Contact Between Low German and Scandinavian in the Late Middle Ages

1 3 2 3 121 The Royal Gustavus Adolphus Academy for Swedish Folk Culture is 121 ACTA ACADEMIAE REGIAE GUSTAVI ADOLPHI CXXI a national academy based in Uppsala. According to its statutes, one of the means by which the Academy is to pursue its object of promoting CONTACT BETWEEN LOW GERMAN AND SCANDINAVIAN IN THE LATE MIDDLE AGES research into Swedish folk culture, understood in a broad sense, is by CONTACT BETWEEN LOW GERMAN AND SCANDINAVIAN IN THE LATE MIDDLE AGES publishing, in its various series, research findings in areas that it is charged with fostering. The main series is the Acta Academiae Regiae Gustavi Adolphi, the first volume of which appeared in 1933. Other series include Folklivsskildringar och bygdestudier (Studies of Folk Life and Local History), Kungl. Gustav Adolfs Akademiens småskrifter Contact between Low German and (Short Publications of the Royal Gustavus Adolphus Academy) and Svenska sagor och sägner (Swedish Folk Tales and Legends). ScandinavianScandinavian in the Late Middle AgesAges

This volume contains the proceedings of a conference held at the Uni- 25 Years of Research versity of Agder in Kristiansand, Norway, in 2010 on the new methods used in and findings emerging from the last twenty-five years of re- Lennart Elmevik and Ernst Håkon Jahr (editors) search into contact between Low German and the Scandinavian languages in the late Middle Ages.

The editor of the various series is the Secretary to the Academy, Associate Professor Maj Reinhammar, [email protected].

Distribution: Swedish Science Press Box 118 SE-751 04 Uppsala ISSN 0065-0897 UPPSALA 2012 E-post: [email protected] ISBN 978-91-85352-97-5

ACTA ACADEMIAE REGIAE GUSTAVI ADOLPHI 121 2 sid2 3

ACTA ACADEMIAE REGIAE GUSTAVI ADOLPHI CXXI

Contact between Low German and Scandinavian in the Late Middle Ages 25 Years of Research

Lennart Elmevik and Ernst Håkon Jahr (editors)

UPPSALA 2012

Kungl. Gustav Adolfs Akademien för svensk folkkultur 4

ISSN 0065-0897 ISBN 978-91-85352-97-5

Printed in Sweden 2012 Textgruppen i Uppsala AB 5 Contents

Preface ...... 7

Lennart Elmevik and Ernst Håkon Jahr: 25 years of research on the contact between Low German and Scandinavian ...... 9

Arnved Nedkvitne: A post-national perspective on the German Hansa in Scandinavia ...... 17

Gro-Renée Rambø: Language contact, communication and change . . . . . 39

Peter Trudgill: Gender reduction in Bergen Norwegian: a North-Sea perspective ...... 57

Agnete Nesse: Norwegian and German in Bergen ...... 75

Kurt Braunmüller: Semi-communication and beyond. Some results of the Hamburg Hanseatic Project (1990–1995) ...... 95

Stefan Mähl: Low German texts from Late Medieval Sweden ...... 113

Birgit Christensen: A survey of Low German loan words in Danish in the medieval period and the transition from Low German to High German as the written language in Tønder in the 17th century ...... 123

Helena Wistrand: Middle Low German loanwords in medieval charters issued in the Swedish province of Närke ...... 137

Kurt Braunmüller and Steffen Höder: The history of complex verbs in Scandinavian languages revisited: only influence due to contact with Low German? ...... 151

Ludger Zeevaert: Low German influence and typological change in Swedish: some results from a research project ...... 171

Erik Simensen: Low German and Nynorsk – a strained relationship? A glimpse into Norway’s most recent language history ...... 191 6 7 Preface

