Historical sociolinguistics

Pluricentricity in language history. Building blocks for an integrated history of Dutch (1550-1850)

Funded by the Research Foundation – Flanders (FWO)
FWO Aspirant: Iris Van de Voorde
2017 – 2021
Joint PhD VUB – Leiden

Supervisors: Rik Vosters, Gijsbert Rutten, Wim Vandenbussche & Marijke van der Wal

The present dissertation examines historical pluricentricity in the Dutch language area. Pluricentricity refers to one language having multiple centers from which language norms spread. Those normative centers often coincide with nation-states, so that each center is associated with a national variety. In Dutch linguistics there is now a consensus that Dutch is a pluricentric language, with national varieties in the Netherlands, Belgium and Suriname, with the first two being the focus of this study. We can discern a center in the Northern Netherlands (the Randstad area) and the Southern Netherlands (around the province of Brabant). The current situation of pluricentricity has produced Dutch-Dutch and Belgian-Dutch language norms. According to traditional language histories of Dutch, this situation of pluricentricity emerged over the course of the 20th century, making it a recent phenomenon. In the preceding centuries, there would have been only one center, shifting according to the political constellation. This is what we label consecutive monocentricity. There are, however, indications to situate pluricentricity earlier in time. This was also the reason for examining Dutch pluricentricity in a broader historical perspective.

The traditional view on the history of Dutch is characterized by consecutive monocentricity. This means that the focus in language histories shifts to the region that is considered the center economically, culturally, and linguistically. This center was located in Brabant in the South in the 16th century. At the end of the 16th century, the center shifted from Brabant in the South to the region of Holland in the North. This also had an impact on the development of a Dutch standard language: after the 16th century, the process of standardization took off in the North of the language area. By the end of the 19th century, the standard language in Flanders became oriented towards Northern Standard Dutch. This standard then underwent autonomous developments in Flanders over the course of the 20th century, leading to the current situation of pluricentricity. Findings from more recent empirical research, however, suggest that that a second normative center was already developing in the South in the late 18th century, with a distinct normative tradition and a developing Southern linguistic identity (Rutten & Vosters, 2010b; Rutten & Vosters, 2011; Rutten & Vosters, 2013; Vosters, Rutten & Vandenbussche, 2012; Vosters, Rutten, Van der Wal & Vandenbussche, 2012). Based on these findings, we could expect to situate pluricentricity at least 100 years earlier in time. With this in mind, we deemed it necessary to focus on pluricentricity from an historical perspective and thus assess the usefulness of the concept in Dutch language history.

To assess the usefulness of the concept of pluricentricity in Dutch language history, the Historical Corpus of Dutch (HCD) was developed. It is a new diachronic corpus with text material from four centuries, namely from the 16th, 17th, 18th and 19th centuries. In addition, the HCD includes four regions from the Northern and Southern Netherlands (Holland, Zeeland, Brabant, Flanders). Within the North and the South, both a central and a peripheral region were chosen. The region of Holland is the center in the North, while Zeeland represents the periphery; in the South, Brabant is the center and Flanders represents the periphery. By integrating Northern and Southern varieties of Dutch, and by mapping the interactions between the different regions, we aim to lay the foundation for an integrated history of Dutch. Finally, the corpus is built around three genres, namely administrative texts, ego-documents and pamphlets. This implies that the corpus contains both handwritten texts (administrative texts and ego-documents) and printed texts (pamphlets). Moreover, with administrative texts, we include formal texts in the study, while ego-documents are considered less formal texts, and printed pamphlets may shift on the continuum between formal and less formal. The corpus amounts to over 460,000 words.

Six linguistic variables were examined in the corpus: the spelling of /a:/ in closed syllables, the spelling of /k/in word-final position, the spelling of /t/ for d-stem verbs, the schwa apocope, the change from bipartite negation to single negation, and the shift from d- to w-forms in relativizers. We will examine how the observed processes of language variation and change in our historical corpus are related to the language-external variables of period, region, and genre. By doing so, our research fits within the research tradition of historical sociolinguistics, a linguistic approach that has gained more and more ground in recent decades. Moreover, by including ego-documents in the corpus, our research also responds to pleas for a language history from below, integrating sources that are conceptually closer to the spoken vernacular into the analysis (Elspaß, 2007; Elspaß, 2012b).

