• Dutch
  • Frisian
  • Saterfrisian
  • Afrikaans
Show all cases

This section is concerned with some special uses of the demonstratives. We start in Subsection I with a number of idiomatic cases in which the demonstrative is used in its case-inflected form. After that, Subsections II to IV discuss a number of constructions that are often characterized by the fact that only one specific type of demonstrative pronoun can be used: we subsequently discuss constructions featuring the D-linked distal, the D-linked proximate, and non-D-linked demonstrative pronouns. We conclude in Subsection V with a discussion of the emphatic element zelf'himself', which is often also considered a kind of demonstrative pronoun.

[+]  I.  Idiomatic case-inflected forms

Dutch demonstratives do not inflect for case (in contrast to German), but there are a large number of historical relics in which inflected demonstratives do occur, and which are used in formal, written language. The examples in (492) must all be considered idiomatic, although we have seen in Section, sub V, that the masculine genitive form diens is still productively used in formal language.

a. bij dezen'by means of this letter'
b. één dezer dagen 'one of these days'
c. in dier voege 'so that ...'
d. met dien verstande dat ... 'provided (that) ...'
e. met alle gevolgen van dien'with all its consequences'
f. te dien einde dat ...'in order that ...'
g. uit dien hoofde 'because of that'
h. wat dies meer zij 'more of similar things'
[+]  II.  Distal demonstrative pronouns

This subsection is concerned with the special uses of the distal pronouns. That these cases are special is clear from the fact that the distal demonstrative cannot be replaced by a proximate one without changing the meaning of the construction.

[+]  A.  Distal demonstratives referring to +human entities

In order to refer to some +human entity in the domain of discourse, normally a personal pronoun is used; using a demonstrative pronoun in this function generally leads to a pejorative connotation; cf. Section, sub II, example (484). This subsection discusses some exceptions to this general rule.

[+]  1.  Distal pronouns with an antecedent in the immediately preceding discourse

One common exception is when the antecedent of the distal demonstrative is mentioned in the immediately preceding discourse, as in the examples in (485a), repeated here as (493). We have added indices to this example in order to unambiguously indicate the intended interpretation of the pronoun.

Heb je Jani/Mariei gezien? Nee, diei is ziek.
  have  you  Jan/Marie  seen  no  (s)he  is  ill
'Did you see Jan? No, he is ill.'

Example (494) shows that, in contrast to referential personal pronouns, distal demonstratives cannot be bound; they must be disjoint in reference to any c-commanding antecedent in the same sentence, and thus they behave like referential noun phrases like het meisje'the girl' in this respect. See Section, sub III, for more discussion of the binding properties of nominal expressions and Section, sub V, for similar but lightly less strict restrictions concerning construal of the possessive pronoun diens.

a. Mariei zei dat ziji/*diei/*het meisjei ziek was.
  Marie said  that  she/that.one/the girl  ill  was
b. Jan vertelde Mariei dat ziji/*diei/*het meisjei ontslagen zou worden.
  Jan  told  Marie  that  she/that.one/the girl  fired  would  be

      Distal demonstratives and referential personal pronouns also exhibit differences in distribution. First, the distal demonstrative is often preferred in contrastive contexts. If the question in (495a) is answered by means of (495b), it is strongly suggested that the person answering the question did see some other person that may be relevant for the given context; a natural continuation would be a report of what Jan has said. The answer in (495b'), on the other hand, is neutral in this respect. Recall that weak pronouns never occur in clause-initial position: die in (495b) can therefore only be replaced by the strong pronoun haar. Replacement of the weak pronoun ’r in (495b') by a distal demonstrative again seems to trigger a contrastive reading.

a. Heb je Mariei nog gesproken?
  have  you  Marie  prt  spoken
  'And, did you talk to Marie?'
b. Nee, diei heb ik niet meer gezien (maar wel Jan).
answer A
  no  that.one  have  no more  seen   but  AFF.  Jan
  'No, I didnʼt see her again, but I did see Jan.'
b'. Nee, ik heb ’r i niet meer gezien.
answer A'
  no  have  her  no more  seen
  'No, I havenʼt seen her again.'

