Borrowing from French into Dutch (1500–1899): Testing the diffusion and graduality assumptions

Gijsbert Rutten; Brenda Assendelft

doi:10.1515/soci-2024-0017

Enjoy 40% off

academic books on De Gruyter Brill *

Article Open Access

Borrowing from French into Dutch (1500–1899): Testing the diffusion and graduality assumptions

Gijsbert Rutten and Brenda Assendelft

Published/Copyright: January 29, 2025

Published by

Become an author with De Gruyter Brill

Submit Manuscript Author Information Explore this Subject

From the journal Sociolinguistica Volume 39 Issue 1

Abstract

The paper tests two assumptions in loanword research: the diffusion assumption and the graduality assumption. The diffusion assumption holds that lexical borrowings typically gain in frequency over time and diffuse across the community. The graduality assumption implies that borrowings enter the recipient language as non-integrated code-switches and eventually integrate into the grammar of the recipient language. These assumptions are inherently diachronic and are here tested on historical data. We investigate loanwords from French as found in the Language of Leiden Corpus. This corpus comprises Dutch textual materials related to the city of Leiden. It covers the sixteenth, seventeenth, eighteenth and nineteenth centuries, and was built specifically to investigate the influence of French on historical Dutch, which is sometimes described in terms of frenchification. The corpus comprises 8,767 established lexical borrowings from French. We show that there is scant empirical evidence for the diffusion and graduality assumptions.

Keywords: borrowing; loanwords; Dutch; French; historical sociolinguistics; language contact

1 Introduction

The Germanic and Romance language areas border each other in the Low Countries and contact between the Dutch-dominant and the French-dominant communities goes back to the earliest sources. From the thirteenth century onwards, when a fully-fledged Dutch writing tradition emerges, borrowings from French can be found in Dutch sources (van der Wal and van Bree 2014: 173–174). Sociolinguistic research on the historical language contact situation has been scarce; recent publications include Peersman et al. (2015), Assendelft (2023), and Verheyden (2023). In this paper, we investigate loanwords from French as found in the Language of Leiden Corpus, a corpus of Dutch textual materials related to the city of Leiden covering the sixteenth, seventeenth, eighteenth and nineteenth centuries. Specifically, we aim to test two assumptions in loanword research identified and criticised by Poplack and Dion (2012, see also Poplack 2017: 121–122). The diffusion assumption holds that borrowings typically gain in frequency over time and diffuse across the community. The graduality assumption implies that borrowings typically enter the recipient language as non-integrated code-switches and eventually integrate into the grammar of the recipient language. This means that they first display donor-language grammar such as inflectional markers, and only later attract recipient language grammar. These assumptions, which are inherently diachronic, have hardly been tested on historical, written data, which is what we aim to do here. We will show that there is scant empirical evidence for these assumptions.

In Section 2, we present the historical and theoretical background of the current study. Section 3 introduces the Language of Leiden Corpus and presents the general diachronic distribution of loanwords from French in the LOL Corpus. Section 4 tackles the diffusion assumption and Section 5 the graduality assumption. We conclude in Section 6.

2 Historical multilingualism and Dutch-French language contact

2.1 Contact settings and results

We focus on the contact setting in the northern Low Countries here, specifically on the city of Leiden in the region of Holland from the sixteenth to the nineteenth century. The Dutch and French languages share a long history with extensive contact effects (Peersman et al. 2015). This includes the use of French alongside Dutch, Latin, and still other languages in the context of trade, diplomacy, administration, and literary culture from the Middle Ages onwards (Frijhoff 2015; Sleiderink 2010; Willemyns 1994). Whereas the southern Low Countries, i.e. present-day Belgium and Luxembourg, are characterised by societal and individual multilingualism with large French- and Dutch-dominant communities (Vanhecke and De Groof 2007; Willemyns 2015), French was often a later-learnt language in the Dutch-dominant north, including the city of Leiden, although migration from the south brought many speakers of French to Leiden. Due to its importance in several social domains, French became a part of the Dutch school system, which intensified in the early modern period after the arrival of tens of thousands of religious and economic migrants, including so-called Huguenots, in the sixteenth and seventeenth centuries (Dodde 2020; Frijhoff 2003). It is usually claimed that the use of French extended into the private domain in the seventeenth and eighteenth centuries, particularly among elites who adopted French in letters and diaries (Frijhoff 1989; Ruberg 2011; van Strien-Chardonneau and Kok-Escalle 2017). As such, the early and late modern Low Countries participated in the European tradition of francophonie (Argent et al. 2014). The intensive and enduring contact with French also affected the Dutch language itself, in which many contact-induced changes at the level of the lexicon and the morphology can be discerned (Assendelft et al. 2023a; Rutten and Assendelft 2023; van der Sijs 2002: 215; van der Wal and van Bree 2014: 173–174).

A significant metalinguistic effect of the contact situation was the emergence of an anti-French discourse of frenchification (Frijhoff 1989; Rutten et al. 2015). The peak of this presumed frenchification is usually located in the seventeenth and even more strongly in the eighteenth century. In the course of the eighteenth century, however, the anti-French discourse was accompanied by the rise of cultural nationalism and the standard language ideology, resulting in a top-down language policy of dutchification around 1800 (Rutten 2019).

The historical Dutch-French contact situation has long been understudied, particularly from a linguistic perspective. Recently, language choice in the northern Low Countries, i.e. in the Netherlands, has been explored in Ruberg (2011), van Strien-Chardonneau (2014, 2017, 2018), Kok Escalle (2017), Puttaert et al. (2022) and Krogull et al. (2023). Most of these publications focus on the eighteenth and nineteenth centuries. Lexical and morphological influence of French on Dutch have been discussed in Rutten et al. (2015), Stevens (2019), Assendelft (2023), Assendelft et al. (2023a, 2023b), and Rutten and Vosters (2023). Verheyden (2023) discusses the frenchification discourse and contact-induced change in the southern Netherlands (Belgium) in the eighteenth and nineteenth centuries. To the best of our knowledge, an extensive discussion of the diffusion and graduality assumptions of loanword research (Poplack and Dion 2012; Poplack 2017) has not been carried out with historical Dutch data.

2.2 Borrowing and the integration of loanwords

The diffusion and graduality assumptions in research on borrowing have been discussed by Poplack and Dion (2012) and Poplack (2017), who set out to analyse borrowing, specifically with respect to the behaviour of individual loanwords in the language of a community. They distinguish nonce borrowings from bona fide loanwords, arguing that bona fide loanwords have become part of the recipient language and are therefore not the main target for research on borrowing: as part of the recipient language lexical stock, they are also available to monolingual speakers. In this view, the diffusion assumption, which states that borrowings typically gain in frequency over time and diffuse across the community, should be investigated on the basis of nonce borrowings in order to find out to what extent such initially unique borrowings increase in frequency at later moments in time. Poplack (2017: 125–127) shows that the diffusion assumption does not hold for the English-origin borrowings in her corpora of nineteenth-, twentieth- and twenty-first century Canadian French. The large majority of nonce borrowings (85 %) occur only once in the data, or recur at a later moment, but again only once, while only 15 % recur more often. Poplack (2017: 127–138) then goes on to show that borrowings should not be treated as incidental code-switches, since they are usually integrated into French grammar, i.e. English verbs receive French inflection, English nouns receive French plural marking and are assigned grammatical gender, and so on. Since this integration happens on the first and often only occurrence of a certain donor language item, Poplack (2017: 137–138) concludes that the graduality assumption does not hold.