Systematic corpus analyses allow us to assess the concept of pluricentricity in Dutch language history, and thus also contribute to more theoretical discussions on pluricentricity. The concept proved to be useful for a range of different languages, but has also been criticized in recent years. Especially in the German language area, pluricentricity is questioned, and researchers critical of the concept have introduced the competing concept of pluriareality. Pluricentricity, in fact, is built around national varieties, separated by national borders. In the case of pluricentric languages in an adjacent language area (e.g. Germany, Austria, and Switzerland, but also the Netherlands and Dutch-speaking Belgium), the political border between the nation-states is in this sense also a language border, with specific national variants on each side of the border (Ammon, 1998; Dollinger, 2019a). The proponents of pluriareality, however, point out that many variants are not exclusively or categorically bound to the nation-state, and spread beyond national borders. The distribution of these variants often follows traditional dialect borders. The pluriarealists thus opt for an areal distribution of variants, across national borders (Elspaß & Dürscheid, 2017; Elspaß, Dürscheid & Ziegler, 2017; Elspaß & Niehaus, 2014). In the HCD, we built in the possibility to explore pluriareality for Dutch as well: the two peripheral regions (Zeeland and Flanders) are adjacent dialect areas, allowing us to investigate areal distributions for shared dialect features. In any case, an important distinction between the two concepts is the focus on national borders for pluricentricity and the focus on dialect borders for pluriareality (Norrby et al., 2020, p. 207). We also used this distinction in operationalizing the concepts for our study. Depending on the distribution pattern, monocentricity, pluricentricity or pluriareality could apply.

Chapter 1, the introduction, first outlines the context of this dissertation. We further address the research design and research questions. The following research questions are central to this dissertation: Can we use the modern concept of pluricentricity to describe the history of Dutch? In that case, does pluricentricity replace monocentricity? Or is pluricentricity actually a recent phenomenon that is added to regional variation and regional writing practices in earlier centuries?

In Chapter 2, we first highlight the language historical context of our research. The general lines of Dutch language history (16th-19th centuries) are set out from a traditional point of view. We describe how the Dutch standard language took shape over the centuries, and how the standard language developed in the North and the South of the language area. Next, the traditional views are supplemented with findings from more recent empirical research. These findings can complement and, in some ways, correct the traditional view of Dutch language history on some important points, from a historical sociolinguistic perspective. Finally, we elaborate on the research tradition of historical sociolinguistics and provide a brief overview of electronic corpora for the Germanic languages, which have been constructed for historical-sociolinguistic research.

In Chapter 3, we provide an overview of the research tradition on pluricentricity, and especially research in the field of the Germanic languages. Pluricentricity is used as a concept to describe multiple national varieties of a language. In the German language area, however, several linguists opt for the concept of pluriareality instead of pluricentricity. We examine, compare and assess both concepts in this chapter. Afterwards, we look at pluricentricity for Dutch. We explain how the current situation of pluricentricity came about, and equally discuss arguments to situate pluricentricity earlier in time than traditionally assumed. Finally, we elaborate on the historical dimension of pluricentricity and describe how the concepts of monocentricity, pluricentricity and pluriareality were operationalized in our study.

Chapter 4 then addresses the composition and structure of the Historical Corpus of Dutch (HCD). As we described above, the HCD consists of three dimensions, namely a genre dimension, a diachronic dimension, and a regional dimension. We describe the construction and size of the corpus per dimension, and overall. Afterwards, we elaborate on the interpretation of the three genres, namely administrative texts, ego-documents and pamphlets. This chapter also explains the method of selection and collection of the texts in the corpus. When we discuss the digitization of the texts, the method of transcription is explained. Finally, in this chapter we discuss the methodology used in the corpus analyses.