This volume grew out of an international conference at the University of Agder in Kristiansand, Norway, held in November 2010 entitled “Contact between Low German and Scandinavian in the late Middle Ages – 25 years of research”. The background to the conference was the intense research activity over the past 25 years or so which has completely revived a traditional field of study: Low German–Scandinavian contact. During this period, new theories of language contact and new methods of studying it have been employed to investigate this field in a fundamentally different way from all previous efforts. From Lennart Elmevik’s 1977 presentation of his project using post-war theoretical ideas about language contact to Gro-Renée Rambø’s 2009 University of Agder socio-historical PhD dissertation on the contact situation throughout Scandina- via in the Hansa period, completely new perspectives and aspects about this historical language-contact period have given rise to a multitude of new insights and results. The study of Low German–Scandinavian contact has now, in our view, been established on a more elaborate level theoretically and meth- odologically. This field’s rapid development has also substantially helped to further and stimulate theoretical developments in the new research field of historical sociolinguistics. We, the organizers of the Kristiansand conference and this volume’s editors, are very pleased that so many of the central researchers in this field, who have made essential contributions to its development over the past 25 years, par- ticipated in the conference and gave papers in which they summarized the main results of their research. We have also included a few papers which were sub- mitted after the conference. Our main objective in this volume is to provide an overview of many of the exciting results from this period, especially for an international audience. Many of these important reports have so far only been published in a Scandinavian language. Therefore, we hope that this volume, in English, will help inform the interested international community of language contact researchers of the numerous new and compelling results from the past quarter century. We thank the University of Agder, Kristiansand, and the Royal Gustavus Adolphus’ Academy for Swedish folk culture, Uppsala, for supporting the conference financially, and the Academy for including the volume in its Acta series. Finally, we thank Jean Hannah for her thorough copy-editing of the papers and Maj Reinhammar, editor of the Acta series, who undertook a final review of the text.

Uppsala and Kristiansand, September 2012

Lennart Elmevik Ernst Håkon Jahr 8 Contact between Low German and Scandinavian in the Late Middle Ages 9 Twenty-five years of research on the contact between Low German and Scandinavian

Lennart Elmevik, Uppsala University and Ernst Håkon Jahr, University of Agder

Language contact between the Scandinavian languages and Middle Low Ger- man in the late Middle Ages is a topic which has concerned a great number of researchers over the last century. A number of lengthy and short dissertations have been written describing the results of the immense influence to which mainland Scandinavian was exposed, particularly in the 14th and 15th centuries, when the German Hanseatic League was at the pinnacle of its power. With Lübeck as their centre, the Hansa merchants controlled a powerful northern European empire, a kind of late medieval EU. (Cf. Nedkvitne’s contribution.) But compared to today’s EU, the Hansa market had one great advantage: Low German was in use over the entire area. From Novgorod in Russia in the east to London in the west, from Bruges, Cologne and Cracow in the south to Bergen in the north, Low German served as the lingua franca of trade. This fact opens up interesting prospects and possibilities for comparative historical studies of language contact. Until about the mid-1970s, however, it was supposed that progress in linguistic research required the assumption of a linguistically homogeneous and self-contained language community. Within the theoretical paradigm of the Neogrammarians as well as within the later paradigms of structuralism and generativism, this has been considered a useful pragmatic assumption. However, in the past two generations it has been shown beyond doubt that a language community is never homogeneous, and hardly ever self-contained. Although most linguists can agree with this statement, it is still a fact that linguists working on language change at earlier times often neglected to reflect this in their work. When working on linguistic change from a diachronic perspective, linguists have often ignored extra-linguistic factors. This has been the case with research on language contact between Middle Low German and the mainland Scandinavian languages in the Late Middle Ages. Within the Neogrammarian framework, which was the dominant paradigm in Scandinavian philology and historical linguistics up until the 1970s, descriptions of the result of language contact are generally restricted to various types of loans. If an element X from language A is transferred (“borrowed”) into language B, this can easily be integrated into a Neogrammarian description. Neogrammarian 10 Lennart Elmevik and Ernst Håkon Jahr