      Second, the anaphoric behavior of distal demonstratives and referential personal pronouns differs if they occur unstressed in clause-initial position; cf. Haeseryn (1997:307-8) and Van Kampen (2009). In (496), the distal demonstratives cannot refer to the subject of the preceding clause but must refer to the object, whereas the referential pronouns are not restricted in this way.

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

This is not related to the syntactic function of the antecedent but instead depends on the information structure of its clause. Consider the following discourse chunk, where the continuations in (497b) and (497b') differ in that in the former but not the latter the distal demonstrative in the second conjunct can refer to the subject de leraar'the teacher' of the first conjunct.

a. Ik zat in de klas.
  was  in the classroom
b. Plotseling kwam de leraari binnen en diei/?hiji zei dat ...
  suddenly  came  the teacher  inside  and  that.one/he  said  that
  'Suddenly, the teacher entered and he said that ...'
b'. De leraari was nog steeds kwaad en hiji/*diei zei dat ....
  the teacher  was  still  angry  and he/that.one  said  that
  'The teacher was still angry and he said that ...'

This difference seems related to the fact that de leraar is preferably interpreted as part of the focus (new information) of the first conjunct in (497b), but as part of the presupposition in (497b'). That information structure is involved is also clear from the fact that the distal demonstrative cannot take a referential personal pronoun as its antecedent; such pronouns never function as the focus of the clause and are therefore unsuitable as antecedents for distal demonstratives. This is illustrated by the two examples in (498), which correspond to, respectively, (496b) and (497b).

a. Jani ontmoette haarj en zej/*diej vertelde hemi dat ..,
  Jan  met  her  and  she/that.one  told  him  that
b. Plotseling kwam hiji binnen en hiji/*diei zei dat ...
  suddenly  came  he  inside  and  he/that.one  said  that
  'Suddenly, he entered and he said that ...'

Information structure may also be the key to the apparent free variation in (496b); scrambling of intonationally unmarked noun phrases is restricted to noun phrases that are part of the presupposition of the clause (see Section 8.1.3), and Van Kampen (2009) claims that such scrambled noun phrases cannot function as antecedents of distal demonstratives. The examples in (499) show, however, that the contrast is not as sharp as one might have hoped: it seems that the distal demonstrative is preferred if the antecedent has not scrambled, but some of our informants also accept the distal pronoun if the antecedent has scrambled.

a. Jani ontmoette gisteren Elsj en diej/ ?zej vertelde hemi dat ..,
  Jan  met  yesterday  Els  and  that.one/she  told  him  that
b. Jani ontmoette Elsj gisteren en zej/%diej vertelde hemi dat ..,
  Jan  met  Els  yesterday  and  she/that.one  told  him  that

The judgments on (499b) may be somewhat blurred, however, by the fact that this example becomes fully acceptable as soon the antecedent or the distal pronoun is assigned accent; in the former case the proper noun Els will receive a contrastive/emphatic focus reading, which cancels the implication that it belongs to the presupposition of the clause; in the latter case the distal demonstrative does not function as a topic-shift device and, as a result, the requirement that its antecedent be part of the focus of its clause is lifted. Given these complications, we put examples such as (499b) aside and leave them for future research.
      The examples in (500) show that the antecedent of the distal pronoun cannot be embedded within a potential antecedent of the distal demonstrative: in example (500a), for example, the antecedent of the distal demonstrative is Peterʼs mother, not Peter. That this is due to the fact that the proper noun is embedded in another noun phrase is clear from the ungrammaticality of (500b) in which the noun phrase embeddding the proper noun cannot function as the antecedent of the demonstrative for reasons related to our knowledge of the world.

a. Plotseling kwam [de moeder [van Peterj]]i binnen en diei/*j zei dat ...
  suddenly  came   the mother   of Peter  inside  and  that.one  said  that
b. * Plotseling zag ik [de auto [van Peter]]i en diei zei dat ...
  suddenly  heard  I    the car   of Peter  and  that.one  said  that