In Chapter 5, we examine the spelling of /a:/ in closed syllables. The traditional practice was to indicate lengthening by adding an -e or -i to the original vowel. This practice is increasingly replaced by doubling of the original vowel, so the variants <ae> and <ai> are opposed to the modern variant <aa>. We could observe that the majority of tokens were spelled with <ae> in our corpus. No major genre differences were uncovered. The language-external variables region and period did exert an important influence on the distribution of variants. The <aa> spelling emerged in Holland and Brabant in the 17th century. By the 18th century, <aa> had convincingly become the prevailing variant in the Northern regions, while the vast majority of tokens in Brabant were still spelled with <ae>, and <aa> was nearly non-existent in Flanders. It was in the 18th century that a clear difference emerged between the North and South of the language area. This difference remained important in the 19th century: whereas Holland and Zeeland had almost completely switched to the <aa> spelling by the 19th century, <ae> remained dominant in Brabant and Flanders.

Chapter 6 focuses on the spelling of word-final /k/. In Middle Dutch, both <c> and <ck> spellings occurred, but these older forms were replaced by a single <k> spelling, as in the modern standard. 52.4% of the tokens were spelled with the modern variant <k>. The number of spellings with <k> was highest in the printed pamphlets. This makes pamphlets the most progressive genre. The more formal administrative texts were the most conservative. For this spelling feature we again noted a North-South difference in the 18th century: while the Northern regions had almost completely adopted the <k> spelling, <ck> was still the prevailing variant in the Southern regions. We inferred from this that the change occurred at a somewhat slower pace in the South of the language area. By the 19th century, however, this North-South difference had disappeared. The Southern regions had also almost completely switched to the <k> spelling by then.

In Chapter 7, we discuss the spelling of 2nd and 3rd person singular and 2nd person plural present indicative endings of d-stem verbs (e.g. worden ‘to become’). The modern standard has a <dt> spelling, but in the past this dental root could be spelled as <t>, <d> or <dt>. Of these three variants, the <dt> spelling was most strongly represented in our corpus. The preference for <dt> was reflected in the administrative texts and in the pamphlets in particular. In the ego-documents, however, less than half of the tokens were spelled with <dt>. In the 16th century, <dt> and <t> were competing variants. Afterwards, the three spelling variants succeeded each other as the dominant variant: <t> in the 17th century, <d> in the 18th century and finally <dt> in the 19th century. When the modern <dt> spelling then prevailed in the 19th century, a North-South difference emerged. The <dt> ending had clearly become the norm in the Northern regions, while a substantial part of the tokens in the South was still spelled with <d>. Within the South, there was still a striking difference between the center of Brabant and the periphery of Flanders: while Brabant had largely adopted the modern variant <dt>, the spelling with <d> was still dominant in peripheral Flanders.

In Chapter 8, we investigate the disappearance of the unstressed vowel in 1st person singular present indicative endings (e.g. ik hebbe > ik heb ‘I have’). This phenomenon is also known as schwa apocope. In this chapter, we have disregarded the administrative texts, as 1st person singular verb forms hardly occur in the administrative documents in our corpus. The final schwa in verbs such as hebbe ‘have’ was ubiquitous in Middle Dutch, but the modern variant is a verb form without final schwa (heb). We were able to distinguish four possible endings in our corpus: no ending (-∅), -e, -en and -t. The last two variants, however, only rarely occurred in the corpus. The verb forms without final schwa were dominant in our corpus (75%). Apocope of final schwa was already visible in the 16th century, and by the 19th century, the schwa ending had almost completely disappeared from language use. The change from verb forms with schwa to verb forms without schwa was led by Holland in the 17th century. By the 18th century, the new variant without final schwa had taken over in the other three regions (Zeeland, Brabant, Flanders) as well. For the genre differences, we limited ourselves to the ego-documents and the pamphlets. We observed that the percentage of the new variant was higher in the pamphlets than in the ego-documents, so the pamphlets were more progressive.