Poplack and Dion (2012) and Poplack (2017) are based on audio recordings; the oldest speakers were born in the nineteenth century. At the same time, quite a few nonce borrowings in the data are from the twenty-first century, and therefore too recent to establish a reliable trajectory through time (see Poplack 2017: 126). Diffusion and graduality are inherently diachronic concepts so that datasets with even more historical depth may shed a valuable light on the matter. Ingham et al. (2021) investigated the integration of French-origin items into medieval English and found that most loanwords were immediately integrated, hence that there was no evidence that they began as code-switches and then gradually integrated into English. Our large-scale corpus with Dutch data from four centuries allows us to trace the trajectory of French loanwords across time, analysing both diffusion and integration.

3 Lexical borrowing from French into Dutch

3.1 Method

We use the Language of Leiden (LOL Corpus) for our research, which was developed specifically to trace the influence of French on Dutch in history. Since previous research had indicated strong regional differences (Rutten et al. 2015), we decided to neutralise the geographical factor by focussing on one city, viz. Leiden, which moreover attracted migrants from French-dominant areas in the sixteenth, seventeenth and eighteenth centuries (Kuijpers 2005: 94–95).

The two main pillars of the corpus are the temporal and the sociohistorical dimension. Since the idea of frenchification is typically associated with the seventeenth and eighteenth centuries (Section 2.1), we designed a corpus covering those centuries including a substantial period before and after, i.e. from 1500 to 1899. Going back further in time would also pose serious problems for data collection, as it was already relatively difficult to find sources for the sixteenth century. The position of French as an international prestige language was gradually taken over by German and English from the nineteenth century onwards. We should thus be able to trace French influence by focussing on the period from 1500 to 1899. We divided this long period into 8 shorter periods of 50 years.

Seven social domains were identified that characterise the social and cultural history of Leiden: the Academy, Charity, Economy, Religion, Literature, Private life, and Public opinion. Textual sources were sought and found for each domain. The data collection resulted in four domains contributing texts of an administrative nature, and three domains with non-administrative texts covering a diversity of genres (plays, private letters, and newspaper articles). All documents were transcribed into txt-files, and all transcriptions were manually checked.

For each domain, we selected c. 5,000 words per 50-year period. Lack of relevant archival materials made this impossible for the domains Religion, Literature and Private life in the period 1500–1549. In 1550–1599, the texts for Private life are just under 5,000 words. The university of Leiden was founded in 1575 so that there are no texts for Academy for the first half of the sixteenth century. The first Leiden-based newspaper came out in the second half of the seventeenth century.

Table 1 gives an overview of the LOL Corpus. The structure of the LOL Corpus has been explained in more detail in Assendelft (2023) and Assendelft et al. (2023a, 2023b).

We tried different ways to select loanwords automatically from the LOL Corpus, but eventually had to rely on an inductive approach of manually extracting French loanwords while reading through the entire corpus (Assendelft 2023: 153–156). All loanwords were looked up in the main historical and etymological dictionaries of Dutch, available on the GTB website (https://gtb.ivdnt.org/search/) and in the Etymologiebank (https://etymologiebank.nl). This resulted in a distinction between possibly French loanwords and established French loanwords. If one of the historical or etymological dictionaries indicates that a word originates from French, it is a possibly French loanword. This category thus also comprises words for which at least one of the dictionaries mentions another donor language, usually Latin. Within the category of possibly French loanwords, there is a smaller category of established French loanwords, for which there is general agreement in the historical and etymological dictionaries that they are borrowings from French. One dissenting voice in the dictionaries suffices to categorise a word as a possibly French loanword.

Table 1:

Overview of the Language of Leiden Corpus (n.a.= not applicable)

Domain	Academy	Charity	Economy	Religion	Literature	Private Life	Public opinion
Genre	Administrative				Non-administrative
	Minutes	Wills	Ordinances Requests	Minutes	Plays	Letters	Newspaper articles
Period								Subtotal period
1500−1549	n.a.	5,027	5,072	-	-	-	n.a.	10,099
1550−1599	5,046	5,229	5,118	5,305	5,116	4,449	n.a.	30,263
1600−1649	5,124	5,131	5,276	5,259	5,138	5,114	n.a.	31,042
1650−1699	5,177	5,111	5,314	5,128	5,143	5,032	5,053	35,958
1700−1749	5,025	5,082	5,189	5,153	5,183	5,421	5,111	36,164
1750−1799	5,067	5,290	5,212	5,128	5,112	5,116	5,095	36,020
1800−1849	5,160	5,114	5,100	5,258	5,173	5,145	5,084	36,034
1850−1899	5,157	5,037	5,052	5,271	5,194	5,038	5,088	35,837
Subtotal domains	35,756	41,021	41,333	36,502	36,059	35,315	25,431	251,417

Our analysis is thus based on etymology. As mentioned above (Section 2.2), Poplack and Dion (2012) and Poplack (2017) argue that bona fide loanwords are part of the normal lexical stock of the recipient language. An important criterion to distinguish bona fide loanwords from recent borrowings is their incorporation into dictionaries (Poplack 2017: 27). In our case, relying on contemporary dictionaries and wordlists, to the extent available, does not seem to be the best choice, since until very recently dictionaries were not based on representative corpora. A purist dictionary such as Lodewijk Meyer’s well-known Woordenschat (1669), for example, does not record 31 loanwords beginning with the letter a attested in the LOL Corpus, which amounts to c. 20 % of all loanwords with a in the corpus, some of which were however already borrowed in the medieval period (e.g. aalmoes ‘alms’, aluin ‘alum’). Obviously, degree of integration and frequency of use in the LOL Corpus, as the topics of the present research, can also not be used to distinguish between bona fide loanwords and recent borrowings. In the absence of criteria to identify bona fide loanwords, we opted for an etymological approach. We thus extracted all possibly French and established French loanwords from the LOL Corpus and traced their frequency across time and social domain (Section 4) as well as their degree of integration at the moment of borrowing (Section 5). We focus here on morphological and syntactic integration, disregarding the even thornier issue of phonetic integration, which depends on different and variable spelling conventions.

3.2 General results

The LOL Corpus comprises 15,186 loanword tokens possibly of French origin. Of these, 8,767 loanword tokens are established French-origin loanwords, while the remaining number of 6,419 tokens are contested: they may be borrowed from French, but they can also be borrowed from Latin or still another language such as Italian. The established French loanwords make up 3 % of the vocabulary in the LOL Corpus (viz. 8,767 out of 251,417 words). The 8,767 tokens correspond to 1,155 unique loanword types. We focus on the established tokens and types here, as these constitute the most reliable evidence of French influence on Dutch.

Figure 1 is based on the 8,767 loanword tokens. It presents the normalised frequency per 1,000 words per 50-year period of the LOL Corpus, divided into administrative and non-administrative use. The general evolution of the French-origin lexical stock, based on administrative and non-administrative use together, exhibits an increase in the sixteenth and seventeenth centuries, a frequency peak in the eighteenth century, and a decrease in the nineteenth century (Assendelft et al. 2023b: 172). Figure 1 shows that the general evolution is different for the administrative and non-administrative domains. The administrative domains display a sharp increase in the sixteenth century, reach a high normalised frequency in the first half of the seventeenth century, remain high until the late eighteenth century, and then drop to the level of 1550–1599. The non-administrative domains are lower overall, increase later, viz. in the seventeenth century, and already decrease in the late eighteenth century.

