Rethinking Historical Multilingualism and Language Contact ‘from Below’. Evidence from the Dutch-German Borderlands in the Long Nineteenth Century

ABSTRACT European language histories, including the history of Dutch, have often been portrayed as broadly linear developments towards one uniform standard language. In this biased account, rooted in the nation-building era around 1800, language diversity and multilingualism were largely rendered invisible. Against the background of clearly segregated spaces, politically and linguistically, border settings have particularly challenged the monolingual ideology of ‘one nation–one language’. Taking a historical-sociolinguistic perspective, this article focuses on the Dutch-German borderlands in the long nineteenth century as an intriguing case to investigate historical multilingualism and language contact ‘from below’. Despite the growing importance of nation-states and their standard languages, it is shown that multilingual practices and contact phenomena can still be traced in handwritten archival documents from the private sphere. Illustrative examples come from various family archives in the border area as well as from a unique collection of letters written by (Low) German labour migrants to their Dutch employer. These sources give evidence of the Dutch-German borderlands as a multi-faceted sociolinguistic space well into the nineteenth century. Moreover, they suggest that established theories of multilingualism and language contact may require rethinking in order to account for less clear-cut and more fluid practices in the past.


Introduction
The phenomena of historical multilingualism and language contact are increasingly acknowledged as essential characteristics of Europe's linguistic past. 1 This historicalsociolinguistic call for a multilingual approach contrasts with, and therefore challenges, traditional language historiography, which has been deeply influenced by monolingual and standard language ideologies since the late eighteenth and nineteenth centuries. Not only are accounts of European language history neatly separated from each other 'as a collection of histories of monolingualism' (the history of Dutch, of German, etc.), they are also 'generally written from the perspective of the modern standard languages and their respective roles as national languages'. 2 Consequently, most language histories are portrayed as broadly linear developments towards one uniform standard language.
This biased account of European language history is rooted in the nation-building era in the long nineteenth century. 3 The emerging ideology of 'one nation-one language' established monolingualism (in the national standard language) as the default in Europe, often leading to an understanding of multilingualism as 'multiple monolingualisms in distinct languages, mastered separately in standard form and kept pure of outside influence'. 4 From the French Revolution (1789-1799) onwards, governments across Europe started to believe that their nation 'needed one single language and set out to homogenize disparate populations into a cohesive linguistic whole'. 5 Diversity in language use, including internal and external multilingualism, 6 was thus largely rendered invisible in the national era. Both geographically and (socio)linguistically, borderlands have constituted 'invisibilised' areas 7 ever since, as they were often characterized by multilingualism and language contact, and thus contested the strongly monolingual ideologies of nation-states.
An intriguing case to investigate historical multilingualism and language contact in the broader context of standard languages, norms and variation 8 is the situation in the Dutch-German borderlands in the long nineteenth century. Taking a multilingual approach to language history, this research seeks to 'look beyond the main language of a text and consider what a holistic overview of all the languages in it reveals, about the "grammar" of non-monolingual writing on the one hand, and individual identity or social practice on the other'. 9 In line with the historical-sociolinguistic tradition 'from below', 10 this article utilizes handwritten documents from the private sphere in order to shed new light on the borderlands. What types of typically underrepresented practices and phenomena can still be traced in archival data, and how do these relate to established theories of (historical) multilingualism and language contact?
The challenges of borderland settings are discussed in the "Borderlands in the Nationstate Era" section. "The Dutch-German Borderlands" section outlines the specific (socio) linguistic space of the Dutch-German borderlands. Data and methodology are briefly introduced in the "Data and Methodology" section. Focusing on archival evidence, the "Examples and Analysis" section contains the analyses of various phenomena and features, as well as theoretical considerations to rethink categories. Some preliminary conclusions are presented in the "Conclusion" section.

Borderlands in the Nation-state Era
In European language historiography, borderlands have widely been neglected, especially since the nation-state era and its 'one nation-one language' ideology. Gaining momentum in the late eighteenth and nineteenth centuries, it regarded multilingualism and minorities as 'deviant and undesirable because they disturb the (presupposed) uniformity and correspondence between a nation and its language'. 11 Not allowing for transition zones or blurred borders, languages were divided structurally through a 'border-making design in which each language is separated and segregated into its own discrete space and time and is not allowed to mix with each other'. 12 While the standard language of a given nation-state was elevated to the one and only legitimate language, 13 contact with other languages or varieties was seen as a threat to the homogeneous national standard.
Therefore, any language contact should be avoided 'for fear that the legitimate language should be contaminated and weakened by influence from those other languages'. 14 Against the ideological background of clearly segregated spaces, both politically and linguistically, the very nature of border settings increases the unwanted 'risk' of languages, varieties and their users coming into contact and mixing with each other. These ideas and beliefs emerged with the introduction of political borders around 1800: The political border as it has existed in Europe since the late 18 th and early 19 th centuries is in many respects a fundamentally new phenomenon. Before this time, the norm in Europe, as in other parts of the world, appears to have been the presence [. . .] of frontiers that were highly permeable both culturally and linguistically. 15 The notion of a 'language border' came into play to strengthen ideological nation-state discourse, as it 'usually evokes -and perhaps even promotes -a belief in ethnic differences, a reliance on monolingualism and the construction of "otherness"'. 16 From the nineteenth-century viewpoint of monolingualism as the norm within national boundaries, borderland populations and their linguistic practices represented 'an alien factor with respect to the nation they were part of and the official language in this nation-state' and, as minorities, were 'often considered potentially illoyal'. 17 With regard to their geographical proximity to 'other' nation-states with 'other' national languages as well as the chance of cross-border exchange, culturally and linguistically, borderlands were thus challenging in many ways. It might be argued that the potential mixing of languages and their speakers, which is 'anathema for standardising ideology', 18 is particularly high in the case of neighbouring and genetically related languages and varieties used at so-called 'contrast-poor' language borders, 19 of which the Dutch-German borderlands are a good example. This also makes it quite distinct from Germanic-Romance contact in the Low Countries. 20

