• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all
13.3.2.Contrastive focus and topic movement

This section discusses focus and topic movement, which are illustrated in (112a) and (112b), respectively. The fact that the movements in (112) involve a PP, which moreover functions as a subpart of a clausal constituent, immediately shows that we are dealing with A'-movement. We will represent the lexical domain of the verb as [LD ... V ...] instead of [vP ... v [VP ... V ...]], and ignore traces of subjects if they are not directly relevant for the discussion; cf. the introduction to Section 13.3.

a. dat Marie [FocP [op Peter]i Foc [LD [AP erg dol ti] is]].
  that  Marie   of Peter  very fond  is
  'that Marie is very fond of Peter.'
b. Ik weet niet wat Marie van Jan vindt, maar ik weet wel ... dat ze [TopP [op Peter]i Top [LD [AP erg dol ti] is]].
  know  not  what  Marie of Jan  considers,  but  know  aff that  she   of Peter  very fond  is
  'I donʼt know how Marie feels about Jan but I do know sheʼs very fond of Peter.'

      The contrastive phrases in (112) are characterized phonetically by a specific accent involving a high pitch followed by a sudden drop in pitch. The two cases differ in that the contrastive focus accent, which is sometimes called A-accent, concludes after the fall in pitch, while the contrastive topic accent, which is sometimes called B-accent, has an additional rise in pitch; cf. Jackendoff (1972: section 6.7), Büring (2007), and Neeleman & Van de Koot (2008). The development of the two pitch accents is represented in (113) by means of lines: in the examples words with an A-accent will be indicated by means of small caps in italics, while words with a B-accent will not be in small caps but will be doubly underlined as well as italicized; cf. (112).


      Semantically speaking, contrastive accent evokes a set of alternative propositions. A common intuition is that contrastive focus involves "some kind of contrast between the Focus constituent and alternative pieces of information which may be explicitly presented or presupposed" (Dik 1997:332). This can be formally represented by assuming that focus adds an additional semantic value (henceforth: focus value) to the regular semantic value (henceforth: ordinary value) of a clause; cf. Rooth (1997). So, while the ordinary value of the sentence Jan bezoekt Marie'Jan is visiting Marie' is simply the proposition given in (114a&b), the added focus values are sets of proposition, as indicated in the primed examples, in which the value of the variables x and y are taken from the set of (contextually defined) individuals E.

a. [Jan bezoekt [Focus Marie]]o = visit(j,m)
ordinary value
a'. [Jan bezoekt [Focus Marie]]F = {visit(j,x) | x ɛ E}
focus value
b. [[Focus Jan] bezoekt Marie]o = visit(j,m)
ordinary value
b'. [[Focus Jan] bezoekt Marie]F = {visit(y,m) | y ɛ E}
focus value

The function of non-contrastive (new information) focus is that the speaker fills in an information gap on the part of the addressee by adding or selecting a proposition to or from the focus value; the speaker crucially does not intend to imply anything for the alternative propositions from the focus value. By using the A-accent on the other hand, the speaker implies that the ordinary value of the clause is counter-presuppositional. An utterance such as Jan bezoekt Marie then opposes the ordinary value of the clause in (114a) to other propositions from the focus value in (114a') that the speaker assumes to be considered true by the addressee, that is, the speaker implies that Jan did not visit at least one individual from E; see also Neeleman & Vermeulen (2012). It should be noted that the nature of the counter-presuppositional relation can be further specified by focus particles like alleen'only' and ook'too'; we will return to this in Subsection IC. By using the contrastive B-accent, the speaker implies that there is at least one other potential discourse topic that could have been addressed. For instance, the plurality of the finite verb in question (115a) implies that the set of contextually defined individuals E contains at least two persons who are expected to be invited for the party. The answer in (115b) does not provide an answer to the question but asserts something about one of the individuals from E; cf. Büring (2007), Neeleman & Van de Koot (2008) and Neeleman & Vermeulen (2012).

a. Wie zijn er uitgenodigd voor het feest?
  who  are  there  invited  for the party
  'Who are invited to the party?'
b. Geen idee. Ik weet alleen dat Peter niet kan komen.
  no idea  know  only  that  Peter not  can  come
  'No idea. I only know that Peter cannot come.'

      The examples in (112) have already shown that contrastive foci and topics are characterized syntactically by the fact that they can be displaced. This property will be investigated in more detail in the following subsections. Subsection I starts with a discussion of focus movement, which is followed by a discussion of topic movement in Subsection II. The investigation of focus and topic movement is relatively recent and it is therefore not surprising that there are still a large number of controversial issues, some of which will be discussed in Subsection III.