      The discussion of examples (496)-(499) above suggests that the unstressed clause-initial distal demonstrative functions as a topic-shift device in the sense that it takes (part of) the focus of the preceding clause as its antecedent and presents it as the new discourse topic. Referential personal pronouns, on the other hand, signal that the discourse topic is maintained. See Van Kampen (2009) for more discussion.
      To conclude this subsection, we want to note that the use of a distal demonstrative is obligatory in (501b), which may be due to the fact that the pronoun must be stressed. Note, however, that instead of O, die! the phrase O, hij weer!'Oh, him again!' could also be used.

a. A: Daar is Jan. B: Wie? A: Jan.
  there  is Jan  who  Jan
b. B: O, die/*hij!
  oh  that one/him
[+]  2.  Left Dislocation

Left Dislocation constructions such as (502) resemble Topicalization constructions, but they differ from them in that they do not involve movement. The left-dislocated element is external to the clause and associated with a resumptive pronoun: if the resumptive pronoun is placed in clause-initial position, as in (502a), it preferably takes the form of a distal demonstrative; if it occupies the middle field of the clause, as in (502b), the referential personal pronoun gives rise to the best result. The demonstrative in constructions such as (502) is normally die, unless the antecedent is clearly neuter, as in Dat meisje[+neuter], dat ken ik niet'That girl, I donʼt know her'.

a. Marie, die/*?haar ken ik niet.
  Marie  that.one/her  know  not
  'Marie, I donʼt know her.'
b. Marie, ik ken haar/?die niet.
  Marie  know  her/that.one  not
  'Marie, I donʼt know her.'

Agreement between the left-dislocated element and the demonstrative does not occur, however, if the demonstrative functions as the logical subject of a nominal predicate; whereas the demonstrative must agree in gender with the dislocated element in the primeless adjectival examples in (503), it cannot agree with it in the primed nominal examples.

a. Jan, die/*dat is aardig.
  Jan  that/that  is nice
  'Jan, heʼs nice.'
a'. Jan, dat/*die is een aardige jongen.
  Jan  that/that  is a nice boy
  'Jan, heʼs a nice boy.'
b. Jan en Piet, die/*dat zijn aardig.
  Jan and Piet  those/that  are  nice
  'Jan and Piet, theyʼre nice.'
b'. Jan en Piet, dat/*die zijn aardige jongens.
  Jan and Piet  that/those  are  nice boys
  'Jan and Piet, theyʼre nice boys.'

The fact that the demonstrative always has the +neuter form dat in copular constructions with a nominal predicate is clearly related to the fact that dat can also appear in such constructions if there is no left-dislocated antecedent, as in the examples in (504). The examples in (505) show that similar facts can be found in the vinden-construction.

a. Dat/*Die is een aardige jongen.
  that/that  is a nice boy
b. Dat/*Die zijn aardige jongens.
  that/those  are  nice boys
a. Jan, dat/*die vind ik een aardige jongen.
  Jan  that/that  consider  a nice boy
  'Jan, I consider him a nice boy.'
a'. Dat/*Die vind ik een aardige jongen.
b. Jan en Piet, dat/*die vind ik aardige jongens.
  Jan and Piet  that/those  consider  nice boys
  'Jan, I consider them nice boys.'
b'. Dat/*Die vind ik aardige jongens.

      Note that the same demonstrative would be used if, instead of the subject, the predicate were left-dislocated. In (506), this is illustrated for the copular construction in the primeless, and for the nominal predicate of the vinden-construction in the primed examples.

a. Een aardige jongen, dat/*die is Jan.
  a nice boy  that/that  is Jan
a'. Een aardige jongen, dat/*die vind ik Jan niet.
  a nice boy  that/that  consider  Jan not
b. Aardige jongens, dat/*die zijn Jan en Peter.
  nice boys  that/those  are  Jan and Peter
b'. Aardige jongens, dat/*die vind ik Jan en Peter niet.
  nice boys  that/those  consider  Jan and Peter not