The Dutch-German Borderlands
Serving as a case to examine multilingualism and language contact, this article focuses on the Dutch-German borderlands in the long nineteenth century. Resulting from the redrawing of the European territorial map after the Congress of Vienna (1814-1815), the language border that separates Dutch from German more or less runs parallel to the political border between the Netherlands and Germany. It stretches from the Dollart Bay in the North Sea to the Dutch-Belgian-German tri-border region in the south (see Figure 1).
Considering the different linguistic constellations along and near the border, Kremer further distinguishes a northern and a southern part, separated by the river Rhine. 21 It is the former, from Emden to Emmerich, that forms the focus of this study. This northern border region, separating the north-eastern Dutch provinces of Groningen, Drenthe, Overijssel and (partly) Gelderland from the German kingdoms of Hanover and Prussia, constitutes a dual language border. 22 On the one hand, it separates Dutch from High German, both of which were codified as standard varieties during the nineteenth century. On the other hand, the Low Saxon and Low German dialects, as the native vernacular of the border population, count as equally important 'players' in this area.
Historically and sociolinguistically, the northern part of the borderlands was by no means homogeneous. Despite the political border, the situation continued to be complex well into the nineteenth century. In north-western Germany, a variety of Dutch sometimes referred to as Grenzniederländisch 23 or border Netherlandic, 24 maintained some functions in certain territories close to the Dutch border. In East Frisia, 25 especially the Calvinist southwest, as well as the county of Bentheim, 26 Dutch was used as a Kultursprache in church, school and administration until the second half of the nineteenth century. Importantly, the primary spoken language in these border areas, across nearly all domains and social classes, was not a spoken variety of Dutch, but the local (Low German) dialect. Compared to the complexity of Dutch as a minority language in Germany, the situation on the Dutch side was relatively coherent. At least from a discursive perspective, the population of the Dutch nation-state was supposed to be monolingual in the newly devised standard from the early 1800s onwards. 27 What makes the sociolinguistic space of the Dutch-German borderlands even more intricate is the fact that the political and (standard) language border was drawn across the continental West Germanic dialect continuum. Although the north-eastern (Low Saxon) Dutch and western Low German dialects historically blended into one another, they have increasingly diverged under the influence of the national state border and their respective standard languages. 28 Given the existence of a political border, the dialects on the German side will thus be referred to as Low German (Niederdeutsch), the neighbouring Dutch dialects as Low Saxon (Nedersaksisch). Since the break in the dialect continuum has been a gradual process, the 'borders' between Low Saxon and Low German were anything but clear-cut in the long nineteenth century, though.
As mentioned, the turn from the eighteenth to the nineteenth century typically marks a change in discussions about Europe's (political) borders and their effects on sociolinguistic spaces. Until c. 1800, the Dutch-German border was porous and indistinct, stimulating intensive cultural and linguistic contacts. In the seventeenth and eighteenth centuries, i.e. prior to the border-making era of European nation-states, contacts between the north-eastern Netherlands and north-western Germany were close and formed 'a symbiotic relationship' 29 -economically, politically, religiously and linguistically. On either side of the border, the area was characterized by high mobility. 30 From 1700 onwards, large numbers of migrants from the German border region, mainly East Frisia, Oldenburg Land and Bentheim, moved to neighbouring regions in the Netherlands. For instance, groups of Protestant ministers from northwest Germany were attracted by the university of Groningen and then settled in the north-eastern Dutch provinces. Moreover, the large-scale phenomenon of seasonal labour migration saw tens of thousands of haymakers and grass-mowers from northwest Germany crossing the border to the Netherlands every year. Generally, intermarriages were not uncommon at all in the borderlands. These contacts testify to an intimate relationship between the north-eastern Netherlands and north-western Germany over centuries.
Relatively minor linguistic and cultural differences facilitated cross-border contact. Due to the dialect continuum, still intact around 1800, the new political border did not constitute a linguistic barrier for cross-border mobility and communication: at least in their old forms, the cross-border dialects [. . .] could be used as varieties for crossborder communication. This function of a 'lingua franca' [. . .] was rather central when the state border did not yet play a big role in everyday life, e.g. because regional working flow was still more important than the more centralized bureaucracy put forward by the modern nation states. 31 While the growing importance of nation-states and their monolingual ideologies can hardly be neglected as interfering factors, the contact situation in the borderlands is unlikely to have changed overnight, and although gradually decreasing, continued well into the nineteenth century (and beyond). This also suggests a continuation of crossborder language practices after 1800. Therefore, the question arises whether and how multilingualism and language contact were still reflected in everyday writing in the long nineteenth century, and to what extent we have access to these practices today. Archival evidence is presented in the "Examples and Analysis" section.