Clause negation is discussed in Chapter 9. We investigated the change from bipartite negation (e.g. Ik en geloofde het niet ‘I did not believe it’) to modern single negation (e.g. Ik geloofde het niet ‘I did not believe it’). 66.3% of the negative sentences in our corpus contained single negation. The change was already visible in the 16th century and nearly complete by the 19th century. Single negation was most advanced in Holland in the 16th century, but also appeared in the other regions. In the 17th century, a notable increase in the percentage of single negation was recorded in Holland, Zeeland and Flanders. These regions were ahead of Brabant, where bipartite negation was still the dominant variant. In the 18th century, we observed a North-South difference: the change had already largely taken place in the North, but in contrast to the 17th century, Brabant followed the Northern regions in the 18th century. In Flanders, however, bipartite negation was clearly the dominant variant in the 18th century. In other words, the North-South difference that emerged in the 18th century is primarily a difference between the Northern regions (with Brabant joining) and Flanders. In the 19th century, the situation stabilized, and there was hardly any variation anymore. In general, single negation was most common in the pamphlets, and the percentage of bipartite negation was highest in the administrative texts. The percentage of single negation was slightly higher in the ego-documents than in the administrative texts.

In Chapter 10 we look into the shift from d- to w-forms in relativizers, i.e. the relative adverb (daar > waar ‘where’) and the relative pronominal adverb (daar + PREP > waar + PREP, e.g. waarmee ‘with which’). The modern w-form proved to be the dominant variant in our corpus (59.8%). The change to w-forms had already clearly taken off by the 16th century and was largely completed by the 19th century. In the 16th century, the w-forms appeared more often in the central regions of Holland and Brabant than in the corresponding peripheries. So, within the North and the South of the language area, we could observe a center-periphery difference in the 16th century. In the 17th century, however, this difference was eliminated in the North, and there was only a small difference between Brabant and Flanders in the South. In the 17th and 18th centuries, we rather observed changing regional tendencies. The situation stabilized in the 19th century when all regions definitely opted for the modern w-forms. The distribution by genre revealed that the administrative texts and pamphlets were slightly more progressive than the ego-documents. When we took the diachronic dimension into account, we found out that the administrative texts adopted the w-forms in large numbers only from the 18th century onwards. In other words, the change took place at a different pace within each genre.

Chapter 11 is the conclusion of this dissertation. We were able to formulate general conclusions regarding the language-external variables of period, genre and region. For the factor period, we observed that the incoming variant spread through the centuries in the form of an s-curve, except for the spelling of /t/. Looking at the different s-curves, we noted that the incoming variant often really broke through in the 18th century. As for the genre differences, we could observe that the incoming variants were generally more likely to be adopted by the pamphlets. The language used in this genre can thus be characterized as the most progressive. This is not surprising, as pamphlets are printed and edited texts, which generally display more standardized language use, whereas administrative documents and ego-documents are handwritten texts. For the factor region, we observed different patterns of diffusion for different variables. The incoming variants did not spread following the same pattern each time, for example, from the North to the South of the language area, or from the center to the periphery within the North and the South. The center-periphery differences did not turn out to be very strong, especially not in the Northern Netherlands. The dynamics between the North and the South of the language area proved more important in our study. Three variables showed a North-South difference in the 18th century: spelling of /a:/, spelling of /k/ and negation. For the spelling of /t/ we observed a North-South difference in the 19th century.

We argued that our results do not support the idea of consecutive monocentricity, which is nonetheless the dominant discourse in traditional language histories. Except for the schwa apocope, there was no clear center from which the incoming variants spread. Pluriareality does not seem to be a useful concept either, as both peripheries usually display fairly different behavior, in spite of their geographic adjacency. For several variables we observed a North-South difference in the 18th century. This North-South divide can be related to pluricentricity in the national sense, as the political border between the Northern and Southern Netherlands seemed to have become more of a linguistic border. This was especially true for the spelling of /a:/ and the spelling of /k/. Hence, pluricentricity in the national sense can be a useful descriptive notion, but it clearly does not always apply to the same degree in each situation and for each feature. In the most basic sense of there being more than one normative center (i.e. without the national component), pluricentricity was even less prevalent. For Dutch, then, we can conclude that there is no reason to put forward pluricentricity as a crucial concept in language history.

Read more about Dr. Van de Voorde’s research: