Syntactic Atlas of Dutch Dialects – Commentary – Volume II
3 Auxiliaries and verb positions
3.1 Introduction
3.1.1 Auxiliary types and forms
3.1.2 Perfective auxiliaries
3.1.2.1 Auxiliary selection
3.1.2.2 Auxiliary doubling
3.1.2.3 Perfective passive
3.1.3 Auxiliary doen ‘do’
3.1.3.1 Periphrastic do
3.1.3.2 Short do replies
3.1.4 Verb positions
3.1.4.1 Complex initials
3.1.4.2 Subject between verb and tense suffix
3.1.4.3 V2 in embedded clauses
3.1.5 Other variation
3.1.5.1 Infinitival absolute with construction
3.1.5.2 Gerund
3.2 Historical developments and discussion of the literature
3.2.1 Auxiliary selection
3.2.2 Auxiliary doubling
3.2.3 Perfective passive
3.2.4 Auxiliary do
3.2.5 Verb positions
3.2.5.1 Complex initials
3.2.5.2 Subject between verb and tense suffix
3.2.5.3 V2 in embedded clauses
3.2.6 Other variation
3.2.6.1 Infinitival absolute with construction
3.2.6.2 Gerund
3.3 Discussion of the maps
3.3.1 Perfective auxiliaries
3.3.1.1 Auxiliary selection (map 40a)
3.3.1.2 Auxiliary doubling (map 40b)
3.3.1.3 Perfective passive (map 41a)
3.3.2 Auxiliary do
3.3.2.1 Periphrastic do
3.3.2.1.1 Declarative clause (map 41b)
3.3.2.1.2 Interrogative clause (map 42a)
3.3.2.1.3 Imperative clause (map 42b)
3.3.2.1.4 Negative clause (map 43a)
3.3.2.1.5 Synthesis periphrastic do (map 43b)
3.3.2.1.6 Perfective periphrastic do (map 44a)
3.3.2.2 Short do replies
3.3.2.2.1 Denial of negative statement (map 44b)
3.3.2.2.2 Denial of positive statement (map 44c)
3.3.3 Verb positions
3.3.3.1 Complex initials (map 45a)
3.3.3.2 Subject between verb and tense suffix (map 45b)
3.3.3.3 V2 in embedded clause (map 46a)
3.3.4 Other variation
3.3.4.1 Infinitival absolute with construction (map 46b)
3.3.4.2 Gerund (map 46c)
3.4 Literature on auxiliaries and verb positions
Chapter 3 Auxiliaries and verb positions
3.1 Introduction
3.1.1 Auxiliary types and forms
Standard Dutch has perfective, passive, modal, aspectual and causative auxiliaries. There are two perfective auxiliaries, hebben ‘have’ and zijn ‘be’. The passive auxiliary is worden ‘become’, which in the perfect tense is replaced by zijn ‘be’. The modals include kunnen ‘can’, moeten ‘must’, mogen ‘may’, zullen ‘will’, hoeven ‘need’, dienen ‘ought’, willen ‘want’. Aspectual auxiliaries are gaan ‘go’, komen ‘come’, blijven ‘remain’. The two causative auxiliaries are laten ‘let’ and doen ‘do’.
Unlike their English counterparts, Dutch auxiliaries are morphologically very similar to main verbs in that they have a complete verbal paradigm consisting both of finite forms, infinitives and participles. The only auxiliaries that do not have a past participle are zullen ‘will’ and dienen ‘ought’. Differences between auxiliaries and main verbs involve the finite paradigms of some of the auxiliaries. The perfective auxiliaries hebben ‘have’ and zijn ‘be’ are irregular. The modals kunnen ‘can’, moeten ‘must’, mogen ‘may’, zullen ‘will’ are preterite-present verbs; they have a present tense conjugation that is similar to the past tense conjugation of ordinary main verbs in not displaying any person distinctions.
The perfective auxiliaries select past participles as their complements. The passive auxiliary selects a passive participle, which has the same form as the past participle, typically consisting of a prefix ge-, the verbal stem and a suffix -d/-t/-en. The modals kunnen‘can’, moeten ‘must’, mogen ‘may’, zullen ‘will’, willen ‘want’, the aspectuals gaan ‘go’, komen ‘come’, blijven ‘remain’ and the causatives laten ‘let’ and doen ‘do’ select bare infinitives consisting of stem+-en. The modals hoeven ‘need’ and dienen ‘ought’ select to-infinitives.
The cross-dialectal variation in the morphological paradigms of auxiliaries, in verbal complementation and in the division of (semantic) labour between the various auxiliaries was not systematically investigated in the SAND-project. Instead, we tested variation in auxiliary selection (3.1.2.1), auxiliary doubling (3.1.2.2), periphrastic do (3.1.3.1) and short do replies (3.1.3.2). We found three further cases of syntactic variation in the verbal system which do not necessarily involve auxiliaries, but shed some light on auxiliary positions. These cases include complex initials, in which two verbs occur in clause-initial position (3.1.4.1), subject intrusion, in which a subject pronoun intervenes between the verbal stem and the past tense suffix (3.1.4.2), and V2 in embedded clauses, in which a finite verb occurs in a position directly following the complementiser and the subject (3.1.4.3). Finally, we found variation with respect to the possibility of to-infinitives in the absolute with construction (3.1.5.1) and the morphology of the gerund (3.1.5.2).
3.1.2 Perfective auxiliaries
Perfectivity in Standard Dutch is expressed by a periphrastic perfective construction involving a perfective auxiliary and a past participle, as in (1).
(1) a. Jan is ge-vall-en.
John is pref.fall.suff
‘John has fallen.’
b. Jan heeft ge-werk-t.
John has pref.work.suff
‘John has worked.’
3.1.2.1 Auxiliary selection
As (1) shows, the perfective auxiliary can be either be orhave. In Standard Dutch, the choice between the two depends on the main verb. be is used when the main verb is unaccusative (2a). have is used with intransitive and transitive main verbs (2c,e) (an exception to this generalisation is the occurrence of past tense have with unaccusatives in counterfactuals, but this phenomenon does not play a role in this atlas). A diagnostic to distinguish unaccusatives from intransitives is prenominal modification. The participle of an unaccusative verb can modify a noun that corresponds to the subject of that verb (2b), but the participle of an intransitive cannot (2d). Since participles of transitives can modify their objects but not their subjects (2f,g), this test shows that the subject of an unaccusative verb patterns with the object of a transitive verb, while the subject of an intransitive verb patterns with the subject of a transitive verb.
(2) a. Het boek is / *heeft gevallen. unaccusative
the book is / has fall.pcp
‘The book has fallen.’
b. het gevallen boek unaccusative
the fallen book
c. Het boek heeft / *is gewerkt. intransitive
the book has / is work.pcp
‘The book has worked.’
d. * het gewerkte boek intransitive
the worked book
e. De man heeft / *is de hond geslagen. transitive
the man has / is the dog beaten
‘The man has beaten the dog.’
f. de geslagen hond transitive
the beaten dog
g. de geslagen man transitive
the beaten man
# ‘The man who beat x.’
In many languages, the choice of perfective auxiliaries is less clear-cut. Here some or all of the unaccusative verbs select have rather than be. In the SAND-project we have tested this for the main verbs sterven ‘die’, vallen ‘fall’, zijn ‘be’ and tegenkomen ‘come across’. The relevant test sentences are presented in (3).
(3) a. Ze weet niet dat Marie gisteren gestorven heeft.
she knows not that Mary yesterday died has
‘She does not know that Mary died yesterday.
b. Geloof je niet dat hij met de fiets gevallen heeft?
believe you not that he with the bike fallen has
‘Don’t you believe that he has fallen with his bike?’
c. Die jongen heb ik mee naar de markt geweest.
that boy have I with to the market been
‘I have been to the market with that boy.’
d. Ik heb hem gisteren tegengekomen.
I have him yesterday across-come.pcp
‘I ran into him yesterday.’
Variation in perfective auxiliary selection is not unexpected in view of recent comparative work on auxiliary selection by Sorace (2000). She shows that the class of unaccusative verbs can be divided into four semantic classes:
(4) Classes of unaccusative verbs
(i) Change of location e.g., go
(ii) Change of state e.g., change
(iii) Continuation of a pre-existing state e.g., survive
(iv) Existence of state e.g., exist
Cross-linguistically, the probability of have increases and the probability of be decreases from (i) to (iv). Telicity seems to be a relevant factor: the less telic an unaccusative verb the more likely it selects have. Intralinguistically, the four classes constitute an implicational scale from (iv) to (i). For example, if a language has be with verbs from the change of state class, it will also use be with verbs from the change of location class. The reverse does not hold.
The SAND data show that there are no dialects in the Dutch language area that select have with sterven ‘die’ (3a), an inherently telic verb of class (ii). There are, however, dialects that select have with vallen‘fall’, zijn‘be’ or tegenkomen ‘come across’. As for vallen‘fall’ (3b), it could be claimed that it is less telic than sterven ‘die’ in that it does not denote an inherent endpoint. Indeed, informal testing shows that speakers tend to find vallen‘fall’ with have less acceptable when a PP is added that denotes the endpoint, as in *Hij heeft op de grond gevallen ‘lit. he has on the ground fallen’.
The selection of have by main verb zijn‘be’ (3c) is not completely surprising, as zijn‘be’ in combination with naar de markt ‘to the market’ does not denote a change of location but the existence of a state of absence (cf. Hij is naar de markt ‘He is off to the market’).
The verb tegenkomen ‘come across’ (3d) is clearly telic, but it selects have in some of the dialects. Since it selects auxiliary be in Standard Dutch and cannot be passivised, it should probably be classified as an unaccusative verb, even though its participle cannot modify a noun corresponding to the subject, unlike other dyadic unaccusative verbs (e.g., bevallen ‘please’). It may be somewhat similar to verbs like vergeten ‘forget’ which also have an experiencer subject and a theme object and can occur both with have and with be in Standard Dutch.
3.1.2.2 Auxiliary doubling
In Standard Dutch the number of perfective auxiliaries per clause is limited to one (5a). Some Dutch dialects allow two perfective auxiliaries per clause (5b,c).
(5) a. Ik heb vandaag nog niet gerookt (*gehad).
I have today yet not smoked had.pcp
‘I have not yet smoked today.’
b. Ik heb vandaag nog niet gerookt gehad.
I have today yet not smoked had.pcp
‘I have not yet smoked today.’
c. Ben je met die fiets gevallen geweest?
are you with that bike fallen been
‘Have you fallen with that bike?’
The question is whether this should be analysed as semantically redundant doubling, as a perfect of the perfect, or as the perfect tense of main verb hebben ‘have’ and zijn ‘be’. This question is discussed in section 3.2.2
3.1.2.3 Perfective passive
The Standard Dutch passive is periphrastic and involves the auxiliary worden ‘become’ and a participle (6a).
(6) a. Het huis wordt vandaag verkocht.
the hous becomes today sold.pcp
‘The house will be sold today.’
b. Het huis is vandaag verkocht.
the house is today sold.pcp
‘The house has been sold today.’
c. Het huis is verkocht geworden.
the house is sold.pcp become.pcp
‘The house has been sold.’
d. Het huis is verkocht geweest.
the hous is sold.pcp been.pcp
‘The house has been sold.’
