of omitting daar poses a problem for the hypothesis that the topicalization constructions above are left-dislocation constructions with a phonetically empty resumptive element, and requires the introduction of some auxiliary hypothesis to regulate the deletion of resumptive pronouns.

    328
    % [Of hij het boek zou kopen] twijfelde Jan over.
      whether  he  the book  would  buy  doubted  Jan  about

          Topicalization of finite argument clauses seems to be quite unrestricted. One exceptional case, taken from Odijk (1998), is given in (329). Although Odijk's judgment on (329b) is correct, it should be noted that example (329a) is an innovation in the language, as is clear from the fact that this use is not included in the latest (14th) edition of the Van Dale dictionary. Furthermore, many of our informants give an affirmative answer to the question as to whether (329a) should be considered an abbreviation of the more regular expression Jan belde om te zeggen dat hij ziek was; compare the translation of (329a) which was taken from Odijk's article. We therefore provisionally conclude that topicalization of finite argument clauses is always possible.

    329
    a. Hij belde [dat hij ziek was].
      he  called   that  he  ill  was
      'He called to say that he was ill.'
    b. * [Dat hij ziek was] belde hij.
    [+]  2.  Om + te- and te-infinitival clauses

    It less clear to what extent om + te- and te-infinitival clauses can be preposed. Koster (1987:129) claims for te-infinitivals that this is "often difficult" and subsequently assigns them an asterisk. Zwart (1993:263) presents a case of topicalization of a te-infinitive as fully acceptable, while Odijk (1995:12) claims that such cases "are always somewhat marginal"; in later work, Zwart (2011:112) assigns two question marks to both topicalized om + te- and te-infinitival clauses. We agree that topicalization of om + te- and te-infinitivals normally gives rise to a marked result, but we also feel that topicalization leads to a markedly worse result in the case of om + te-infinitivals; this is what we try to express by means of our diacritics on the two singly-primed examples in (330). The left-dislocation constructions in the doubly-primed examples seem fully acceptable (although speakers again seem to vary somewhat in their judgments). Observe that the contrast between the singly- and doubly-primed examples is unexpected on the hypothesis that topicalization constructions are left-dislocation constructions with a deleted (phonetically empty) resumptive pronoun.

    330
    a. Jani weigert [(om) PROi weg te gaan].
    om + te-infinitival
      Jan refuses  comp  away  to go
      'Jan refuses to leave.'
    a'. *? [(om) PROi weg te gaan] weigert Jani.
    a''. [(om) PROi weg te gaan], dat weigert Jani.
    b. Jani probeert al tijden [PROi de auto te repareren].
    te-infinitival
      Jan tries  already  ages  the car  to repair
      'Jan has been trying for ages to repair the car.'
    b'. ? [PROi de auto te reparen] probeert Jani al tijden.
    b''. [PROi de auto te reparen], dat probeert Jani al tijden.

    The examples in (330) involve direct object clauses. In (331), we give similar examples with a verb selecting a prepositional object.

    331
    a. Jani klaagde (erover) [PROi niet te kunnen komen].
      Jan complained  about.it  not  to be.able  come
      'Jan complained about not being able to come.'
    b. * [PROi niet te kunnen komen] klaagde Jani (erover).
    c. [PROi niet te kunnen komen] daar klaagde Jani over.

    Example (332) shows that omission of the pronominal part of the discontinuous PP daar ... over in the left-dislocation construction (331b) gives rise to a quite marked result for most speakers. This is again problematic for the claim that topicalization constructions are left-dislocation constructions with a phonetically empty resumptive element.

    332
    % [Niet te kunnen komen] klaagde Jan over.
      not  to be.able  come  complained  Jan about

    The discussion above is typical for opaque and semi-transparent infinitival clauses which may occur in extraposed position; cf. Section 5.2.2.3. There are a number of additional, complicating issues for transparent te-infinitivals, that is, infinitivals that exhibit verb clustering and the infinitivus-pro-participio effect. However, because topicalization of te-infinitival normally gives rise to a marked result and we can discuss the same issues by means of fully acceptable cases in which a bare infinitival clause is topicalized, we will address these issues in the next subsection.

    [+]  3.  Bare infinitivals

    At first sight, topicalization of bare VPs seems easily possible, but closer scrutiny soon reveals that there are at least two complicating issues. The first issue is related to the fact that om general bare infinitival clauses are obligatorily split as a result of verb clustering. This phenomenon is illustrated in (333a) for the bare infinitival complement of the modal main verb willen'to want'. When we now consider the corresponding examples in (333b&c) notice to our surprise that clause splitting is optional (although we should note that dat hij graag die problemen oplossen wil is possible as a marked order). The primed examples are added to show that both topicalization constructions alternate with a left-dislocation counterpart, as predicted by the hypothesis that the topicalization constructions are left-dislocation constructions with a deleted (phonetically empty) resumptive pronoun.

    333
    a. dat hij <die problemen> graag wil <*die problemen> oplossen.
      that  he  those problems  gladly  wants  prt.-solve
      'that he dearly to solve those problems.'
    b. Die problemen oplossen wil hij graag.
    b'. Die problemen oplossen, dat wil hij graag.
    c. Oplossen wil hij die problemen graag.
    c'. Oplossen, dat wil hij die problemen graag.

    A second problematical factor is related to the Infinitivus-Pro-Participio (IPP) effect. Example (334a) first shows that in perfect-tense constructions the matrix verb does not appear as a past participle but as an infinitive. The singly-primed examples in (334) show that the IPP-effect disappears in the topicalization constructions, regardless of whether the infinitival clause is split or not. The primed examples show the same for the corresponding left-dislocation constructions.

    334
    a. Hij had die problemen graag willen/*gewild oplossen.
      he  had  those problems  gladly  want/wanted  prt.-solve
      'He had wanted to solve those problems very much.'
    b. Die problemen oplossen had hij graag gewild/*willen.
    b'. Die problemen oplossen, dat had hij graag gewild/*willen.
    c. Oplossen had hij die problemen graag gewild/*willen.
    c'. Oplossen, dat had hij die problemen graag gewild/*willen.

    The set of data in (333) and (334) thus shows that the core properties of constructions with transparent infinitives (clause splitting and IPP) disappear if the infinitival clause is topicalized. Although this has been known for a long time, there are still no theoretical accounts of it that meet with general acceptance. This is related to the current state of theories for these two phenomena. First, there are many competing theories on verb clustering that are more or less successful in describing the core data (see Section 7.5), but these are often quite different in nature and therefore also require quite different approaches to the (b)- and (c)-examples in (333). Second, there are only a few theories available for the IPP-effect, and most of these are highly controversial, so that we can at best conclude from the data in (334) that the IPP-effect only arises if the embedded main verb is physically located in the verbal cluster, a suggestion supported by examples such as (335), which show that the IPP-effect must be preserved if the full (non-finite part of the) verb cluster is topicalized.

    335
    Willen/*Gewild oplossen had hij die problemen graag.
      want/wanted  prt.-solve  had  he  those problems  gladly
    'He had dearly wanted to solve those problems very much.'