“Stammbaum” theory could only explain clear and identifiable loans, be it on the lexical, morphological or syntactic level. Results of the contact situation which could not be traced to the influencing language, in our case Low German, were simply outside of the scope of what this paradigm could incorporate into the linguistic description. Up to only a few decades ago, most students of this transition- al period have concerned themselves mainly with problems of loan word adoption, i.e. they have tried to map the Low German loanwords which were adopted for use in different domains. Other types of linguistic consequences that the contact situation may have had have not been given much attention. There is a long tradition in Scandinavian historical linguistics of studying and classifying Middle Low German loanwords in the Scandinavian languages. A substantial amount of sound philological work has been carried out in this field over the past hundred years or so, and a great deal of empirical evidence has been presented. Thousands of loanwords have been identified, explained and classified. (Cf. Simensen’s contribution; Skancke 2001.) During the past generation, however, it has been established that language contact can result in a multitude of different linguistic and sociolinguistic phenomena, and that various mechanisms of language contact may bring about language change, short-term as well as long-term. It has been emphasized that a language contact situation has to be studied and described as an interplay between linguistic and non-linguistic factors. Language and language use are social phenomena that are influenced by extra-linguistic as well as intra-linguistic factors. Direct loans, previously the sole focus of interest, are thus only one of the many consequences of language contact. Thanks to post-war developments within language contact theory, it has now been possible to pick up this old discipline and reanalyse the sources by means of methods, theories and view- points other than those offered by the 19th century Neogrammarian paradigm. Equipped with new knowledge of what results language contact may actually yield in addition to loanwords, one can now ask new questions and discover outcomes and answers other than those presented in earlier research. Therefore, projects over the past 25 years have aimed at describing the language changes which can be observed in the Scandinavian languages in the 14th and 15th centuries as a combined result of language contact and social and language system-internal causes. One very important aspect that was soon brought to light was that linguistic interaction took quite different shapes in the various parts of the vast area of the Hanseatic trading empire, and as a consequence of this, the linguistic and sociolinguistic results of contact were not the same in all places. The conditions governing linguistic interaction were completely different in Russia and Estonia, for example, from those in Scandinavia. Even within the Scandinavian area we do not find the same type of contact everywhere. The sociolinguistic conditions, which defined the bounds of interaction everywhere, varied enormously from town to town. (On the differences within Scandinavia, see Ram- bø’s contribution; Rambø 2009.) Contact between Low German and Scandinavian in the Late Middle Ages 11

The new start or approach for investigating medieval Low German/Scandi- navian contact using post-war theories and perspectives can be dated to the late 1970s and 1980s. One of the first was a project in Stockholm led by Professor Lennart Elmevik (cf. Elmevik 1977, 1979), resulting inter alia in Lena Moberg’s 1989 dissertation, and another was the Mannheim-based Swedish professor Sture Ureland’s initiative which resulted in the conference volume Sprachkontakt in der Hanse (Ureland 1987). The five conferences in the series “Niederdeutsch in Skandinavien” (between 1985 and 1993) also provided important input (cf. Elmevik & Schöndorf 1992). Then, from about 1990, came the innovative Hamburg project led by Professor Kurt Braunmüller (cf. Braun- müller’s contribution; Braunmüller & Diercks 1993; Braunmüller 1995; Jahr 1995), and from 1997 the project “Språkhistoriske prinsipp for lånord i nordiske språk” [Language-historical principles for loan words in the Scandina- vian languages], financed by the Nordic Council and led by Professor Ernst Håkon Jahr, then at the University of Tromsø (cf. Jahr 1998a, 2000, 2006, Skancke 2001, Nesse 2002, Rambø 2009). Most articles and dissertations through the years (starting in the 19th century) on contact between Low German and mainland Scandinavian have, as mentioned above, discussed loanwords, i.e. direct loans from Low German into Scandinavian. One was Clara Holst’s dissertation in 1903, when she became the first woman to defend a PhD in Norway (Holst 1903; cf. Jahr 2006). There is no doubt that the new words which entered the Scandinavian word stock within only a couple of centuries changed the character of these languages completely. At the same time, a typological change from synthetic to analytic language structure took place, so that by the beginning of the 16th century the Scandinavian languages had reached their modern stage. This volume brings together contributions which summarize a great many of the results and perspectives developed and followed up during the last 25 years of intensive research on contact between the Low German of the Hanse- atic League and the Scandinavian languages.

One major result of the past 25 years of research: the refutation of the “mixed language” theory (cf. Jahr 1997, 1998b, 2001) The influence of Low German on the mainland Scandinavian languages, both on the lexical and structural level, was enormous and unparalleled in Scandi- navian language history, and may be compared with the huge influence that Norman French exerted on Old English following the Norman Conquest. A typological shift occurred simultaneously in Danish, Swedish and Nor- wegian, all of which had close contact with Low German, but it did not take place in Icelandic, which had much less contact with Low German. 12 Lennart Elmevik and Ernst Håkon Jahr