Figure 1:

Established loanwords from French in the LOL Corpus, normalised frequency per 1,000 words, per 50-year period and across domain group

Assendelft (2023) and Assendelft et al. (2023a, 2023b) show that the diachronic pattern is even more complex, and less clear, if the results are split up at the level of individual domains. While some domains (Academy, Charity, Economy) follow the big diachronic trend represented by the administrative line in Figure 1, other domains display low numbers of French-origin lexical and morphological items throughout the centuries. An example is Literature, which comprises still quite a few French-origin items in the sixteenth century, after which they virtually disappear from the literary register. It has become clear in our previous research that the difference between administrative and non-administrative is one of the strongest factors influencing French-origin items in historical Dutch. As Figure 1 shows, a first wave of French influence emerges in the sixteenth century and is largely restricted to the administrative domains. A second wave is more strongly associated with the seventeenth and eighteenth centuries and affects the non-administrative domains. All domains show a frequency decrease of French-origin items in the late eighteenth or nineteenth century, when the Dutch language is ideologically constructed as a symbol of the Dutch nation-state (Rutten 2019; Rutten and Assendelft 2023).

4 The diffusion assumption

4.1 Type frequency

The diffusion assumption states that other-language items typically gain in frequency and diffuse across the community. The results in Figure 1 appear to confirm the frequency increase, which is however halted around 1800, when French-origin lexical items decrease. Diffusion across the community is also seen in Figure 1, which shows that French loanwords first occur in the administrative domains and later diffuse to non-administrative domains. Figure 1, however, shows normalised frequencies of all established loanwords in the LOL corpus. The diffusion assumption is about the trajectory of individual borrowings. In this section, we delve deeper into the issue of diffusion, taking also into account the actual loanword types involved.

It appears that an increasing frequency of the French-origin lexical stock does not at all imply an increasing frequency at the level of individual loanword types. Table 2 presents the token frequency of the 1,155 established loanword types from French. It shows that 454 types occur only once in the entire LOL Corpus and another 528 types occur between 2 and 10 times in the corpus. Together, these types make up 85 % of all loanword types. The LOL Corpus comprises 8 periods of 50 years. This means that a gradual diachronic increase of words with only 1 to 10 tokens in total is a priori hardly demonstrable. Instead, most words with relatively low token frequencies occur only once or perhaps a few times in isolated 50-year periods. While the results in Figure 1 may thus present powerful generalisations at the community level, the borrowing of individual types needs to be investigated on the basis of highly frequent types. We will focus on the 31 frequent types with more than 50 tokens in the remainder of this section.

Table 2:

Frequency of established loanword types

Token frequency in LOL Corpus	N Types	% Types
1	454	39
2–10	528	46
11–20	77	7
21–50	65	6
51–100	23	2
> 100	8	1
Total	1,155	100

4.2 Frequent types

In order to uncover potential diffusion across time and across domains, we adopt a series of methods. We first investigate the absolute frequencies of the most frequent types across time to find out whether there is a general pattern of increasing frequency as a sign of diffusion. A second way to tackle this issue is by zooming in on the types that occur in each of the 50-year periods. A third way is to look at the diffusion across social domains by tracing the occurrence in domains across time, focusing on types that occur in each domain in the corpus, in order to find out whether there is a general pattern of an increasing number of domains a type is used in. Finally, we compare the distribution of the most frequent types in administrative and non-administrative domains as this distinction appears to be highly relevant (see above, Figure 1).

We first look at the 31 most frequent types, which have an absolute token frequency of more than 50 in the entire LOL Corpus. Table 3 lists these 31 types alphabetically along with the absolute token count. The following columns give the absolute token count per 50-year period. The two periods with the highest token count are marked grey. If the second highest token count occurs in more than one period, all 50-year periods with this token count are grey.

Table 3 shows that some frequent types follow the perhaps expected pattern from a lower frequency to a higher frequency over time, indicating diffusion; examples are faculteit and missive. Such an increase may be followed by a decrease, as in the case of accorderen and passeren. Typically, the highest frequencies occur after 1650, although earlier frequency peaks are also attested (saai, sint). The 50-year periods with the highest frequencies are often adjacent, but certainly not always (e.g. akte, dispositie, suppliant, testatrice). Thus, diffusion in the sense of increasing frequency is only attested in few cases.

The second way to approach the issue is by looking at the loanwords that occur in each 50-year period. In the entire LOL Corpus, there are 15 established loanwords from French occurring from 1500 to 1899. There is a lot of overlap with the previous analysis: 10 out of these 15 types are also among the 31 most frequent types, viz. leveren, passeren, plaats, prijs, prins, publiek, rekwest, sint, som and suppliant. We will not repeat those in Table 4, where we restrict ourselves to the five additional types.

Table 3:

Absolute token frequencies per 50-year period of the 31 established loanword types with >50 tokens in the LOL Corpus

	English gloss	N tokens	1500–1549	1550–1599	1600–1649	1650–1699	1700–1749	1750–1799	1800–1849	1850–1899
accorderen	agree	51	1	1	2	10	10	17	10	0
akte	deed	114	0	14	26	9	14	19	11	21
comparant	appearer	232	0	12	50	18	26	71	9	46
curateur	curator	146	0	50	34	31	18	12	1	0
dispositie	disposition	71	0	4	7	14	12	9	11	14
faculteit	faculty	54	0	2	4	7	3	3	16	19
florijn	florin	96	0	0	0	2	18	15	18	43
Frans	French	54	0	1	7	18	2	5	15	6
leveren	deliver	54	2	9	12	9	5	2	12	3
majesteit	majesty	57	0	0	3	4	21	13	11	5
missive	letter	148	0	1	6	4	7	15	52	63
order	order (noun)	83	0	3	7	21	28	19	2	3
ordonnantie	ordinance	86	5	7	8	13	16	20	17	0
ordonneren	order (verb)	73	4	22	13	17	11	6	0	0
passeren	pass	98	1	3	15	12	22	20	19	6
plaats	place	324	4	20	23	44	42	36	81	74
prijs	price	56	14	8	3	9	4	6	7	5
prins	prince	63	3	3	17	16	15	3	1	5
publiek	public	86	4	5	5	26	19	21	3	3
rapport	report	55	0	7	9	4	5	4	17	9
regent	regent	89	0	1	23	26	8	13	5	13
rekwest	request	59	3	1	3	3	13	5	23	8
resideren	reside	71	0	0	3	12	15	16	17	8
resolutie	resolution	108	0	7	8	19	30	37	6	1
saai^[1]	say	66	1	6	22	27	0	8	2	0
sint	saint	97	15	22	4	22	7	12	7	8
som	sum	159	1	19	25	28	20	14	24	28
suppliant	supplicant	165	2	28	28	23	9	58	16	1
testateur	testator	76	0	11	1	24	30	2	0	8
testatrice	testatrix	72	0	15	0	8	10	1	23	15
toneel^[2]	theatre	74	0	0	0	1	19	17	18	19

Table 4:

Absolute token frequencies per 50-year period of the 15 established loanword types occurring in all 50-year periods of the LOL Corpus

	English gloss	N tokens	1500–1549	1550–1599	1600–1649	1650–1699	1700–1749	1750–1799	1800–1849	1850–1899
conditie	condition	40	1	7	6	9	10	4	2	1
consent	consent	31	1	8	9	5	3	2	2	1
continueren	continue	38	5	5	7	5	8	5	2	1
kopie	copy	48	1	5	4	4	11	3	15	5
manier	manner	50	8	10	5	2	14	7	2	2

Table 4 gives a similar impression of variability as Table 3: conditie displays a gradual frequency increase (followed by a decrease), some types have a peak after 1650 (e.g. continueren), others earlier (e.g. manier), some have their highest frequencies in adjacent periods (e.g. consent), while others do not (manier).