Data and Methodology
This article takes a historical-sociolinguistic perspective 'from below' in utilizing archival sources, and ego-documents or first-person writings in particular, to reconstruct actual language practices in past societies. 32 Previous studies on the Dutch-German border, especially its post-1800 history, mainly focused on the evolution of the respective standard languages on each side of the border, and their influence on the dialect continuum. 33 This study shifts the focus to handwritten texts by potentially multilingual individuals in the area and period under investigation. By examining authentic sources relatively close to the everyday lives of 'common' people, it seeks to give important evidence of practices beyond the alleged norm of monolingualism and standard languages.
For this purpose, data was collected in various Dutch archives in the northern and north-eastern provinces, i.e. Groningen, Drenthe, Overijssel and Friesland. 34 The aim was to unearth handwritten documents from individuals and families from various social ranks, primarily letters and diaries as particularly suitable ego-documents, but also cookbooks and housekeeping books as handwritten evidence from the private domain in the wider sense. The fragmentary nature of these sources, scattered over numerous archives, does not allow the compilation of a corpus at this point. Instead, a looser collection of archival sources is examined, cohering in terms of time (long nineteenth century), space (Dutch-German borderlands) and material (handwritten archival sources). Whenever possible and/or necessary, genealogical information about the writers is provided, in order to assess their language use against their socio-geographical background. Methodologically, this data-driven article analyses archival documents primarily on a qualitative level by close reading of these sources and identifying potentially multilingual individuals and their handwritten practices. 35

Examples and Analysis
Based on archival evidence of multilingual practices and language contact, this section contains the analyses of various phenomena and features: lexical transference, morphosyntactic transference, code-switching, and practices beyond (traditional) categories. It should be emphasized that these categories are not straightforward or unproblematic, which manifests itself in theoretical discussions about varying models and terminology. 36

Lexical Transference
One of the linguistic outcomes typically found in a language contact situation involving multilingual repertoires is the transference 37 of lexemes, i.e. of words in form and content. Alternatively, the insertion of content words 'into a surrounding passage in the other language' can be referred to as insertional code-switching. 38 Lexical transfers or insertions may be integrated to different degrees.
In private letters from the Dutch-German borderlands, the use of High German content words in otherwise Dutch writing appears to be a recurring phenomenon during the long nineteenth century. In what follows, this type of contact-induced lexical transference is illustrated by evidence from different letter writers. Examples (1)-(4) derive from the archives of the Winsinghhof te Roden in Drenthe (Drents Archief), which comprise letters by relatives across the border in East Frisia. One of them is Wilhelmus Lefferdus Wübbena (1808-1894), who was born in the East Frisian village of Midlum, but studied theology at the university of Groningen in the 1820s, before returning to East Frisia, where he became a minister in Veenhusen. In a letter written in 1856, Wübbena describes his journey across the Dutch-German border. Addressed to relatives in Drenthe, it is perhaps unsurprising that the letter is written in Dutch, largely in line with contemporary norms. Given that Dutch was an official school language in East Frisia until the mid-nineteenth century, 39 and the fact that he studied at a Dutch university, Wübbena was probably familiar with the written standard. Strikingly, though, in (1)  The highlighted words derive from the semantic field of 'railway', most notably Baanhof 40 (High German Bahnhof), two attestations of Zug, and Haltestelle. Adopting a more Dutch spelling, yet capitalized, and the neuter article in line with Dutch plan, we find Vaarplan (HG Fahrplan), as well as vaarbillet (HG Fahrbillet). Interestingly, Wübbena's use of vaar-/varen (compare HG fahren) in Vaarplan, vaarbillet, but also in the past tense forms voeren and afvoer is probably influenced by the homophonous High German words Fahrplan, Fahrbillet, (ab)fahren. Semantically, it appears to be extended to 'travel by train', as opposed to the more restricted meaning of Dutch varen as 'travel on water'. Since these words had equivalents in contemporary Standard Dutch, the High German transfers in Wübbena's letter do not fill lexical gaps. It is also unlikely that they can be explained by the writer's '(momentary) incompetence in the established languageof-interaction', 41 although Dutch rail transport (and its terminology) was only introduced after he had returned to East Frisia in the 1830s. Wübbena seems to emphasize that this part of his journey took place on the German side, not only implying his command in two (standard) languages, but also awareness of his mobility across the border. 42 Another letter from the Winsinghhof archives was written by Wübbena's sister-in-law Aaltje Anthonetta Pannenborg, née Mecima (1807-1885). Born in the Drenthe village of Anloo, she married East Frisian minister Eerke Pannenborg and settled in Esklum. In this letter from 1880, A.A. Pannenborg keeps her relatives across the border informed about family life in Germany: (2) Albertus die Lehrer is -en in Aurich 6 Jaar geweest is, is tegenwoordig in Göttingen aan het Gÿmnasium als Lehrer. [. . .] de Gÿmnasien worden hier nu ook veel bezogt.
Like the use of railway-related lexemes from High German in (1), the two occurrences of Lehrer in (2) seem to highlight that her son is indeed a teacher at a school in Germany.
Orthographically, Gÿmnasium and Gÿmnasien (compare Dutch gymnasia) are interesting forms as they combine the capitalized German spelling with a double-dotted <ÿ> (not <y>) typically found in Dutch handwriting. 43 Similar evidence from the same family archives is found in a letter from 1873, written by W.L. Wübbena's daughter and A.A. Pannenborg's niece Heidina Wübbena (*1844). In (3), she uses Bankdirektor (not bankdirecteur), but, unlike her aunt in (2), Dutch onderwijzer instead of Lehrer:  (6) is in line with contemporary norms and clearly distinct from Dutch gezelschap, occurring twice in the letter. In (7), it is even more remarkable that Groszmŭtter contains a so-called u-Bogen (ŭ), typical of German Kurrentschrift. 47 This signals the writer's knowledge of spelling and script conventions in High German, while separating the two linguistic systems.
To sum up, this section has illustrated the recurring phenomenon of lexical transference in private letter writing, more specifically High German content words in standardoriented Dutch texts. All writers have in common that they come from relatively well-todo families (ministers, textile manufacturers). Language contact appears to affect the two standard varieties only, which, although integrated to different degrees, are in most cases distinguishable as two linguistic entities. Some hybrid-like forms do occur, as does transference on the lexicosemantic level, suggesting that Dutch-High German contact did not only result in the intentional use of High German lexemes, but also in more subtle transfers.