In the perfect tense, passive auxiliary worden ‘become’ is replaced by the perfective auxiliary zijn ‘be’ (6b). If the building blocks of the periphrastic perfect and the periphrastic passive were simply combined to get a perfective passive, we would expect (6c), in which is ‘is’ is the perfective auxiliary and geworden ‘become’ the passive auxiliary. The construction in (6c) is possible in certain dialects but not in Standard Dutch. As a variant of (6c), some dialects have (6d), in which the participle of the passive auxiliary is replaced by the participle of be. In Standard Dutch, (6d) is also possible but then it can only mean that the house is no longer sold, i.e., it is for sale again. In such cases, (6d) should probably be analysed as the perfect tense of an adjectival predicative construction denoting a property of the subject, not an event. If so, then Standard Dutch sentences of type (6d) do not involve semantically redundant doubling. Similarly, the auxiliary doubling illustrated in (6c,d) is not necessarily redundant, which makes it mysterious why many varieties of Dutch do not allow it.
3.1.3 Auxiliary doen ‘do’
The Standard Dutch verb doen ‘do’ occurs as a dynamic main verb and as an auxiliary. As an auxiliary, it can be causative (7a), a semantically empty dummy (7b) or anaphoric (7c).
(7) a. Deze gedachte deed iedereen huiveren.
this thought did everyone shiver
‘This thought made everyone shiver.’
b. Werken doet Jan niet.
work.inf does John not
‘John does not work.’
c. Jan werkt hard en Marie doet dat ook.
John works hard and Mary does that too
‘John works hard and so does Mary.’
We have only investigated dialectal variation for dummy doand anaphoric do and therefore leave variation in causative do constructions undiscussed here.
It may be helpful to discuss dummy doin relation to the verbal positions of the Dutch clause. In chapter 1, three positions for verbs in Standard Dutch were identified: the clause-initial position (8a), the second position of the clause (8b), and a right-peripheral position (8c). There is no evidence that Dutch dialects are different in this respect. In main clauses, the finite verb occurs in first or second position, other verbs occur at the right-periphery (8c). In embedded clauses, all verbs including the finite verb occur in right-peripheral positions (8d), except when there is no complementiser, in which case the finite verb occurs in clause-initial position (8e). The number of verbs in right-peripheral position can be more than two.
(8) a. Eet Jan veel appels? verb-initial
eats John many apples
‘Does John eat a lot of apples?’
b. Jan eet veel appels. verb second
John eats many apples
c. Jan heeft veel appels gegeten. verb-final
John has many apples eaten
‘John has eaten a lot of apples.’
d. Als Jan veel appels gegeten heeft, dan ... verb-final
if John many apples eaten has
‘If John has eaten a lot of apples, then ...
e. Heeft Jan veel appels gegeten, dan ... has John many apples eaten then
‘If John has eaten a lot of apples, then ...
The examples in (8d,e) show that there is a dependency between the verb-initial position and the verb-final position. The position of the finite verb depends on whether the verb-initial position is already filled by a complementiser. Similarly, the position of the main verb in main clauses depends on whether the verb-initial position or verb-second position is already filled by an auxiliary. If so, then the main verb occurs in clause-final position (8c), but if not, the main verb occurs in second (8b) or first position (8a). Put differently, it is impossible to predict the position of a verb in Dutch without taking all three verb positions into consideration.
Given these distributional, but also morphosyntactic and selectional dependencies between auxiliaries and main verbs it is uncontroversial to take all verbs of a clause together as a unit denoting the temporally, aspectually and modally modified event. In generative grammar, the three verb positions are taken to be related by movement.
The occurrence of do in (7b) (Werken doet Jan niet lit. ‘work.inf does John not’) can now be explained as the result of the requirement that a declarative main clause should have a finite verb in the V2 position. In a clause that has no auxiliaries, this would be the main verb. However, in (7b) the infinitival main verb has been fronted into the position preceding the V2 position, a position where finite verbs cannot occur. A semantically empty verb, do, will then rescue the structure. It is merely there to fulfill the V2 requirement. In the next section, we will see that in some varieties of Standard Dutch the main verb does not have to be preposed in order for dummy do to occur, the relevant cases having in common that the main verb fails to be in the required position.
3.1.3.1 Periphrastic do
At first sight periphrastic do in Dutch dialects looks very similar to English do support, another case in which do is just there to fill a syntactic position. Like Dutch, English has three positions for verbs, a clause-initial position used in questions (9a), a position following the subject (9b) and, abstracting away from the different positions of objects etc. in English and Dutch, a position at the right-periphery of the clause (9b).
(9) a. Will John win the race?
b. John will certainly not win the race.
English differs from Dutch in that only auxiliaries, not main verbs can occur in the verb-initial position and the position between subject and negation (10a,b). In declarative sentences, these verb-initial and clause-medial positions simply remain empty, and the main verb occurs in a right-peripheral position (10c). In cases such as (10a,b) where the V1 position or the V2 position needs to be filled, a problem arises. The main verb cannot fill these positions. Just like in the Dutch case (7b) where the main verb was focused and preposed and therefore could not fill the V2 position, semantically empty do saves the structure (10d,e). In emphatic declarative clauses in which the verb second position needs to, but cannot, be supported by the main verb, do can also be inserted (10f). It is clear that do does not denote a dynamic action in this use and hence can be considered semantically empty, as it can co-occur, e.g., with stative raising verbs (10g).
(10) a. * Wins John the race?
b. *John wins not the race.
c. John certainly wins the race.
d. Did John win the race?
e. John did not win the race.
f. John did win the race!
g. It does not seem to rain.
In many varieties of Dutch spoken in the Netherlands, the occurrence of semantically empty do in the V1 or V2 position is an alternative to placing the main verb in these positions. An example is given in (11). The main verb occurs in clause-final position, similar to what happens when other types of auxiliaries fill the V1 or V2 position. It is important to note that (11) is an unmarked, non-emphatic clause, showing that periphrastic do or do support in dialects of Dutch is not restricted to interrogative, negative and emphatic clauses, unlike in English.
Importantly, a conceivable alternative analysis according to which do in sentences such as (11) is the main verb and [de kopjes afwassen] a nominalised constituent cannot be generally correct, since in some dialects de kopjes precedes the adverbs wel even, which would be impossible if [de kopjes afwassen] were one constituent. In test sentence (26), the main verb and its object are separated too.
(11) Ik doe wel even de kopjes afwassen.
I do aff just the cups wash
‘Let me wash the cups.’
For the atlas, we tested periphrastic do in declaratives, interrogatives, imperatives, negative clauses with a dynamic verb and negative clauses with a stative verb. The maps show that declaratives have the widest geographical distribution and that dialects differ with respect to the type of clauses that they allow periphrastic do to occur in, yielding implicational relations between the constructions (cf. section 3.3.2.1.5). They also show that the very limited occurrence of periphrastic do with stative verbs suggests that in the majority of dialects periphrastic do has retained some of its dynamic meaning.
As in the case of perfective and passive auxiliaries, the question arises as to whether the periphrastic do construction can be perfectivised. This is indeed the case in a number of north-eastern dialects. Strangely enough, most of these are dialects that do not have periphrastic do in non-perfective tenses.
(12) Ik heb heel wat lopen gedaan.
I have whole what walk.inf done.pcp
‘I did a lot of walking.’
3.1.3.2 Short do replies
In the previous section we saw that dummy do can be used anaphorically in Standard Dutch, e.g., in coordinative constructions with ellipsis in the second conjunct, as in example (7c) (Jan werkt hard en Marie doet dat ook; lit. ‘John works hard and Mary does that too’). In such cases do+that is a stand-in for the verb phrase in the first conjunct. Unlike English (13a), Standard Dutch does not allow the use of dummy doin short elliptical answers (13b). In a number of dialects spoken in Belgium, short do replies are possible, as exemplified in (13c,d).
(13) a. A. Marie admires Pierre. B. She does not.
b. A. Marie bewondert Pierre B. * Zij doet niet
c. A. Marie zie Pierre geirn. B. Z’en doet.
Mary sees Pierre gladly she.neg does
‘Mary loves Pierre.’ ‘No, she does not.’
d. A. Marie zie Pierre nie geirn. B. Ze doet.
Mary sees Pierre not gladly she does
‘Mary does not love Pierre.’ ‘She does.’
Van Craenenbroeck (2005) has shown that the Belgian short doconstruction should not be identified with the English short do construction. Some properties distinguishing a Belgian short do reply from its English counterpart are: (i) In most dialects, it is only possible as a contradictory reply to declarative (positive or negative) statements; (ii) there are strong restrictions on the types of subjects that can occur in short do replies; (iii) unlike English, it is impossible to use an auxiliary other than do in a short do reply, even when the antecedent clause contains a modal, perfective or aspectual auxiliary; (iv) do always occurs in the present tense, even when the antecedent clause is past tense; (v) only evidential and modal adverbs can occur in short do replies, temporal, frequency and manner adverbs cannot; (vi) a short do reply cannot be a WH question. Nor should the Belgian short do reply be considered a subcase of the periphrastic do construction discussed in the previous paragraph, as there are no dialects in the Dutch language area that have both short do replies and periphrastic do.
3.1.4 Verb positions
This section introduces three dialectal constructions in which, from a Standard Dutch point of view, the finite verb occurs in an unexpected position.
3.1.4.1 Complex initials
As we have seen above, the V1 and V2 position in Dutch and its dialects can only contain one verb per position. Thus, in imperatives the finite auxiliary occurs in the V1 position, and the infinitival main verb remains in the Vfinal position (14a). In some southern dialects, it is possible for both the main verb and the auxiliary to occur in the V1 position (14b,c). This construction is referred to as complex initial, and it is well known to occur frequently in Afrikaans (14d). As opposed to Dutch, Afrikaans allows the auxiliary – verb complex to occur in V2 position in non-imperative clause types (14e).
(14) a. Ga die bestelling nu maar op-halen!
go.fin that order now just up-get.inf
‘Go get that order now!’
b. Gaan haalt die bestelling nu maar op!
go.inf get.fin that order nu just up
‘Go get that order now!’
c. Komen helpt eens!
come. inf help. fin once
‘Come and help me!’
d. Bly sit jy hier! Afrikaans
remain sit you here
‘Remain seated here!’
e. Sy kom help die kinders. Afrikaans
she come help the children
‘She comes to help the children.’
Strikingly, in the Dutch dialect cases (14b,c) theauxiliary has an infinitival form, while the main verb has a finite form, exactly the opposite of the Standard Dutch situation (14a). As Afrikaans lacks a morphological distinction between finite and non-finite verb forms, the status of two verbs remains unclear in that language.
3.1.4.2 Subject between verb and tense suffix
In Dutch and its dialects, the past tense is formed by attaching the suffix –te or –de to the verbal stem, as illustrated in (15a). Since we are in the realm of inflectional morphology here, it should be impossible for constituents to intervene between the stem and the tense suffix. However, in some dialects in Zuid-Holland and Utrecht, a subject pronoun can intervene. This is illustrated in (15b). The phenomenon has also been reported for Dutch child language, mainly with the first person singular pronoun, but other pronouns occur occasionally as well.
(15) a. Gisteren wandel-de die door het park.
yesterday walk.past he through the park
‘He walked through the park yesterday’.
b. Gisteren wandel-die-de door het park.
yesterday walk-he-past through the park
‘He walked through the park yesterday’.
This construction is relevant for a hotly debated issue concerning the clause structure of Dutch, i.e., whether Dutch has two clause-initial verb positions, like English and French (cf. 3.1.3), or one. Proponents of the first view would assume that the verbal stem in (15b) is in V1 position and the past tense morpheme in V2 position, while proponents of the second view would assume that [wandel-die-de] is in the V1 position in its entirety.