[+]  I.  Focus movement

The direct objects in answers such as (116b&c) are assigned regular sentence accent (indicated by small caps) and are therefore part of the new-information focus. They can nevertheless be construed as contrastive foci in the sense that they exclude values of the variable x other than Marie. It should be noted, however, that in these cases the contrastive interpretations are entirely pragmatic in nature, as Grice’s cooperative principle requires the answers in (116) to be complete; cf. Neeleman & Vermeulen (2012).

a. Wie heeft Jan/hij bezocht?
question: ?x(Jan/he has visited x)
  who  has  Jan/he  visited
  'Who has Jan/he visited?'
b. Hij heeft een vriendin bezocht: Marie.
  he  has  a friend  visited  Marie
  'He has visited a lady friend: Marie.'
c. Jan heeft Marie bezocht.
  Jan has  Marie  visited
  'Jan has visited Marie.'

The cases of contrastive foci that will be discussed in this subsection are different in that they are characterized as contrastive by the phonetic property of carrying a contrastive A-accent and the syntactic property that they can be moved leftward by focus movement. Subsection A starts by discussing the landing site of focus movement, Subsection B will argue that focus movement is A'-movement, and Subsection C will conclude by arguing that focus movement is obligatory, just like other semantically motivated movements.

[+]  A.  The landing site of focus movement

This subsection discusses the landing site of focus movement. Following the line of research in Rizzi (1996) and Haegeman (1995), one option would be to postulate a focus phrase (FocP) in the middle field of the clause, the specifier of which is a designated landing site for focus movement. Neeleman & Van de Koot (2008) assume that focus movement is motivated by the need to assign scope to the focus phrase or, in their formulation, to distinguish contrastive foci from the backgrounds against which they are evaluated; see Barbiers (2010) for an alternative proposal. Since we have seen that contrastive foci evoke a set of alternative propositions, we may safely conclude that the background at least contains the lexical domain of the main verb: this entails that FocP is part of the verb’s functional domain.

... [FocP XPi Foc [ ... [LD ... ti ...]]]

Neeleman & Van de Koot argue against hypothesis (117), in as far as it postulates a designated target position for focus movement, and claim that focus movement can target any position from which the contrastively focused phrase may take scope over its background. The advantage of their proposal is that we can easily account for examples such as (118) by saying that the word order difference between the two examples reflects a scopal difference between the focused phrase and the modal adverb waarschijnlijk'probably': the adverb is in the scope of the focus in (118a), but not in (118b).

a. dat ze [op Peter]i waarschijnlijk [LD [AP erg dol ti] is].
  that  she   of Peter  probably  very fond  is
  'that she is probably very fond of Peter.'
b. dat ze waarschijnlijk [op Peter]i [LD [AP erg dol ti] is].
  that  she probably   of Peter  very fond  is
  'that she is probably very fond of Peter.'

A potential problem for the hypothesis that the contrastively focused phrase can target any position from which it may scope over the lexical domain of the clause is that it seems to overgenerate. The examples in (119b&c), for instance, show that the target position of focus movement cannot follow negation or precede a weak subject pronoun in the regular subject position.

a. dat ze [op Peter]i niet [LD [AP erg dol ti] is].
  that  she   of Peter  not  very fond  is
  'that she probably isnʼt very fond of Peter.'
b. * dat ze niet [op Peter]i [LD [AP erg dol ti] is].
  that  she  not   of Peter  very fond  is
c. * dat [op Peter]i ze niet [LD [AP erg dol ti] is].
  that   of Peter  she  not  very fond  is

The schematic representation in (120a) summarizes the positions in which the contrastively focused PP op Peter can or cannot be found. On the assumption that focus movement targets the specifier position of a FocP, we can account for this in at least two ways. One option is to adopt the representation in (120b), according to which there are two FocPs, one relatively high and one relatively low in the middle field of the clause; cf. Belletti (2004), Aboh (2007), and Zubizarreta (2010). Another option is that there is just a single FocP but that the modal adverb can be placed either above or below FocP depending on its scope relative to the contrastive focus, as in (120b').

a. dat <*PPi> hij <PPi> waarschijnlijk <PPi> niet <*PPi> [LD [AP erg dol ti] is].
b. dat hij .. [FocP .. Foc [.. waarschijnlijk [FocP .. Foc [NegP .. Neg [LD ...]]]]]
b'. dat hij <waarschijnlijk> [FocP .. Foc [.. <waarschijnlijk> [NegP .. Neg [LD ...]]]]