But, of course, we cannot appeal to this fact in order to account for the contrast between the primeless and primed examples in (503) given that the demonstrative dat is also used with the left-dislocated adjectival predicate in (507).

a. Aardig, dat/*die is Jan.
  kind  that  is Jan
b. Aardig, dat/*die zijn Jan en Peter.
  kind  that  are  Jan and Peter
[+]  B.  Distal demonstratives in conversation and narratives

Distal demonstratives are frequently used in conversations or narratives to introduce discourse entities which are not part of the topic of the discourse, but which are nevertheless presented as “familiar”. In a discussion about corruption in the army, an example such as (508) could be used to introduce another example of corruption that was not mentioned earlier but can be expected to be known to the participants in the conversation. Using a proximate demonstrative in this context would lead to an infelicitous result.

En dan is er nog die/??deze kwestie van drugssmokkel door die mariniers.
  and then  is there  prt  that/this case  of smuggling drugs by those marines
'And then there is that case of smuggling drugs by those marines.'

Note that the PP-modifier of the noun phrase in (508) also contains a distal demonstrative. It seems that this is a hallmark of this type of noun phrases; especially in colloquial Dutch, PP-modifiers containing another distal demonstrative are frequently used to identify certain persons. The PP is usually introduced by van'of', but other prepositions are possible as well. In example (509a), for instance, some person is identified as the man who plays in a certain TV-commercial, in (509b) reference is made to a friend the speaker met at a trip to Rome, in (509c) the person in question is uniquely identified by mentioning the kind of car he drives.

a. Hé, dat is die man van dat reclamespotje!
  hey  that  is that man  from the commercial
  'Hey, that is the man from this commercial!'
b. Die vriendin van die reis naar Rome komt vanavond eten.
  that friend  of that trip to Rome  comes  tonight  eat
  'This friend I met on this trip to Rome is coming to dinner tonight.'
c. Die vent met die BMW is ook weer thuis.
  that bloke  with the BMW  is also again  home
  'This bloke with the BMW is back again.'

      If the demonstrative is used anaphorically, that is, if the discourse topic is not physically present, the use of the distal demonstrative is also much preferred. For example, when talking about a certain man or boy, who is not present, one does not use the proximate demonstrative deze'this' in (510a&b); this would only be possible if the speaker is able to point at that person. Similar observations can be made if reference is made to a certain time (span): one would not use the proximate demonstrative deze in an example such as (510c) unless one were be able to point to a certain day on a calendar, and using the proximate demonstrative dit'this' in dit moment in (510d) is only possible if the phrase refers to the speech time, which is incompatible with the use of the past tense in this example.

a. Zegt die/#deze vent ineens ...
  says  that/this guy  suddenly
  'Suddenly, this guy says ...'
b. Die/#Deze jongen werd natuurlijk erg boos.
  that/this boy  became  of course  very angry
c. Die/#Deze dag kom ik niet.
  that/this day  come  not
d. Hij zei op dat/*dit moment even niets.
  he  said  at that/this time  for.a.moment  nothing
[+]  C.  Distal demonstratives in imperatives

In imperatives, the choice of the demonstrative depends on the position of the noun phrase in the clause. First, consider the examples in (511), which shows that a direct object normally precedes the particle of a particle verb like neerleggen'to put down'.

a. Jan legt de/deze/die bal neer.
  Jan puts  the/this/that ball  down
b. * Jan legt neer de/deze/die bal.

If the verb takes the imperative form, the direct object can, of course, also precede the particle, as is shown in (512a), but (512b) shows that the object can also follow the particle provided that the determiner is the distal demonstrative die; if the determiner is the definite article or the proximate demonstrative deze, this order is unacceptable.

a. Leg de/deze/die bal neer!
  put  the/this/that ball  down
b. Leg neer die/*deze/*de bal!
  put  down  that/this/the ball

In imperative constructions in which the verb has the infinitival form, the particle can also be followed by a direct object headed by a distal demonstrative, as in (513a), and something similar happens in imperative constructions in (513b) without a verb form; observe that in the latter case the direct object can be optionally preceded by the preposition met'with'. For a more extensive discussion of these imperative constructions, see Section V11.2.3.

