Eric Hoekstra


Future venues of research and the SAND (volume 1)


0. Introduction *)


The award-winning Syntactic Atlas of Dutch Dialects, 'Syntactische Atlas van de Nederlandse Dialecten, abbreviated as SAND (volume 1) features the latest results of a dialect survey spanning The Netherlands and Dutch (Flemish) speaking Belgium. I will point out two types of future research, for and with the SAND. On the one hand, I will point out research which can be done with the material that has already been collected within the SAND project, but which must be dug out and analysed. An example of this type of research involves testing the hypothesis of Hoekstra & Smits (1997, 1998) on complementiser agreement (section 1.2). On the other hand, I will point out research which entails sending out new questions to informants, hence a continuation of the SAND. An example of this is given in section 4.1. on reciprocals with singular antecedents.


Volume 1 of the SAND contains the following five chapters, corresponding to five areas of syntactic research: complementisers (chapter 1), subject pronouns (chapter 2), subject doubling (chapter 3), reflexive and reciprocal pronouns (chapter 4) and fronting (chapter 5). My review has the same substructure, dealing with each of the chapters of the SAND in turn in a separate section. It is not my aim to present a full-fledged critical review of the SAND. I have done that elsewhere (Hoekstra 2006, to appear); there is some slight overlap, though, between that article and the present article. In this contribution, I will focus on future venues of research for and with the SAND.


I will pay special attention to complementisers. Hence the sections dealing with complementisers and complementiser agreement are more elaborate than the others.


The contents of my review, which correspond to the contents of the SAND, are given below:


1. Complementisers

2. Subject pronouns

3. Subject doubling and subject clitics following ja ('yes') and nee 'no'.

4. Reflexive and reciprocal pronouns

5. Fronting

6. Conclusion


1. Complementisers

1.1. Temporal complementisers

1.1.1. Temporal complementiser voor 'before' + tensed clause


The complementiser dat 'that' can be optionally present following various other complementisers such as relative complementisers, Wh-complementisers, temporal complementisers, among others.1) Dialects differ to which extent they allow doubling of complementisers by dat, whereas the standard language tends to disallow doubling. Curiously, the standard language allows doubling in the case of voor 'before' and voordat 'before that'. The commentary raises the question of how this doubling complementiser should be analysed syntactically. Two analyses are proposed, of which we quote the first (p.10):


"This would mean that voordat consists of two elements, namely, a preposition voor 'before' and a complementiser dat 'that'. The latter can be left out because it does not contribute to the meaning of the clause."


There are two problems here. First, it is not generally the case that meaningless elements are optional in syntax; thus there is not necessarily a causal connection between the optionality of dat and its alleged lack of a semantic contribution to the sentence. Secondly, as it turns out, the presence or absence of dat does have an effect on the sentence's meaning, in the case of the alternation between voor and voordat, as noted by Jarich Hoekstra in unpublished work.

Jarich Hoekstra refers to Geis (1970), who notes that the following sentence is ambiguous in English, as indicated here by the paraphrases in (1a) and (1b):


(1) I saw Mary in New York before she claimed that she would arrive.

(1a) I saw her before the time at which she made the claim about her time of arrival.

(1b) I saw her before the time at which she would arrive according to her.


Jarich Hoekstra notes that only voordat is ambiguous in Dutch, whereas voor is not:2)


(2a) Ik zag Marie in Leeuwarden, voordat ze zei dat ze daar zou zijn.

I saw Mary in Leeuwarden before she said that she there would be

"I saw Mary in Leeuwarden before she said that she would be there."

(2b) Ik zag Marie in Leeuwarden, voor ze zei dat ze daar zou zijn.

I saw Mary in Leeuwarden before she said that she there would be

"I saw Mary in Leeuwarden before she said that she would be there."


n (2a), the seeing can take place either before Mary's saying or before Mary's being there. In (2b), the seeing can only take place before her saying so. If these judgments hold water, then the presence or absence of the complementiser dat does affect the sentence's meaning. Jarich Hoekstra reports similar facts for Frisian:


(3a) Ik seach Martsje yn Amsterdam, foardat se sei dat se dr wze soe.