Since the debate on the landing site of focus movement is just in its initial stage, we will not evaluate the three proposals any further, but simply assume for concreteness’ sake that focus movement targets the specifier of FocP.

[+]  B.  Focus movement is A'-movement

This subsection reviews a number of arguments for assuming that focus movement is A'-movement. A first, and conclusive, argument is that focus movement can affect non-nominal categories. It has also been argued that focus movement may violate certain word order restrictions that constrain A-movement, but we will see that there are certain difficulties with this argument. A third argument found in the literature is that focus movement is not clause-bound.

[+]  1.  Categorial restrictions

A-movement is restricted to nominal categories. The fact illustrated again in (121b) that focus movement may also affect PPs is therefore sufficient for concluding that we are dealing with A'-movement. Example (121c) further supports this conclusion by providing an example in which an adjectival complementive has undergone focus movement.

a. dat Jan waarschijnlijk [het boek]i niet ti wil kopen.
  that  Jan probably  the book  not  wants  buy
  'that Jan probably doesnʼt want to buy the book.'
b. dat Jan waarschijnlijk [op vader]i niet ti wil wachten.
  that  Jan probably  for father  not  wants  wait
  'that Jan probably doesnʼt want to wait for father.'
c. dat Jan deze zaak waarschijnlijk [zo belangrijk]i niet ti vindt.
  that  Jan this case  probably  that important  not  considers
  'that Jan probably doesnʼt consider this case that important.'

The conclusion that focus movement is A'-movement is in line with the conclusion that focus movement may target a position to the right of the modal adverbs because Section 13.2 has shown that nominal argument shift targets a position to the left of the modal adverbs. This contrast can be highlighted by the VP-topicalization constructions in (122), which show that the direct object can only be stranded in a position after the clause adverbials if it is contrastively focused.

a. [VPti Kopen] wil Jan <het boeki> waarschijnlijk <*het boeki> tVP.
  buy  wants  Jan    the book  probably
b. [VPti Kopen] wil Jan waarschijnlijk het boekitVP.
  buy  wants  Jan probably  the book

It can also be illustrated quite nicely by means of the placement of strong (phonetically non-reduced) referential personal pronouns like zij'she' en haar'her'; such pronouns may only occur after the modal adverbs if they carry an A-accent.

a. dat <zij/zij> waarschijnlijk <zij/*zij> het boek gekocht heeft.
  that    she/she  probably  the book  bought  has
  'that she/she probably has bought the book'
b. dat Jan <haar/haar> waarschijnlijk <haar/*haar> wil helpen.
  that  Jan    her/her  probably  wants  help
  'that Jan probably wants to help her/HER.'

Furthermore, that nominal argument shift and focus movement target different landing sites is highlighted by the fact that -human referential personal pronouns can never occur after the modal adverbs, for the simple reason that they are obligatorily reduced phonetically; in order to contrastively focus an inanimate entity, the demonstrative deze/die'this/that' is needed.

a. dat hij <de auto> waarschijnlijk <de auto> gekocht heeft.
  that  he   the car  probably  bought  has
  'that he probably has bought the car.'
b. dat hij <ʼm/die> waarschijnlijk <die/*hem/*ʼm> gekocht heeft.
  that  he    him/dem  probably  bought  has
  'that he probably has bought that one.'
[+]  2.  Word order

Another argument in favor of an A'-movement analysis of focus movement has to do with word order. Section 13.2, sub IC, has shown that nominal argument shift cannot affect the unmarked order of nominal arguments (agent > goal > theme) in Dutch. Neeleman & Van de Koot (2008), Van de Koot (2009) as well as Neeleman & Vermeulen (2012) claim that focus movement is able to change the order of nominal arguments, as illustrated in (125), and that this supports the claim that we are dealing with A'-movement.

a. % Ik geloof [dat dit boeki Jan Marie ti gegeven heeft].
  believe   that  this book  Jan Marie  given  has
b. % Ik geloof [dat Jan dit boeki Marie ti gegeven heeft].
  believe   that  Jan this book  Marie  given  has
c. Ik geloof [dat Jan Marie dit boek gegeven heeft].
  believe   that  Jan Marie  this book  given  has
  'I believe that Jan has given Marie this book.'