Morphosyntactic Transference
Multilingual practices and language contact in the Dutch-German borderlands are by no means restricted to the lexical and lexicosemantic levels. Analysing evidence of morphosyntactic transference in private letter writing, two recurring transfer phenomena are discussed: the pronominal system and past participle forms. Unlike the Dutch-High German standard constellation illustrated above, this section focuses on contact between Dutch on the one hand, and Low German and/or Low Saxon on the other.

Pronominal System
The first morphosyntactic phenomenon concerns the pronominal system, more specifically the second person singular forms of personal pronouns and possessive determiners. The paradigms in (Standard) Dutch, High German as well as in Low German and Low Saxon dialects contain different forms, as outlined in Table 1. It should be noted that more pronominal forms of address occur in Dutch letter writing, including ul 'your love' and ue 'your honour'. 48 There is also more variation across the Low Saxon area than presented in this simplified overview. 49 Kept in the Winsighhof archives (Drents Archief), these texts are written by members of the (extended) Foget family from the East Frisian town of Jemgum to relatives across the border in Drenthe. One interesting ego-document with regard to pronominal forms is a letter dated 1807 by Margaretha Helmers van Baden (1742-1817). Born in the border village of Bellingwolde in Groningen, she married Hinderk Harms Foget, a baker from Jemgum, in 1762 and lived in East Frisia until her death. In letters to her children, we find morphosyntactic phenomena that reflect the language contact situation typical of the northern Dutch-German borderlands, adding Low German and/or Low Saxon to the standard dichotomy of Dutch and High German. While these letters are generally intended to be written in Dutch, Helmers van Baden's use of personal and possessive pronouns of the second person singular displays remarkable variation, with several forms deviating from nineteenth-century Dutch norms. Examples (8)- (11) illustrate her use of pronominal forms, which derive from Low German and/or Low Saxon: (8) dat du en dyn man niet te Hujs waaren (9) ik kan mij an dy niet zeggen Hoe ik gestelt ben (10) broers schryven en dyn schryven waaren my nuttig (11) ik ben dyn lieve moeder The writer's cross-border mobility across her lifespan (i.e. growing up on the Dutch side, but spending most of her adult life on the German side, where Low German was the common spoken language, while Dutch was also used) makes it difficult to draw a line between Low German and Low Saxon. However, it is debatable whether a distinction is relevant at all, as the dialect continuum was still very much intact in the early 1800s, before political borders and standard languages started to exert influence on these border dialects. Strikingly, though, these forms are not in line with the pronominal system of the main language in this letter (Dutch). We find two attestations of nominative du, three oblique dy and six possessive dyn. Alongside these Low German/Saxon forms, her writing also contains standard Dutch forms, illustrated by (12)-(14): two nominative gij, one oblique u, as well as five possessive uwe and non-standard u (for uw): (12) ik denk als gy deese krijgt zult gij die ontvangen Hebben (13) ons welzyn zy u met deesen bekend (14) uwe brieven van oom en van dij myn lieve kint It is evident that the linguistic repertoires of this writer comprise various pronominal systems, and they appear to be used alongside each other, rather than separated.
In the Foget family archives, further attestations of non-standard pronouns are found for the second person singular. In a letter signed by Hind k H. Foget, possibly Helmers van Baden's son Hinderk (1798-1825), written in the East Frisian town of Leer in 1823 to his cousin in Drenthe, plenty of Low German/Saxon forms occur in the nominative (2x doe), oblique (at least eight unambiguous tokens of die) and possessive (4x dien), as shown in (15)- (18): (15) Het is mijn hartelijken wenschen verlangen, dat doe die mogst daartoe schicken kunnen (16) daar aan dunkt my, kanst doe niet twijfelen (17) ik verlange om wat van die te horen (18)

ik moet, en zal gerust dien besluiten afwagten
Compared to the forms du, dy, dyn in his mother's writing, the spelling variants doe, die, dien in H.H. Foget's letters are more Dutch-oriented and 'phonetic' (from a Dutch perspective). Moreover, he uses these forms invariably, with the exception of one token of Dutch u in the postscript sentence.

Past Participle
The second phenomenon concerns transference on the level of inflectional morphology, namely the past participle. Table 2 presents a simplified overview of variants occurring in (Standard) Dutch and High German as well as Low German and Low Saxon. While past participles in both Dutch and High German take the prefix ge-, as in geschreven or geschrieben, forms without ge-are found in most Low German dialects. 50 Dialects in the Low Saxon area of the Netherlands display even more variation. Unlike Standard Dutch, almost none of the north-eastern dialects has ge-in the past participle, but e-or a øvariant without ge-or e-. 51 The latter is the only occurring variant in Groningen and (eastern) parts of Drenthe.
As in the case of pronominal forms ( (8)- (14)), we find non-standard variants of past participles in Helmers van Baden's letters. Recall that these were broadly written in Dutch, but not consistently in accordance with contemporary norms. Examples (19)-(22) illustrate her use of ø-forms, which omit Standard Dutch ge-:
More attestations of this transference type (examples (23)- (25)) are found in two letters by another family member, Gepke Foget from Oldersum, who also maintained close contacts with relatives across the border in Drenthe:
Summarizing, the archival sources utilized in this section display a tangible distance from nineteenth-century Dutch norms, showcasing morphosyntactic features that reflect transference from Low German and/or Low Saxon, i.e. resources other than the two standard languages. Unlike the intentional use of High German lexemes in Dutch letters, the writers' variable use of linguistic repertoires here indicates that the resources in contact were not always neatly separated from each other in actual writing, as becomes evident from variation in the pronominal system and past participles. It does suggest, though, that multilingualism and language contact were inherent parts of everyday practices in the borderlands well into the nineteenth century.

Code-switching
Turning to yet another facet of multilingual practices in the Dutch-German borderlands, this section focuses on code-switching, which involves 'the use of more than one language in a specific communicative event', 52 excluding however switches of single words (see "Lexical Transference" above). Although code-switching was initially studied as a typical feature of speech, research on historical code-switching has offered plenty of examples in written records of language use. 53 Exploring code-switching in the borderlands, the Heerspink family archives (Drents Archief) prove to be a promising source. This family consisted of ministers working in various municipalities in Drenthe, while the female line descended from East Frisia. One particularly intriguing document is a (private) cookbook 54 by a female offspring, titled Recepte von Dien Heerspink (cover) or Recepten van D.H. Heerspink (stamped inside). It is attributed to Diddina Haukea Hitjer (1821-1889), the wife of minister Johannes Bernardus Folkerus Heerspink, and dated around 1860. On closer inspection, though, it seems more likely that this book was written by her granddaughter Diddina Henriëtte Heerspink . Born in Oudeschans near the German border, D.H. Heerspink was the eldest daughter of Herman Hitjer Heerspink, minister in Eelde, and remained unmarried. Irrespective of the writer's exact attribution, this cookbook provides evidence of a multilingual individual from the borderlands, and of Dutch-High German codeswitching in particular.
The first 27 pages contain handwritten recipes of puddings and pastries in High German, throughout which we find a few (subtle) transfers from Dutch as in Mandelen (< SD amandelen, but HG Mandeln, the latter of which also occurs) or various attestations of Schaale with typically Dutch <aa> alongside Schale. The hybrid-like spelling of susjes (SD soesjes) contains the High German representation of /u/ as <u> rather than its Dutch equivalent <oe>, while keeping Dutch <sj> for /ʃ/, rather than High German <sch>. Further transfers include a self-correction of the Dutch prefix af-to High German ab-in ab-gebacken, and the recurring construction soviel . . . als (e.g. soviel Zucker als es fassen kann), which appears to be a High German calque of Dutch zoveel . . . als. Although limited in number, these transfers indicate that the main writer, 55 probably D.H. Heerspink, was indeed multilingual, since elements of Dutch shine through her use of written High German. No transference from Low German/ Saxon could be identified, therefore involving only standardized languages. What is more, we find few traces of two competing writing systems. While the High German recipes are largely written in Latin script, 56 there are some inconsistencies. For instance, thut (p. 1) is partly written in German Kurrentschrift, whereas most other forms of the verb thun are not.
The most striking feature of this document, however, is found on page 27, where we encounter a case of (intersentential) code-switching from High German to Dutch, occurring in the recipe of Bremer Pudding, as shown in (26)