Still another way to discover potential diffusion is by delving into the established loanword types that occur in each of the 7 domains in the LOL Corpus. There are 12 types that occur in all domains. Table 5 gives these 12 types with the number of domains they occur in per 50-year period.

It is immediately clear that hardly any of these types occurs in all 7 domains in one 50-year period. In fact, only plaats reaches the score of 7 and even multiple times. It is also not the case that the number of domains in which these types occur gradually increases. Such a pattern may be found in some cases (officier, order), which also show a subsequent decline in the nineteenth century. In most cases, the number of domains varies from 0 to 4 across time with only a few exceptions. Note also that the 50-year period with the highest token frequency (see Table 3) is not necessarily the same as the one with the highest number of domains: whereas plaats and ordonneren show this overlap, other types do not have the highest token frequencies in the period in which they occur in the highest number of domains (e.g. akte, sint, prijs).

Our final attempt to detect diffusion focuses on the distribution of the 31 most frequent established loanword types across the major groups of domains, i.e. administrative v. non-administrative. Table 6 presents the absolute frequencies across time and domain group.

Table 6 resembles Figure 1 in that it shows that administrative domains adopt loanwords from French earlier and to a higher extent than non-administrative domains. This means that these frequent loanwords were first borrowed in administrative domains and then diffused to non-administrative domains. In both groups, the frequency gradually increases, although the administrative domains rise much faster and remain high until 1899, whereas the non-administrative domains increase more gradually and display the typical eighteenth-century peak followed by a decrease.

Table 5:

Number of domains in which loanwords occur per 50-year period, for all 12 established loanword types occurring in all 7 domains of the LOL Corpus

	English gloss	1500–1549	1550–1599	1600–1649	1650–1699	1700–1749	1750–1799	1800–1849	1850–1899
accorderen	agree	1	1	2	3	4	3	3	0
akte	deed	0	4	3	3	4	5	3	4
Frans	French	0	1	3	2	2	1	6	2
kwartier	quarter	0	4	3	1	4	1	1	2
officier	officer	0	1	2	1	3	5	3	4
order	order (noun)	0	1	3	3	4	5	1	1
ordonneren	order (verb)	2	5	2	3	4	4	0	0
passeren	passe	1	1	4	3	4	2	3	1
plaats	place	2	5	4	7	6	7	7	7
prijs	price	1	4	2	3	4	3	5	4
sint	saint	2	3	1	2	1	3	4	3
som	sum	1	4	4	4	4	3	3	4

In some cases, the diffusion across domain groups can be established at the level of individual words. For example, akte and missive first occur in the administrative domains in 1550–1599 and remain in use until the final period, whereas they first occur in the non-administrative domains in 1700–1749; both remain in non-administrative use until 1899, although missive does not occur in 1800–1849. The type resolutie also first occurs in the administrative domains in 1550–1599 and remains in use until the final period, whereas it only occurs in non-administrative domains in 1750–1799. The highly frequent word plaats first occurs in the administrative domains in 1500–1549 and remains in use until the final period; it occurs once in non-administrative domains in 1550–1599, but not in 1600–1649, and then reappears in 1650–1699 and remains in use until 1899. The word som first occurs in the administrative domains in 1500–1549 and remains in use until 1899; it occurs once in the non-administrative domains in 1550–1599 and once in 1600–1649, but not in 1650–1699, and then appears in 1700–1749 and is used until 1899. While these individual cases demonstrate the initial use in administrative domains and the subsequent use in non-administrative domains, generally with lower frequencies, there are also highly frequent types that do not diffuse to the non-administrative domains and instead remain restricted to administrative use: comparant is almost exclusively used in the domain of Charity, curateur almost only in the domain Academy, and suppliant is restricted to Economy.

Table 6:

Absolute frequencies of the 32 established loanword types with >50 tokens in the LOL Corpus, per 50-year period and across groups of domains

	1500–1549	1550–1599	1600–1649	1650–1699	1700–1749	1750–1799	1800–1849	1850–1899
Administrative	60	272	363	398	355	398	368	356
Non-administrative	0	10	5	83	104	101	86	78

4.3 Conclusions about the diffusion assumption

Figure 1 shows diffusion in terms of increasing frequency and from the administrative to the non-administrative domains. Nonetheless, the diffusion assumption, which is about the trajectory across time and domains of individual loanwords, can hardly be confirmed for individual loanword types in the LOL Corpus. The large majority of established French loanwords occurs only once or a handful of times in the corpus. Only a few highly frequent lexical items follow the trajectory of increasing frequency across time and of diffusion from administrative to non-administrative domains. In sum, the idea of diffusion across time and domain is visible at the level of aggregate loanword frequencies in the language community, represented by the LOL Corpus, but it is not the typical trajectory of a concrete lexical borrowing from French into Dutch in our corpus.

5 The graduality assumption

Assendelft (2023: 217–225) analyses the morphological integration of established loanword tokens in the LOL Corpus. The corpus comprises 8,767 loanword tokens from French (Figure 1), of which many, however, were borrowed from French in the Middle Ages, well before the beginning of the LOL Corpus. Excluding those forms, Assendelft restricts the analysis to 3,046 loanword tokens borrowed between 1500 and 1899, i.e. during the timespan of the corpus. She demonstrates that 1,186 or 39 % of these forms have Dutch inflectional marking and are thus integrated into the Dutch morphological system. Another 1,675 tokens (55 %) do not show any morphological marking and cannot be said to be morphologically integrated. A group of 185 nouns (6 %) have the plural suffix -s, which is identical to the French plural suffix. This means that there is no grammatical conflict site (Poplack 2017: 26), so that it cannot be determined which grammatical system operates when the plural suffix -s is selected.

The aforementioned 1,186 tokens comprise 415 inflected verb forms. Verbs borrowed from French receive the suffix -en in the infinitive and in plural present tense forms, for example feliciter-en ‘congratulate’; this means they combine two etymologically infinitival suffixes. Other inflectional markers include the prefix ge- and a dental suffix for past participle formation as in ge-insisteer-d ‘insisted’, and the suffix -ende for present participle formation as in amplier-ende ‘amplifying’. There are 426 borrowed nouns with the plural marker -en, for example address-en ‘addresses’. 270 adjectives and nouns have case or gender marking as in respectabel-en ‘respectable’ with the masculine accusative singular suffix -en, or as in magazijn-s ‘of the shop’, which has the genitive suffix -s. The suffix -(e)lijk occurs 45 times to make adverbs from adjectives as in respectiev-elijk ‘respectively’. Finally, there is a small number of diminutive, comparative, superlative, nominalising, and adjectival suffixes (30 in total).

The analysis by Assendelft (2023: 217–225) thus suggests quite extensive integration into Dutch, since a considerable minority of 39 % of the French-origin items borrowed between 1500 and 1899, during the timespan of the LOL Corpus, display Dutch morphology. Nonetheless, many tokens do not show any inflection unambiguously demonstrating morphological integration. In addition, the LOL Corpus covers four centuries, so that integrated items may be restricted to the more recent periods, after having entered the Dutch language as non-integrated switches.

We therefore focus on loanwords that first occur in the LOL Corpus in the period in which they are also first attested in the Dutch language in general, based on the historical and etymological dictionaries (see Section 3). For example, the oldest examples of presideren ‘preside’ in the historical dictionaries (GTB, s. v. presideeren) are from the second half of the sixteenth century, and it also first occurs in the LOL Corpus in the period 1550–1599. The noun souper ‘supper’ is dated to the first half of the eighteenth century (Etymologiebank, s. v. souper) and the first occurrence in the LOL Corpus is likewise from 1700–1749. In total, there are 122 established French-origin loanword types with the first token in the corpus being contemporaneous with the first attestation in the historical and etymological dictionaries of Dutch (Assendelft 2023: 226–227).

For each of these 122 first occurrences, we investigated the morphological and syntactic integration. The analysis by Assendelft (2023: 217–225) was restricted to morphological integration. Dutch does not have a rich morphology in the nominal domain: by the early modern period, case and gender had been in decline in northern varieties of Dutch for centuries, including the language of Leiden. The analysis thus resulted in a large number of items lacking overt morphological integration, including many nouns (Assendelft 2023: 219). Dutch singular nouns that lack nominal inflection, however, may still be assigned grammatical gender. The postmedieval system in northern varieties of Dutch comprises the common gender marked by the definitive article de ‘the’, which combines the former masculine and feminine gender, and the neuter gender marked by het ‘the’. The de/het distinction does not map onto the French le/la distinction so that there is a conflict site in the sense of Poplack (2017: 26): gender assignment, made visible on adnominal words such as definite articles, thus indicates integration into Dutch syntax.

Table 7:

Morphological and syntactic integration of first attestations of loanword types

	N Types	Morphological	Syntactic	Not integrated	Ambiguous
Preposition	1	0	0	0	1
Verb	25	25	n.a.	0	0
Adjective	23	12	11	3	6
Noun	73	20	36	0	21
	122

Table 7 presents the 122 loanwords and the frequency of morphological and syntactic integration per part of speech. In the case of verbs, we have refrained from analysing syntactic integration since all 25 verbs are already morphologically integrated on the first occurrence in the LOL Corpus. They can be considered integrated into the grammar of Dutch so that a further exploration of their syntactic integration may be interesting, but also superfluous for the present purposes. Some adjectives and nouns are both morphologically and syntactically integrated and are thus counted twice in Table 7, while other words are neither morphologically nor syntactically integrated, or cannot be said to be either because the Dutch and French constructions are structurally similar.

There is one preposition, à, which is morphologically nor syntactically integrated (example 1). The syntactic position within two numerals is identical to the position in French so it cannot be claimed the preposition is integrated into a Dutch syntactic frame.

(1)	een	wooningh	van	4	a	5	plancke [Private, 1650–1699]
	a	house	of	4	to	5	planks

As said, all 25 verbs are morphologically integrated. There are 15 past participles with both the ge-prefix and a dental suffix; examples include ge-surceer-t ‘postponed’ (Economy, 1500–1549), ge-chagrineer-d ‘grieved’ (Public opinion, 1700–1749), ge-motiveer-d ‘motived’ (Academy, 1800–1849). There are also 6 infinitives with the suffix -en attached, for example assopier-en ‘pacify’ (Religion, 1550–1599) and pousser-en ‘push’ (Academy, 1600–1649). Finally, there are 4 finite verb forms with Dutch present or past tense suffixes, such as insereer-t ‘inserts’ (Charity, 1500–1549) and regaleer-de ‘regaled’ (Private life, 1700–1749).

There are 23 adjectives, 12 of which are morphologically integrated and 11 syntactically (note that this is not a breakdown of the total of 23 adjectives as they can be morphologically and syntactically integrated at the same time). The measure for morphological integration is the presence of Dutch morphological material. The measure for syntactic integration is the typical Dutch noun phrase structure with the adjective in prenominal position, as opposed to the canonical French noun phrase structure with the adjective after the noun. In historical French, prenominal adjectives were more common than today, but the strong tendency to place adjectives after the noun already exists in Middle French (1300–1550, see Marchello-Nizia et al. 2020: 997). In total, there are 9 inflected, prenominal adjectives, which we thus consider morphologically and syntactically integrated. An example is the adjective in (2), which has the Dutch inflectional suffix -e and is placed between the article and the noun.

(2)	de	correctionele	regtbank	van	Parijs [Public opinion, 1800–1849]
	the	correctional	court	of	Paris

Another 2 adjectives do not show Dutch inflection but do follow the Dutch word order of adjective before noun, such as atroce in (3).

(3)	sulke	atroce	Innurien	ende	offencien [Religion, 1550–1599]
	such	cruel	injustices	and	offences

A total of 3 adjectives are used adverbially and are morphologically marked as such by the adverbialising suffix -(e)lijk: ordinaer-lyk ‘ordinarily’ (Religion, 1600–1649), serieus-elyk ‘seriously’ (Private life, 1600–1649), and succesiv-elyk ‘successively’ (Charity, 1600–1649).

In addition to these 14 morphologically and/or syntactically integrated adjectives, there are 3 interesting cases of adjectives that are not morphologically integrated, but which moreover display French word order so that they are also, unambiguously, not syntactically integrated. Instead, they maintain French word order. One example is in (4), where clausulen Derogatoir are borrowings.

(4)	eenige	andere	clausulen	Derogatoir [Charity, 1550–1599]
	any	other	clauses	repealing
	‘any other repealing clauses’

According to the Dutch historical dictionaries (GTB, s. v. derogatoir), the adjective derogatoir is characteristic of legal language and often occurs in postpositie, i.e. after the noun. Example (4) shows two inflected pronouns with the suffix -e in front of the plural noun clausulen, which is followed by the uninflected adjective Derogatoir. Uninflected adjectives in postposition occasionally occur in Middle Dutch (1100–1500, van der Wal and van Bree 2014: 134), but disappear later except in legal terminology, where they are undoubtedly contact phenomena from French. A second example from the present dataset is dispositie testamentair ‘testamentary disposition’ (Charity, 1550–1599). Still other examples from the LOL Corpus include notaris publiek ‘public notary’ and testament solemneel ‘solemn testament’ (Rutten and Assendelft 2023). It seems likely that these legal expressions were borrowed as entire phrases, reintroducing postnominal adjectives into Dutch, although only in these fixed phrases.

Finally, there are 6 adjectives that cannot be said to be morphologically and/or syntactically integrated for lack of evidence. This means it can also not be claimed they are not integrated; they are ambiguous. Examples include adverbial usage, where the loanwords constitute a part of speech by themselves so that syntactic integration is difficult to ascertain. A sentence adverb as in (5) can be variably placed both in French and in Dutch (in some cases, commas are even used to demarcate the adjective as a separate adverbial phrase).

(5)	es	com~cative	ooc	voorgeslagen [Academy, 1550–1599]
	is	communicatively	also	proposed
	‘has been proposed by mutual agreement’

For the 73 nouns (see Table 7), we first look at Dutch inflectional marking, which occurs in 20 cases. In 14 cases, the plural suffix -en (or -n if the singular ends in -e) shows the integration into Dutch morphology, as in emolument-en ‘revenues’ (Academy, 1600–1649) and complaisance-n ‘pleasantries’ (Academy, 1700–1749). In a few other cases, diminutive suffixes, nominalising suffixes, and case endings occur; an example is soúpee-tje-s ‘(lit. ‘little suppers’, ‘convivial suppers’; Private life, 1700–1749), which has the diminutive suffix -tje along with the plural suffix -s. There is also one case where the borrowed noun has a genitival -s: testatricis vterste wille ‘testatrix’s last will’ (Charity, 1550–1599).

This means that 53 out of 73 nouns are not morphologically marked, but some of these are of course syntactically integrated, as is the case for 36 nouns in total (Table 7). The primary measure to attest syntactic integration is the combination with definite or indefinite articles or with other adnominal words. In 15 cases, a singular noun is combined with a definite article or a demonstrative pronoun (the historical origin of the definite article). As argued above, the (northern) Dutch system of common gender de ‘the’ and neuter het ‘the’ constitutes a conflict site in the sense of Poplack (2017: 26) with the French masculine le and feminine la, if only because nouns referring to persons are usually either masculine or feminine in French, while receiving de in Dutch. The common/neuter distinction is also visible in Dutch demonstratives: the common gender forms are deze ‘this’ and die ‘that’, and the neuter forms are dit ‘this’ and dat ‘that’. Examples of Dutch gender assignment include de nominatie ‘the nomination’ (Charity, 1550–1599), de Generaliteyt ‘the generality, the general government’ (Academy, 1550–1599), dat droevige chapitre ‘that sad chapter (i.e. business)’ (Private life, 1700–1749), which also has an inflectional -e on the adjective, and tpretecxt ‘the pretext’ (Economy, 1550–1599) with a cliticised definitive article.

There are 9 instances of singular nouns in combination with the indefinite article een ‘a, an’, which also constitutes a conflict site. Historically, the form een is the masculine and the neuter form, while eene is the feminine form. In the early modern northern Netherlands, as in the present-day standard, een is the only form. The system was, however, still in flux in early modern times, so that forms inflected according to gender can still be found in the LOL Corpus. Whatever the exact phase is in this change in the gender system, there is always a grammatical difference with the French un/une distinction. Examples include een qúadrilletje ‘a quadrille, a card game for four people’ (Private life, 1700–1749), which also has the diminutive suffix -tje, and een occasie ‘an occasion’ (Economy, 1500–1549).

There are another 9 cases of singular and plural nouns preceded by other adnominal words inflected in accordance with Dutch grammar. An example is tot grote prosperite ‘to great prosperity’ (Economy, 1500–1549) with the inflectional -e attached to the adjective, indicating the common gender assigned to the noun.

In addition to these 33 instances, there are 2 partitive examples that also indicate syntactic integration, see example (6).

(6)	de	drie	manties	met	de	kan	stroop [Private life, 1650–1699]
	the	three	baskets	with	the	jug	sirup
	‘the three baskets with the jug of sirup’

In Dutch, partitive constructions are characterised by bare nouns (stroop), which constitutes a conflict site with the French partitive article de (e.g. une cruche de sirop), which is firmly grammaticalised by the end of the Middle Ages (Marchello-Nizia et al. 2020: 1562–1563). Finally, there is the aforementioned prenominal genitive testatricis vterste wille ‘testatrix’s last will’ (Charity, 1550–1599), where French would also have a postnominal construction with de.

A total of 37 nouns cannot be said to be syntactically integrated, but lack of linguistic evidence does also not imply the opposite; the examples are syntactically ambiguous. These ambiguous cases include plural nouns with the definite article de as in de respondenten ‘the respondents’ (Academy, 1550–1599). Both in French and in Dutch, there is one plural definite article for all genders, which means that the languages are structurally similar so that there is no conflict site. Nonetheless, 16 out of the 37 ambiguous cases are morphologically integrated. This leaves 21 nouns where it cannot be said whether they are morphologically or syntactically integrated. They do not show any inflectional marking, and they are also not combined with any adnominal words (including articles): they are bare nouns. There are clear patterns in this group of bare nouns. In 6 cases, they are attached to a preposition, as in French, so that there is no grammatical conflict site. In (7), in vigoer is used as French en vigeur, and in (8), the preposition omtrent is followed by the bare noun desinfectie.

(7)	te	dragen	In	vigoer [Economy, 1500–1549]
	to	carry	in	strength
(8)	de	voorschriften	omtrent	desinfectie [Public opinion, 1850–1899]
	the	regulations	concerning	disinfection

The use of bare nouns decreases in postmedieval French (Larrivée and Goux 2024), but since a lot of variation in the presence and absence of determiners remains, not unlike Dutch, and since extensive knowledge of the factors conditioning this variation is lacking, we decided to treat the combination of a preposition and a bare noun as a structural option in historical French and Dutch. This implies that there is no conflict site, so that syntactic integration cannot be established.

In 8 examples, the bare noun is a personal noun used to identify or specify a participant in the communicative situation. In (9), the word maseur ‘sister’, a lexicalisation of French ma soeur ‘my sister’, is used twice as a personal address. The spelling reflects the origin in the French expression consisting of two separate words. In (10), testateur is an apposition specifying the person first indicated by hy.

(9)	aen	ma soeur	susanna	ende	ma soeur	sara
	to	sister	Susanne	and	sister	Sara
	geschreven [Private life 1550–1599]
	written

(10)	heeft	hy	testateur	gelegateert [Charity, 1550–1599]
	has	he	testator	bequeathed

In 7 cases, the bare noun is the nominal part of a subject complement or an abstract noun such as sinceriteyt (Private life, 1600–1649). The nominal complements are also personal nouns, as in te wesen solliciteur ‘to be advocate’ (Literature, 1550–1599). In all these cases, the context provides hardly, if any clues to justify integration into Dutch syntax. The predominance of personal nouns lacking any adnominals strongly suggests that these words may also be treated as clause-peripheral alternations in the sense of Muysken (2013; see also Treffers-Daller 2023), or non-integrated borrowings in the sense of Poplack (2017).

In this section, we have analysed the first occurrences of 122 loanwords from French in the LOL Corpus. It turned out that 1 preposition, 6 adjectives and 21 nouns cannot be said to be morphologically or syntactically integrated or not, so that 28 loanword types are ambiguous. Another 3 adjectives are not syntactically integrated as they display French word order. The remaining 91 loanwords (75 % of 122) are morphologically and/or syntactically integrated on their first occurrence in the LOL Corpus, so that these results do not lend much probability to the graduality assumption.

6 Conclusions

We have sought to test the diffusion and graduality assumptions defined and criticised by Poplack and Dion (2012) and Poplack (2017) on the basis of a large historical corpus of Dutch covering four centuries. The lexical stock originating from French in the LOL Corpus increases over time, from the sixteenth to the eighteenth century, and then decreases. This decrease is probably linked to ideological dutchification around 1800 (Rutten 2019; Rutten and Assendelft 2023), following a previous phase of so-called frenchification. Administrative texts use more loanwords from French throughout the centuries. This means that diffusion can be attested in two perspectives: increasing frequency across time as well as a spread from administrative to non-administrative language. The diffusion hypothesis, however, targets the trajectory of individual borrowings. A large majority of 85 % of the loanword types only occur between 1–10 times in the LOL Corpus. It is only for a handful of frequent types that we can establish both diffusion in terms of increasing frequency and in terms of a spread to other, i.e. non-administrative domains. An analysis of the first occurrences of loanwords contemporaneous with the first attestations found in the historical and etymological dictionaries of Dutch reveals that many types were immediately morphologically and/or syntactically integrated into Dutch. This means that these words do not enter the Dutch language as code-switches and then gradually become loanwords. Instead, they are borrowed and directly integrated into Dutch grammar. It could, of course, be argued that the written language often follows the spoken language and that loanwords integrated into written Dutch may have gone through a stage of non-integration in the spoken language before, but that is a claim that can be neither verified nor falsified for historical periods without oral data. It would also still need to be explained why Poplack and Dion (2012) find similar results for spoken language data.

While Poplack and Dion (2012: 310) argue that knowledge of the donor language is not necessary for the use of integrated nonce borrowings, the social situation that we study may be quite different. In our case, it seems that we are, initially, dealing with a multilingual, Dutch-dominant community, in which some language users, often for professional reasons, were quite proficient in French, reading and writing included. They borrowed words from French, mostly in administrative texts in certain domains. We normally cannot identify the individuals responsible for the production of concrete textual materials, but it is safe to assume that these were relatively well-educated men: notaries, secretaries, administrators with at least domain-specific knowledge of French and/or Latin. In this respect, it is relevant that legal and administrative language were among the main targets of purist discourse from the sixteenth century onwards (van der Wal and van Bree 2014: 195–198). The reason these writers apply Dutch grammar to loanwords is thus not a lack of knowledge of French. A plausible hypothesis would be that language mixing is generally disfavoured within Early and Late Modern language ideologies, in which ideals of purism emerge along with selection and codification (Ayres-Bennett 2021: 43; Langer and Nesse 2012: 611–614). Krogull and Rutten (this issue), who analyse language mixing practices and who also focus on the Dutch-French contact situation in the northern Netherlands, show that multilingual discourse is usually around 10 % or even lower, based on a large, diachronic dataset of personal correspondence. This means that up to 90 % of these letters are either predominantly Dutch or predominantly French. The social acceptability of language mixing in different genres and domains is a topic that requires more research in historical sociolinguistics (Ingham 2018: 331).

In the current study, the increasing frequency of French loanwords in administrative texts shows the sociolinguistic background to the aforementioned purist discourse as well as its lack of success. These loanwords from French, and the act of borrowing from French in general, extend wider into the community in subsequent periods when French loanwords also increase in frequency in non-administrative texts, along with the increased importance and visibility of French in the Dutch society of the time (Frijhoff 2015). While this type of diffusion at the community level is highly relevant for our sociolinguistic research on presumed frenchification and the social embedding of language contact more generally, it cannot count as evidence for the diffusion assumption, which concerns the word level.

Our study confirms the scepticism of Poplack and Dion (2012) with respect to the diffusion and graduality assumptions. Adapting their methodology to historical, written data has enabled us to replicate their results (see also Ingham et al. 2021). Intuitive as the diffusion and graduality assumptions may seem to be, our results suggest that the borrowing process is quite different in our case. We hope our study will inspire other historical sociolinguists to apply the same methodology to their datasets, so that future theories of borrowing will incorporate extensive historical data.

Historical and etymological dictionaries

Etymologiebank. Institute for the Dutch Language. https://etymologiebank.nl

GTB = Geïntegreerde Taalbank. 2007–2018. Institute for the Dutch Language. https://gtb.ivdnt.org/search/#

References

Argent, Gesine, Vladislav Rjéoutski & Derek Offord (eds.). 2014. European francophonie. The social, political and cultural history of an international prestige language. Oxford: Lang.Search in Google Scholar

Assendelft, Brenda. 2023. Verfransing onder de loep. Nederlands-Frans taalcontact (1500–1900) vanuit historisch-sociolinguïstisch perspectief. Amsterdam: LOT.Search in Google Scholar

Assendelft, Brenda, Gijsbert Rutten & Marijke van der Wal. 2023a. Tracing Frenchification: a sociolinguistic analysis of French loan suffixes in a historical corpus of Dutch. In Rita Franceschini, Matthias Hüning & Peter Maitz (eds.), Historische Mehrsprachigkeit: Europäische Perspektiven, 37–55. Berlin: De Gruyter.10.1515/9783111338668-003Search in Google Scholar

Assendelft, Brenda, Gijsbert Rutten & Marijke van der Wal. 2023b. Franse woorden in het Nederlands. Een corpusanalyse van de lexicale invloed van het Frans op het Nederlands, 1500–1900. Tijdschrift voor Nederlandse Taal- en Letterkunde 139. 160–179.10.5117/TNTL2023.2/3.003.ASSESearch in Google Scholar

Ayres-Bennett, Wendy. 2021. Modelling language standardization. In Wendy Ayres-Bennett & John Bellamy (eds.), The Cambridge handbook of language standardization, 27–64. Cambridge: Cambridge University Press.10.1017/9781108559249.002Search in Google Scholar

Dodde, Nan. 2020. Franse scholen in Nederland. Ontstaan en ontwikkeling vanaf de vijftiende eeuw tot het midden van de negentiende eeuw. Oud-Turnhout & ’s-Hertogenbosch: Gompel & Svacina.Search in Google Scholar

Frijhoff, Willem. 1989. Verfransing? Franse taal en Nederlandse cultuur tot in de revolutietijd. Bijdragen en Mededelingen betreffende de Geschiedenis der Nederlanden 104. 592−609.10.18352/bmgn-lchr.3170Search in Google Scholar

Frijhoff, Willem. 2003. Uncertain brotherhood. The Huguenots in the Dutch Republic. In Bertrand Van Ruymbeke & Randy Sparks (eds.), Memory and identity. The Huguenots in France and the Atlantic diaspora, 128−171. Columbia: University of South Carolina Press.Search in Google Scholar

Frijhoff, Willem. 2015. Multilingualism and the challenge of frenchification in the early modern Dutch Republic. In Catharina Peersman, Gijsbert Rutten & Rik Vosters (eds.), Past, present and future of a language border: Germanic-Romance encounters in the Low Countries, 115−140. Berlin: De Gruyter.10.1515/9781614514152-008Search in Google Scholar

Ingham, Richard. 2018. Medieval bilingualism in England: On the rarity of vernacular code-switching. In Päivi Pahta, Janne Skaffari & Laura Wright (eds.), Multilingual practices in language history, 319–337. Berlin: De Gruyter.10.1515/9781501504945-015Search in Google Scholar

Ingham, Richard, Louise Sylvester & Imogen Marcus. 2021. Lone other-language items in later medieval texts. Journal of Historical Sociolinguistics 7. 179–2015.10.1515/jhsl-2019-0030Search in Google Scholar

Kok Escalle, Marie-Christine. 2017. Les Souvenirs du jeune Harco Hora Siccama (1854–1858). De l’exercice de style à l’expression de soi. In Madeleine van Strien-Chardonneau & Marie-Christine Kok Escalle (eds.), French as language of intimacy in the Modern Age, 143–164. Amsterdam: Amsterdam University Press.10.2307/j.ctt1s476vm.10Search in Google Scholar

Krogull, Andreas, Jill Puttaert & Gijsbert Rutten. 2023. Verfransing in de achttiende-eeuwse Noordelijke Nederlanden? Empirisch onderzoek naar taalkeuze in egodocumenten en pamfletten Tijdschrift voor Nederlandse Taal- en Letterkunde 139. 238–258.10.5117/TNTL2023.2/3.007.KROGSearch in Google Scholar

Kuijpers, Erika. 2005. Migrantenstad: immigratie en sociale verhoudingen in 17e-eeuws Amsterdam. Hilversum: Verloren.Search in Google Scholar