The argument is not entirely convincing; the fact that this type of order preservation does not hold for German nominal argument shift shows that it is not a defining property of nominal argument shift; cf. Section 13.2, sub IC. Furthermore, the judgments given by Neeleman and his collaborators are controversial, as some speakers of Dutch (including the authors of this work) reject the examples in (125a&b) with the indicated intonation pattern; see also Neeleman & Van de Koot (2008:fn.2) and Van de Koot (2009:fn.4). A simpler example –which is likewise rejected by some of our informants– is given in (126). In our view, the unclear acceptability status of (125a&b) and (126a) makes it impossible to draw any firm conclusion from them; in fact, it remains to be seen whether these examples should be considered part of the standard variety of Dutch, but we will leave this issue for future research.

a. % Ik geloof [dat dit boeki Jan ti gelezen heeft].
  believe   that  this book Jan  read  has
b. Ik geloof [dat Jan dit boeki gelezen heeft].
  believe   that  Jan this book  read  has
  'I believe that Jan has read this book.'

In order to avoid confusion, we should note that the examples marked with % become acceptable if the contrastively accented phrases are given a B-accent, in which case we are dealing with a contrastive topic; Subsection II will provide more data showing that topic movement may indeed affect the unmarked order of nominal arguments under certain conditions.
      Example (127a) shows that focus movement is able to change the unmarked order of nominal and prepositional objects: while prepositional indirect objects normally follow direct objects, focus movement of the former can easily cross the latter. It should be noted, however, that this requires the direct object to follow the modal adverb: the examples in (127b&c) show that object shift of het boek has a degrading effect on focus movement regardless of whether the focused phrase precedes or follows the modal adverb; we added the adverb niet to (127c) to make focus movement visible. Observe that (127b) becomes fully acceptable if the PP is assigned a B-accent, which shows that topic movement may cross a shifted object.

a. dat Jan <aan Els> waarschijnlijk <aan Els> het boek zal geven.
  that  Jan    to Els  probably  the book  will  give
  'that Jan will probably give the book to Els.'
b. ?? dat Jan aan Els het boek waarschijnlijk zal geven.
  that  Jan  to Els  the book  probably will  give
  'that Jan will probably give the book to Els.'
c. dat Jan het boek waarschijnlijk <??aan Els> niet <aan Els> zal geven.
  that  Jan the book  probably       to Els  not   will  give
  'that Jan probably will not give the book to Els.'

This subsection has shown that the claim that focus movement is able to change the unmarked order of nominal arguments in Standard Dutch is controversial; whether this property could be used as an argument in favor of the claim that focus movement is A'-movement is not clear either, as order preservation seems to be an accidental property of nominal argument shift in Dutch.

[+]  3.  Focus movement is not clause-bound

A'-movement differs from A-movement in that it allows extraction from finite clauses under certain conditions. Neeleman (1994a/1994b) and Barbiers (1999/2002) have shown that this also holds for focus movement: the examples in (128) illustrate that foci can target a focus position in the middle field of a matrix clause. The percentage signs are used to indicate that this type of long focus movement is normally not found in writing but can be encountered in colloquial speech; cf. Zwart (1993:200).

a. % Ik had [in de tuin]i gedacht [dat het feest ti zou zijn].
  had  in the garden  thought   that  the party would  be
  'I had thought that the party would be in the garden.'
b. % Ik had [een boek]i gedacht [dat Jan ti zou kopen].
  had   a book  thought   that  Jan  would  buy
  'I had thought that Jan would buy a book.'

That the landing site of the foci is external to the embedded clause is clear from the fact that the foci precede the clause-final main verb of the matrix clause. Because the examples in (129) show that embedded topicalization is impossible in Dutch (cf. Section 11.3.3, sub II), it is even impossible for foci to follow the verbs in clause-final position.

a. * Ik had gedacht [[in de tuin]i dat het feest ti zou zijn].
  had thought    in the garden  that  the party would  be
b. * Ik had gedacht [[een boek]i dat Jan ti zou kopen].
  had  thought     a book  that  John  would  buy

Although examples such (128) may be objectionable to certain speakers, the sharp contrast with the examples in (129) show that they are at least marginally possible in standard Dutch. This conclusion is also supported by the fact that the examples in (128) are clearly much better than the corresponding examples in (130) with the factive verb betreuren'to regret'. This contrast shows that long focus movement is only possible in specific bridge contexts.

a. * Ik had [in de tuin]i betreurd [dat het feest ti zou zijn].
  had  in the garden  regretted   that  the party  would  be
b. * Ik had [een boek]i betreurd [dat Jan ti zou kopen].
  had   a book  regretted  that  John  would  buy