3.1.4.3 V2 in embedded clauses
In Standard Dutch and most of its dialects, all verbs occur in right-peripheral position in embedded clauses. In some varieties in and close to Friesland, it is possible for the finite verb to occur in a position following the complementiser and the subject (16a).
(16) (Jan vindt) dat je moet dit niet geloven.
(John finds) that you must this not believe
‘In John’s opinion, you should not believe such things.’
This construction is relevant as well for the debate mentioned in section 3.1.4.2.
3.1.5 Other variation
In this section we introduce two further cases of variation having to do with the verbal system.
3.1.5.1 Infinitival absolute with construction
Examples of the Standard Dutch absolute with construction are given in (17a,b). It involves a constituent consisting of the preposition met ‘with’, a subject and a predicate. In Standard Dutch, the predicate can only be adjectival (17a) or prepositional (17b), verbal predicates are impossible (17c). In many dialects in Belgium it is possible for an infinitive to occur in the absolute with construction, and in a subset of these dialects, the subject then occurs in the nominative, suggesting that met introduces a clause (17d).
(17) a. [Met Cruyff ziek] winnen we niet. Standard Dutch
with Cruyff sick win we not
‘We will not win with Cruyff sick.’
b. [Met Cruyff in het doel] winnen we niet. Standard Dutch
with Cruyff in the goal win we not
‘We will not win with Cruyff in the goal.’
c. * [Met Cruyff ziek te zijn] winnen we niet. Standard Dutch
with Cruyff sick to be win we not
d. Met hij / hem te werken, moest zij thuisblijven Belgian dialects
with he / him to work, must she home-stay
‘She had to stay home because he was working.’
3.1.5.2 Gerund
The final construction discussed here involves a Belgian-Limburgian gerund with a special ending, illustrated in (18a). By comparison, the Dutch equivalent is given in (18b). It is not clear how the Belgian-Limburgian gerund should be analysed (but see section 3.2.6.2 for a suggestion).
(18) a. Loop-entere kwam ik hem tegen
walk-gerund came I him across
‘I came across him while walking.’
b. Loop-end kwam ik hem tegen.
walk-gerund came I hime across
‘I came across him while walking.’
3.2 Historical developments and discussionof the literature
3.2.1 Auxiliary selection
Auxiliary selection is a well-studied phenomenon. The common generalisation is that have is used with intransitive and transitive main verbs, whereas be is used with unaccusative main verbs. A central debate in the literature is about the question whether semantic or syntactic principles govern the choice of the auxiliary (e.g., Hoekstra 1984, 1999; Kayne 1993; Den Dikken 1994; Lieber and Baayen 1997; Sorace 2000; Bentley and Eythórsson 2003).
As mentioned in section 3.1.2.1, some languages, among which (dialects of) Dutch, vary between taking have or be in the context of unaccusative verbs. The generalisation seems to hold that terminative/telic unaccusative verbs select be, whereas durative/atelic unaccusatives select have (cf. Den Dikken 1994:77; Zaenen 1988,1993; Van Valin 1990; Sorace 2000). The preference for using have with intransitive verbs that express duration and be with mutative, non-durative intransitive verbs can already be observed in Middle Dutch and 16th and 17th century Dutch (cf. Stoett 1923; Kern 1912).
For an overview of auxiliary have/be selection in Middle Dutch we refer to Duinhoven (1997: 329-349). For an overview of auxiliary selection by main verb zijn ‘be’ in older stages of Dutch, see Kern (1912) and Kooiman (1954). In Middle Dutch texts, hebben geweest ‘have been’ is found much more frequently than zijn geweest ‘be been’ (Kern 1912:101). From the 15th century on, the use of have decreased relative to the use of be. De Rooij (1988:29) notes that in the 17th century the use of have was typical for Zuid-Holland, and in the 19th century the use of be in written (and spoken) standard language became the norm in the Dutch-speaking language area.
Kooiman (1954) found that in modern Dutch dialects the combination hebben geweest ‘have been’ still occurs frequently and has a wide geographical distribution. Whereas in the main part of the Dutch speaking language area both constructions hebben geweest ‘have been’ and zijn geweest ‘be been’ appear, in the northeastern part of the Netherlands, and in particular in the dialects spoken in Groningen, only the construction with have is found to occur. Compare in this respect Van der Woude (1971) for the dialects spoken in Friesland, which predominantly use have. Kooiman (1954: 214) further notes that there is also dialectal variation in auxiliary selection by other verbs, such as vallen ‘fall’ and komen ‘come’. He suggests that these verbs develop in the same direction as the main verb zijn ‘be’.
An important study is De Rooij (1988) - this study includes a comprehensive overview of the existing literature on the topic. Mainly on the basis of data from a written questionnaire, which was offered to informants of the Meertens Instituut in 1970 and in 1985, and on the basis of a corpus of spontaneous speech from dialect speakers in most areas of the Netherlands, De Rooij (1988) provides a detailed overview of the use of the auxiliaries have and be in combination with the participles of zijn ‘be’ and gaan ‘go’, and of the transitive verbs vergeten ‘forget’ and verliezen ‘lose’ in the Dutch standard language, in older stages of Dutch and in dialect speech. Interestingly, De Rooij found that, even over a period of just 15 years (1970-1985), an increase of approximately 8% can be observed in the use of auxiliary be with geweest in the entire Dutch-speaking language area.
With respect to auxiliary selection of have or be by geweest ‘been’, De Rooij found that the Dutch-speaking language area can be divided into three areas. In the first area, consisting of West- and Oost-Vlaanderen, have is most common. A second relevant area consists of Limburg, the east of Noord-Brabant, and the main part of Gelderland. be is predominantly used in this area. The third area, in which both have and be occur, comprises the remaining part of the Netherlands and Dutch-speaking Belgium.
3.2.2 Auxiliary doubling
Kern (1912:36) and Stoett (1923:181) have noted that the auxiliary doubling construction as found in the dialects already occurred in late Middle Dutch (i.e., 15th century Dutch). Duinhoven (1997:346-348) describes the origin and development of auxiliary doubling as follows. Initially, Middle Dutch hebben ‘have’ occurs as a main verb with the meanings ‘get’ and ‘possess’. A participle occurring with main verb have does not depend on have but has an adverbial function, it describes how the state of possession was achieved. The sentence in (19a), for example, means that they acquired the land by conquering it. In other words, a sentence like (19a) focusses on the resulting state, not on the process.
(19) a. [Si hebben dat lant] gewonnen.
they have.main verb that land won.pcp
‘They have acquired that land by conquering it.’
b. Si hebben dat lant gewonnen gehad.
they have.pl that land won.pcp had.pcp
‘They have acquired that land by conquering it’
Next, the participle in constructions like (19a) is increasingly reanalysed as the main verb, expressing a perfect tense and focussing on the completion of the process. At this point, in late Middle Dutch, auxiliary doubling starts to arise. An example is given in (19b). According to Duinhoven, the auxiliary doubling construction temporarily restores the old construction in which have is the main verb and the participle an adverbial expression. Thus, the construction [[hebben gehad] + past participle] involves the perfect tense of main verb have, focussing on the resulting state. The adverbial past participle (gewonnen ‘won’ in (19)) describes how the state of possession was achieved. Finally, when the construction in (19a) dies out, the auxiliary doubling construction disappears as well.
Auxiliary doubling is found most frequently in the southeastern part of the language area (see map 40b). Similar constructions occur in dialects of German, Italian (Poletto 2006) and French. For Franco-Provençal, it has been claimed that the construction is a passé surcomposé expressing an unspecified moment in the remote past (see Carruthers 1994 for an overview). This possibility was tested for the relevant Dutch dialects by Barbiers, Koeneman and Lekakou (BKL 2007) but no meaning difference was found with the simple periphrastic perfect construction, suggesting that the double occurrence of have is truly redundant semantically.
Another possibility would be that auxiliary doubling involves the perfective of the undative construction in (20). That option was rejected in BKL as well, where it was shown that the dialects with perfective auxiliary doubling do not always have the undative construction. Also, the construction in (20) necessarily involves possessive semantics, which is not the case for auxiliary doubling constructions.
(20) Ik heb de broek gewassen gehad.
I have the trousers washed had.pcp
‘I have had my trousers cleaned.’
3.2.3 Perfective passive
Kern (1912:290-291) and Duinhoven (1997:365) note that the addition of geworden ‘become.pcp’ to the perfect tense of the passive starts in the 15th century and that this construction occurred mostly in eastern texts. The latter is not surprising in view of modern Standard German that systematically has the form worden (a participle like form lacking the prefix ge-) in this construction. Stoett (1923:190) notes that the addition of geworden to the perfective passive was found more often in the irrealis mood than in combination with the indicative mood. According to Duinhoven (1997:366) and Kern (1912), the addition of geweest ‘been.pcp’ to the perfective passive is older and was more common in Middle Dutch than the addition of geworden ‘become.pcp’. See Van der Wal (1986,1988) for an overview of the historical development of the (perfective) passive construction.
The ANS (1997:959-960) notes that additional geworden ‘become.pcp’ in the perfective passive construction is normally not expressed in the standard language and that the presence or absence of geworden might be related to a difference in meaning, i.e., the addition of geworden ‘become.pcp’ attributes to the sentence more the character of a ‘dynamic passive’. Moreover, due to its redundancy, geworden ‘become.pcp’ can be used as a stylistic variant.
The perfective passive construction with geworden/geweest has a widespread geographical distribution. It is found in both the north and the south of the Dutch-speaking language area (De Schutter 1990:24). Weijnen (1971:108) notes that the form with geweest ‘been.pcp’ still occurs very often in southern Dutch, whereas the form with geworden ‘become.pcp’ occurs more in the northern part of the language area - western-Gronings and Frisian. Pauwels (1949) observes that, although geweest ‘been.pcp’ used to be the only auxiliary in the perfective passive in the Aarschot dialect (Vlaams-Brabant), in current dialect speech, both geweest ‘been.pcp’ and geworden ‘become.pcp’ may occur in the construction. Contrary to Pauwels’s observations, Vanacker (1948:42) only documents the form without geweest/geworden for the dialect of Aalst, which is located in Oost-Vlaanderen close to the province of Vlaams-Brabant. On the basis of oral interviews in Friesland and a written questionnaire in Groningen and Drenthe (the results of which are presented on a map) Hoekema (1963) found that the geworden-construction sporadically occurs in Groningen and is missing in Drenthe. In the (north)eastern part of Friesland, however, the construction is frequently found. It should be noted that these results do not correspond to our results as presented on map 41a.
3.2.4 Auxiliary do
In the literature various perspectives have been developed on the historical development of periphrastic do. Especially for English a great deal of attention has been paid to the origin and the development of this construction (cf. Visser 1969; Hausmann 1974; Tieken 1990; Schütze 2004). Moreover, the relation between English (periphrastic) do, German tun and Dutch doen has been the subject of research (cf. Tieken et al. 1998). It is an intriguing fact that the use of the periphrastic auxiliary in negative, interrogative, elliptical and emphatic clauses is standardised in English (the so-called NICE properties of English auxiliaries), whereas do periphrasis in German/Dutch is restricted to dialect speech and colloquial language (cf. par. 3.1.3.1).