Op een lood gaan vier of vyf bladen gelatine. Een half fleschje water weegt vier honderd gram.
After the High German-to-Dutch entry, all (culinary) recipes are written in Dutch. 57 Interestingly, the 'inter-recipe' code-switch is preceded by a mention of place and date on page 26, stating that the recipe was written down in the German city of Bremen in 1890. This is important evidence, from which we can deduce that, against the archives' attribution, Diddina Hitjer, who died in 1889, cannot be the writer of this book. On page 28, we find a second indication of place and date, Eelde 9 Augustus '90, written in the same hand as Bremen Mei 1890 before. This indicates the writer's cross-border mobility, as she must have spent some time in (northwest) Germany, although the exact circumstances, including her language learning experience and the possible trigger for her code-switch, remain obscure.
Apart from the cookbook, the Heerspink archives provide more insights into the language practices of the extended family. Although the preserved documents are fairly heterogeneous, they showcase further evidence of Dutch-German code-switching in handwriting. In a (primarily) High German letter card from Bremen, dated 1893, written by a certain J. Timmerman to his cousin, the closing formula and postscript in (27) displays (intrasentential) code-switching to Dutch and (intersententially) back to High German (Wer lacht da?, i.e. 'Who is laughing?'), possibly referring to his effort to write in Dutch: (27)  These code-switching practices, both in D.H. Heerspink's cookbook and the more informal letter card, suggest that High German and Dutch played a role in private writing of border families, and therefore are important counter-evidence to the persistent idea of 'either/or' as the linguistic norm in nineteenth-century nation-states.

Beyond Categories
Up to here, the phenomena discussed were predominantly cases of various languages and varieties alongside each other yet identifiable as distinct entities (though less clear-cut in the case of morphosyntactic transference from Low German/Saxon). However, the final and most intricate example of cross-border language contact is one that not only challenges the monolingual bias in history writing, but also traditionally established theoretical categories in multilingualism and language contact research.
As briefly mentioned earlier, the phenomenon of seasonal labour migration (Hollandgängerei) saw large numbers of German labourers, often from the Low Germanspeaking parts in the Northwest, crossing the border to the Netherlands well into the nineteenth century. Though limited to harvest seasons, more regular working relationships were not uncommon and many hannekemaaiers (e.g. haymakers, grass-mowers) returned to the same employers every year. While the social history of this large-scale migration has been thoroughly investigated, 58 linguistic evidence is extremely sparse. Luckily, a small collection of letters, written by German labour migrants to their Dutch employer, has survived 59 and is now kept in the Fries Scheepvaart Museum. These hannekemaaiersbrieven 60 were addressed to the well-to-do dairy farmer's family Bootsma in the Frisian village of Loënga near Sneek. Most letters (15 out of 17) were sent from Brual in the northern Emsland, which makes them a fairly coherent collection in their own right. It is these texts which are discussed, focusing on two comparatively prolific writers: haymakers Johann Evers (1842-1920; three letters from 1868-1871) and Johann Roskam (1855Roskam ( -1937six letters from 1879-1906. To begin with, examples (28)-(30) written by Evers, and (31) by Roskam, illustrate the typical negotiation practices between harvest seasons, giving a glimpse into their linguistic nature: (28) ŭnse afsprak is gewest dat wie ü schriwen zŭllen en dan ie ŭns ok wer terüg schriwen wŭllen (29) Jk kan det Jar nit bäi ü kommen mar ik hop op en ander Jar wer bai ü te wessen om te heŭen (30) ik heb 10 Jaren bei ü gewest in de heŭinge en nŭ schreift gei mi of ik det 11 Jahr ok wer bei ü komen sal in de heŭinge (31) Ick hebbe ü brief den 27 Maarz richtig ontfangen en darüt gelesen dat ick ü twe vlenke Mannen mitbrengen sel twe vlenke Mannen kann ick ü wel mitt bringen as naŭ bei ü mar gŭtt Gras groeit dat wie alle drie bei ü mar werk häben The first impression is that of a 'mixed' language 61 waiting to be unravelled in detail.
Given the limited scope of this article, however, this subsection focuses on the value of these remarkable sources for the study of historical multilingualism and language contact.
Unlike previous examples, which can be categorized as Dutch with transfers from other languages or varieties, the labourers' letters do not fit any linguistic category.
Previous attempts by social historians to label the 'language' as 'Low German', 62 i.e. the (spoken) language of the labourers, fail to do justice to its complexity. This neat label assumes that these letters are written representations of northern Emsland dialects. Historical-sociolinguistic research has disproved the assumption that writing 'from below' is simply a transliteration of dialects. 63 Also bearing in mind that the labourers addressed their employer, implying a slightly asymmetrical relationship, the intended written language is more likely to be a form of Dutch, not Low German: Die Briefschreiber haben sich [. . .] ganz offensichtlich bemüht, adressatenorientiert zu schreiben und eine Art Niederländisch oder Stadtfriesisch zu verwenden -Sprachen, mit denen sie allem Anschein nach in der täglich Kommunikation mit ihren Arbeitgebern konfrontiert waren und die ihnen von ihrem eigenen niederdeutschen Dialekt auch nicht allzu viel Verständnisprobleme bereiteten. 64 This orientation towards Dutch does not at all disregard other languages or varieties. The Brual writers had various linguistic resources at their disposal, which they utilized to communicate with the Bootsma family. Irrespective of their level of competence, the labourers' multilingual repertoires must have included their local Low German dialect, High German, and Dutch. 65 Importantly, though, none of these fully qualified for crossborder written communication in this specific setting. Low German was the labourers' native language and used for everyday spoken communication. Presumably, it was also used in face-to-face situations on Dutch farmlands, based on the principle of receptive multilingualism. 66 However, they were unfamiliar with Low German as a written language, which is why different linguistic resources were required for letter writing. High German, mainly written, was acquired through the education system, probably even in village schools in Low German-speaking area. 67 In fact, two similar letters in the Bootsma archives, both from Wechold near Hoya, give evidence of labour migrants' competence in (standard) High German. In speech, however, High German was most likely limited to communication with socially superior contacts. 68 Finally, Dutch, though never acquired formally, can be considered the intended, receiver-oriented language in these letters. In contrast to writers from East Frisia, where Grenzniederländisch was used well into the nineteenth century, the Brual labourers' access to Dutch was much more restricted. This is clearly reflected in their highly variable spelling, partly illustrated in (28)- (31), which deviate from nineteenth-century Dutch norms. For instance, long vowels in closed syllables, typically represented by double graphemes in Dutch (e.g. <aa>, <ee>, <oo>), are represented by single graphemes (<a> in mar, jar; <e> in wer, hel; <o> in ok, hop) or with German lengthening-h (Jahr, vehl). As regards Dutch diphthongs, we encounter much variation in the representation of West Germanic *ī (monophthong in Low German), both diphthongal <ai>, <äi>, <ei>, <ej>, <eij> and monophthongal <i>, <ie>. 69 Generally, the use of High German umlaut characters, absent from the Dutch alphabet, and the u-Bogen (ŭ), signal that they fall back on their High German resources, at least orthographically. They tend to adopt High German spelling conventions, phonetically 'mimicking' Dutch words that they probably heard on the farmlands, but never learned to write down. Striking examples are äut (for uit), mitschin (for misschien) and sniŭ (for sneeuw), but similarly illustrative forms can be found across all letters. 70 The challenges of separable languages and varieties become obvious in Niebaum's attempt to categorize individual (yet closely related) lexical items, 71 which is problematic when a substantial number potentially derives from more than one resource. The typological proximity in the Dutch-German borderlands even increases their inseparability. Methodologically and theoretically, it appears that the labourers' letters require rethinking. First and foremost, we need to address the general issue whether any linguistic categorization is appropriate or possible at all. Moreover, we need to discuss how these letters can best be described and theorized by considering both their linguistic characteristics and socio-historical contact setting. Traditional concepts prove challenging, as the established definitions do not capture the linguistic hybridity and fluidity of these writing practices, not least because they are grounded on the deeply monolingual ideal of languages as separable entities. 72 A more current concept in multilingualism studies worth considering is languaging, 73 which rejects the established assumption about languages as separable, coherent entities, 74 particularly in the ideological sense of national languages as the monolingual norm. According to Jørgensen, this term refers to practices, originally in speech, where 'language users employ whatever linguistic features are at their disposal with the intention of achieving their communicative aims'. 75 Sebba argues that '"languaging" in the broadest sense is possible in the written mode as well as the spoken', adding that 'languagers' [. . .] sometimes draw on languages of which we have only a limited knowledge. That this behaviour is available to humanity as a whole, rather than a subset known as 'bilinguals' who have sufficient knowledge of two languages, distinguishes it from codeswitching as it has been traditionally conceived. 76 The varying degrees of knowledge are reflected in the labourers' writing, particularly their limited competence in (written) Dutch, 77 yet they draw on these partially available linguistic resources to communicate across a (standard) language border. 78 While it is beyond the scope of this article to discuss the suitability and shortcomings of these concepts for historical-sociolinguistic cases in detail, they seem to offer an interesting alternative to account for less categorizable practices in the past.