I saw Mary in Amsterdam before she said that she there be would

"I saw Mary in Amsterdam before she said that she would be there."

(3b) Ik seach Martsje yn Amsterdam, foar't se sei dat se dr wze soe.

I saw Mary in Amsterdam before she said that she there be would

"I saw Mary in Amsterdam before she said that she would be there."


In Frisian, the complementiser dat may be present, or it may cliticise onto the complementiser foar. In case cliticisation takes place, the downstairs reading (compare (1b)) is unavailable. It would be interesting to investigate this semantic effect in the various dialects of Dutch.


1.1.2. Temporal complementisers na 'after' and alvorens 'before' + infinitival clause


It is noted in the commentary that infinitival clauses can be introduced by an allegedly meaningless complementiser like om "for" or by meaningful complementisers like na 'after', alvorens 'before' and zonder 'without' (p.10). Here I would like to point out the remarkable behaviour of na in Standard Dutch. This complementiser requires the presence of the verb hebben 'have' in infinitival clauses:


(4a)* Na lekker te slapen werden we wakker.

after well to sleep became we awake

"After sleeping well we woke up."

(4b) Na lekker te hebben geslapen werden we wakker.

after well to have slept were we awake

"After sleeping well we woke up."


Is this the case in the dialects as well? Spanish certainly allows any verb to occur after despus 'after'. Furthermore, we note that the complementiser 'before' may change shape in Standard Dutch depending on whether it introduces a finite or a non-finite clause. The complementisers voordat and voor are used to introduce finite clauses, but they cannot be used to introduce infinitival clauses. The complementiser alvorens, however, can introduce both finite and non-finite clauses. Frisian only features the complementiser foar, which is followed by dat or the cliticised variant 't. Interestingly, foar is not allowed to introduce non-finite clauses in Frisian, and there is no direct Frisian equivalent for alvorens. How is this in the dialects of Dutch?


1.2. Complementiser agreement


Several maps are relevant for, and tend to confirm, the generalisation made by Van Haeringen (1958), which says that the agreement on the complementiser is identical to that of the verb in inversion: 23b, 24b, 29b, 31b, 34b. The maps also test the narrower hypothesis suggested by Goeman (1979) (and discussed in subsequent work such as De Vogelaer, Neuckermans en Vanden Wyngaerd (2002)), who relates complement agreement to the agreement of tensed monosyllabic verbs in inversion. However, the notion 'monosyllabic verb' is not unproblematic. First of all, monosyllabic verbs tend to behave as (regular) bisyllabic verbs in the present tense, but they hardly ever do so when used as infinitives. So the verb gaan is often regularised in the present tense plural from gaan to gane. The infinitive is hardly ever thus regularised. The conceptual question is, why should complementiser agreement 'look' at the infinitival form of verbs? The complementiser itself introduces finite, not infinitival sentences. Furthermore, Goeman (1979:299) does not seem to distinguish between complementiser agreement proper and suffixed and infixed clitics in forms such as astie ('when + he') and merktiede (merk 'perceive' = verb stem, die = clitic, de = past tense suffix). Let us quote the relevant passage from Goeman's article in full:


Person agreement on subordinators also occurs in the past tense, but not by means of the ordinary verbal past tense suffixes but with the clitic induced suffixes of the present tense. Person agreement on subordinators is therefore unspecified for [+ tense], so astie maar goed werkte (when he merely good worked) is paralleled by infixed forms as merktiede (merk verb stem, die pronoun, de past tense suffix: 'remarked he") ... while forms such as ase die werkte (when he worked) do not exist.


