A brilliant piece of work: Syntactic Atlas of the Dutch Dialects
Sjef Barbiers, Hans Bennis, Gunther de Vogelaer, Magda Devos and Margreet van der Ham. Volume I. 'Kaarten' (maps, 95 pages) and 'Commentaar' (commentary, 79 pages). Amsterdam University Press.
At the end of the nineties, Prof. Hans Bennis of the Meertens Institute (Royal Netherlands Academy of Arts and Sciences) initiated the research project known as the Syntactic Atlas of Dutch Dialects ('Syntactische Atlas van de Nederlandse Dialecten", abbreviated as SAND). The other parties participating in this venture were: the Universities of Gent and Antwerpen (Belgium), the Universities of Leiden and Amsterdam (The Netherlands) and the Frisian Academy (The Netherlands). This cooperation guaranteed the presence of sufficient knowledge not only of the Dutch dialects in The Netherlands, but also of the Dutch dialects in Belgium (sometimes termed Flemish) and of the Frisian dialects in The Netherlands.
An atlas is a collection of maps. This implies that a selection of the original data has been made accessible, in the form of the book that is currently being reviewed. The original data can be accessed through the internet: www.meertens.knaw.nl/sand/).
In the past, atlases tended to concentrate on lexical maps, maps dealing with isolated lexical items, and on historical-phonological maps. Recent decades, beginning with the seminal work of Chomsky in the sixties, have witnessed an increase of linguistic interest in syntax. Apart from this, linguistics has always been a branch of science in which The Netherlands and Belgium did particularly well. Thus it was perhaps only a matter of time before an undertaking such as the SAND would come into existence.
Sometimes linguistic atlases are somewhat superficial. They deal only with one aspect of a complex phenomenon. A strong point of the SAND is that it provides in-depth data and discussion of several phenomena. Thus volume I of the SAND contains the following chapters:
2) Subject pronouns.
3) Subject doubling and subject clitics following ja 'yes' and nee 'no'.
4) Reflexive and reciprocal pronouns.
The atlas has both an English and a Dutch title. The example sentences going with the maps have English glosses and translations. The comments have been written in Dutch in a separable booklet that can be detached from the rest of the book. Thus one can have the comments next to the maps. Although I could not find any information on this, I suppose that there will also appear an English version of the detachable booklet of commentaries, seeing that the glosses were written in English. Because of this, and in order to advertise the SAND to the international scientific community, I have chosen to write this review in English as well.
In this review, I will pay special attention to the chapter on complementisers, in order to illustrate in detail all the problems which are involved in an in-depth analysis of facts from a variety of geographically ordered language varieties.
The commentary opens with an introduction. We are told that volume 2 of the SAND will appear in 2006, and that it will contain maps and commentaries involving the verbal system, negation and quantification.
The introduction also tells us that dialects seem to vanish faster than ever before, leaving only regiolects behind, as a result of "decreasing isolation and increasing mobility and communication". Nonsense! It is an untrue cliché to claim that in the past there was less mobility than now. In the middle ages, people often went abroad to gain a living: the fishermen sailed to England and Scotland, farm hands' gained a living in the North of Germany, and monks and scientists travelled through Europe for years. At least during the period of the wellfare state (let's say from 1960 onwards), unemployed people and scientists travel much less than in the middle ages. Mobility is in part a function of wellfare. Furthermore, many dialects have given way before a standard language in the past as well: the Celtic dialects of France (Gaul) gave way before Latin, and a few memorial inscriptions is all that is left of them. Although it is certainly true that dialects are disappearing, the cause behind this process involves the expanding power of a culturally centralised state (or other form of organisation): Latin pushed out French Celtic, Dutch is pushing out Frisian, and English may push out Dutch.