a. Neerleggen die/*deze/*de bal!
  put.downinfinitive  that/this/the ball
b. Het huis uit (met) die/?deze/?de bal!
  the house  out.of   with  that/this/the ball
[+]  D.  Distal demonstratives in evaluative contexts

The distal demonstratives die/dat can also be used to express a (mostly negative) evaluation. Under the evaluative reading, the examples in (514) require a distal demonstrative: the proximate demonstratives deze and dit are only compatible with a truly deictic meaning.

a. Die/#Deze rotmol[-neuter] heeft weer gaten in het gazon gemaakt!
  this/these  damnʼd mole  has  again  holes  in the lawn  made
  'That damnʼd.mole has made holes in the lawn again!'
b. Dat/#Dit rotbeest[+neuter] heeft weer gaten in het gazon gemaakt!
  this/these  damnʼd.animal  has  again  holes  in the lawn  made
  'That damnʼd animal has made holes in the lawn again!'

      In these evaluative contexts, the noun phrase may also contain a proper noun provided that the latter is modified by a non-restrictive adjectival phrase expressing some subjective evaluation on the part of the speaker, as in (515a&b); if the modifier does not allow such an evaluative interpretation, as in (515c), the result is infelicitous.

a. Die/*deze vreselijke Van Dijk!
negative subjective evaluation
  that/this  horrible  Van Dijk
b. Die/*deze lieve Peter!
positive subjective evaluation
  that/this  sweet Peter
  'Sweet Peter!'
c. * Die grote Peter!
no subjective evaluation
  that  big  Peter

In those cases in which an article is part of the name, as in De Amstel'the Amstel' or De Alpen'the Alps', the non-restrictive interpretation of the modifying AP is available both with the definite article and the D-linked demonstrative determiner, as shown in (516). The difference between the two constructions is subtle: with the definite article the modifier serves a descriptive role, whereas with the demonstrative it takes on a more evaluative role.

a. Die/De prachtige, blauwe Amstel!
  that/The  splendid  blue  Amstel
  'That splendid, blue Amstel!'
b. Die/De adembenemende Alpen!
  those/The  breathtaking  Alps
  'Those breathtaking Alps!'
[+]  E.  Spurious distal demonstratives

In the examples in the previous subsection, the demonstrative force of the demonstratives seems considerably weakened. This subsection discusses some cases in which the pronoun die has lost its demonstrative force entirely. One example is the use of the distal demonstrative in front of a proper noun (which normally resists a determiner) in amicable greeting formulas like (517a). Note that the pronoun does not agree in gender with the proper noun following it: diminutives like Jantje are neuter, and hence we would expect the neuter demonstrative dat, which does indeed appear in examples such as (517b), where we are dealing with true demonstratives. This suggests that die is a spurious demonstrative in (517a).

a. Ha, die/*dat/*deze Jantje!
  hey  that/that/this  Jandim
b. Praten we nu over dít of dát Jantje?
  talk  we  now  about this or that Jandim
  'Are we talking about this or that Jantje?'

      A similar spurious use of the distal demonstrative can be found in exclamations of the type in (518). Constructions of this kind express (positive) surprise on the part of the speaker, along the lines of “Wim prime minister; who would have thought it!”. Interestingly, there is gender agreement between determiner and noun in (518a&b), but not with the diminutive proper noun in (518c): here the non-neuter demonstrative determiner die is combined with a neuter noun.

a. Die Wim toch! Minister-president!
  that Wim  prt  prime minister
b. Dat Duitsland toch! Zomaar wereldkampioen!
  that Germany prt  like.that  world champion
c. Die/*Dat Marietje toch! In een keer geslaagd!
  that Mariedim  prt  in one time passed

      A final case of a spurious demonstrative die is given in (519). Actually, in this example we are also dealing with a spurious preposition van; the van-PP occurs in a position in which normally only noun phrases can be used, as is shown in (519b). For more discussion of this construction, see Section, sub IB, and Section P1.4.