    We will return to the problem of clause splitting illustrated in examples (333b&c) in Subsection C on VP-topicalization, but have to leave the other questions and issues to future research.
          The remainder of this subsection is devoted to an issue regarding topicalization of bare infinitival argument clauses that is more specifically related to accusativus-cum-infinitivo constructions such as (336a), in which we have again italicized the full complement clause. Example (336b'') shows that it is impossible to topicalize the full bare infinitival clause: the subject die man must remain in the middle field of the matrix clause. The two remaining (b)-examples show that the direct object de boeken can but need not be part of the topicalized phrase. Observe that we added the negative adverb niet'not' to (336b''), as these topicalization constructions are natural only if the middle field contains some material next to the subject of the matrix verb.

    336
    a. dat hij die man de boeken niet zag stelen.
      that  he  that man  the books  not  saw  steal
      'that he didn't see the man steal the books.'
    b. Stelen zag hij die man de boeken niet.
    b'. De boeken stelen zag hij die man niet.
    b''. * Die man de boeken stelen zag hij niet.

    One way of accounting for contrast between (336b') and (336b'') might be to appeal to the fact that while the object of the infinitival clause can be assigned accusative case by the infinitival verb stelen'to steal', the subject of the infinitival clause must be assigned accusative case by the matrix verb zien'to see', as is clear from the fact that it can be replaced by the object pronoun hem'him'; cf. Section 5.2.3.3. It might be that topicalization as in (336b'') makes the latter, exceptional form of case assignment impossible; see Lasniks (1999) discussion of "raising to object" in English for a line of thinking that may indeed have this effect. A potential (but not insurmountable) problem for this suggestion is that it is sometimes claimed that the subject can be part of the topicalized clause if it is indefinite, as in (337b''); cf. Odijk (1998:204). We again added the negative adverb niet'not' to this example in order to make it more natural, but even then many speakers find examples like these highly questionable, for which reason we have assigned it a percentage sign.

    337
    a. dat hij iemand de boeken zag stelen.
      that  he  someone  the books  saw  steal
      'that he saw someone steal the books.'
    b. Stelen zag hij iemand de boeken.
    b'. De boeken stelen zag hij iemand.
    b''. % Iemand de boeken stelen zag hij niet.
    [+]  B.  Adverbial clauses

    From a syntactic point of view, topicalization of adverbial clauses seems quite unrestricted; we illustrate this in (338) for finite adverbial clauses of various kinds.

    338
    a. Voordat ik vertrek, bezoek ik mijn moeder.
    temporal
      before  I  leave visit  I  my mother
      'Before I leave, I will visit my mother.'
    b. Omdat/Doordat Jan ziek is, gaat het feest niet door.
    reason/cause
      because/because  Jan ill  is  continue  the party  not  prt.
      'Because Jan is ill, the party is cancelled.'
    c. Als je op deze knop drukt, gaat de computer aan.
    conditional
      if  one  on this button  presses  goes  the computer  on
      'If one presses this button, the computer starts up.'
    d. Ondanks dat hij ziek was, was hij aanwezig.
    concessive
      despite  that  he  ill  was  was he present
      'Despite his illness, he was present.'

    This does not mean, however, that anything goes. Topicalization of an adverbial result clause such as (339a'), for instance, is distinctly odd. We marked this example with a dollar sign in order to indicate that its markedness is probably of a non-syntactic nature, and simply reflects the general tendency to present eventualities in the order of their actual occurrence: cf. Jan stond op en kleedde zich aan'Jan got up and dressed' versus $Jan kleedde zich aan en stond op. Example (339b') presents another marginal case of topicalization that can potentially be accounted for in a similar way.

    339
    a. Jan ging naar buiten zodat hij meer licht had.
    result
      Jan went  to outside  so.that  he  more light  had
      'Jan went outside so that he would have more light.'
    a'. $ Zodat hij meer licht had, ging hij naar buiten.
    b. Je mag komen, mits je je gedraagt.
    conditional
      you  may  come  provided  you  refl  behave
      'You may come provided that you behave.'
    b'. $ Mits je je gedraagt, mag je komen.

    The examples in (340) show that infinitival adverbial clauses are like finite ones in that they normally can be topicalized easily. Note in passing that goals differ from results in that they can be topicalized, which may be due to the fact that a goal comes into existence before the action that aims at realizing it.

    340
    a. Alvorens te vertrekken, bezoek ik mijn moeder.
    temporal
      before  to leave  visit  I  my mother
      'Before leaving, I will visit my mother.'
    b. Om meer licht te krijgen, ging Jan naar buiten.
    goal
      comp  more light  to get  went Jan to outside
      'In order to get more light, Jan went outside.'

    For completeness' sake, we want to note that it is generally not easy to left-dislocate adverbial clauses; this is illustrated in (341a) for the temporal adverbial clause in (338a). Conditional clauses are a notable exception; this is illustrated in (341b) for the conditional clause in (338c).

    341
    a. * Voordat ik vertrek, dan bezoek ik mijn moeder.
    temporal
      before  I  leave then  visit  I  my mother
      'Before I leave, I will visit my mother.'
    b. Als je op deze knop drukt, dan gaat de computer aan.
    conditional
      if  one  on this button  presses  then  goes  the computer  on
      'If one presses this button, the computer starts up.'

    The unacceptability of examples such as (341a) suggests that the hypothesis formulated for argument clauses that topicalization constructions are actually left-dislocation constructions with a deleted (phonetically empty) resumptive pronoun cannot readily be applied to adverbial clauses.

    [+]  C.  Complements of non-main verbs (VP-topicalization)

    The previous subsections have shown that (finite and infinitival) clauses functioning as clausal constituents can normally be topicalized. This subsection shows that the same holds for verbal complements of non-main verbs. We will discuss the three cases in (342), that is, non-main verbs that take a complement headed by a past/passive participle, a te-infinitive and a bare infinitive. Because these cases all involve contrastive accent on the topicalized phrase and all receive a contrastive interpretation, we may safely assume that we are dealing with focus constructions.

    342
    a. Hij heeft nooit geschaakt.
    past/passive participle
      he  has  never  played.chess
      'He has never played chess.'
    a'. Geschaakt heeft hij nooit.
    b. Hij zit daar te schaken.
    te-infinitive
      he  sits  there  to play.chess
      'He is playing chess over there.'
    b'. ? Te schaken zit hij daar.
    c. Hij gaat morgen schaken.
    bare infinitive
      he  goes  tomorrow  play.chess
      'He is going to play chess tomorrow.'
    c'. Schaken gaat hij morgen.
    [+]  1.  Perfect tense and passive constructions

    Perfect tense constructions like (343a) easily allow topicalization of the perfect participle. The resulting construction in (343a') is potentially problematic as topicalization seems to affect a single word, while wh-movement normally affects phrases. The (b)-examples show, however, that it is also possible to topicalize verb phrases.

    343
    a. Hij heeft dat boek nog nooit gelezen.
      he  has  that book  yet  never  read
      'He has never read that book.'
    a'. Gelezen heeft hij dat boek nog nooit.
    b. Hij heeft nog nooit boeken gelezen.
      he  has  yet  never  books  read
      'He has never read aby books.'
    b'. [Boeken gelezen]heeft hij nog nooit.