      There are reasons for assuming that long focus movement is like long wh-movement in that it has to pass through the initial position of the embedded clause. A weakish argument in favor of this claim is that the direct object een boek'a book' in (128) can easily cross the subject, as this is a well-established property of A'-movements that target the clause-initial position. A stronger argument is that long focus movement cannot co-occur with long wh-movement, as is illustrated by the examples in (131): the examples in (131b&c) first show that wh-phrases and foci can be extracted from the embedded clause in (131a), while (131d) shows that they cannot be extracted simultaneously. This would follow immediately if long movement must proceed via the clause-initial position of the embedded clause: long wh-movement would then block long focus movement (or vice versa) because this position can be filled by a single (trace of a) constituent only; see Barbiers (2002) for a slightly different account.

a. Ik had gedacht [dat Jan morgen in de tuin zou werken].
  had thought   that  Jan tomorrow  in the garden  would  work
  'I had thought that Jan would work in the garden tomorrow.'
b. Waari had jij gedacht [dat Jan morgen ti zou werken]?
  where  had  you  thought   that  Jan tomorrow  would  work
  'Where had you thought that Jan would work tomorrow?'
c. % Ik had morgenj gedacht [dat Jan tj in de tuin zou werken].
  had tomorrow  thought   that  Jan  in the garden  would  work
  'I had thought that Jan would work in the garden tomorrow.'
d. * Waari had jij morgenj gedacht [dat Jan tjti zou werken]?
  where  had  you  tomorrow  thought  that Jan  would  work
[+]  C.  Is focus movement obligatory?

There is good reason for assuming that A'-movement is obligatory because it is needed to derive structures that can be interpreted by the semantic component of the grammar. Section, sub II, argued, for instance, that wh-movement in wh-questions is obligatory because it derives an operator-variable chain in the sense of predicate calculus. And Section 13.3.1, sub II, has argued that negation movement is obligatory in order to assign scope to sentence negation. In view of this we may hypothesize that focus movement is needed to assign scope to contrastively focused phrases (unless there is some other means to indicate scope). Languages such as Hungarian, where contrastive foci are obligatorily moved into a position left-adjacent to the finite verb, seem to support this hypothesis; cf. É. Kiss (2002:ch4). Languages such as English, which seem to mark contrastive focus by intonation only, are potential problems for the hypothesis, but since it has been argued that English does have focus movement in at least some constructions (cf. Kayne 1998), it remains to be seen whether languages like English constitute true counterexamples. This subsection argues that focus movement is normally obligatory in Standard Dutch by appealing to constructions featuring focus particles of two types: counter-presuppositional focus particles (alleen'only', ook'also', etc.) and scalar focus particles (al'already', nog'still', maar'just', etc.).

[+]  1.  Constituents with an A-accent that remain in situ

One potential problem for the hypothesis that focus movement is obligatory in Standard Dutch is that it is sometimes possible to leave constituents with an A-accent in their original position. This is illustrated by the two examples in (132), which suggests that focus movement is optional. Of course, this conclusion is valid only if the two examples are semantically equivalent; this does not seem to be the case, however.

a. dat Jan [FocP [op Peter]i Foc [LD [AP erg dol ti] is]].
  that  Jan   of Peter  very fond  is
  'that Jan is very fond of Peter.'
b. dat Jan [AP erg dol op Peter] is.
  that  Jan  very fond  of Peter  is
  'that Jan is very fond of Peter.'

Before showing that the two examples in (132) are not fully equivalent, we will first consider example (133a), which clearly has two readings: contrastive focus may be restricted to the direct object only, in which case the sentence expresses that there are certain other things in the domain of discourse that Jan did not read, or it may extend to the verb phrase, in which case the sentence expresses that there were certain things that Jan did not do. The examples in (133b&c) show that the two readings evoke different word orders if the negative adverb niet is present. The clearest case is (133b), in which contrastive focus is restricted to the moved direct object. Example (133c) is somewhat more complicated, as it again allows two readings, one with contrastive focus on the verb phrase, and one with contrastive focus on the noun phrase. This can be accounted for if we assume that in both cases we are dealing with constituent negation: Hij heeft niet de roman gelezen, maar het gras gemaaid'he hasnʼt read the novel, but mowed the grass' versus Hij heeft niet de roman gelezen, maar het gedicht'he didnʼt read the novel but the poem'.

a. dat Jan waarschijnlijk de roman gelezen heeft.
  that  Jan  probably  the novel  read   has
  'that Jan ha