a. Jan kocht van die lekkere koekjes.
  Jan bought  of these tasty cookies
  'Jan bought these tasty cookies.'
b. Jan kocht lekkere koekjes.
  Jan bought  tasty cookies
[+]  III.  Proximate demonstrative pronouns

Generally speaking, proximate demonstratives are always used deictically: the speaker must be able to point at the referent of the noun phrase containing the demonstrative. There are only two exceptions to this rule. First, the proximate demonstrative can be used anaphorically if it has a linguistic antecedent in the immediately preceding discourse. The pronoun must then refer to the noun phrase immediately preceding it; in other words, deze in (520a) can only refer to Peter, not to Jan. The latter would be preferred if the personal pronoun hij'he' or the distal demonstrative is used, as in (520b).

Jan ontmoette gisteren Peter en ...
  Jan met  yesterday  Peter and
'Jan met Peter yesterday ...'
a. ... deze vertelde hem dat ...
  ... this one (= Peter)  told  him (= Jan)  that
b. ... hij/die vertelde hem dat ...
  ... he/that one (= Jan)  told  him (= Peter)  that

      Second, proximate (but not distal) demonstratives can also be used with an anticipatory function: in (521) the demonstrative functions as an anticipatory pronoun referring to what follows the colon. Note that, as is to be expected in a copular construction, the demonstrative agrees in gender with the noun phrase de zaak/het geval.

a. De zaak is deze/*die: Jan is ontslagen en ...
  the issue  is this/that:  Jan is fired and
b. Dit/*Dat is het geval: Jan is ontslagen en ...
  this/that  is the case:  Jan is fired and

The examples in (522) are similar in that the demonstrative is an anticipatory pronoun introducing the clausal modifier following the noun. Observe that in the fixed expression in (522b), the case-inflected distal demonstrative dien is used.

a. met dit/??dat verschil dat Jan het vrijwillig doet en Els gedwongen.
  with this/that difference  that  Jan it  voluntarily  does and  Els forced
  'With this difference that Jan does it voluntarily, whereas Els is forced to do it.'
b. met dien verstande dat ...
  with  that provision  that
  'provided (that) ...'
[+]  IV.  The non-D-linked demonstratives zoʼn and zulke used as amplifiers

Although zoʼn and zulke'such (a)' normally function as demonstrative determiners, they can also be used as amplifiers with the loss of their original demonstrative function. In the examples in (523), for instance, zoʼn and zulke do not refer to particular, identifiable type(s) of headache, hunger or ideas, but indicate that the headache(s) and hunger are quite severe, or that the ideas are very weird.

a. Ik heb zoʼn pijn in mijn hoofd.
  have  such a pain  in my head
  'I have such a terrible headache.'
b. Ik heb zoʼn honger.
  have  such a hunger
  'Iʼm so hungry.'
c. Hij heeft soms zulke hevige hoofdpijnen.
  he  has  sometimes  such fierce headaches
  'He sometimes has such terrible headaches.'
d. Jan heeft soms zulke rare ideeën.
  Jan has  sometimes  such weird ideas
  'Sometimes Jan has such weird ideas.'
[+]  V.  The emphatic modifier zelf'himself'

The emphatic element zelf'himself' is traditionally also considered a demonstrative pronoun. However, it does not occur in prenominal position and it can be used as a modifier not only of full noun phrases, but also of a proper nouns and pronouns. Although this is not so clear from (524a'), example (524b') shows that the pronoun and the emphatic modifier can at least sometimes be considered a constituent (the constituency test).

a. Die man/Jan wil dat boek zelf lezen.
  that man/Jan  wants  that book  himself  read
  'That man likes to read that book himself.'
a'. ?? Die man zelf wil dat boek lezen.
b. Hij wil dat boek zelf lezen.
  he  wants  that book  himself  read
  'He wants to read that book himself.'
b'. Hij zelf wil dat boek lezen.