Now, for pronominal clitics it is trivially true that they are unmarked for tense. Verbal affixes, however, may conflate tense and agreement information: the -t of Standard Dutch comprises synthetically both agreement information (3SG) and tense information (present). The schwa of Standard Dutch, though spelled differently in present and past, only gives agreement information (PL), no tense information. Goeman's incidental remark quoted above, however, may be taken as the first explicit speculation indicating that complementiser agreement is never synthetic, though he does not develop it further. That is done in Hoekstra & Smits (1997). They propose a generalisation which states that complementiser agreement never bears tense information, and test this claim on a number of dialects by comparing the present tense and past tense paradigm of (inverted) verbs to the complementiser paradigm. Thus Hoekstra & Smits can account for the absence of complementiser agreement in Frisian: they note that the verbal agreement in the present tense is distinct from the past tense (-E versus -EN). Unfortunately, the hypothesis of Hoekstra & Smits, which is supported by data from several dialects, was not tested, which is obviously a desideratum for future research. In fact, the data in the SAND already make it possible to make this comparison, but it is quite a job and it will also require knowledge of the dialect in question.


Incidentally, the Frisian data also contradict the idea that monosyllabicity is a conditioning factor for complementiser agreement. Frisian features monosyllabic gean 'go'; it shows up as geane in inversion in the present tense plural, hence complementiser agreement datte should be found in the plural, on the hypothesis of Goeman that verbal monosyllabicity is somehow relevant. But Frisian does not feature complementiser agreement in the plural.


Complementiser agreement has been related to the agreement of tensed verbs in inversion. But verbs constitute a large class. It is hardly conceivable that a grammatical condition checks on all members of the class of verbs. It is far more likely that only a characteristic closed subset of all verbs is concerned. Hence Hoekstra & Smits (1998, see suggested that it is the agreement of auxiliaries (a functional category) which is relevant. Since monosyllabic verbs like gaan 'go' are auxiliaries, the special similarity with monosyllabic verbs now follows. See Hoekstra & Smits (1998) for further arguments in defense of this claim. The material of the SAND could be used to test whether the proposed refinement of Hoekstra & Smits is empirically adequate. Again, this is a task for future research. The large amount of variation in the field of complementiser agreement indicates that the syntactic agreement mechanism must be very powerful (compare Barbiers 2006).


2. Subject pronouns

2.1. Premodified subject pronouns 2SG


Chapter 2 of the SAND deals with various aspects of subject pronouns. One of the most peculiar properties of Standard Dutch is that the verb loses its inflectional ending -T in inversion in the present tense 2SG:


(5a) Je gaat. Ga je? Je ging. Ging je?

You go. Go you? You went. Went you?

(5b) Je leeft. Leef je? Je leefde. Leefde je?

You live. Live you? You lived. Lived you?


It would be interesting to investigate which dialects share this property with Standard Dutch, and also to investigate whether there are dialects which feature this phenomenon with other person/number combinations of the paradigm (cf. De Wulf & Taeldeman 2006).


In addition, I found out (Hoekstra 1994) that the pronoun 2SG tends to resist premodification just in case there is no preceding -T.


(6a) (*) Morgen ga zelfs jij braaf naar school.

tomorrow go even you well-behaved to school

"Tomorrow even you will go to school like a good boy / girl."

(6b) Gister ging zelfs jij braaf naar school.

yesterday went even you well-behaved to school

"Yesterday even you went to school like a good boy / girl."


In Standard Dutch, premodification of the second person pronoun is for many speakers ungrammatical, just in case the (zero) flection of the second person in inversion is distinct from the (non-zero) flection used in the non-inverted order. In other language varieties, deviant behaviour of the 2SG can also be found. Take Frisian for example:


(7a) * Moarn giest sels do braaf nei skoalle ta.

tomorrow go-2SG even you well-behaved to school to

"Tomorrow even you will go to school like a good boy / girl."

(7b) * Ik tink datst sels do moarn braaf nei skoalle ta giest.

I think that-2SG even you well-behaved to school to go-2SG

"Yesterday even you went to school like a good boy / girl."


Here premodification of the second person is never allowed in inversion. This phenomenon seems well worth investigating and will doubtlessly yield a wealth of interesting material.