In Middle Dutch, periphrastic do was common (cf. Giesbers 1983-84; Weijnen 1971:106; Stoett 1923; Duinhoven 1997:472-478). On the basis of Old and Early Middle Dutch texts, Van der Horst (1998) concludes that the verb do originated as a full verb, that is to say, during the 11th century it did not occur in combination with other verbs. In the course of the 12th and 13th centuries, do (also) developed into an auxiliary. This development is very similar to the development of the English auxiliary do, but whereas in English the auxiliary developed NICE properties, the Dutch auxiliary did not. Rather, out of the early construction a new causative construction evolved (the modern English auxiliary does not have this causative function).
To be able to account for the occurrence of periphrastic do in English, Dutch and German, and at the same time for the fact that these languages have different uses of it, Hausmann (1974) put forward the hypothesis that in each language in which a periphrastic like do occurs, it developed independently. A somewhat similar idea can be found in Tieken (1990), who suggests, on the basis of evidence from English, Dutch, German and Frisian, that the periphrastic use of do probably has emerged independently as a simplification device, which initially was mostly used by children and/or L2 speakers. Causative do, habitual do and aspectual do would have developed out of this periphrastic do construction. Others have argued the opposite: periphrastic do would have developed from causative do. See Visser (1969) and Poussa (1998) for an overview of different approaches to the development of English auxiliary do, and Fischer (1998), Eroms (1998) and Erb (2001) for periphrastic tun in German. For a short discussion of possibly relevant external factors, see Nuijtens (1962). A typology of periphrastic do constructions is given in (Jäger 2006).
The ANS (1997:1021) notes that the occurrence of the periphrastic do construction is restricted to the Netherlands, and that it is mostly used in colloquial or dialectal speech (cf. Nuijtens 1962; Weijnen 1966). Giesbers (1983-84:59) did not find a clear geographical pattern, in contrast to what our data show (cf. map 43b). It should be noted that most of the literature only mentions the occurrence of periphrastic do in declarative clauses. Little is known about periphrastic do in other clause types.
The occurrence of periphrastic do in declarative sentences has been documented in various locations: Giesbers (1983-84) mentions Gronings, Zeeuws, Drents, Rotterdams and Twents (cf. Nuijtens 1962), Stapelkamp (1948:181) mentions the eastern Bommelerwaard (Gelderland) and Tieken (1990:11) mentions Nijmegen (Gelderland), Noordwijk, Hoogmade (Zuid-Holland) and Brabant (cf. Ryckeboer 1986). Tieken (1990:13) combines these studies and provides a map of periphrastic do in the Netherlands.
Periphrastic do does not only occur in dialects of Dutch but also quite frequently occurs in Standard Dutch child language, mainly in declarative and negative sentences (cf. Schaerlaekens 1977; Giesbers 1983-84; Tieken 1990; Duinhoven 1994a:111, Van Kampen 1997). In the literature the construction as it occurs in child language and in the dialects are taken to be similar. Some linguists have claimed that the periphrastic do construction is a strategy to avoid complex syntactic and/or morphological structures (cf. Nuijtens 1962:155-156) and is semantically redundant (cf. Duinhoven 1994a, 1997:477). It is therefore not surprising that the periphrastic do construction has been subject to stigmatisation (cf. Giesbers 1983-4; Tieken 1990; Duinhoven 1994a).
Cornips (1994, 1998) argues that the periphrastic do construction in Heerlen Dutch is not without semantic content but expresses habitual aspect. It is clear, however, that most of the periphrastic do sentences tested for this atlas are not habitual (cf. example 11). Erb (2001) analyses German tun as a semantically empty functional auxiliary.
The construction in which periphrastic do has perfect tense, i.e., in which do has past participle morphology and occurs in combination with the auxiliary have and an infinitive, is attested in the northern language area only. Ter Laan (1953:156) and Cornips (1998:88) mention the occurrence of the construction in Groningen, Pannekeet (1995:375) mentions West-Frisian.
Whereas the distribution of periphrastic do is restricted to the Netherlands, the distribution of short do replies is found only in some southern Dutch dialects, i.e., in the southwestern part of Flanders (cf., Ryckeboer 1986, 1998). Recently, Van Craenenbroeck (2004:127) has given a very detailed overview of the properties of this construction (cf. 3.1.3.2). He argues against the claim frequently found in the literature that the construction is parallel to English VP-ellipsis in combination with do support (e.g., Haegeman (1995, 2002) and proposes that the unpronounced part of the short do reply does not have any internal syntactic structure.
Ryckeboer (1998) presents a rather detailed overview of the dialect-geographical literature on substitute do in short replies (and tag questions) and sketches the historical development of the construction. Short do replies already occurred quite frequently in Middle Dutch (Stoett 1923:155). Whereas the construction used to occur also in Holland, Zeeland and Noord-Brabant (cf. Ryckeboer 1998:70), over time the phenomenon has completely disappeared there and is nowadays only found in some southern Dutch dialects.
In Middle Dutch, short do replies used to occur frequently in various types of constructions: short contradictory replies, tag questions and short questions expressing surprise (Van Craenenbroeck 2004:129). The latter two uses have become more or less extinct and, according to Ryckeboer (1986:327), are clearly archaic and strongly marked, although there are exceptions. One dialect in which they are alive and well is the West-Flemish dialect of Wingene (Willy Vandeweghe, p.c.). In the areas where the use of do in denying replies is still available, decay of the paradigm is found. According to Ryckeboer (1998) the most common form of substitute do in short replies is found in conjunction with the personal pronoun het ‘it’: ’t (en) doet (see also maps 44b,c).
3.2.5 Verb positions
3.2.5.1 Complex initials
Ponelis (1993:330) notes that two clause-initial verbs in imperative clauses is a widely attested phenomenon in early Dutch and dialect speech, besides the frequent occurrence of the construction in different clause types in Afrikaans. Regarding the historical development of the construction, it has been proposed that verbal hendiadys constructions and IPI-constructions (cf. 2.1.2.3) may have served as possible antecedents (De Vos 2005:132). Ponelis (1993:330) suggests that the Dutch complex initial pattern is the result of imperative-clause union.
On the basis of literary Dutch texts from the 16th and 17th century, Paardekooper (1991) shows that the construction in which komen ‘come’, lopen ‘walk’or gaan ‘go’combine with another verb in clause-initial position, occurred at least in Holland. It should be noted that Paardekooper’s examples mostly consist of two adjacent finite verbs. Colinet (1896:147) documents for the dialect of Aalst (Oost-Vlaanderen) the forms komenét (kom eten, ‘come eat’), gonét (ga eten ‘go eat’), gonspéld (ga spelen, ‘go play’) and gonold (ga halen, ‘go get’). According to Colinet this construction is some sort of compound, comparable to gaan eten. This suggests that the Aalst construction consists of an infinitival verb and an imperative. In current Dutch, the construction with an infinitival verb and a finite verb is still found in Vlaanderen, see map 45a.
In the generative literature the most common approach has been to assume that complex initials involve incorporation of the lexical verb into the restructuring verb (Ponelis 1993, Robbers 1997, Den Besten 2002). Recently, De Vos (2005) has argued for a coordination analysis of Afrikaans complex initials.
The unexpected morphology of complex initials in Dutch dialects may seem to support an analysis of Dutch complex initials in terms of coordination of two imperatives as well, with the conjunction en ‘and’ reduced to n and cliticised onto the first imperative, such that gaa-n haalt would not be ‘go.inf get.imp’ but rather ga en haalt ‘goimp and getimp’. However, as Magda Devos and Gunther De Vogelaer p.c. note, a strong argument against such an analysis is that in most of the relevant Flemish dialects the imperative of go is gaat, which in a coordination would yield gaadenhaalt rather than gaan haalt. Also, under such an analysis the cross-dialectal variation attested and the absence of the construction with aspectual verbs such as blijven ‘stay’ and zitten ‘sit’ are unexpected.
3.2.5.2 Subject between verb and tense suffix
The phenomenon of subject intrusion has a very restricted distribution. It is mainly found in some of the dialects in Zuid-Holland (cf. Kooiman 1950,1958; Goeman 1978:294). The occurrence of the construction is also documented by Aussems (1953:87) for the dialect of Culemborg, by Goeman (1984:101) for the dialect of Zoetermeer and by De Wilde-Van Buul (1943:301) and Van Oostendorp (2002:46-47) for Rotterdams. De Boer (1950:122-123) documents the occurrence of a comparable construction with the first person plural enclitic pronoun we for the dialect of Hindeloopen in Friesland. Subject intrusion also occurs in (Standard Dutch) child language (cf. Flikweert 1994). Van Oostendorp (2002) suggests a prosodic explanation of subject intrusion (somewhat similar to the analysis in De Wilde-Van Buul 1943). Kooiman (1950:329) suggests that enclitic die was reanalysed as part of the verb, giving rise to forms like wandel-die-de. Barbiers & Van Koppen (2006) provide an analysis of the phenomenon stated within the generative framework. They argue that the past tense morpheme is in a clause-medial position in which full finite verbs can occur in languages such as English and French. For a typological study of the order of agreement and tense markers see Siewierska (1995).
3.2.5.3 V2 in embedded clauses
Some Germanic varieties allow the embedded verb to appear in a position directly following the subject of the embedded clause, even when a complementiser is present, i.e., the embedded clause introduced by a complementiser exhibits the main clause pattern (cf. De Rooij 1965; Den Besten 1986,1989:137-138). In the Dutch language area, the occurrence of this word order phenomenon is documented for Frisian and Gronings (cf. Overdiep 1932; De Haan and Weerman 1986; De Haan 1990; Van der Meer 1988, 1991). De Haan (1983:45) notes that the possibility of fronting the embedded finite verb in the subordinate clause with a lexical complementiser is strongly dependent on the nature of the matrix verb. The matrix verb should be positive, cannot be accompanied by modals and it should denote a statement, a feeling, and/or an observation (cf. Overdiep 1932).
Overdiep (1932) argues that the subordinate clause with main clause word order introduced by the complementiser dat ‘that’ is in fact a main clause. Moreover, he argues that the two clauses are in a coordinating, instead of a subordinating, relation. More recently, De Haan (1990) has shown that Overdiep’s analysis of this construction can be maintained. Zwart (1993, 2001) shows that embedded V2 and complementiser agreement are in complementary distribution, i.e., complementiser agreement only occurs when the embedded verb is in clause-final position (cf. Van der Meer 1991:71). If complementiser agreement is analysed as the combination of the complementiser in V1 position and the inflectional features of the verb located in the intermediate V2 position and if embedded V2 is analysed as movement of the verb to the intermediate verb position, the fact that complementiser agreement and embedded V2 are mutually exclusive is predicted. This would then provide further empirical evidence for the analysis that there are three different positions for finite verbs in Dutch and its dialects (see also 3.1.4.3).
3.2.6 Other variation
3.2.6.1 Infinitival absolute with construction
In the generative literature, two competing analyses exist of the absolute with construction. Van Riemsdijk (1978) argues that there is no implicit verb present in such constructions, whereas Klein (1983) proposes to analyse them as derived from a sentential structure containing hebbend(e), the present participle of have. Beukema & Hoekstra (1983) provide strong arguments against Klein’s analysis. For a detailed overview of different analyses of the absolute withconstruction, see Haslinger (2007).
While an overt verb is impossible in absolute with constructions in Standard Dutch, De Schutter (1990:18) notes that such constructions frequently occur with to-infinitives in West- and Oost-Vlaanderen (cf. also Hoebeke 1975 for southeastern Flemish; Haegeman 1986 for West-Flemish) and in the main part of the provinces of Antwerpen and Vlaams-Brabant. In the Netherlands, the distribution of the construction is restricted to some locations in Zeeuws-Vlaanderen and Noord-Brabant. On the basis of data from a written questionnaire, De Vriendt (1975) provides a map of the southern Dutch language area that makes a distinction between constructions in which the pronoun has the subject form or the object form. In general, the distribution of the infinitival absolute with construction is comparable to our results as presented on map 46b. De Vriendt found several occurrences in Zeeuws-Vlaanderen and one in Noord-Brabant but for our atlas the construction was only tested in Dutch speaking Belgium.
Hoebeke (1975) provides an analysis of the infinitival absolute with construction in terms of transformation rules. Haegeman (1986) focuses on the nominative case assignment in this construction in West-Flemish and argues that the preposition introduces a finite clause. The analysis of the construction given by Haslinger (2007) is based on data from the dialect of Wambeek (Vlaams-Brabant) and basically states that in dialects that have it the preposition met ‘with’ has grammaticalised into an aspectual marker that is able to license the infinitive and the nominative or accusative subject.
3.2.6.2 Gerund
Verbs of the form verb stem + -entere (or: infinitive + -tere) in the function of present participle are typically found in Belgisch Limburg. It should be noted that in this area this form of the gerund is certainly not the only one occurring (cf. Goossens 1964). In a small research project carried out in 1922-23, the results of which are presented on a modest map, Grootaers (1948) found a lot of diversity in the form of the suffix of the Limburgian present participle, among which -etèère, -entéére, -e(n)teerent, -e(n)tierent. Belemans and Keulen (2004:70-71) note that the gerund with suffix -(en)tère and ending in -nt (or -nd) was the normal form in the 19th and a large part of the 20th century, but forms ending in -nt are no longer used nowadays. In the SAND-project, the Limburgian gerund was not tested for variation in the form of the suffix. Belemans and Keulen (2004:70-71) note that the distribution of the gerund is restricted to the south of Belgisch Limburg, but it used to be more widespread, that is to say, some more northern dialects used to have it. De Vries (1923) suggests that the Limburg gerund should not be analysed as a combination of the infinitival form of the verb and some mysterious suffix -tere, but rather involves the combination of an infinitive plus the Middle Dutch noun tiere meaning ‘manner’.
3.3 Discussion of the maps
3.3.1 Perfective auxiliaries
3.3.1.1 Auxiliary selection (map 40a) (map in dynaSAND)
Map 40a presents the distribution of auxiliary have with the perfect tense of the verb zijn ‘be’, the verb tegenkomen ‘come across’ and the unaccusative verb vallen ‘fall’. These verbs would select the auxiliary be in the standard language.
(21) a. Die rare jongen ben/heb ik mee naar de markt geweest
that weird boy am/have.fin I with to the market
‘I have been to the market with that weird boy.’
b. Ik heb hem gisteren tegengekomen. I have.fin him yesterday across-come.pcp
‘I came across him yesterday.’
c. Geloof je niet dat hij met de fiets gevallen heeft?
believe you not that he with the bike fallen.pcp has.fin
‘Don’t you believe that he fell off his bike?’
Sentence (21a) was tested in all locations of the oral interviews. It also tests preposition stranding, a phenomenon that is irrelevant to the issue of auxiliary selection. In dialects that do not allow preposition stranding (cf. SAND 1, map 93b), (21a) was tested in a different way to get the auxiliary selection data. Sentence (21b) was tested in a total of 251 locations (the oral and telephone interviews taken together). The informants were asked whether it occurs in their local dialects and how it should be translated. Sentence (21c) was part of the telephone interviews in most sampling points (245 locations in total).
Map 40a shows that auxiliary have occurs most frequently with main verb be in the perfect tense (21a). It was found in 85 locations. This sentence also has the widest geographical distribution. Besides its frequent occurrence in West-Vlaanderen, Oost-Vlaanderen and the southern part of Frans-Vlaanderen, the construction also occurs in the northern Dutch language area - frequently in Friesland and occasionally in Overijssel and Noord-Holland. These data coincide with the data on this construction as presented by De Rooij (1988), cf. section 3.2.1.
The distribution of the sentence in which the verb tegenkomen ‘come across’ selects the auxiliary have - sentence (21b) - is largely restricted to Vlaanderen (68 occurrences in total). It occurs in almost every sampling point in West- and Oost-Vlaanderen, and it is found frequently in Antwerpen and Vlaams-Brabant. In the Netherlands the construction is only found sporadically.
The number of occurrences of have with the perfect participle of vallen ‘fall’ (21c) is even smaller (46). Whereas sentences like (21a) and (21b) occur fairly often in Antwerpen and Vlaams-Brabant, little to no instances of sentence type (21c) are found in these provinces. The distribution of type (21c) is restricted to Oost-Vlaanderen and West-Vlaanderen, where it occurs in almost every sampling point.
The general picture that emerges is that selection of the auxiliary hebben ‘have’ by the perfect tense of zijn ‘be’, tegenkomen ‘come across’ and vallen ‘fall’, frequently occurs in Oost-Vlaanderen and West-Vlaanderen. Furthermore, selection of hebben ‘have’ by the perfect tense of the verb zijn ‘be’ is the only variant that occurs commonly in (the northern part of) the Netherlands.
3.3.1.2 Auxiliary doubling (map 40b) (map in dynaSAND)
Map 40b shows the geographical distribution of auxiliary doubling, as exemplified in the sentences in (22a-c). Sentence (22a) was part of the oral interviews in
Belgisch Limburg, Vlaams-Brabant, and the eastern part of the province Antwerpen only (39 sampling points). Sentences (22b) and (22c) on the other hand were part of the telephone interviews in almost all sampling points (245 locations). The sentences were tested by asking the informants whether they occurred in their dialects.
(22) a. Ik heb dat gezegd gehad.
I have.fin that said.pcp had.pcp
‘I have said that.’
b. Ik heb vandaag nog niet gerookt gehad.
I have.fin today yet not smoked.pcp had.pcp
‘I have not smoked yet, today.’
c. Ben je met die fiets gevallen geweest?
aAre.fin you with that bike fallen.pcp been.pcp
‘Have you fallen with that bicycle?’
The distribution of participle doubling is largely restricted to the southeastern part of Dutch speaking Belgium, including Belgisch Limburg, Vlaams-Brabant and the eastern part of Antwerpen. In the Netherlands, the construction is only found sporadically. Interestingly, although sentence (22a) was only tested in 39 locations, whereas the other sentences were part of the interviews in almost all locations, the number of instances of (22a) is the highest, 29 versus 22 and 12. The relatively small number of occurrences of sentence (22c) might be attributed to the fact that the verb in this sentence – vallen ‘fall’ – is intransitive, whereas the verbs in the other two sentences are not.
If we compare map 40b with the map in Barbiers, Koeneman and Lekakou (BKL 2007), we find a discrepancy in the distribution and the number of cases of the construction in Noord-Brabant. Whereas our data show little to no auxiliary doubling in this province, the map provided by BKL shows that doubling is attested in 6 locations in the central and eastern parts of Noord-Brabant. BKL also found doubling in the western part of Antwerpen, an area for which we have not attested doubling.
3.3.1.3 Perfective passive (map 41a) (map in dynaSAND)
Map 41a sketches the distribution of the perfective passive construction. In Standard Dutch the perfect tense of the passive is formed with the verb be in combination with a past participle. In some Dutch dialects this construction may be accompanied by an additional perfective auxiliary: geworden or geweest. The phenomenon is exemplified in the sentences in (23).
(23) a. Het huis is verkocht geworden.
the house is.fin sold.pcp become.pcp
‘The house has been sold.’
b. Het huis is verkocht geweest.
the house is.fin sold.pcp been.pcp
‘The house had been sold.’
In the telephone interviews both sentences were presented to the informants in almost all sampling points, 245 and 247 locations respectively - data from Frans-Vlaanderen are missing. In section 3.1.2.3 it was mentioned that sentence (23b) is also possible in Standard Dutch in a specific semantic context when it means that the house is no longer sold, i.e., it is for sale again. In the telephone interviews the informants had to indicate whether the sentence occurs in their local dialects, but crucially, the informants were not always asked about the meaning of the construction, and therefore, the status of the geweest ‘been’ construction is unclear.
The construction verkocht geweest ‘sold.pcp been.pcp’ is attested 120 times and is found in most locations in Dutch-speaking Belgium (cf. Weijnen 1971:108), with the exception of the province of Antwerpen and the central part of Vlaams-Brabant. In the Netherlands this construction is found in the (south)east of Noord-Brabant and, roughly speaking, along the eastern border, i.e., attestations are found in the south of Gelderland, in the Achterhoek, in Twente and in the east of Drenthe and Groningen. The construction occurs sporadically in the remaining part of the Netherlands.
The amount of occurrences of the perfective passive construction drastically decreases when the participle geworden ‘become.pcp’ instead of geweest ‘been’ is involved: we found 48 cases. The geographical distribution of the geworden ‘become.pcp’ construction is not very clear, although we find a concentration of locations in Belgisch Limburg, the east of Vlaams-Brabant and the northeast and the south of Dutch Limburg in which both forms verkocht geweest and verkocht geworden are attested. In Belgium and southern Dutch Limburg the generalisation seems to hold that geworden is found in a particular location only when the form geweest is attested there as well. In the Netherlands, on the other hand, we find some locations in which only the form geworden is attested.
The results do not coincide with those presented in Hoekema (1963). A striking difference with the map in Hoekema is the frequent occurrence of the geworden-construction in the (north)eastern part of Friesland. Our data show a total absence of the perfective passive with additional perfective auxiliary in Friesland (the geworden-construction does occur in Lies, Terschelling).
3.3.2 Auxiliary do
3.3.2.1 Periphrastic do
In this section we consider the occurrence of periphrastic do in different clause types. The first five maps deal with a finite form of the verb do in combination with an infinitival verb, whereas the sixth map involves do with participial morphology. The geographical distribution of the do+infinitive construction in declarative clauses is presented on map 41b, in interrogative clauses on map 42a and in imperative clauses on map 42b. Map 43a finally, presents the distribution of periphrasticdo in negative clauses with dynamic and stative verbs. Map 43b combines these four maps. The geographical distribution of the participial auxiliary do is presented on map 44a.
3.3.2.1.1 Declarative clause (map 41b) (map in dynaSAND)
Map 41b shows the distribution of sentence (24), in which we have an instance of periphrastic do in a declarative clause. The sentence was presented to the informants in the local dialect in every sampling point (258 in total). The informants were asked to indicate whether the sentence occurs in their local dialect.
(24) Ik doe wel even de kopjes afwassen.
I do.fin affirm just the cups off-wash.inf
‘I’ll do the washing up.’
It is immediately clear from the map that the construction is completely absent in the northern dialects - Friesland, Groningen and Drenthe - and in Belgium. We do find the construction in West-Friesland, however. These results do not coincide with the map presented in Tieken (1990:13), which shows that do periphrasis in declarative clauses occurs in Groningen and Drenthe. In the remaining part of the Dutch-speaking language area the construction is attested 61 times. It seems to be the case that periphrastic do in a declarative clause becomes less available the more we go to the north of the Netherlands, in other words, the density of the occurrence of periphrastic do in this clause type is the highest in the southern part of the Netherlands.
3.3.2.1.2 Interrogative clause (map 42a) (map in dynaSAND)
Map 42a presents the geographical distribution of periphrastic do in interrogative clauses. Because the written pilot study had shown that periphrastic do was completely absent in Belgium, sentence (25) was only tested in the sampling points in the Netherlands, which means that the construction was part of the oral interviews in 156 locations.
(25) Doet Marie elke avond dansen?
does.fin Marie Mary every night dance.inf
‘Does Marie Mary dance every night?’
The construction is attested 31 times, and its geographical distribution is mostly restricted to the southern part of the Netherlands consisting of Limburg, Brabant, the islands of Zuid-Holland and Zeeland, with the exception of Zeeuws-Vlaanderen. This region together with Vlaanderen does not seem to exhibit the periphrastic do construction. The construction also occurs in the southwestern part of Gelderland and it is found in Twente.
3.3.2.1.3 Imperative clause (map 42b)(map in dynaSAND)
Periphrastic do in imperative clauses is exemplified in (26). This sentence was presented to the informants in the local dialect in every sampling point in the Netherlands, yielding a total of 157 locations. On map 42b we see that the geographical distribution of this construction is parallel to the distribution of periphrastic do in interrogative clauses. It occurs mainly in Limburg, Noord-Brabant, the south of Zuid-Holland and Zeeland (except for Zeeuws-Vlaanderen) and it is occasionally found in the south and east of Gelderland and one single instance in Twente.
(26) Doe het brood even snijden!.
do.imp the bread just slice.inf
‘Could you slice the bread, please.!’
3.3.2.1.4 Negative clause (map 43a) (map in dynaSAND)
The declarative sentences in (27) and (28) both involve do support in combination with the negative element niet ‘not’. The difference between the two sentences is the nature of the main verb. Although they both express durative/atelic aspect, the verb voetballen ‘play-football’ in (27) is a dynamic verb expressing an action, whereas branden ‘burn’ in (28) is a stative verb expressing a state. Both in the oral and the telephone interviews the sentences were presented to the informants in a total of 184 locations of which most are located in the Netherlands. The sentences were also tested in most of the sampling points in Oost-Vlaanderen. The informants were asked to indicate whether the sentences are used in their local dialects.
(27) De kinderen doen hier niet voetballen.
the children do.fin here not playfootball-playfootball.inf
‘The kids do not play soccer here.’
(28) De lamp doet niet meer branden.
the lamp does.fin not anymore burn.inf
‘The lamp does not work anymore.’
Map 43a shows in which dialects either one or both of the sentences in (27)-(28) occurs. As becomes clear from the map, there is an apparent difference between the two sentences with respect to the distribution and the number of occurrences. The construction involving a dynamic verb is attested 25 times, whereas the construction with a stative verb is found only 13 times. Moreover, the latter construction is only found in locations where periphrastic do in combination with a dynamic verb is found, the exceptions being Liempde in Brabant and Druten in Gelderland where we only find sentence (28). This result (i.e., more instances of the sentence with a dynamic verb) suggests that do still has some of its dynamic meaning.
Interestingly, when we compare the distribution and the number of occurrences of sentence (27) with sentence (24), which also involves a verb that expresses an action (map 41b), we observe a striking difference: the number of cases with a positive declarative sentence is much higher. There are two differences between (24) and (27): the sentence in (27) is negative and the most salient interpretation is habitual. One of these factor could explain the different numbers of locations. It is also interesting to note that the dialects outside of the southern core area, i.e., in Noord-Holland and Zuid-Holland, Utrecht and Overijssel, do not have periphrastic do on map 43a.
3.3.2.1.5 Synthesis periphrastic do (map 43b)
Map 43b is a synthesis of periphrastic do in different clause types, combining the maps 41b, 42a,b and 43a. do periphrasis occurs exclusively in the Netherlands, and the phenomenon shows the highest density in Limburg, Noord-Brabant, Zeeland with the exception of Zeeuws-Vlaanderen, and the southeastern part of Gelderland. We find that periphrastic do in declarative clauses is the most frequently occurring variant: it is attested 61 times. Comparing this with periphrastic do in interrogative and imperative clauses, we observe that the number of occurrences of latter is considerably lower: 31 and 34 versus 61 attestations. Periphrastic do in combination with a stative verb is attested even less. We found only 13 occurrences.
The map shows some implicational relations between the different clause types. First, if a dialect has do periphrasis in a negative clause involving a stative verb, then it also allows the construction with a dynamic verb (two exceptions). Second, if a dialect has the construction do+negation+dynamic verb, it is quite likely to also allow do-periphrasis in a declarative or an interrogative clause (two exceptions). Third, do-periphrasis in an interrogative clause occurs in a dialect only when the construction occurs in another clause type (one exception only). Finally, if a dialect allows the do+infinitive construction in an imperative clause, then it most likely also has the do+infinitive construction in a declarative clause. There are some exceptions to this latter implication in the southern part of Limburg. In the remaining part of the language area there are two exceptions. For convenience, the four implications are given in (29).
(29) a. negative+stative > negative+dynamic
b. negative+dynamic > declarative/interrogative
c. interrogative > declarative/negative/imperative
d. imperative > declarative
3.3.2.1.6 Perfective periphrastic do (map 44a)
In addition to different clause types in which finite do appears as an auxiliary and combines with an infinitival verb, the SAND questionnaire also included a sentence in which periphrastic do has participial morphology and combines with the perfective auxiliary have and an infinitival verb; this is exemplified in (30).
(30) Ik heb heel wat lopen gedaan.
I have.fin whole what walk.inf done.pcp
‘I have walked quite a lot.’
The Standard Dutch variant of (30) would be Ik heb heel wat gelopen (lit.: ‘I have whole what walked’). Given the results of the written pilot study, sentence (30) was part of the oral interviews in 86 sampling points located exclusively in the Netherlands and mostly along the eastern border of the language area.
Map 44a shows that we found 21 occurrences of the construction: in Groningen, Drenthe, Overijssel and occasionally in Gelderland. As mentioned in paragraph 3.2.4, Pannekeet (1995) noted that the construction also occurs in West-Frisian. Unfortunately, we do not have data for this area.
Comparing this distribution with the geographical distribution of do periphrasis in the present tense (cf. map 43b), we find a striking difference. Putting some sampling points in Gelderland and Twente aside, the two areas where the constructions are found, do not overlap, strongly indicating that we need to distinguish these two types of do support.
3.3.2.2 Short do replies
In this section two maps are presented that deal with so-called short do replies: short elliptical answers that consist of a pronoun accompanied by an inflected form of the verb do. Map 44b presents the distribution of contradicting replies to a negative statement and map 44c shows the distribution of contradicting replies to a positive statement. Both maps are primarily based on the following test sentences.
(31) A zegtsays: Hij zal niet komen B deniesontkent: Hij doet.
A says he will not come B denies he does
‘A: He won’t come. B: Yes, he will.’
(32) a. A sayszegt: Hij slaapt B deniesontkent: Hij en doet.
A says he sleeps B denies he negpart does
b. A sayszegt: Hij slaapt B deniesontkent: ‘t en doet.
A says he sleeps B denies it negpart does
Both: ‘A: He is sleeping. B: No, he is not.’
These sentences were tested in the entire Dutch-speaking part of Belgium. Data from the oral and the telephone interviews are combined here.
3.3.2.2.1 Denial of negative statement (map 44b) (map in dynaSAND)
Map 44b presents the distribution of short do replies as denials of negative statements. The geographical distribution of the phenomenon is restricted to Dutch- speaking Belgium. We find the construction frequently in Frans-Vlaanderen, West-Vlaanderen, Oost-Vlaanderen and Vlaams-Brabant, and the construction is found sporadically in the province of Antwerpen.
In contradicting a negative statement, short do replies without a negation element are attested very frequently (98 instances), whereas contradicting replies with the negation element en are found only 6 times. The latter construction is roughly found in an area around the Schelde river in Oost-Vlaanderen and a single case is found in the north of West-Vlaanderen. Interestingly, in 5 of the 6 locations where this construction is found, the contradicting reply without a negative element is also attested.
The variant of short do replies that occurs most often in the context of a negative statement is toetoet or ’t doet (cf. Ryckeboer 1998). We found 64 cases. The variant with the third person singular masculine pronoun (hij doet ‘he does’) is found in 34 locations, and in general it seems to be the case that this latter variant only occurs in locations where the form toetoet or ’t doet is found. Short do replies with ‘t ‘it’ do not require a third person singular subject in the preceding clause, they can also be used when this clause contains a first or second person subject. The large majority of locations that have hij doet is found in Frans-Vlaanderen, West-Vlaanderen and Oost-Vlaanderen.
In a limited number of locations short do replies can also be used to confirm a negative statement (A: He won’t come. B: That’s right, he won’t.). These include Ninove, Strijpen, Aalter and Eeklo in Oost-Vlaanderen and Kortrijk in West-Vlaanderen. In all of these locations Hij en doet ‘he negpart does’ is used for this purpose. In Ninove and Strijpen, ‘t en doet ‘it negpart does’ occurs in addition.
3.3.2.2.2 Denial of positive statement (map 44c) (map in dynaSAND)
Map 44c shows the geographical distribution of short do replies as denials of positive statements. This distribution is restricted to the provinces Frans-Vlaanderen, West-Vlaanderen and Oost-Vlaanderen. As opposed to short do replies as denials of a negative statement, presented on map 44b, short do replies as denials of positive statements predominantly consist of a pronominal element, a form of the verb do, and a negative element: usually en negpart, sometimes niet ‘not’ or both. Forms without negation, ’t doet or toetoet ‘it does’, are attested in 6 sampling points, of which 5 are located in Oost-Vlaanderen.
The forms hij en doet ‘he negpart does’ and ’t en doet‘it negpart does’ are attested most often, 32 and 19 instances respectively. In the northeastern part of West-Vlaanderen we found some isolated occurrences with a specific form. In Oostende, a form with the reduced pronoun’t ‘it’ and two instances of negation are found: ’t en doet niet ‘it negpart does not’. A similar construction is found in Diksmuide, but here the pronoun is hij: hij en doet niet ‘he negpart does not’. In Veurne we found the form ’t doet niet ‘it does not’, and in Poelkapelle we found hij doet niet ‘he does not’. In the latter two locations, short do replies without a negative element are also attested.
A relatively large number of dialects also uses short do replies to confirm a positive statement (meaning: ‘Indeed’), in most cases with the form ’t doet ‘it does’. In Brabant this includes the locations Kampenhout and Overijse, in Oost-Vlaanderen Zaffelare, Eeklo and Aalter, and in West-Vlaanderen Nieuwpoort, Tielt, Moorsele en Menen. The latter two also have hij doet ‘he does’ with this meaning. This form is also found in three locations in Oost-Vlaanderen: Sint-Laureins, Ninove and Geraardsbergen.
When we compare the geographical distribution of short do replies (map 44b and 44c) with the periphrastic do construction discussed in 3.3.2.1, we see a striking difference in that no dialect has both constructions. This suggests that the Flemish short do replies are not a subcase of the periphrastic do construction.
3.3.3 Verb positions
3.3.3.1 Complex initials (map 45a)(map in dynaSAND)
The imperative clauses in (33) and (34) both consist of two clause-initial verbs - an infinitival aspectual verb and a finite verb. Sentence (33) was tested in 89 sampling points: in Frans-Vlaanderen, West-Vlaanderen, Oost-Vlaanderen, Zeeuws-Vlaanderen, Noord-Brabant, the western part of Vlaams-Brabant and the province of Antwerpen except for the eastern area. In the oral interviews, sentence (34), or some variant of this sentence, was included as a control question. It was presented to the informants in 34 locations spread over West- and Oost-Vlaanderen.
(33) Gaan haalt die bestelling nu maar op!
‘Go get that order, now!’
(34) Komen helpt een keer!
come.inf help.fin a time
‘Come and help me for once!’
Map 45a shows that the construction in (33) is attested 41 times and is mainly found in Frans-Vlaanderen, West- and Oost-Vlaanderen. Although sentence (34) was tested in a smaller number of locations than (33), there seems to be a difference between the distribution and the number of occurrences of the two sentences. In all the locations in West- and Oost-Vlaanderen in which (33) occurred, sentence (34) was part of the oral interviews, but it is has only 9 instances and mostly occurs in the area around and to the north of the Schelde river. In addition, it has two isolated attestations: one in the south of West-Vlaanderen, and another in the south of Oost-Vlaanderen.
3.3.3.2 Subject between verb and tense suffix (map 45b) (map in dynaSAND)
Sentence (35) shows the phenomenon of subject intrusion: the pronominal subject appears in a position between the verb and its past tense suffix.
(35) Gisteren wandel-die-de door het park.
yesterday walk-he-edpast through the park
‘He walked through the park yesterday.’
In the oral interviews, sentence (35) was tested in most locations (202 in total) - with the exception of Frans-Vlaanderen and the main parts of the provinces Antwerpen and West-Vlaanderen - by asking the informants whether the sentence occurs and if so, how it should be translated into their local dialects. It becomes immediately clear from map 45b, that the geographical distribution of this construction is extremely limited. It occurs in 3 locations in Utrecht and in 3 locations in Zuid-Holland only. De Boer (1950:122-123) documents the occurrence of subject intrusion with the first person plural enclitic pronoun we for the dialect of Hindeloopen, spoken in the western part of Friesland. We do not have data for this area.
3.3.3.3 V2 in embedded clauses (map 46a) (map in dynaSAND)
Sentence (36) contains the subordinate clause dat je moet zulke dingen niet geloven ‘that you must such things not believe’ in which the finite modal verb moet occurs in a position directly following the subject, instead of in a clause-final verb cluster, as would be the case in Standard Dutch: dat je zulke dingen niet [moet geloven].
(36) Jan vindt dat je moet zulke dingen niet geloven.
John thinks that you must.FIN such things not believe.INF
‘Jan John thinks that you should not believe such things.’
From the literature it is known that the distribution of the construction is restricted to the northern provinces. Therefore, sentence (36) was part of the oral interviews in every sampling point to the north of the major rivers in the middle of The Netherlands, yielding a total of 89 locations. The informants had to indicate whether the sentence occurs in their own local dialects. On map 46a we see that the construction is only attested 8 times and that its geographical distribution is restricted to the northern provinces including Friesland, Groningen and Overijssel.
3.3.4 Other variation
3.3.4.1 Infinitival absolute with construction (map 46b) (map in dynaSAND)
Sentence (37) involves a with-infinitive, a special instance of the absolute with construction. The preposition met ‘with’ is followed by a pronominal subject, which, as indicated in (37), can have either the subject- or the object form (hij or hem), and the pronoun in turn is followed by a to-infinitive, te werken. There is abundant evidence in the literature that the construction is mostly found in Dutch speaking Belgium (cf. par. 3.2.5.4), and therefore, the sentence was tested in Belgium exclusively: in almost every location with the exception of Limburg (89 locations in total).
(37) [Met hij / hem te werken] moest zij de hele dag thuis blijven.
with he.nom/him.acc to work.inf must she she the the whole day day at-home blijven.
‘Because he was workingworked, she had to stay home
the whole dayall
day long.’
Map 46b sketches the geographical distribution of sentence (37). Since we did not systematically test the variation between nominative and accusative subjects of the absolute withconstruction, this variation is not represented on this map. The map shows that the construction is most frequently found in Oost-Vlaanderen and the east, southeastern and southwestern West-Vlaanderen. The construction furthermore occurs in the middle-south and eastern parts of Vlaams-Brabant, occasionally in the southwest of Belgisch Limburg and sporadically in the province of Antwerp. On De Vriendt’s (1975) map the construction also occurs in Zeeuws-Vlaanderen, an area for which we do not have data.
3.3.4.2 Gerund (map 46c) (map in dynaSAND)
Sentence (38) involves a special form of the Dutch present participle, lopentere, a form in which the verbal stem loop ‘walk’ is followed by –entere ‘-ing’.
(38) Lop-en-tere kwam ik hem tegen.
walk-ing came I him acrossagainst
‘I met him walking.’
It was clear from the written pilot study and the literature that this form of the gerund is typical for Belgisch Limburg. Therefore, in the oral interviews sentence (38) was tested in the sampling points in Belgisch Limburg only (13 locations in total). The sentence was tested by asking the informants to indicate whether it occurs in their local dialect. The suffix of the gerund was not tested for further morphological variation.
As map 46c shows, the gerund lopentere is attested 9 times in the southeast of Belgisch Limburg. If we compare this result with the data that Grootaers (1948) found on the basis of a research project from 1922-23, the main difference we find is that the Limburgian gerund was more wide-spread. On Grootaers’s map there are also attestations of the gerund in the (south)west of Vlaams-Brabant and the gerund was found frequently in the south of Dutch Limburg. If the results of the written pilot are reliable, these differences seem to indicate that this specific form of the gerund is rapidly disappearing.
3.4 Bibliography auxiliaries and finite verb positions
ANS (1997). W. Haeseryn et al. (eds.) Algemene Nederlandse Spraakkunst.
Groningen: Wolters-Noordhoff.
Aussems, Th. (1953). Klank- en vormleer van het dialect van Culemborg. Assen.
Barbiers, S. and J.M. van Koppen (2006). ‘Een plaats voor tijd in het middenveld van
het Nederlands’. Taal en & Tongval, themanummer 17, 24-39.
Barbiers, S., O. Koeneman & M. Lekakou (2007). ‘Syntactic Doubling and the
structure of chains’. To appear in Proceedings of WCCFL 2007.
— Barbiers, S., O. Koeneman & M. Lekakou (2007). ‘Perfect Doubling’. Handout of a
talk presented at the Grand Meeting for
Belemans, R. and R. Keulen (2004). Belgisch-Limburgs. Taal in stad en land. Tielt:
Uitgeverij Lannoo.
Bentley, D. and T. Eythórsson (2003). ‘Auxiliary selection and the semantics of
unaccusativity’. Lingua 114: 447-471.
Beukema, F. and T. Hoekstra (1983). ‘Met met PRO of met zonder PRO’. De nieuwe
taalgids 76. 532-548.
Boer, B. de (1950). Studie over het dialect van Hindeloopen. Assen.
Carruthers, J. (1994). ‘The passé surcomposé régional: Towards a definition of its
function in contemporary spoken French’. Journal of French Language Studies 4:2, 171-190.
Colinet, Ph. (1896). ‘Het dialect van Aalst’. Leuvensche Bijdragen 1.
Cornips, L. (1994). ‘De hardnekkige vooroordelen over de regionale doen+infinitief-
constructie’. Forum der Letteren 35,4: 282-294.
— Cornips, L. (1998). ‘Habitual doen in Heerlen Dutch’. In I. Tieken-Boon van
Ostade, M. van
Craenenbroeck, J. van (2004). Ellipsis in Dutch Dialects. Ph. D. Dissertation, Leiden
University.
De Haan, G.J. (1983). ‘The position of the finite verb in modern West Frisian’. In:
Danielsen, N. et al. (eds.), Friserstudier III. Odense: Odense Universitetsforlag. 37-48.
De Haan, G. and F. Weerman (1986), ‘Finiteness and Verb Fronting in Frisian’. In H.
Haider and M. Prinzhorn (eds.), Verb Second Phenomena in Germanic Languages. Dordrecht: Foris.
De Haan, G. (1990). ‘Hoofd- en bijzinnen: Traditie en progressie’. Handelingen van
het 40ste Nederlandse Filologencongres, 203-213. The Hague.
Den Besten, H. (1986). ‘Decidability in the Syntax of Verbs of (Not Necessarily)
West-Germanic languages’. Groninger Arbeiten zur germanistischen Linguistik 28, 232-256. Reprinted in Den Besten (1989).
— Den Besten, H. (1989). Studies in West Germanic Syntax. Ph. D. Dissertation,
University of Tilburg.
—Den Besten, H. (2002). ‘Khoekhoe syntax and its implications for L2 acquisition of
Dutch and
Den Dikken, M. (1994). ‘Auxiliaries and Participles’. In M. Gonzalez (ed.),
Proceedings of NELS 24, GLSA, Amherst, 65-79.
De Rooij, J. (1965). Als-Of-Dat. Ph. D. Dissertation, University of Nijmegen. Assen:
Van Gorcum & Comp. N.V.
— De Rooij, J. (1988) Van hebben naar zijn. Het gebruik van hebben en zijn in de
voltooide tijden
De Schutter, G. (1990). ‘De studie van de syntaxis van Nederlandse en Friese
dialecten’. Taal en & Tongval, Themanummer 3, 10-47.
De Vos, M.A. (2005). The syntax of pseudo-coordination in English and Afrikaans.
Ph. D. Dissertation, Leiden University.
De Vriendt, S. (1975). ‘Een nominativus-cum-infinitivo in Zuidnederlandse
dialecten’. In: R. Jansen-Sieben, S. de Vriendt & R. Willemyns (eds.), Spel van Zinnen. Album A. van Loey. Brussel: Editions de l’Université de Bruxelles. 71-78.
De Vries, W. (1923). ‘Het Zuidlimburgs ptc. op –enteere’. Tijdschrift voor
Nederlandse taal- en letterkunde 42. 79 .
De Wilde-van Buul, G. (1943). ‘Het enclitisch pronomen personale van de tweede en
derde persoon singularis in het Rotterdams’. Tijdschrift voor Nederlandse Taal- en Letterkunde 62, 290-301.
Duinhoven, A.M. (1994a). ‘Het hulpwerkwoord doen heeft afgedaan’. Forum der
Letteren 35 (2): 110-131.
— Duinhoven, A.M. (1994b). ‘Doen en laten in beweging’. Forum der Letteren 35 (4),:
272-276.
— Duinhoven, A.M. (1997). Middelnederlandse syntaxis, synchroon en diachroon. Deel 2. De Werkwoordgroep. Groningen: Martinus Nijhoff.
Erb, Marie Christine (2001). Finite auxiliaries in German. Ph. D. Dissertation,
Katholieke Universteit Brabant.
Eroms, H.W. (1998). ‘Periphrastic tun in present-day Bavarian and other German
dialects’. In I. Tieken-Boon van Ostade, M. van der Wal and A. van Leuvesteijn (eds.), DO in English, Dutch and German. History and present-day variation, Amsterdam/ Münster: Stichting Neerlandistiek/ Nodus Publikationen. 139-157.
Fischer, A. (1998). ‘Tunperiphrasis in Early New High German’. In I. Tieken-
Boon van Ostade, M. van der Wal and A. van Leuvesteijn (eds.), DO in English, Dutch and German. History and present-day variation, Amsterdam/ Münster: Stichting Neerlandistiek/ Nodus Publikationen. 121-138.
Flikweert, M. (1994). Wat hoor-ik-te jou zeggen? Over het verschijnen van een
pronomen tussen werkwoordsstam en flexie-morfeem, in kindertaal en enkele Nederlandse dialecten. Unpublished MA Thesis, Utrecht University.
Giesbers, H. (1983-84). ‘“Doe jij lief spelen?” Notities over het perifrastisch doen’.
Mededelingen van de Nijmeegse Centrale voor Dialect- en Naamkunde 19, 57-64.
Goeman, A. (1978). ‘COMP-agreement?’ In: W. Zonneveld and F. Weerman (eds.)
Linguistics in the Netherlands 1977-1979. 291-306.
— Goeman, A. (1984). Klank- en vormverschijnselen in het dialect van Zoetermeer.
Amsterdam.
Goossens, J. (1964). ‘Een tweede Limburgs gerundium’. Taal en & Tongval 16. 103-
106.
Grootaers, L. (1948). ‘Het verspreidingsgebied van het Limburgs gerundium’.
Miscellanea Gessleriana. 561-565.
Haegeman, L. (1986). ‘INFL, COMP and nominative case assignment in Flemish
infinitivals’. In Muysken, P. and H. van Riemsdijk (eds.), Features and projections. Dordrecht: Foris. 123-137.
—Haegeman, L. (1995). The Syntax of Negation. Cambridge: Cambridge University
Press.
—Haegeman, L. (2002). ‘West-Flemish negation and the derivation of SOV-order in
West Germanic’. Nordic Journal of Linguistics 25. 154-189.
Haslinger, I. (2007). The Syntactic Location of Events; Aspects of Verbal
Complementation in Dutch. Ph. D. Dissertation, University of Tilburg.
Hausmann, R. B. (1974). ‘The Origin and Development of Modern English
Periphrastic Do’. In J. M. Anderson and C. Jones (eds.). Historical Linguistics. Vol. I. Amsterdam: North-Holland Publishing Co. 159-189.
Hoebeke, M. (1975). ‘Voorzetselconstructies met infinitief + te in een Zuidnederlands
dialect’. In: G. Kazemier & P.P.J. van Caspel (eds.), Taal- en letterkundig gastenboek voor Prof. Dr. G.A. van Es. Groningen: Archief voor de Nederlandse Syntaxis. 67-84.
Hoekema, T. (1963). ‘Drie syntagmen uit oostelijk Friesland’. Driemaandelijkse
bladen 15, 72-76.
Hoekstra, T. (1984). Transitivity: Grammatical Relations in Government-Binding
Theory. Dordrecht: Foris.
— Hoekstra, T. (1999). ‘Auxiliary selection in Dutch’. Natural Language and Linguistic
Theory 17:1. 67-84.
Honselaar, W. (1987). ‘Zijn vs. Hebben in het samengesteld perfectum.’ ’. In: De
Nieuwe Taalgids 80. 55-68.
Jäger, A. (2006). ‘Typology of periphrastic ‘do’-constructions.’ In: Diversitas
Linguarum 12. Bochum: Brockmeyer.
Kayne, R. (1993). ‘Toward a modular theory of auxiliary selection’. Studia
Linguistica 47. 3-31.
Kern, J.H. (1912). De met het Participium Praeteriti omschreven Werkwoordsvormen
in 't Nederlands. Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, Afdeeling Letterkunde, 12.
Klein, M. (1983). ‘Over de zgn. absolute met-constructie’. De Nieuwe Taalgids 72
(2).151-164.
Kooiman, K. (1950). ‘Hij’. De Nieuwe Taalgids 43. 324-333.
— Kooiman, K. (1954). ‘Ik heb geweest, ik ben geweest’. De Nieuwe Taalgids 47. 209-
214.
— Kooiman, K. (1958). ‘Zuidhollands ‘hoordiede’ en ‘lachtiede’’. De Nieuwe Taalgids
51, 240 .
Lieber, R. and H. Baayen (1997). ‘A semantic principle of auxiliary selection in
Dutch’. Natural Language and Linguistic Theory 15. 789-845.
Nuijtens, E. (1962). De tweetalige mens. Een Taalsociologisch Onderzoek naar het
Gebruik van Dialect en Cultuurtaal in Borne. Assen: Van Gorcum.
Overdiep, G. (1932). ‘Syntaxis en dialectstudie II’. Onze Taaltuin 1, 46-52.
Paardekooper, P.C. (1991). ‘Gaet souckt een ander medecijn’. De Nieuwe Taalgids.
84. 334-339.
Pannekeet, J. (1995). Het Westfries: inventarisatie van dialectkenmerken.
Wormerveer: Noord-Holland.
Pauwels, J. L. (1949). ‘De lijdende vorm in het Aarschots dialect’. Taal en & Tongval 1.
27-29.
Poletto, C. (2006). ‘Doubling as a “spare-movement” strategy’. Handout of a talk
presented at the Workshop on Doubling. March 16-18, 2006. Meertens
Institute, Amsterdam.
Ponelis, F. (1993). The development of Afrikaans. Frankfurt am Main: Die Deutsche
Bibliothek.
Poussa, P. (1998). ‘Do and dialect death’. In I. Tieken-Boon van Ostade, M.
van der Wal & A. van Leuvesteijn (eds.), DO in English, Dutch and German. History and present-day variation, Amsterdam/ Münster: Stichting Neerlandistiek/ Nodus Publikationen.7-24.
Robbers, K. (1997). Non-finite Verbal Complements in Afrikaans: A Comparative
Approach. Den Haag: Hollands Academic Graphics. HIL Ph. D. Dissertation.
Ryckeboer, H. (1986). ‘Het hulpwerkwoord doen in replieken’. In M. Devos & J.
Taeldeman (eds.). Vruchten van z'n akker, Opstellen van (oud-)medewerkers en oud-studenten voor Prof. dr. V.F. Vanacker. Seminarie voor Nederlandse Taalkunde en Vlaamse Dialectologie. 321-337.
—Ryckeboer, H. (1998). ‘Substituting doen in tag questions and short replies in
southern Dutch dialects’. In I. Tieken-Boon van Ostade, M. van der Wal and A. van Leuvesteijn (eds.), DO in English, Dutch and German. History and present- day variation, Amsterdam/ Münster: Stichting Neerlandistiek/ Nodus Publikationen. 65-81.
Schaerlaekens, A.M. (1977). De Taalontwikkeling van het Kind. Groningen:
Wolters-Noordhoff.
Schuermans, L.W. (1865). Algemeen Vlaamsch Idioticon. Leuven, Vanlinthout.
Bijvoegsel aan het Algemeen Vlaamsch Idioticon, Leuven, 1883.
Schütze, C.T. (2004). ‘Synchronic and diachronic microvariation in English do’.
Lingua 114. 495-516.
Shannon, T.F. (1993). ‘To be or not to be in Dutch: A cognitive account of some
puzzling perfect auxiliary phenomena.’ In: R.S. Kirsner (ed.), The Low Countries and Beyond. Lanham: University Press of America. 85-96.
Siewerska, A. (1995). ‘On the Origins of the Order of agreement and Tense Markers.’
In: J.C. Smith and D. Bentley (eds.). Historical Linguistics 1995. Volume 1: General issues and non-Germanic Languages (Selected papers from the 12th International Conference on Historical Linguistics, Manchester, August 1995). 377-392.
Sorace, A. (2000). ‘Gradients in auxiliary selection with intransitive verbs’. Language
76. 859-890.
Stapelkamp, Chr. (1948). ‘“Bliken dwaen” – “Gjin lichten dwaen” – “Writen
dwaen”’. It Beaken 10. 180-192.
Stoett, F.A. (1923). Middelnederlandse Spraakkunst; Syntaxis. 's Gravenhage:
Martinus Nijhoff.
Ter Laan, K. (1953). Proeve van een Groninger Spraakkunst. Winschoten: Van der
Veen.
Tieken-Boon van Ostade, I. (1990). ‘The origin and development of periphrastic
auxiliary do: a case of destigmatisation.’. Nowele 16, 3-52.
Tieken-Boon van Ostade, I., M. van der Wal and A. Van Leuvensteijn (eds.) (1998).
do in English, Dutch and German. History and present-day variation. Stichting Neerlandistiek VU Amsterdam, Nodus Publikationen Münster.
Travis, L. (1984). Parameters and Effects on Word Order Variation. Ph. D.
Dissertation, MIT.
Vanacker, V.F. (1948). Syntaxis van het Aalsters dialect. Michiels, Tongeren.
Van der Horst, J.M. (1998). ‘Doen in Old and Early Middle Dutch: A comparative
approach.’. In I. Tieken-Boon van Ostade, M. van der Wal and A. van Leuvesteijn (eds.), DO do in English, Dutch and German. History and present-day variation, Amsterdam/ Münster: Stichting Neerlandistiek/ Nodus Publikationen. 53-64.
Van der Meer, G. (1988). ‘Reported Speech and the Position of the Finite Verb (Some
Facts from West Frisian)’. Leuvense Bijdragen 77. 301-324.
— Van der Meer, G. (1991). ‘The “Conjugation” of Subclause Introducers: Frisian –st’.
Nowele 17. 63- 84.
Van der Wal, M. (1986). Passiefproblemen in oudere taalfasen. Middelnederlands
sijn/ werden + participium praeteriti en de pendanten in het Gotisch, het Engels en het Duits. Ph. D. Dissertation Leiden University. Dordrecht.
— Van der Wal, M. (1988). ‘Variatie en verandering in de Middelnederlandse
passiefsystematiek. Een beschrijving binnen een strukturalistisch getint kader.’ In J.A. van Leuvensteijn (ed.) Uitgangspunten en toepassingen. Taalkundige studies over Middelnederlands en zestiende- en zeventiende-eeuws Nederlands, Amsterdam: VU Uitgeverij. 35-58.
Van der Woude, G. (1971). ‘De wjerljocht hat ynslein - de bliksem is ingeslagen’.
Driemaandelijkse bladen 23. 137-144.
Van Kampen, J. (1997). First Steps in Wh-movement. Ph. D. Dissertation, Utrecht University.
Van Oostendorp, M. (2002). Rotterdams. Taal in stad en land. Den Haag: Sdu
uitgevers.
Van Riemsdijk, H. (1978). A case study in syntactic markedness: the binding nature
of prepositional phrases. Lisse: Peter de Ridder Press.
Van Valin, R. (1990). ‘Semantic parameters of split intransitivity’. Language 66. 221-
60.
Verhagen, A. (1994a)a. ‘Taalverandering en cultuurverandering: doen en laten sinds
— Verhagen, A. (1994b)b. ‘Doen of laten: woordbetekenis of (ook) structuur?’. Forum
der Letteren 35 (4). 277-281.
—Verhagen, A. (1998). ‘Changes in the use of Dutch doen and the nature of semantic
knowledge’. In I. Tieken-Boon van Ostade, M. van der Wal and A. van Leuvesteijn (eds.), doDO in English, Dutch and German. History and present-day variation, Amsterdam/ Münster: Stichting Neerlandistiek/ Nodus Publikationen. 103- 119.
Visser, F.Th. (1969). An Historical Syntax of the English Language. Leiden: E.J.
Brill.
Weijnen, A. (1966). Nederlandse dialectkunde. Assen.
— Weijnen, A. (1971). Schets van de geschiedenis van de Nederlandse syntaxis. Assen:
Van Gorcum & Comp. N.V.
Zaenen, A. (1988). Unaccusative Verbs in Dutch and the Syntax-Semantics Interface.
CSLI Reports. CSLI, Stanford University. 88-123.
— Zaenen, A. (1993). ‘Unaccusativity in Dutch: Integrating syntax and lexical
semantics’. In
Zwart, C.J.W. (1993). Dutch syntax. A minimal approach. Ph. D. Dissertation,
Groningen: Groningen Dissertations in Linguistics.
—Zwart, C.J.W. (2001). ‘Syntactic and Phonological Verb Movement’. Syntax 4.1. 34-
62.