      Let us provisionally assume that it holds for all occurrences of zelf that it forms a constituent with the noun phrase it modifies at at least some stage of the derivation, that is, that it functions as a kind of floating quantifier. Such an assumption would account for the fact illustrated by the contrast in (525) that zelf requires a noun phrase associate to be present: if zelf is generated as the modifier of a noun phrase, the presence of the former of course implies the presence of the latter.

a. De man leest het boek zelf.
  the man  read  the book  himself
  'The man is reading the book himself.'
b. * Het boek wordt zelf gelezen.
  the book  is  himself  read
  'The book is read himself.'

      If the suggested proposal is on the right track, the fact that the modifier zelf and its noun phrase associate can be discontinuous leads to the conclusion that they can be split in the course of derivation by movement. The structure of the topicalization constructions in the primed examples in (524) would then be something like that given in (526).

Die man/Hiji wil dat boek [ ti zelf] lezen.
  that man/he  wants  that book  himself  read

Furthermore, we have to assume that scrambling may also result in a split pattern. This is clear from the primed examples in (527), in which the surface position of the direct object is the result of scrambling.

a. Ik heb gisteren de directeur zelf gesproken.
  have  yesterday  the director  himself  spoken
  'Yesterday I spoke to the director himself.'
a'. Ik heb de directeuri gisteren [ ti zelf] gesproken.
b. Ik heb gisteren hem zelf gesproken.
  have  yesterday  him  himself  spoken
  'Yesterday I spoke to him himself.'
b'. Ik heb hemi gisteren [ ti zelf] gesproken.

      An argument in favor of the movement analysis above is that the split cannot occur if zelf modifies a noun phrase that is the complement of a preposition: on this analysis the noun phrase and zelf form a constituent, so that movement of the string met de directeur would involve movement of a non-constituent and (528b) would correctly be predicted to be ungrammatical under the intended reading.

a. Ik heb gisteren [PP met [[de directeur] zelf]] gesproken.
  have  yesterday  with the director himself  spoken
  'Yesterday, I spoke with the manager director himself.'
b. # Met de directeur heb ik gisteren zelf gesproken.

      Note that example (528b) is acceptable under a reading where zelf modifies the subject pronoun. This may give rise to the idea that the intervention of the subject blocks the intended reading. It seems, however, that this kind of intervention effect does not occur with zelf: in (529a) zelf can be equally well construed with the subject as with the object pronoun. Note, however, that if zelf is placed in front of the adverbial phrase, as in (529b), the sentence is unambiguous; zelf can then only be construed with the direct object, which indicates that its position is the result of pied piping.

a. Iki heb hemj gisteren [ ti/j zelf] gesproken.
  have  him  yesterday  myself/himself  spoken
  'Iʼve spoken to him myself/himself, yesterday.'
b. Ik heb [[hem] zelf] gisteren gesproken.
  have    him  himself  yesterday  spoken
  'Iʼve spoken to him himself, yesterday.'

Note in passing that the fact that (529a) is ambiguous apparently supports the assumption in traditional grammar (cf. Haeseryn et al. 1997: 1185) that zelf functions as a kind of supplementive, which exhibits the same kind of ambiguity. This assumption is, however, undermined by the fact that zelf can also occur within the PP in (528), an option that does not arise with supplementives, which are only predicated of subjects and direct objects. The analysis according to which zelf is generated as a modifier within the noun phrase seems therefore superior.
      Example (530a) shows that it is also possible to topicalize the modifier zelf in isolation, provided that it is assigned emphatic/contrastive accent. It seems, however, that this option is more or less restricted to those cases in which zelf modifies the subject of the clause; the reading of (530b'), where zelf is construed with the direct object, seems degraded, whereas the reading in (530b), where it is construed with the subject, sounds completely natural.

a. Zelf heb ik dat boek niet gelezen (maar Jan wel).
  myself  have  that book  not  read   but  Jan  aff
  'I didnʼt read that book myself (but Jan did).'
b. Zelf heb ik de directeur nooit gesproken (maar Jan wel).
  myself  have  the director  never  spoken   but  Jan  aff
  'I never spoke the managing director myself (but Jan did).'
b'. ?? Zelf heb ik de directeur