3. Subject doubling and subject clitics following ja ('yes') and nee 'no'.

3.1. Premodification and doubled pronouns


The premodification tests of the previous section can also be done with doubled pronouns. Furthermore, De Haan (1997) reacted to Van der Meer (1991) on the subject of distinguishing cliticisation from agreement. Van der Meer had argued that the Frisian 2SG, which observationally features pro-drop, should be analysed theoretically as cliticisation. De Haan's article contains a wealth of tests distinguishing clitics from agreement. Those tests could also be the basis for further surveys in the SAND. See also De Schutter (1997).


4. Reflexive and reciprocal pronouns

4.1. Elkaar 'each other' with singular antecedent


The SAND features one map with the reciprocal, which has the form elkaar 'each other' in Standard Dutch. The example sentence used is one in which elkaar has a plural antecendent. However, Standard Dutch features a remarkable phenomenon, which is absent in English and French. The reciprocal can also take a singular antecedent:


(8a) Het horloge viel uit elkaar.

the watch fell out each other

"The watch fell apart."

(8b) Dat stel is uit elkaar.

that couple is out each other

"That couple is divorced."

(8c) Ik zet het horloge in elkaar.

I put the watch in each other

"i put the watch together."


Frisian features the construction in much the same way as Dutch does. It would be interesting to investigate the extent to which the southern dialects of Dutch have this construction, regardless of whether the explanation for this phenomenon is going to be syntactic or semantic.


4.2. Distribution of z'n eigen 'his own' meaning 'himself'


As is clear from the SAND, many central Dutch dialects feature the non-Standard reflexive z'n eigen. The maps of the SAND already contain the results of this reflexive in a number of syntactic contexts. Interestingly, the reflexive z'n eigen 'his own' is systematically absent in this example sentence (map 70a), as noted in the commentary:


(9) The timmerman heeft geen spijkers bij zich.

the carpenter has no nails with him

"The carpenter has no nails with him."


This is curious, as the reflexive does show up with inherent reflexice verbs like zich herinneren / z'n eigen herinneren 'remember'. This calls for an in-depth investigation of this reflexive.


Preliminary research on this reflexive was performed by a trainee of mine who investigated the Kempisch dialect (Breukels 1997), spoken in the province Noord-Brabant. According to Breukels, Kempisch uses z'n eigen ('his own') both for men and for women. Map 74b of the SAND reports the results for a female antecedent third person. The map only features the female form of the reciprocal, d'r eigen ('her own'), which is absent in Kempisch according to Breukels. Strangely enough, the form z'n eigen with female antecedent (a gender neutral form) is not found in the results of the SAND. Now, Breukels notes that Kempisch lacks the female form d'r eigen simply because it lacks the possessive pronoun d'r 'her'. However, the SAND had to abstract away from this in order to produce a map, so the Kempisch data were (probably) assimilated to the data from dialects which had distinct forms for reflexives with male and female third person antecents. Thus we see that an atlas, by presenting a map in which only a restricted number of subcategories in the data have been made visible, sometimes makes abstractions which are misleading. This is inherent to the making of maps.


5. Fronting


With respect to fronting, the results of the SAND do not seem to raise questions for future research, except for the lack of verb-second in French Flemish, as shown on map 95a. It would be interesting to find out which elements do trigger verb-second in those dialects.


6. Conclusion


The SAND offers an overwhelming amount of Dutch dialect data. It can be used as an excellent starting point for further research, and there is still lots of work to be done. I also checked out the digital version of the SAND (, and I found it quite useful. I suspect that assistance is required when one wants to do some complex data-mining. For example, I suspect that some outside help would be needed if one wanted to test the hypothesis of Hoekstra & Smits (1997) on complementiser agreement (see section 1.2). Nevertheless, the digital SAND (dynaSAND) is well worth an effort on the part of the user.





*) I would like to thank two anonymous reviewers for providing useful comments.