Conclusion
This article sought to challenge the monolingual bias in language history by exploring the potential of archival (ego-)documents in the Dutch-German borderlands to gain insights into multilingual practices and language contact in the long nineteenth century. Traditionally, this period is associated with the emergence of nation-states and standard languages. By taking a historical-sociolinguistic perspective 'from below', the unearthed handwritten sources indicate that the Dutch-German borderlands were a multi-faceted sociolinguistic space, in which multilingualism and language contact occurred in many different forms and shades.
In terms of 'named' languages and varieties, written practices from individuals and families in the border area involved Dutch alongside High German (as illustrated by lexical transference and code-switching), but also Dutch alongside Low German and/or north-eastern Dutch (Low Saxon) dialects (as illustrated by morphosyntactic transference). The latter gives important evidence of transfers from linguistic resources other than the two national standard languages. This highlights the ongoing importance of spoken and non-standard resources well into the nineteenth century, which have commonly remained invisible. In Dutch private letters across the border, most lexical transfers from High German seemed intentional and were clearly differentiated (e.g. through spelling, capitalization), indicating the writers' awareness of distinct linguistic systems. This also holds true for code-switching in a private cookbook. Hybrid forms and lexicosemantic transfers were identified, too, as more subtle results from Dutch-High German contact. Morphosyntactic transference from Low German/Saxon paradigms (pronominal system, past participles), on the other hand, was probably unintentional, with non-standard forms co-occurring with variants typical of (Standard) Dutch. Overall, the investigated features and phenomena could be described by established concepts of multilingualism and language contact, based on the idea of languages as separable entities.
However, this research also presented evidence of cross-border writings that challenge traditional categories, both linguistically and theoretically. Letters by German labour migrants to their Dutch employer displayed a remarkable 'mixture' of linguistic resources for the functional purpose of communication. It is the fluidity of these practices, utilizing the writers' full multilingual repertoires (irrespective of competence), where traditional concepts fail to capture the blurred lines between languages and varieties in use. Exploring the opportunities brought by more recent notions in multilingualism studies, it is yet to be discussed whether the labourers' letters indeed constitute a case of what could be called historical languaging, and what the limitations are to apply it to historicalsociolinguistic settings. These archival sources show that the field of historical multilingualism and language contact is still undertheorized. Particularly for less clear-cut practices 'from below', we may need alternative, more flexible terms than code-switching, code-mixing or even umbrella terms like transference, in order to do justice to the fluid multilingual realities of the past -in reference to the 'fluid multilingual realities of today's world'. 79 At the same time, all sources investigated in this article signify that multilingualism and language contact, in all their reflections and manifestations, were indeed essential characteristics of the past, even in the ideologically-driven environment of nation-states and standard languages in the long nineteenth century. For a more complete language history of the Low Countries specifically, this means that a multilingual and transnational approach is needed, which also considers language practices beyond political and (standard) language borders. Notes 39. Reershemius,Niederdeutsch in Ostfriesland,. The Dictionary of the Dutch Language (WNT) notes that Baanhof, modelled on HG bahnhof was sometimes used for stationsgebouw. 41. Auer,"Dynamic Typology," 314. 42. This also becomes evident when he writes: Wat kan men thans toch schielÿk reizen! Donderdag morgen om 7 uur nog te Roden in Drenthe, en 's namiddags 5 ½ uur weder te Veenhuizen in Oostfriesland! 43. Krogull,Policy versus Practice,Ch. 9. 44. A common transfer is the recurring use of the modal particle ja, as found in letters by A.A.
The letter reveals lexicosemantic transference in inladend (< HG einladend, rather than SD uitnodigend) and tegenzijdige bezoeken (< HG gegenseitig, rather than SD wederzijds [i]n nineteenth-century informal writing, code-switching appears to be very rare on the whole'. Langer and Havinga, "Invisible Languages," 28. 54. Although recipe or cookbooks are not categorized as ego-documents in the strictest sense, they constitute equally unpublished, handwritten evidence from the private domain, especially of women. See however, Auer et al., "Historical Sociolinguistics," 7. 55. An unidentifiable second hand also contributes recipes, all written in Dutch. 56. For the use of writing systems in the nineteenth-century borderlands, see Elspaß,Sprachgeschichte von unten,149. 57. From page 38 onwards, Heerspink's book contains non-culinary entries in High German (e.g. Probates Mittel, die Männer bei guter Laune zu erhalten), possibly copied from other publications. The handwritten register also reflects the multilingual nature of the book and its main writer. 58. See for example, Lucassen,Migrant Labour;Eiynck et al.,Werken over de grens;and Gladen et al.,Hollandgang. 59. It is estimated that thousands or perhaps even tens of thousands of these letters were written, of which only a handful has been preserved. Lucassen, "Hannekemaaiersbrieven," 202. 60. Lucassen, "Hannekemaaiersbrieven." See also Niebaum, "Hollandgänger-Briefe." 61. The similarly 'mixed' letters by both writers shows that we are not dealing with idiosyncratic practices, although there is inter-writer variation. 62. 'De briefschrijvers maakten gebruik van het Nederduits, hun dialect, in het besef dat hun werkgever zonder de minste moeite hun boodschap zou verstaan'. Lucassen, "Hannekemaaiersbrieven," 206.

Data Availability Statement
The data that supports the findings of this research is available from the author upon reasonable request.