Universiteit Leiden

nl en


Retired and Kicking: An LUCL Symposium

Wednesday 25 October 2023
Retired and Kicking series
Cleveringaplaats 1
2311 BD Leiden

This is the programme of the event. 


A recent development in the study of argumentative discourse is the systematic attention for its ‘argumentative style’ (van Eemeren et al. 2022). The notion argumentative style captures the idea that arguers, in their attempts to convince an audience, do that in a particular way.

Argumentative style is a more comprehensive notion than linguistic style, because it also takes into account  the choice of specific (types of) argument, the specific way(s) arguers try to adapt to their audience and the (in)coherence between these three dimensions.

An argumentative style can be highly determined by institutional goals and demands. In this talk I discuss and analyse the argumentative style used in a Dutch legal dispute. On the basis of a case study I will examine several institutionally determined aspects of Dutch legal argumentative style.


Eemeren, F. H. van, Garssen. B., Greco, S., Haaften, T. van, Labrie, N., Leal, F., & P. Wu (2022). Argumentative Style. A Pragma-Dialectical Study of Functional Variety in Argumentative Discourse. Amsterdam-Philadelphia: John Benjamins. Argumentation in Context 20.

To reduce the risk of misperception, the words in the lexicon of any language should differ substantially from one another, and they generally do. Nevertheless, even if we exclude homonyms/homophones, pairs of words can easily be found that differ only marginally (e.g., minimal pairs such as cat-rat). For use in forensic applications, I have determined the statistical distribution of several lexical similarity measures for all pairs of words sampled from the 3000 most frequently occurring lemmas of content words, separately in Dutch and in English. Each word was listed in normal spelling (orthography) and in a broad phonemic transcription (one symbol per phoneme, except long vowels and diphthongs, which were represented as two symbols; one symbol for primary word stress). After elimination of homophones, some 3 million word pairs remain per language for statistical analysis. Similarity measures were the percentages of symbol bigrams (bi%) and trigrams (tri%) shared between the members of word pairs, as well as the length-normalized plain and feature-weighted string-edit distance (a.k.a. Levenshtein Distance, Heeringa 2004, Heeringa & Braun 2003), resp. pLD% and wLD% – no feature weighting was attempted for orthographic distance).

In my talk I will briefly explain these measures and then present the correlations among them. The least correlated phonetic similarity measures in both languages are bi% and wLD%. Using a weighted combination of these two measures allowed us to correctly predict 82.4% the court’s decisions in trademark litigation cases in the USA (Disner & Van Heuven 2023). For each of the similarity measures, bivariate distributions were then computed for the orthographic versus phonetic distance. The results show rather poor correlations between orthographic and phonetic distance. The difference in pronunciation between two Dutch words can be predicted from their orthographic difference with an accuracy of no more than 33%. For (American) English, the prediction accuracy is a mere 21%, which confirms the general intuition that the spelling-to-sound correspondences are more regular in Dutch than in English. For either language, the conclusion follows that the generally accepted legal practice (both in the EU and in the USA) to equate spelling and pronunciation in court decisions about confusing similarity between trademarks (i.e., product names) is naïve if not outright wrong (Torenbosch & Van Heuven 2023).


Disner, Sandra F. & Vincent J. van Heuven (2023). Some measures of phonetic similarity for use in legal trademark disputes. Proceedings of the 20th International Congress of Phonetic Sciences, Prague, pp. 3840–3844. https://www.researchgate.net/publication/371340744.

Heeringa, Wilbert J. (2004). Measuring dialect pronunciation differences using Levenshtein distance. Center for Language and Cognition Groningen. https://pure.rug.nl/ws/portalfiles/portal/9800656/thesis.pdf.

Heeringa, Wilbert J. & Angelika Braun (2003). The use of the Almeida-Braun system in the measurement of Dutch dialect distances. Computers in the Humanities, 37, 257–271. 

Torenbosch, Jorn R. & Vincent J. van Heuven (2023). Visuele en auditieve overeenstemming van merknamen: fabels en feiten [Visual and auditory similarity between trademarks: fables and facts]. Intellectuele Eigendom en Reclamerecht (accepted).

In my talk, I will present the main results of the Linguistic History of East Africa project (www.lheaf.org). We refined reconstructions of subgroups of Cushitic resulting in new insights into the (sub)classification of South Cushitic. Using these, we examined contact between the major languages families of East Africa in various subregions. This results in a new and radically different view on their spread urging re-examination of earlier proposals for links with archaeology. Re-examining borrowing scenarios based on more and better reconstructions of lexical material in these language groups suggest early settlement of South Cushitic in Kenya but late spread to Tanzania, intense contact between various subgroups of East Cushitic, South Cushitic, and various subgroups of South Nilotic, and influence of South Cushitic on East African Bantu cattle terminology. It also urges us to rethink the presence of pre-existing populations and their linguistic and cultural dynamics.

From the sixteenth until the eighteenth centuries, the present-day Netherlands were an attractive destination for migrants of various professions from abroad. A large group originated from complex linguistic areas such as Germany and Scandinavia, which had been dominated by multilingualism for centuries. In my paper I will focus on a selection of letters written by migrants who were sailing aboard Dutch ships in the late eighteenth century and who communicated with relatives, spouses and friends in both The Netherlands and their places of origin. Their letters were preserved among the so-called Prize Papers and made available in my Letters as Loot corpora (https://brievenalsbuit.ivdnt.org and https://brievenalsbuit2.ivdnt.org). I intend to explore the multilingual migrants’ command of Dutch and address the question whether their writings show Dutch features from below or from above, features that in our previous research were found characteristic of either handwritten or printed texts and either lower- and middle-ranks or upper-class writers (Rutten and van der Wal 2014). I will concentrate in particular on maritime migrants who originated from East Frisia and the North-Frisian Isles and were thus familiar with Frisian, Low German, High German and Dutch.

Firstly, the language choice of the letters and the methodological issue of how to establish and characterise the letter writers’ command of Dutch will be discussed. Next, the selected letters will be examined with respect to the question of what Dutch variants were adopted, and to establish the number of deviating, non-Dutch features. Finally, an explanation will be given for the different levels of Dutch proficiency attested and for the presence in the letters of a variable number of non-Dutch features.


Gijsbert Rutten & Marijke van der Wal (2014), Letters as Loot: A Sociolinguistic Approach to Seventeenth- and Eighteenth-Century Dutch. Amsterdam & Philadelphia: John Benjamins (OA).

This website uses cookies.  More information.