1) A reviewer suggests that dat, if it occurs, is obligatory in such cases, rather than optional. I suspect that this depends on the dialect involved. When analysing dialect material from South Holland, North Holland, Friesland and Groningen, I noticed that dat tended to be optional. This is in line with my native speaker's intuitions about regional Dutch and about Frisian. Possibly, the reviewer bases his judgment on different (southern?) dialects.


2) This contrast is not shared by everybody: one reviewer shares J. Hoekstra's judgments, the other does not and reports that both voor and voordat are ambiguous. Apparently, different dialects (or even idiolects) are involved. Perhaps the following example, suggested by a reviewer, makes it easier to distinguish between the two readings:


(ia) De trein vertrok voor de spoorbeamte zei dat hij zou vertrekken

the train left before the railway official said that it would leave

"The train left before the railway official said that it would leave.

(ib) De trein vertrok voordat de spoorbeamte zei dat hij zou vertrekken

the train left before the railway official said that it would leave

"The train left before the railway official said that it would leave.


Complexities may be due to the interaction of a temporal complementiser with the past tense, and the use of the past tense could also be due to dialect variation.




Barbiers, S. (2006) Er zijn grenzen aan wat je kunt zeggen. Inaugurele rede. Universiteit Utrecht.


Breukels, F. (1997) Anaforen en quantifiers. Een benadering van de anaforendistributie aan de hand van het Kempisch, Nederlands en Engels. Trainee report Meertensinstituut. Unpublished.


Geis, M. (1970) Adverbial subordinate clauses in English. Dissertation, MIT.


Goeman, A. (1979) COMP-agreement?. In: W. Zonneveld & F. Weerman (eds.), Linguistics in the Netherlands 1979. Dordrecht: Foris Publications. 291-306.


Haan, G.J. de (1997) Voegwoordcongruentie in het Fries. In: E. Hoekstra & C. Smits (eds.), Vervoegde voegwoorden. Amsterdam: P.J. Meertensinstituut. 50-67.


Haeringen, C.B. van (1958) Vervoegde voegwoorden in het Oosten. Driemaandelijkse bladen NS 10, 115-124.


Hoekstra, E. (1994) Agreement and the Nature of Specifiers. In: J.-W. Zwart (ed) Minimalism and Kayne's Asymmetry Hypothesis. Groninger Arbeiten zur Germanistischen Linguistik 37, 159-168.


Hoekstra, E. & C. Smits (1997) Vervoegde voegwoorden in de Nederlandse dialecten: een aantal generalisaties. E. Hoekstra en C. Smits (eds) Vervoegde Voegwoorden. Cahiers van het P.J. Meertens Instituut 9. Amsterdam.


Hoekstra, E. & C. Smits (1998) Everything you always wanted to know about complementiser agreement. Proceedings of Western Conference on Linguistics (WECOL 1998).


Hoekstra, E. (2006, to appear) Review of the SAND. It Beaken 55.


Hoekstra, J. (1993) The split CP hypothesis and the Frisian complementiser system. Ms. Fryske Akademy, Leeuwarden. Unpublished.


Meer, G. van der (1991) The 'conjugation' of subclause introducers: Frisian -st. North-Western European Language Evolution 17, 63 - 84.


Schutter, G. de (1997) Incorporatie-in-C in de Vlaamse en Brabantse dialecten. In: E. Hoekstra & C. Smits (eds.), Vervoegde voegwoorden. Amsterdam: P.J. Meertensinstituut. 31-49.


Vogelaer, G., A. Neuckermans en G. vanden Wyngaerd (2002) Complementizer Agreement in the Flemish Dialects. In: Barbiers, Sjef, Leonie Cornips & Suzanne van der Kleij (eds.). Syntactic Microvariation (Meertens Institute Electronic Publications in Linguistics (MIEPiL) II). Amsterdam: Meertens Instituut. (


Wulf, C. de & J. Taeldeman (2006) T-deletie in de Nederlandse dialecten: een globaal overzicht. In L. Cornips, M. van Oostendorp en G. de Schutter (red.) Taal & Tongval themanummer 19. Het morfologisch landschap van het Nederlands, 244-272.