The atlas has disregarded the language spoken in the IJsselmeerpolders (the land reclaimed from the IJsselmeer after the Second World War). This is a pity. In an atlas spanning dialects as different as West Flemish, Frisian and Limburgian ones, there would surely have been space for the language varieties spoken in the IJsselmeerpolders, which differ only mildly from what is spoken in the adjacent areas (compare Scholtmeijer 1992).
The introduction also provides an account of the methodology. The most important point that is made is that there should be as little accommodation as possible. Hence the interviews were not made by people speaking Standard Dutch, but by people speaking the local dialect. Various elicitation techniques were used: indirect grammaticality judgments, drawings accompanied by incomplete sentences which the informant had to complete, translation tasks.
Sentences from the interviews were transcribed orthographically. Lexical morphemes were transcribed as Standard Dutch, functional morphemes were literally transcribed. Of course, this distinction is not unproblematic, but it is hard to see how the data could have been made accessible otherwise.
Chapter 1 discusses several complementiser phenomena. The commentary, somewhat surprisingly, discusses an issue which does not seem to be addressed directly in the maps, namely the issue of whether a complementiser like "voordat" (lit. 'before that', meaning "before") consists of a preposition + complementiser or of complementiser + complementiser. Sometimes the 'dat' can be omitted, as in "voordat / voor", sometimes it can't, as in "zodat / * zo". Strangely enough, it is claimed in 1.1.2. that the presence of combinations in which "dat" cannot be dropped supports the complementiser + complementiser analysis. I fail to see why this should be the case. Furthermore, no maps are presented which involve this alternation; hence this passage could have been dropped alltogether. Section 1.1.1, which introduces the chapter, is a bit pedestrian, going as far as to leisurely explain the distinction between coordinating and subordinating conjunctions. Otherwise, the chapter is well-written and the pace (information density per sentence) increases as we are propelled down the ins and outs of complementiser agreement.
However, other aspects of complementisers are also investigated. In 1.1.4. it is claimed that the comparative complementiser "of" ("as if") introduced sentences with main clause word order. Map (14b) shows that this is still the case in Antwerpen. The map specifies that the main clause order is V/2, but the commentary should have emphasised this, since the same complementiser is also found with V/1 order, as is still the case in present-day Frisian. It is not until section 188.8.131.52 presents example sentence (11) that it becomes clear that the map deals exclusively with V/2, and that V/1 has not been taken into account here. This could have been written down more clearly.
In general, it is important to explicit ask for variants which occur in the dialect if the standard language variant also occurs in the dialect. To exemplify, if a dialect allows both V-final (standard language) and V/2 order, the dialect-specific V/2 order will hardly surface in a translation task unless it is specifically asked for; it will of course surface if it is the only possible order which the dialect allows. In case a word-for-word translation is grammatical in the dialect, much more Standard Dutch, that is V-final orders, will be scored than is justified in such a situation. From e-mail contact with Prof. Barbiers, I understood that dialect-specific phenomena were explicitly asked for. This is methodologically sound.
It can be noticed that the geographical spread described in two articles by myself and Caroline Smits (Hoekstra & Smits 1997, 1998) shows remarkable similarities with the spread represented in the SAND-project. If anything, this shows that various methodologies tend to converge as far as the research results are concerned. Of course, the empirical basis of the SAND project is much more detailed; nevertheless, clever arm chair linguistics seems to have given a pretty accurate picture of the main geographical spread. This has also been my experience as a dialect researcher: the extensive enquiries which we sent out usually (not always) confirmed what an 'armchair linguist' had concluded on the basis of his own or somebody else’s native speaker knowledge of a particular dialect and dialect area.
The chapter contains a lot of information on the relation between complementiser agreement and the agreement which the verb in inversion has. Several maps are relevant for 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. Hoekstra & Smits (1997) proposed a generalisation which states that complementiser agreement never bears tense information. This accounts for example for the lack of complementiser agreement in Frisian. They note that the verbal agreement in the present tense plural is distinct from the past tense (-E versus -EN); correspondingly there is no complementiser agreement in the plural. 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.
4. The other chapters
4.1. Subject pronouns
The second chapter deals with subject pronouns. The author warns us (and himself) that mere variation in the form of the pronoun does not belong to syntax, but to phonology or the lexicon, and hence does not deserve a place in the SAND. Nevertheless, the discussion in this chapter focuses very much on the phonological form of the pronoun. This is only interesting for syntax if the form correlates with some syntactic property. For example, the difference between a strong form like “hun” and a weak form like “ze” becomes interesting when we observe, for example, that only the weak form occurs in idioms like “werk ze” (‘have a nice work’), but not the strong form: * “werk hun”. Of course, in order to uncover such properties one must first know what the form is of the strong and weak pronouns. However, research in this chapter did not move very far beyond the inventarisation of strong and weak pronouns. A minor point is that the chapter (p.26 top) refers to Hoeksema (2000) and Zwart (2002), but these references are absent from the bibliography.
The general discussion is also from a non-syntactic point of view. For example, the author notes correctly that new clitic pronouns are easily created in enclitic position: a pronoun in enclitic position may be assimilated to the flectional ending and thus a clitic pronoun, distinct from the ‘normal’ form, is born. Or the flectional ending may be reanalysed as part of the clitic, or vice versa. Although it is true that phonology plays a role, it is at least as important to realise that all these reanalyses are allowed for in syntax. Apparently, the syntax of flection, pronouns and agreement is an extremely powerful mechanism; otherwise these ‘reanalyses’ could not have come into existence in the first place, for, to quote Shakespeare, “nothing comes of nothing”. For the rest, the chapter contains interesting data about variation in expletive subjects, and about the distribution of the various pronouns. The data will be directly useful for linguists taking an historical interest in the development of the various forms, such as the development of the polite pronouns, the development of “je / ge” (second person pronouns), and so on and so forth.
I found it a pity that no questions were asked about the second person pronoun in inversion, when it is premodified. To illustrate, I have in mind sentences like those below:
(1a) (*) Morgen ga zelfs jij braaf naar school.
Tomorrow go even you well-behaved to school
(1b) Gister ging zelfs jij braaf naar school.
Yesterday went even you well-behaved to school
In Standard Dutch, premodification of the second person pronoun is for many speaker 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 (remember the paradigm of the present tense is: “je gaat”, ‘you go-2SG’; “ga je” ‘go you’). In other language varieties, deviant behaviour of the 2SG can also be found. Take Frisian for example:
(2a) * Moarn giest sels do braaf nei skoalle ta
tomorrow go-2SG even you well-behaved to school to
(2b) * Ik tink datst sels do moarn braaf nei skoalle ta giest
I think that-2SG even you well-behaved to school to go-2SG
Here premodification of the second person is never allowed in inversion, it seems (see Hoekstra 1994 for the Dutch and Frisian facts). This phenomenon seems well worth investigating and would have yielded a wealth of interesting material.
4.2. Subject doubling and subject clitisation after “ja” (‘yes’) and “nee” (‘no’)
Many Flemish dialects exhibit the phenomenon of subject doubling or even tripling. The phenomenon is extensively dealt with. It is noted that doubling in inversion is parallel to doubling after a complementiser, except in the second person. The author of this chapter (p.41, right column bottom) speculates that doubling indicates that the weak pronouns are evolving into flectional suffixes. This seems doubtful to me, especially in the light of the general tendency of the Germanic languages towards deflection. The question of a principled explanation of doubling phenomena is still to be answered.
Language-genetically, I cannot help thinking that the doubling phenomena of Flemish are somehow related to the doubling phenomena found in French. It would be extremely interesting to see what is going on in the Wallonian dialects, seeing that Flemish speakers used to be pretty bilingual, if only because Flemish was forbidden in parts of the public domain for centuries. It is also possible that the similarities between Flemish and Wallonian dialects can not only be traced back to the influence of Standard French, but also to the existence of a substratum that is common to both Wallonian and Flemish dialects. For example, De Jong (1998) has shown that some apparent phonological peculiarities of Picardian (nothern French dialects) make perfect sense, since they are also found in the adjacent Flemish dialects, thus forming homogenuous geographical area.
Flemish dialects and some ‘old’ dialects from The Netherlands (like Volendam) also exhibit a subject clitic on the affirmative and negative markers “ja” (yes) and “nee” (nee). Some examples are given below:
(3a) (Heb jij al gegeten?) Jaa-k
have you already eaten yes-I
Did you eat already? Yes I did.
(3b) (Hebben wij al gegeten) Jaa-w
have you already eaten yes-we
Did we eat already? Yes we did.
Sometimes an agreement marker appears in between the affirmative or negative marker and the clitic. Cross-linguistically, the phenomenon seems to be quite rare. If I’m allowed a speculation, I wonder whether there is any similarity with the Celtic languages here, and whether this can ultimately be related to Celtic substrate.
4.3. Reflexives and reciprocals
Chapter 4 deals with reflexive and reciprocal phenomena, and with one-pronominalisation. English exhibits the latter phenomenon as well, as is clear from the contrast between a good one alongside * a good; Dutch dialects show intricate patterns of variation here (see Barbiers 2005). It is clear and straightforward well-written chapter. It is remarkable that a small country like The Netherlands, or Belgium, for that matter, features such a large amount of variation in the type and distribution of reflexives. It cannot but be the case that surveys similar to the SAND will uncover similar variation elsewhere.
The chapter on fronting subsumes various fronting phenomena, many in the domain of question formation and relativisation, but also including fronting in imperatives, fronting in declarative main clauses without inversion, preposition-stranding and split fronting (type: “books I have no / none”). Map (89b) in all likelihood contains a severe processing error. According to this map, the possessive relatives used in Frisia are “wie/die z’n” and “wiens”. However, “wiens” is basically absent from spoken and written Frisian. Perhaps the informants were reporting on their Dutch when they okayed “wiens”. I strongly suspect something went terribly wrong here. If it did, the second volume of the map will need to have an erratum.
5. Concluding comments
I found it a pity that the authors of the separate chapter did not completely uniformize their subsection division, but each chapter by and large has the same overall structure. First the phenomenon in general is discussed, then each map is discussed. I sometimes felt that it was not always effective to separate these two information chunks. Sometimes I need more information while reading a specific map commentary, sometimes I felt that the map commentary repeated information already given earlier, and sometimes the map commentary stated the obvious, that is, the mere distribution of forms which could be gleaned more accurately from the map itself.
Although the maps are beautiful, the use of colours is not very distinctive. All map symbols have the same shape, they differ by colour only. I sometimes found it hard to distinguish yellow and brown, and to distinguish the various shades of blue from green. Complex maps must be viewed under strong light.
The atlas, and the data collection underlying it, is a valuable contribution to the study of syntactic variation, both as studied for its own sake and for the historical investigation of the Dutch dialects, in relation to historical processes. One would wish to have similar data collections and atlases for Germany, France, French-speaking Belgium, England and Denmark.
Once more, it has been proven that linguistics is a science in which The Netherlands and Belgium are excellent. In fact, project leader dr. Sjef Barbiers received the prestigious ESF-prize (European Science Foundation) for the SAND: a prize worth 1,25 million euro to be spent on research. He fully deserved the European prize he won for this piece of work.
Barbiers, S. (2005) Variation in the morphosyntax of ONE. Journal of Comparative Germanic Syntax 8, 159-183.
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).
Jong, T. de (1998) The evolution of o in open position. In: M.S. Schmid, J.R. Austin en D. Stein (eds) Selected Papers from the 13th International Conference on Historical Linguistics 1997. Benjamins, Amsterdam, 163-174.
Scholtmeijer, H. (1992) Het Nederlands van de IJsselmeerpolders. Dissertation, University of Leiden.