Langer, Nils & Agnete Nesse. 2012. Linguistic purism. In Juan Manuel Hernández-Campoy & Juan Camilo Conde Silvestre (eds.), The handbook of historical sociolinguistics, 607–625. Chichester: Wiley-Blackwell.10.1002/9781118257227.ch33Search in Google Scholar

Larrivée, Pierre & Mathieu Goux. 2024. The evolution of bare nouns in the history of French. The view from calibrated corpora. Journal of French Language Studies. 1–28.10.1017/S0959269524000061Search in Google Scholar

Marchello-Nizia, Christiane, Bernard Combettes, Sophie Prévost & Tobias Scheer (eds.). 2020. Grande grammaire historique du français (GGHF). Berlin: De Gruyter.10.1515/9783110348194Search in Google Scholar

Meyer, Lodewijk. 1669. Woordenschat. In drie deelen ghescheiden, van welke het I. Bastaardtwoorden, II. Konstwoorden, III. Verouderde woorden beghrijpt. Amsterdam: De Weduwe van Jan Hendricksz. Boom.Search in Google Scholar

Muysken, Pieter. 2013. Language contact outcomes as the result of bilingual optimization strategies. Bilingualism: Language and Cognition 16. 709–730.10.1017/S1366728912000727Search in Google Scholar

Peersman, Catharina, Gijsbert Rutten & Rik Vosters (eds.). 2015. Past, present and future of a language border: Germanic-Romance encounters in the Low Countries. Berlin: De Gruyter.10.1515/9781614514152Search in Google Scholar

Poplack, Shana & Nathalie Dion. 2012. Myths and facts about loanword development. Language Variation and Change 24. 279–315.10.1017/S095439451200018XSearch in Google Scholar

Poplack, Shana. 2017. Borrowing. Loanwords in the community and in the grammar. Oxford: Oxford University Press.10.1093/oso/9780190256388.001.0001Search in Google Scholar

Puttaert, Jill, Andreas Krogull & Gijsbert Rutten. 2022. Towards a methodological framework for historical language choice: the case of Dutch and French in the Netherlands (1800–1899). Slovo a Slovesnost 83. 251–278.Search in Google Scholar

Ruberg, Willemijn. 2011. Conventional correspondence. Epistolary culture of the Dutch elite, 1770–1850. Leiden: Brill.10.1163/ej.9789004209732.i-281Search in Google Scholar

Rutten, Gijsbert. 2019. Language planning as nation building. Ideology, policy and implementation in the Netherlands, 1750–1850. Amsterdam: Benjamins.10.1075/ahs.9Search in Google Scholar

Rutten, Gijsbert, Rik Vosters & Marijke van der Wal. 2015. Frenchification in discourse and practice. Loan morphology in Dutch private letters of the eighteenth and nineteenth centuries. In Catharina Peersman, Gijsbert Rutten & Rik Vosters (eds.), Past, present and future of a language border: Germanic-Romance encounters in the Low Countries, 143−170. Berlin: De Gruyter.10.1515/9781614514152-009Search in Google Scholar

Rutten, Gijsbert & Brenda Assendelft. 2023. The rise and fall of French borrowings in postmedieval Dutch. Linguistica LXIII. 337–352.10.4312/linguistica.63.1-2.337-352Search in Google Scholar

van der Sijs, Nicoline van der. 2002. Chronologisch woordenboek. De ouderdom en herkomst van onze woorden en betekenissen. Amsterdam & Antwerpen: Veen.Search in Google Scholar

Sleiderink, Remco. 2010. From francophile to francophobe: The changing attitude of medieval Dutch authors towards French literature. In Christopher Kleinhenz & Keith Busby (eds.), Medieval multilingualism: The Francophone world and its neighbours, 127−143. Turnhout: Brepols.10.1484/M.TCNE-EB.3.4606Search in Google Scholar

Stevens, Emmy. 2019. ‘Adieu, Vaarwel, groet mijne zo Waarde Ouders’: Het gebruik van Franse leensuffixen en leenwoorden in een negentiende-eeuwse familiecorrespondentie. Taal en Tongval 71. 137−156.10.5117/TET2019.2.STEVSearch in Google Scholar

van Strien-Chardonneau, Madeleine. 2014. The use of French among the Dutch elites in eighteenth-century Holland. In Gesine Argent, Vladislav Rjéoutski & Derek Offord (eds.), European francophonie. The social, political and cultural history of an international prestige language, 145–173. Oxford: Lang.Search in Google Scholar

van Strien-Chardonneau, Madeleine. 2017. Correspondance de Gijsbert Karel van Hogendorp (1762–1834) avec sa mère Carolina van Haren (1741–1812). Une formation morale, intellectuelle et sociale par le biais du français langue seconde. In Madeleine van Strien-Chardonneau & Marie-Christine Kok Escalle (eds.). 2017. French as language of intimacy in the Modern Age, 67–86. Amsterdam: Amsterdam University Press.10.2307/j.ctt1s476vm.6Search in Google Scholar

van Strien-Chardonneau, Madeleine. 2018. Practice and functions of French as a second language in a Dutch patrician family. The van Hogendorp family (eighteenth–early nineteenth centuries). In Vladislav Rjéoutski & Willem Frijhoff (eds.), Language choice in Enlightenment Europe. Education, sociability, and governance, 65–85. Amsterdam: Amsterdam University Press.10.1515/9789048535507-004Search in Google Scholar

Treffers-Daller, Jeanine. 2023. The simple view of borrowing and code-switching. International Journal of Bilingualism. Online first. https://doi.org/10.1177/1367006923116853510.1177/13670069231168535Search in Google Scholar

Vanhecke, Eline & Jetje De Groof. 2007. New data on language policy and language choice in 19th-century Flemish city administrations. In Stephan Elspaß, Nils Langer, Joachim Scharloth & Wim Vandenbussche (eds.), Germanic language histories ‘from below’ (1700–2000), 449–470. Berlin: De Gruyter.10.1515/9783110925463.449Search in Google Scholar

Verheyden, Charlotte. 2023. ‘Ge moet zien dat gij uw vlaamsche taal niet vergeet’. Een historisch-sociolinguïstisch onderzoek naar de invloed van het Frans op het Zuidelijke Nederlands. Vrije Universiteit Brussel, PhD Dissertation.Search in Google Scholar

van der Wal, Marijke & Cor van Bree. 2014. Geschiedenis van het Nederlands. Houten: Spectrum.Search in Google Scholar

Willemyns, Roland. 1994. Taalpolitiek in de Bourgondische tijd. Verslagen en Mededelingen van de Koninklijke Academie voor Nederlandse Taal – en Letterkunde. 162−177.Search in Google Scholar

Willemyns, Roland. 2015. Trilingual tug-o’-war: Language border fluctuations in the Low Countries. In Catharina Peersman, Gijsbert Rutten & Rik Vosters (eds.), Past, present and future of a language border: Germanic-Romance encounters in the Low Countries, 39−60. Berlin: De Gruyter.10.1515/9781614514152-005Search in Google Scholar

Published Online: 2025-01-29

Published in Print: 2025-05-30

Dieses Werk ist lizensiert unter einer Creative Commons Namensnennung 4.0 International Lizenz.

Articles in the same Issue

https://doi.org/10.1515/soci-2024-0017

Keywords for this article

borrowing; loanwords; Dutch; French; historical sociolinguistics; language contact

Creative Commons

BY 4.0