Conference Agenda

In the Introduction session of this workshop, we will give a general overview of the use of agent-based models in historical sociolinguistics, by discussing how to construct a model given a research question, what are different paradigms that exist within agent-based modelling and which sociolinguistic data can be integrated.

Around 1930, the Standard Dutch pronunciation in the Netherlands sounded remarkably similar to the Standard Dutch still spoken today in Belgium. Interestingly, the two standards were aligned at one point, but after 1930, the Dutch standard started to shift away (H. Van de Velde 1996). The most notable aspects of this shift were, among others, the diphthongisation of the [e] and [o] sounds towards [ei] and [oʊ] and the devoicing of syllable-initial fricatives. Speakers in Belgium did not follow this shift, which led to the divergence we see today. There have been several attempts to explain this divergence on the basis of social factors: (i) there was not enough contact between the two countries (van den Toorn 1997) (ii) the Dutch standard evolved too fast, and Belgian speakers could not follow its pace (F. Van de Velde 2019) (iii) Belgian speakers did not want to sound like the Dutch (‘ethnocentrism’, Deprez 1985) (iv) a difference in language norms in Dutch and Belgian public broadcasting companies (H. Van de Velde 1996) accelerated change in The Netherlands and inhibited change in Belgium. Since there is very little data available – and we cannot go back in time to collect evidence – it is difficult to investigate these sociolinguistic tendencies in a historical context. Therefore, the theories are examined using an agent-based simulation model. The results show that a lack of contact between both countries can indeed lead to divergence in the model, but only if abroad travel is at least 5000 times less likely than domestic travel. The pace of language change in the Netherlands does not have a sizeable impact on convergence or divergence tendencies in Belgium in the model. High values for ethnocentrism in Belgian agents are able to lead to divergence in the model, as long as these high values are shared by the entire population. Media receptiveness (assumed to be a powerful force in this model) in agents always kickstarts convergence in the model and it accelerates this convergence as well.

References

Deprez, Kas. 1985. ‘De aard van het Nederlands in Vlaanderen’. Heibel: onafhankelijk kritisch-satirisch-polemisch tijdschrift 19 (4): 101–27.

Toorn, Maarten Cornelis van den. 1997. Geschiedenis van de Nederlandse taal. Amsterdam: university press.

Van de Velde, Freek. 2019. ‘External History of Dutch’. Lecture, Leuven, May 8.

Van de Velde, Hans. 1996. Variatie En Verandering in Het Gesproken Standaard-Nederlands (1935-1993). https://repository.ubn.ru.nl/handle/2066/146159.

One of the best-known long-term changes in the Germanic languages is the rise of the weak past tense (e.g. kick ~ kicked) at the expense of the strong past tense (e.g. sing ~ sang). This change has been progressing at different speeds in various languages, with the strong inflection having declined more in English than in Dutch or German, for instance. We propose that such differences can be explained in part through demographic developments, namely migration and urbanization (compare Lupyan and Dale 2010; Dale and Lupyan 2012: Bentz and Winter 2013, although see Cuskley et al. 2015; De Smet et al. 2022). We begin by investigating the behavior of an agent-based simulation that implements both inflections as competitors, with the strong inflection initially dominant in terms of frequency, and still consisting of several regular ablaut classes. Its crucial distinction with the weak inflection in the simulation only lies in the weak inflection’s general applicability, i.e. the ability of the weak inflection to form past tense forms for any verb, in principle. The simulation predicts that an influx of new agents indeed drives up the success of the weak inflection: the greater the demographic upheaval, the stronger the surge of the weak inflection.

Next, we test this prediction with empirical data about historical language changes and demography. In particular, data from English and German are taken respectively from Lieberman et al. (2007) and Cuskley et al. (2014), and we replicate these studies for Dutch. These data are then cross tabulated with historical urbanization rates from Bairoch et al. (1988). It is found that (i) both urbanization and weakening rates are highest in English, followed by Dutch and German, and (ii) English urbanization best predicts English weakening, Dutch urbanization best predicts Dutch weakening, and German urbanization best predicts German weakening.