    Den Besten & Webelhuth (1987) argue that the contrast between the two primed examples in (343) is only apparent and that they both involve topicalization of a verb phrase (VP); the difference in size of the topicalized VP is merely a side effect of some other phenomenon of Dutch, viz. scrambling. When we consider the two primeless examples in (343) we see that the direct objects occupy different locations: the definite object dat boek precedes the adverbial phrase nog nooit while the indefinite object boeken'books' follows it. Den Besten & Webelhuth argue that this is due to leftward movement (scrambling) of the definite object to some position external to the VP; the two primeless examples in (343) thus have the (simplified) structures given in the primeless examples in (344). If these are the input for VP-topicalization, we end up with the structures in the primed examples.

    344
    a. Hij heeft dat boeki nog nooit [VPti gelezen].
    scrambling of object
      he  has  that book  yet never  read
    a'. [VPtigelezen]j heeft hij dat boeki nog nooit tj.
    VP-topicalization
    b. Hij heeft nog nooit [VP boeken gelezen].
    no scrambling of object
      he  has  yet never  books  read
    b'. [VPboeken gelezen]j heeft hij nog nooit tj.
    VP-topicalization

    On this view the apparent movement of the participle is the result of movement of the remnant of the VP after scrambling, and Den Besten & Webelhuth therefore refer to this type of topicalization as remnant VP-topicalization. There are various empirical arguments in favor of an analysis of this kind. First, we predict that elements that are difficult to scramble normally cannot be stranded by VP-topicalization either. This holds, e.g., for the complementive AP ziek'ill' in the copular construction in (345); the examples in (345b&b') show that it must be taken along under VP-topicalization. For completeness' sake, we added (345b'') to show that the actual position of the complementive in the middle field does not affect the acceptability judgments.

    345
    a. Hij is <*ziek> gelukkig niet <ziek> geworden.
      he  is      ill  fortunately  not  become
      'Fortunately, he hasn't become ill.'
    b. [Ziek geworden]j is hij gelukkig niet tj.
    b'. * [ti Geworden]j is hij zieki gelukkig niet tj.
    b''. * [ti Geworden]j is hij gelukkig niet ziekitj.

    The examples in (346) show essentially the same for complementives like the AP paars'purple' and the PP in zijn spaarpot'in his money box' in resultative constructions (although it should be noted that these examples improve if the complementives are given emphatic accent). For completeness' sake, note that the structures in the primed examples are somewhat simplified, e.g., by not indicating the movement of the direct object; cf. (349) below.

    346
    a. Hij heeft het hek <*paars> gisteren <paars> geverfd.
      he  has  the gate    purple  yesterday  painted
      'Yesterday he painted the gate purple.'
    a'. [Paars geverfd]j heeft hij het hek tj.
    a''. * [tiGeverfd] heeft hij het hek paarsitj.
    b. Hij heeft het geld <*in zijn spaarpot> gisteren <in zijn spaarpot> gestopt.
      he  has  the money    in his money.box  yesterday  put
      'Yesterday he put the money in his money box.'
    b'. [In zijn spaarpot gestopt]j heeft hij het geld tj.
    b''. *? [ti Gestopt]j heeft hij het geld [in zijn spaarpot]itj.

    Second, we expect that elements that normally scramble into some more leftward position in the middle field must be stranded by VP-topicalization. The examples in (347) show that this prediction is borne out for weak (phonetically reduced) pronouns like he t'it'.

    347
    a. Hij heeft <het> nog nooit <*het> gelezen.
      he  has     it   yet never  read
      'He has never read it uet.'
    b. [tiGelezen]jheeft hij heti nog nooit tj.
    b'. * [Het gelezen]jheeft hij nog nooit tj.

    Third, example (348a) shows that scrambling of the definite noun phrase de auto is optional (or, more precisely, depends on whether or not it introduces new information) and we therefore expect that it can optionally be stranded (again depending on its information-structural status). The (b)-examples show that this is again borne out.

    348
    a. Ik heb <de auto> gisteren <de auto> gerepareerd.
      I  have     the car  yesterday  repaired
      'Yesterday I repaired the car.'
    b. [ti Gerepareerd]j heb ik de autoi wel tj (maar nog niet gewassen).
      repaired  have  I  the car  aff   but  yet  not  washed
      'I have repaired the car (but I haven't washed it yet).'
    b'. [De auto gerepareerd]j heb ik wel tj (maar nog niet [het hek geverfd]).
      the car  repaired  have  I  aff   but  yet  not   the gate  painted
      'I have repaired the car (but I haven't painted the gate yet).'

    Finally, since scrambling need not affect all VP-internal elements equally, we expect that VP-topicalization may strand some of these elements while taking along some of the others. This was in fact already illustrated in (346), in which VP-topicalization takes along the complementive while stranding the direct object, which is base generated as the logical subject of the complementive phrase in a so-called small clause configuration; cf. Section 2.2. The remnant VP-topicalization approach would thus assign to these examples the structural representations in (349).

    349
    a. [[small clausetiPaars] geverfd]j heeft hij het hekitj.
    b. [[small clauseti In zijn spaarpot] gestopt]j heeft hij het gelditj.

    The same can be shown by means of the double object construction in (350b): while the indirect object is stranded in the middle field of the clause, the direct object is still part of the topicalized VP.

    350
    a. De gemeente heeft de koning nog niet eerder een concert aangeboden.
      the municipality  has  the king  yet  not  before  a concert  prt.-offered
      'The municipality hasn't yet offered the King a concert before.'
    b. [ti Een concert aangeboden]j heeft de gemeente de koningi nog niet eerder tj.

    Example (351) provides one more example with the verb beveiligen'to safeguard' that selects a direct and a prepositional object: in the primed example the PP-object is taken along under VP-topicalization, while the direct object is stranded.

    351
    a. Hij heeft zijn huis nog niet tegen inbraak beveiligd.
      he  has  his house  yet  not  against burglary  safeguarded
      'He hasn't safeguarded his house against burglary yet.'
    b. [ti Tegen inbraak beveiligd]j heeft hij zijn huisi nog niet tj.

    At first sight, it seems that extraposed complement clauses can optionally be stranded under VP-topicalization, which would be surprising given that such clauses normally do not scramble. This impression may be deceptive, however, because postverbal complement clauses can also be right-dislocated, as is clear (352a). They can be introduced by the anticipatory pronoun het. That pied piped clauses are extraposed while stranded clauses are left-dislocated is suggested by the fact that the former do not allow the addition of the anticipatory pronoun while the latter actually prefer it to be present.

    352
    a. Jan wil (het) niet beloven [dat hij komt].
      Jan wants  it  not  promise   that  he  comes
      'Jan doesn't want to promise (it) that he will come.'
    b. [Beloven [dat hij komt]] wil hij (*het) niet.
      promise   that  he  comes  wants  he      it  not
    b'. [Beloven] wil hij ?(het) niet [dat hij komt].
      promise  wants  he     it not  that  he  comes

          The discussion above has shown that the remnant VP-topicalization approach is quite successful in accounting for a number of core properties of VP-topicalization. There are, however, also a number of potential problems. We restrict our discussion here to one problem that can be illustrated on the basis of Standard Dutch, and refer the reader to Haider (1990) for a number of potential problems more specifically related to German. The problem in question, which was signaled by Den Besten & Webelhuth (1990), concerns the position of stranded prepositions. Section P5.3 has shown that stranded prepositions must be adjacent to the verb(s) in clause-final position. This suggests that they occupy a VP-internal position and we therefore expect that they must be taken along under VP-topicalization, but this is not borne out. The (b)-examples in (353) illustrate this by means of the discontinuous pronominal PP er ..op'on it'.

    353
    a. Hij had er niet op gerekend.
      he  had  there  not  on  counted
      'He hadn't counted on it.'
    b. Gerekend had hij er niet op.
    b'. * Op gerekend had hij er niet.

    It must be noted, however, that this problem only occurs on the traditional assumption that PP-complements are base-generated as complements of verbs, but that it has been argued on independent grounds that PP-complements are actually base-generated external to the lexical projection of the verb (cf. Barbiers 1995:ch.4), or perhaps even created in the course of the derivation (cf. Kayne 2004). If we adopt one of these proposals, the pattern in (353) is in fact expected: see Broekhuis (2008:115ff.) and references cited there.
          The primed examples in (354) show that passive constructions also allow topicalization of the participle. Again we may assume that we are dealing with topicalization of VPs in both cases, although the VP-internal traces are now coindexed with the noun phrase that has been promoted to subject.

    354
    a. Dat boek wordt niet meer gelezen.
      that book  is  no longer  read
      'That book isn't read any more.'
    a'. [tiGelezen]j wordt dat boeki niet meer tj.
    b. Zijn huis is nog niet tegen inbraak beveiligd.
      his house  is  yet  not  against burglary safeguarded
      'His house is not yet burglarproof.'
    b'. [ti Tegen inbraak beveiligd]j is zijn huisi nog niet tj.

    It seems that the subject of a passive construction can sometimes marginally remain VP-internal if it is indefinite, but then the regular subject position is normally filled by the expletive er if the middle field does not contain any presuppositional material.

    355
    a. Er worden bijna geen boeken meer gelezen.
      there  are  almost  no books  any.more  read
      'Books are hardly read any more.'
    b. [Boeken gelezen] worden *(?er) bijna niet meer.
      books  read  are  there  almost  not  any.more

    Haider (1990) claims for German that indefinite subjects of active monadic verbs can also be taken along by VP-topicalization (which would be in line with the current view that such subjects are base-generated in a VP-internal position). This gives rise to a rather marginal result in Dutch, as is illustrated in the (a)-examples for the intransitive verb spelen'to play' and in the (b)-examples for the unaccusative verb sterven'to die'.

    356
    a. Er hebben hier nog nooit kinderen gespeeld.
    intransitive
      there  have  here  yet  never  children  played
      'Children have never played here.'
    a'. * [Kinderen gespeeld]i hebben (er) hier nog nooit ti.
    b. Er zijn daardoor nog nooit patiënten gestorven.
    unaccusative
      there  are  by.that  yet  never  patients  died
      'Patients have never died because of that so far.'
    b'. * [Patiënten gestorven]i zijn (er) daardoor nog nooit ti.

    Haider also claims that indefinite subjects of dyadic verbs can be taken along under VP-topicalization provided that the object is stranded. The primed examples in (357) show that this is impossible in Dutch if the verb is transitive, while it gives rise to a marginal result if it is unaccusative (that is, a nom-datverb).

    357
    a. Er heeft nog nooit een buitenlander die derby gewonnen.
    transitive
      there  has  yet  never  a foreigner  that derby  won
      'A foreigner has never won that derby so far.'
    a'. * [Een buitenlander ti gewonnen]j heeft [die (derby)]i nog nooit tj.
    b. Er is hem nog nooit een ongeluk overkomen.
    nom-dat verb
      there  is him  yet  never  an accident  happened
      'He has never had an accident so far.'
    b'. ?? [ti Een ongeluk overkomen]j is hemi nog nooit tj.

    The acceptability contrast between the two primed examples in (357) can probably be attributed to the fact that Standard Dutch does not allow the object of a transitive verb to scramble across the subject, while is quite normal for the object to precede the subject in clauses headed by nom-dat verbs; cf. Section 2.1.3. This contrast can therefore be taken as support for the remnant VP-topicalization approach.

    [+]  2.  Te-infinitives

    Subsection A has shown that topicalization of te-infinitival clauses is normally at least somewhat marked. The same seems to hold for te-infinitival complements of the semi-aspectual verbs like zitten'to sit'. As in perfect and passive constructions, the direct object of the main verb may be taken along with VP-topicalization or be stranded in the middle field, depending on whether it expresses "new" or presupposed information. Judgments on the primed examples seem to vary from speaker to speaker and range from marked to ungrammatical; the examples seem to improve if some element in the middle field of the clause can be assigned emphatic accent: cf. ?Te lezen zit hij dat boek altijd hier.

    358
    a. Hij zit hier altijd boeken te lezen.
      he  sits  here  always  books  to read
      'He is always reading books here.'
    a'. ? [Boeken te lezen] zit hij hier altijd.
    b. Hij zit dat boek altijd hier te lezen.
      he  sits  that book  always  here  to read
      'He is always reading that book here.'
    b'. ? [ti Te lezen] zit hij dat boeki altijd hier.

    It seems that VP-topicalization of verbal projections headed by a te-infinitive exhibits more or less the same properties as topicalization of verbal projections headed by a past/passive participle, but we will not illustrate this here because all examples are minimally perceived as marked.

    [+]  3.  Bare infinitives

    Topicalization of bare infinitival complements of non-main verbs like the aspectual verbs inchoative gaan'to go', komen'to come', and blijven'to stay' is easily possible. As in perfect and passive constructions, the direct object of the main verb may be taken along with VP-topicalization or be stranded in the middle field of the clause, depending on whether it expresses new or presupposed information.

    359
    a. Hij gaat vandaag bloemen plukken.
      he  goes  today  flowers  pick
      'He is going to pick flowers today.'
    a'. [Bloemen plukken]i gaat hij vandaag ti.
    b. Hij gaat de bloemen vandaag plukken.
      he  goes  the flowers  today  pick
      'He is going to pick the flowers today.'
    b'. [tiPlukken]j gaat hij de bloemeni vandaag tj.

    It seems that the remnant VP-topicalization approach is also descriptively adequate for cases of this type. The examples in (360) first show that elements such as the complementives paars and in zijn spaarpot, which are normally not scrambled, are taken along by the preposed VP. The doubly-primed examples improve a little bit if the complementive is assigned emphatic stress.

    360
    a. Hij gaat het hek paars verven.
      he  goes  the gate  purple  paint
      'He is going to paint the gate purple.'
    a'. paars verven gaat hij het hek.
    a''. * Verven gaat hij het hek paars.
    b. Hij gaat het geld in zijn spaarpot stoppen.
      he  goes  the money  in his money.box  put
      'He is going to put the money in his money box.'
    b'. In zijn spaarpot stoppen gaat hij het geld.
    b''. * Stoppen gaat hij het geld in zijn spaarpot.

    Because the direct objects originate as logical subjects of the complementives, the singly-primed examples also show that VP-topicalization may take along some VP-internal element while stranding other VP-internal elements (here: the direct object) in the middle field of the clause. This is shown in the simplified structures of these examples in (361).

    361
    a. [[small clausetiPaars] verven]j gaat hij het hekitj.
    b. [[small clauseti In zijn spaarpot] stoppen]j gaat hij het gelditj.

    A similar apparent VP-split can be shown by means of the double object construction in (362b): while the indirect object is stranded in the middle field of the clause, the direct object is still part of the topicalized VP.

    362
    a. De gemeente gaat de koning volgende week een concert aanbieden.
      the municipality  goes  the King  next week  a concert  prt.-offer
      'The municipality is going to offer the King a concert next week.'
    b. [ti Een concert aanbieden]j gaat de gemeente de koningi volgende week tj.

    Example (363) provides one more example with the verb beveiligen'to safeguard' that selects a direct and a prepositional object: in the primed example the PP-object is taken along under VP-topicalization, while the direct object is stranded.

    363
    a. Hij gaat zijn huis snel tegen inbraak beveiligen.
      he  goes  his house  soon  against burglary  safeguard
      'He is going to safeguard his house against burglary soon.'
    b. [ti Tegen inbraak beveiligen]j gaat hij zijn huisi snel tj.

    The fact that stranded prepositions cannot be part of the preposed VP is again potentially problematic for the remnant VP-topicalization approach, but we have already mentioned that this is in fact expected under some more recent hypotheses concerning the nature of PP-complements.

    364
    a. Hij gaat er niet op wachten.
      he  goes  there  not  for  wait
      'He is not going to wait for it.'
    b. Wachten gaat hij er niet op.
    b'. * Op wachten gaat hij er niet.

    The examples in (365) show that indefinite subjects can only marginally be part of topicalized VPs if the main verb is monadic, and the examples in (366) show that in the case of dyadic verbs we find again a contrast in this respect between transitive and unaccusative (nom-dat) verbs.

    365
    a. Er komen volgende week kinderen spelen.
    intransitive
      there  come  next week  children play
      'Children are going to play here next week.'
    a'. * [Kinderen spelen]i komen (er) volgende week ti.
    b. Er gaan daardoor patiënten sterven.
    unaccusative
      there  go  by.that  patients  die
      'Patients are going to die because of that.'
    b'. * [Patiënten sterven]i gaan (er) daardoor ti.
    366
    a. Er gaat nooit een buitenlander die derby winnen.
    transitive
      there  goes  never  a foreigner  that derby  win
      'A foreigner is never going to win that derby.'
    a'. * [Een buitenlander ti winnen]j gaat [die (derby)]i nooit tj.
    b. Er gaat hem iets naars overkomen.
    nom-dat verb
      there  goes  him  something nasty  happen
      'Something nasty is going to happen to him.'
    b'. ?? [ti Iets naars overkomen]j gaat hemi niet tj.

    The examples above make it clear that VP-topicalization of bare infinitives exhibits more or less the same behavior as VP-topicalization of past/passive participles.

    [+]  4.  A special case

    The examples in (367) provide a special case of VP-topicalization: these examples show that VP-topicalization can also occur if there is no non-main verb, but that in the resulting structure the verb-second position must be filled by the "dummy" verb doen'to do'. As the second position of the clause can only be occupied by finite verbs, tense and agreement cannot be expressed by the main verb but must be transferred to a finite form of doen. We refer the reader to Section 6.4.3 for more discussion of this "dummy" use of doen.

    367
    a. Hij verkoopt zijn postzegels beslist niet.
      he  sells  his stamps  definitely  not
      'He definitely won't sell his stamps.'
    a'. [ti Verkopen]j doet hij zijn postzegelsi beslist niet tj.
      sell  does  he  his stamps  definitely  not
    b. Hij verzamelt geen postzegels meer.
      he  collects  no stamps  any.more
      'He doesn't collect stamps any more.'
    b'. [Postzegels verzamelen]i doet hij niet meer ti.
      stamps  collect  does  he  not  any.more
    [+]  5.  VP-topicalization and left dislocation

    This subsection has provided a discussion of VP-topicalization based on Den Besten & Webelhuth's (1987) remnant VP-topicalization approach, according to which VP-topicalization can be preceded by scrambling of VP-internal material. The attractive appeal of this approach is that it immediately accounts for the fact that the elements stranded in the middle field of the clause can be semantically licensed by the verb heading the VP in clause-initial position, as these stranded elements are base-generated within this VP. The fact that the noun phrase dat boek in (368a) is interpreted as the theme argument of lezen'to read' is simply due to the fact that this thematic role is assigned to the position occupied by its trace ti, that is, the position originally held by this noun phrase. The two examples in (368) therefore do not differ in any crucial way when it comes to the assignment of thematic roles.

    368
    a. [VPti Gelezen]j heeft hij dat boeki nog nooit tj.
      read  has  he  that book  yet  never
      'He has never read that book yet.'
    b. [VP Boeken gelezen]i heeft hij nog nooit ti.
      books  read  has  he  yet  never
      'He has never read books yet.'

    All acceptable VP-topicalization constructions discussed in the previous subsections alternate with left-dislocation constructions; the addition of the resumptive pronoun dat'that' never affects the acceptability judgments given in the previous subsection for the topicalization construction. This is illustrated in (369) for the examples in (368).

    369
    a. [VPti Gelezen], dat heeft hij dat boeki nog nooit tj.
      read   that  has  he  that book  yet never
      'He has never read that book yet.'
    b. [Boeken gelezen], dat heeft hij nog nooit.
      books  read  that  has  he  yet never
      'He has never read books yet.'

    Such left-dislocation constructions potentially undermine the argument in favor of the remnant VP-topicalization approach based on the assignment of thematic roles because they may force us to introduce some special mechanism to account for the fact that the "stranded" elements are interpreted as part of the clause-external, left-dislocated VP. So, if we introduce a special mechanism to account for the fact that the noun phrase dat boek'that book' in (369a) is interpreted as the direct object of the participle gelezen, we do not have to appeal to scrambling in order to account for the fact that the same holds for example (368a). It should be noted, however, that there are also proposals according to which left-dislocation is simply a special case of topicalization; in such analyses, which will be discussed in Section 14.2, sub VII and Section 14.2, sub VIII, the argument based on the assignment of thematic roles can probably be maintained in full force.
          Moreover, much is still not well-understood. Müller (1998:221), for instance, approvingly cites unpublished work by Truckenbrodt that shows that German behaves more in accordance with what is predicted by Den Besten & Webelhuth's (1987) proposal in allowing constructions like (369b), in which the left dislocated phrase is a "full" VP, but prohibiting constructions like (369a), in which the left-dislocated phrase is a remnant VP. In fact, some speakers report the same for cases in which a bare infinitival complement is topicalized/left-dislocated. The examples in (370a) first show that topicalization is fully acceptable to all speakers of Dutch, whereas the corresponding left-dislocation construction (370b) seems degraded. Note further the left dislocation becomes acceptable to all speakers if we insert the "dummy" verb doen; the topicalization construction, on the other hand, is not compatible with doen in Standard Dutch.

    370
    a. Lezen gaat hij die boeken niet (*doen).
      read goes  he  those books  not     do
      'He isn't going to read those books.'
    b. Lezen, dat gaat hij die boeken niet ??(doen).
      read that  goes  he  those books  not     do
      'He isn't going to read those books.'

    Second, the examples in (371) show that most speakers consider left dislocation at least marginally acceptable if the preposed VP contains the direct object (see Odijk 1995), although they may still prefer the addition of the "dummy" verb doen. Adding doen to the topicalization construction is impossible. For completeness' sake, note that (371a) can be construed as a topic drop construction, provided that the phrase boeken lezen is followed by an intonation break.

    371
    a. Boeken lezen gaat hij niet (*doen).
      books  read  goes  he  not    do
      'He isn't going to read books.'
    b. Boeken lezen, dat gaat hij niet ?(doen).
      books  read  that  goes  he  not   do
      'He isn't going to read books.'

    VP-topicalization constructions such as (372) with a finite form of "dummy" doen'to do' in second position do not seem to raise similar restrictions as constructions with a non-main verb: the two left-dislocation constructions in (372) seem to be equally acceptable for most speakers.

    372
    a. Verkopen doet hij zijn postzegels beslist niet.
      sell does  he  his stamps  definitely  not
      'He is definitely not selling his stamps .'
    a'. Verkopen, dat doet hij zijn postzegels beslist niet.
      sell that  does  he  his stamps  definitely  not
      'He is definitely not selling his stamps.'
    b. Postzegels verzamelen doet hij niet meer.
      stamps  collect  does  he  not  any.more
      'He doesn't collect stamps anymore.'
    b'. Postzegels verzamelen, dat doet hij niet meer.
      stamps  collect  that  does  he  not  any.more
      'He doesn't collect stamps anymore.'

    The discussion above shows that more research is needed in order to clarify the relation between VP-topicalization and left-dislocation, as well as its implications for the remnant VP-topicalization approach adopted in the discussion above. We refer the reader to Müller (1998) for additional arguments in favor of this approach, and Haider (1990) and Fanselow (2002) for arguments against it based on German.

    [+]  D.  Conclusion

    The previous subsections have shown that, broadly speaking, it is possible to topicalize finite clauses. Infinitival clauses exhibit a more varied behavior: om + te-infinitivals seem to resist topicalization, while topicalization of bare infinitivals is fully acceptable; judgments on topicalization of te-infinitivals seem to vary from speaker to speaker but these topicalizations are normally considered marginal, or marked at least. VP-topicalization, that is, topicalization of the complements of non-main verbs, is possible if these are headed by a past/passive participle or a bare infinitive, and again marked in the case of te-infinitives. To our knowledge there are no theoretical proposals that aim at accounting for this pattern.
          We also investigated whether the topicalization constructions discussed in the previous subsections alternate with left-dislocation constructions, since it is often claimed that the former are actually derived from the latter by deletion of the resumptive element. This claim is not fully supported by the empirical facts, which are summarized in Table 1.

    Table 1: Topicalization and left dislocation of clauses and other verbal projections
      topicalization left dislocation
    finite argument (SU, DO)
    argument (PP)     		✓
    *     		✓
    ✓
      adverbial
    adverbial (conditional)     		✓
    ✓     		*
    ✓
    infinitival om + te-infinitive * ✓
      te-infinitives ? ✓
      bare infinitives ✓ ✓
    VP-
    topicalization    		 past/passive participle ✓ ✓
      te-infinitival ? ?
      bare infinitival (full)
    bare infinitival (remnant)     		✓
    ✓     		✓
    ?
    [+]  VII.  Some differences between English and Dutch topicalization

    English topicalization and Dutch topicalization seem to differ in various non-trivial respects. First consider the English examples in (373). The (a)-examples show that while wh-movement of the interrogative object which book triggers subject-verb inversion (or do-support if there is no auxiliary verb) in main clauses, topicalization of the direct object this book does not. The (b)-examples show that while wh-phrases like which book cannot co-occur with a complementizer in embedded clauses, topicalized phrases can although the result is marginal for some speakers. Note that whereas the wh-phrase is normally assumed to precede the empty complementizer position, the topicalized phrase must follow the complementizer.

    373
    a. Which book should I read?
    a'. This book, you should read.
    b. I wonder [which book (*that) I should read].
    b'. % I believe [that this book you should read].

    Chomsky (1977) proposed that topicalization in English is like question formation in that it is derived by means of wh-movement (but see Baltin 1982 and Lasnik & Saito 1992 for alternative proposals). His analysis is given in a slightly adapted form as (373), where "Topic" refers to the topicalized phrase, which is associated to the following clause by being coindexed with a phonetically empty operator that is wh-moved into clause-initial position.

    374
    Topici [CP OPi C [TP .... ti ....]]

    If we assume that the topicalized phrase indicated by "Topic" is a sentence-external element, the structural representation in (374) is able to account for a number of characteristic properties of Dutch topicalization. First, the Dutch counterparts of the (a)-examples in (373) given in (375) show that Dutch topicalization behaves like question formation in that it obligatorily triggers subject-verb inversion.

    375
    a. Welk boeki moet ik ti lezen?
      which book  must  I  read
    b. Dit boeki [CP OPi moet je ti lezen].
      this book  must  you  read

    Second, topicalization exhibits the typical qualities of wh-movement: example (376) shows that it is not clause-bound but nevertheless island-sensitive in that it cannot be extracted from an embedded question or an adjunct clause.

    376
    a. Dit boeki [CP OPi denk ik [CP dat ik ti moet lezen]].
      this book  think  I  that  I  must read
    b. * Dit boeki [CP OPi vraag ik me af [CP of ik ti moet lezen]].
      this book  wonder  I  refl  prt.  if  I  must  read
    c. * Dit boeki [CP OPi huil ik [CP omdat ik ti moet lezen]].
      this book  cry  I  because  I  must  read

    Third, the examples in (377) show that Dutch topicalization differs from question formation in that it is categorically rejected in embedded clauses (contrary to what has been shown for English in the (b)-examples in (373)). This would follow immediately if we assume that the topicalization structure in (374) cannot be embedded: this is illustrated in (377b) for an embedded clause with the finite verb in clause-final position and in (377b') for an embedded clause with verb-second (which is an acceptable option in German).

    377
    a. Ik vraag me af [welk boek (of) ik ti moet lezen].
      I  wonder  refl  prt.  which book  comp  I  must  read
    b. * Ik denk [<(dat)> dit boeki (<dat>) je ti moet lezen].
      I  think      that  this book      that  you  must  read
    b'. * Ik denk [dit boeki moet je ti lezen].
      I  think this book  must  you  read

    The analysis in (374) treats topicalization in essentially the same way as the left-dislocation constructions in (378); the only difference is that topicalization involves a phonetically empty operator or, alternatively, derives it from examples like (378) by deletion of the phonetic content of the wh-moved element.

    378
    a. Dit boeki [CP dati moet je ti lezen]].
    cf. (375b)
      this book  that  must  you  read
      'This book you should read it.'
    b. Dit boeki [dati denk ik [dat ik ti moet lezen]].
    cf. (376a)
      this book   that  think  I  that  I  must read
      'This book, I think I should read it.'

    The strongest hypothesis would therefore be that left dislocation of the type in (378) and topicalization alternate freely. This hypothesis does not seem to be tenable, however, given that there are certain differences between the two constructions. The examples in (379), for instance, show that while topicalization of quantified expressions like iedereen'everyone' and niemand'no one' can easily be realized, they cannot occur in left-dislocation constructions. This requires the additional ad hoc stipulation that the empty operator and the overt pronoun differ in that only the latter prohibits a quantified antecedent; we refer the reader to Section 14.2 for more examples.

    379
    a. Iedereeni [OPi/*diei heb ik ti gezien (behalve Peter)].
      everyone   OP/them  have  I  seen   except Peter
      'I have seen everyone (except Peter).'
    b. Niemandi [OPi/*diei heb ik ti gezien (behalve Peter)].
      no.one   OP/them `have  I  seen   except Peter
      'I have seen no one (except Peter).'

    Example (380) further shows that the analysis in (374) requires that we assume that the wh-moved empty operator cannot strand a preposition. This is again ad hoc since examples like (380b) show that empty operators are normally able to do this; see Section A6.5, sub IVA for an extensive discussion of such constructions.

    380
    a. Deze schoeneni [daari/*OPi voetbalt Peter [ti mee]].
      these shoes  there/OP  plays.soccer  Peter  with
      'These shoes, Peter plays soccer with them.'
    b. Deze schoeneni zijn zeer geschikt [OPi om PRO [ti mee] te voetballen].
      these shoes  are  very suitable  comp  with  to  play.soccer
      'These shoes are very suitable suitable for playing soccer.'

    The contrasts in (379) and (380) casts serious doubts on the analysis in (374), especially because they follow without much ado under the alternative analysis, according to which topicalization involves wh-movement of the topicalized phrase itself. The unacceptability of examples (381a'&b'), for example, can simply be accounted for by the independently motivated assumption that demonstrative pronouns like die normally cannot refer to quantified phrases (if we ignore so-called bound variable readings); cf. *Niemandi was aanwezig, maar diei werd niet gemist. And the contrast between the two (c)-examples follows from the well-established fact that prepositions can only be stranded if the PP undergoes R-pronominalization.

    381
    a. [Iedereeni heb ik ti gezien (behalve Peter)].
    a'. * Iedereeni [diei heb ik ti gezien (behalve Peter)].
    b. [Niemandi heb ik gezien (behalve Peter)].
    b'. * Niemandi [diei heb ik gezien (behalve Peter)].
    c. * Deze schoeneni voetbalt Peter [ti mee]].
    c'. Deze schoeneni [daari voetbalt Peter [ti mee]].

    Of course, rejection of (374) as the proper structural representation of topicalization constructions also has its problems. For example, it would result in the loss of the elegant account for the fact that the Dutch topicalization is a root phenomenon, that is, that it applies in main clauses only. For this reason, it seems premature to make a definitive choice between the two options and we therefore leave this issue to future research. The reader is referred to Section 14.2 for a more extensive discussion of left dislocation.

    References:
    • Ankelien Schippers2012Variation and change in Germanic long-distance dependenciesUniversity of GroningenThesis
    • Baltin, Mark1982A landing site theory of movement rulesLinguistic Inquiry131-38
    • Baltin, Mark1982A landing site theory of movement rulesLinguistic Inquiry131-38
    • Barbiers, Sjef1995The syntax of interpretationThe Hague, Holland Academic GraphicsUniversity of Leiden/HILThesis
    • Barbiers, Sjef1995The syntax of interpretationThe Hague, Holland Academic GraphicsUniversity of Leiden/HILThesis
    • Barbiers, Sjef, Bennis, Hans, Vogelaer, Gunther de, Devos, Magda & Ham, Margreet van de2005Syntactic atlas of the Dutch dialectsAmsterdamAmsterdam University Press
    • Besten, Hans den & Webelhuth, Gert1987Remnant topicalization and the constituent structure of VP in the Germanic SOV languagesGLOW Newsletter1815-16
    • Besten, Hans den & Webelhuth, Gert1987Remnant topicalization and the constituent structure of VP in the Germanic SOV languagesGLOW Newsletter1815-16
    • Besten, Hans den & Webelhuth, Gert1987Remnant topicalization and the constituent structure of VP in the Germanic SOV languagesGLOW Newsletter1815-16
    • Besten, Hans den & Webelhuth, Gert1990StrandingGrewendorf, Günter & Sternefeld, Wolfgang (eds.)Scrambling and barriersAmsterdam/PhiladelphiaJohn Benjamins77-92
    • Booij, Geert1995The phonology of DutchOxfordClarendon Press
    • Bouma, Gerlof J2008Starting a sentence in Dutch. A corpus study of subject- and object-frontingUniverisity of GroningenThesis
    • Bouma, Gerlof J2008Starting a sentence in Dutch. A corpus study of subject- and object-frontingUniverisity of GroningenThesis
    • Bouma, Gerlof J2008Starting a sentence in Dutch. A corpus study of subject- and object-frontingUniverisity of GroningenThesis
    • Bouma, Gerlof J2008Starting a sentence in Dutch. A corpus study of subject- and object-frontingUniverisity of GroningenThesis
    • Broekhuis, Hans2008Derivations and evaluations: object shift in the Germanic languagesStudies in Generative GrammarBerlin/New YorkMouton de Gruyter
    • Chomsky, Noam1977On <i>wh</i>-movementCulicover, Peter W., Wasow, Thomas & Akmajian, Adrian (eds.)Formal syntaxNew YorkAcademic Press71-132
    • Chomsky, Noam1977On <i>wh</i>-movementCulicover, Peter W., Wasow, Thomas & Akmajian, Adrian (eds.)Formal syntaxNew YorkAcademic Press71-132
    • Dik, Simon C1978Functional grammarDordrechtForis Publications
    • Erteschik-Shir, Nomi2007Information Structure. The syntax-discourse interfaceOxford/New YorkOxford University Press
    • Fanselow, Gisbert2002Against remnant VP-movementAlexiadou, Artemis, Anagnostopoulou, Elena, Barbiers, Sjef & Gaertner, Hans-Martin (eds.)Dimensions of movement. From features to remnantsAmsterdam/PhiladelphiaJohn Benjamins Publishing Company91-125
    • Florijn, Arjen1992Beregeling van Nederlandse woordvolgordeUniversity of AmsterdamThesis
    • Haegeman, Liliane1995The syntax of negationCambridge studies in linguistics 75CambridgeCambridge University Press
    • Haider, Hubert1985V-second in GermanHaider, Hubert & Prinzhorn, Martin (eds.)Verb second phenomena in Germanic languagesDordrecht/RivertonForis Publications49-75
    • Haider, Hubert1990Topicalization and other puzzles of German syntaxGrewendorf, Günther & Sternefeld, Wolfgang (eds.)Scrambling and barriersAmsterdam/PhiladelphiaJohn Benjamins Publishing Company93-112
    • Haider, Hubert1990Topicalization and other puzzles of German syntaxGrewendorf, Günther & Sternefeld, Wolfgang (eds.)Scrambling and barriersAmsterdam/PhiladelphiaJohn Benjamins Publishing Company93-112
    • Haider, Hubert1990Topicalization and other puzzles of German syntaxGrewendorf, Günther & Sternefeld, Wolfgang (eds.)Scrambling and barriersAmsterdam/PhiladelphiaJohn Benjamins Publishing Company93-112
    • Haider, Hubert2010The syntax of GermanCambridgeCambridge University Press
    • Hoeksema, Jack1991Theoretische aspekten van partikelvooropplaatsingTabu2118-26
    • Hoekstra, Eric & Zwart, Jan-Wouter1994De structuur van CP. Functionele projecties voor topics en vraagwoorden in het NederlandsSpektator23191-212
    • Jansen, Frank1981Syntaktische konstrukties in gesproken taalAmsterdamHuis aan de drie grachten
    • Kayne, Richard S2004Prepositions as probesBelletti, Adriana (ed.)Structures and beyondNew YorkOxford University Press192-212
    • Kiss, Katalin É2002The syntax of HungarianCambridge syntax GuidesCambridge University Press
    • Kiss, Katalin É2002The syntax of HungarianCambridge syntax GuidesCambridge University Press
    • Kiss, Katalin É2002The syntax of HungarianCambridge syntax GuidesCambridge University Press
    • Kooij, Jan1978Fronting in DutchJansen, Frank (ed.)Studies on frontingLissePeter de Ridder
    • Koster, Jan1978Why subject sentences don't existKeyser, S. Jay (ed.)Recent transformational studies in European languages53-64
    • Koster, Jan1987Domains and dynasties. The radical autonomy of syntaxDordrecht/ProvidenceForis Publications
    • Lasnik, Howard1999Minimalist analysisOxfordBlackwell
    • Lasnik, Howard & Saito, Mamoru1992Move α. Conditions on its applications and outputLinguistic Inquiry MonographsCambridge, MA/London
    • Lasnik, Howard & Saito, Mamoru1992Move α. Conditions on its applications and outputLinguistic Inquiry MonographsCambridge, MA/London
    • Maienborn, Claudia2001On the positon and interpretation of locative modifiersNatural Language Semantics9191--240
    • Müller, Gereon1998Incomplete category fronting. A derivational approach to remnant movement in GermanDordrecht/Boston/LondonKluwer Academic Publishers
    • Müller, Gereon1998Incomplete category fronting. A derivational approach to remnant movement in GermanDordrecht/Boston/LondonKluwer Academic Publishers
    • Müller, Gereon1998Incomplete category fronting. A derivational approach to remnant movement in GermanDordrecht/Boston/LondonKluwer Academic Publishers
    • Odijk, Jan1998Topicalization of non-extraposed complements in DutchLinguistic Inquiry16191-222
    • Odijk, Jan1998Topicalization of non-extraposed complements in DutchLinguistic Inquiry16191-222
    • Odijk, Jan1998Topicalization of non-extraposed complements in DutchLinguistic Inquiry16191-222
    • Odijk, Jan2009Topicalization of sentences in Dutch is contrastive dislocation
    • Odijk, Jan2009Topicalization of sentences in Dutch is contrastive dislocation
    • Odijk, Jan2009Topicalization of sentences in Dutch is contrastive dislocation
    • Riemsdijk, Henk C. van1978A case study in syntactic markedness: the binding nature of prepositional phrasesPeter de Ridder Press
    • Rizzi, Luigi1997The fine structure of the left peripheryHaegeman, Liliane (ed.)Elements of grammar. Handbook of generative grammarDordrecht/Boston/LondenKluwer Academic Publishers
    • Rizzi, Luigi1997The fine structure of the left peripheryHaegeman, Liliane (ed.)Elements of grammar. Handbook of generative grammarDordrecht/Boston/LondenKluwer Academic Publishers
    • Salverda, Reinier1982On the problem of topicalization in DutchVooruit3244-255
    • Salverda, Reinier1982On the problem of topicalization in DutchVooruit3244-255
    • Salverda, Reinier1982On the problem of topicalization in DutchVooruit3244-255
    • Salverda, Reinier1982On the problem of topicalization in DutchVooruit3244-255
    • Salverda, Reinier2000On topicalization in Modern DutchShanno, Thomas F. & Snapper, Johan P. (eds.)The Berkeley conference on Dutch linguistics 1997. The Dutch language at the millenniumLanham/New York/OxfordUniversity Press of America93-111
    • Salverda, Reinier2000On topicalization in Modern DutchShanno, Thomas F. & Snapper, Johan P. (eds.)The Berkeley conference on Dutch linguistics 1997. The Dutch language at the millenniumLanham/New York/OxfordUniversity Press of America93-111
    • Salverda, Reinier2000On topicalization in Modern DutchShanno, Thomas F. & Snapper, Johan P. (eds.)The Berkeley conference on Dutch linguistics 1997. The Dutch language at the millenniumLanham/New York/OxfordUniversity Press of America93-111
    • Salverda, Reinier2000On topicalization in Modern DutchShanno, Thomas F. & Snapper, Johan P. (eds.)The Berkeley conference on Dutch linguistics 1997. The Dutch language at the millenniumLanham/New York/OxfordUniversity Press of America93-111
    • Salverda, Reinier2000On topicalization in Modern DutchShanno, Thomas F. & Snapper, Johan P. (eds.)The Berkeley conference on Dutch linguistics 1997. The Dutch language at the millenniumLanham/New York/OxfordUniversity Press of America93-111
    • Sturm, Arie1996Over functionele projectiesNederlandse Taalkunde1191-206
    • Travis, Lisa1984Parameters and effects of word order variationMITThesis
    • Werner Frey2006Contrast and movement to the German prefieldMolnár, Valéria & Winkler, Susanne (eds.)The Architecture of FocusBerlinMouton de Gruyter235--64
    • Zwart, Jan-Wouter1992Dutch expletives and small clause predicate raisingK. Broderick (ed.)Proceedings of North East Linguistic Society (NELS)22477-491
    • Zwart, Jan-Wouter1993Dutch syntax. A minimalist approachGroningenUniversity of GroningenThesis
    • Zwart, Jan-Wouter1997Morphosyntax of verb movement. A minimalist approach to the syntax of DutchDordrechtKluwer Academic Publishers
    • Zwart, Jan-Wouter2011The syntax of DutchCambridgeCambridge University Press
    • Zwart, Jan-Wouter2011The syntax of DutchCambridgeCambridge University Press
    • Zwart, Jan-Wouter & Hoekstra, Eric1997Weer functionele projectiesNederlandse Taalkunde2121-132
    About this article
    Author(s): Hans Broekhuis, Norbert Corver
    Category: Dutch Syntax
    Publication date: 22 November 2023
    Keywords: AccentAdverbial ModifierAnticipatory PronounBare InfinitivalBridge VerbComplementiveEmbedded Verb-secondExpletiveFocusIntonationInversionModalParticiplePied PipingPreposition StrandingResumptive ProlepsisSentence AccentStrandingSubject-initialSubject-verb InversionTopic ShiftTopicalizationVerbVerb-secondVerb ClusterWh-movement
    report errorprintcite
    Jump to:
    X