The main question here, is this: what role, if any, did language contact play in the Late Middle Ages with respect to the typological change that the mainland Scandinavian languages underwent during this period? The most important period of change coincided exactly with the hundred years or so when contact between Low German and the Scandinavian languages was most intense. It is difficult to imagine that this contact had no effect whatsoever on the typological shift. The question of how the decline in inflectional morphology came about is, of course, not new, and Scandinavian philologists and historical linguists raised the issue relatively early on. The explanation adhered to by most Scandinavian philologists of the Neogrammarian persuasion was to postulate a grammatically simplified Scandinavian language used originally by the German merchants when communicating with the natives, or a mixed-language type of idiom consisting of Scandinavian (Swedish, Danish or Norwegian) and Low German elements. These varieties, which never have been given a precise description – for obvious reasons – were then thought to have caused or at least been a cata- lyst for the overall deconstruction of the Scandinavian case system. This view has been especially popular among Swedish scholars. As early as 1889, Tegnér suggested the possibility of a written mixed variety among the upper classes, but it was Wessén (1929) who developed the theory: When towards the end of the Middle Ages the rich inflectional system of the old language dissolved and underwent simplification, this has, with good reason, been explained partly by the fact that the German immigrants could never learn to use the old case forms and inflections correctly; their simplified morphological system was gradually adopted by the country’s own children. (Wessén [1929] 1954, p. 27; translated from the Swedish original.) According to Wessén, then, the Germans used a morphologically simplified Swedish, not necessarily a mixed language with Low German and Swedish/ Scandinavian elements. In later publications, however, this distinction is not maintained. Törnquist (1954) for example, assumes a mixed language, a kind of pidgin, which was used between German and Swedes, and subsequently among Swedes themselves. On the other hand, he also claims that Swedish merchants acquired Low German and that their Swedish was affected by this because Low German carried high social prestige. Such a concrete mixed language was postulated in 1913 by Cederschiöld, who claimed that the Low German loanwords in Swedish had probably been a feature of the mixed language spoken by merchants – travelling or resident in Sweden – as well as by artisans from Northern Germany and their assistants, mercenaries, etc., when they needed to make themselves understood by the Swedes, and which was later adopted by the Germans’ Swedish colleagues, the inhabitants of the town. (Cederschiöld 1913, quoted by Elmevik 1979, p. 227; translated from the Swedish original.) Ronge (1967, p. 275) expressed the following opinion: Contact between Low German and Scandinavian in the Late Middle Ages 13

In towns with a strong German element, the linguistic communication between Ger- mans and Swedes probably resulted in some kind of mixed language, a variable spoken language, internalized very differently from speaker to speaker, and thus difficult to pin down. [...] It is possible that this hypothetical mixed language, which arose in certain ‘core areas’, has served as the direct link in the linguistic influence which Middle Low German had had on Old Swedish. This influence has resulted in a vast number of German loanwords, even functional words, as well as a number of prefixes and suffixes, and it has also changed Swedish morphology and syntax in various ways. (Translated from the Swedish original.) In two dissertations (Moberg 1989; Winge 1992) the mixed language theory is discussed. Winge (1992, p. 17) claims that even though in her view it can be proved that the upper strata in Denmark were bilingual at certain points in time, it was also the case that “the socially lower classes earlier communicated through a mixed language or simply (like the Scandinavians do today) through semi-communication” (translated from the German original; on semi-communication, cf. Braunmüller 1994). As for the Swedish situation, Moberg (1989, p. 12) writes: In easy linguistic transactions Swedes and Germans must obviously have had relatively good chances of making themselves understood across linguistic barri- ers without much knowledge of the other language. Working from this assumption, Scandinavian linguistics has not infrequently described the code or codes which were used in the interaction between Swedes and Germans in terms which make one think of what we today refer to as pidgin. (Translated from the Swedish original.) And, finally – with respect to the Norwegian situation – Skard (1973, p. 132) writes that the Norwegians have in part used a kind of lingua franca – a mixed language – in their interaction with the Germans. [...] As a result, they [i.e. the Germans] did not need to learn the complex peculiarities of the Norwegian inflectional system but could instead speak ‘Norwegian’ as they pleased. This simplification of Norwegian could easily have spread. There were already forces inherent in the language supporting such a tendency, which would then have been helped on its way by foreign influence, resulting in an even more radical departure from earlier language traditions. (Translated from the Norwegian original, cf. also Dalen 1994.) What then are we to make of all the discussions concerning a mixed language during the Hanseatic period, which come up time and again? Are they all un- founded? It would appear so. The most concrete reference to the existence of a mixed language of the pidgin type was presented by Wahl (1927, p. 226, note), who translated a directive to the Bergenhus Fort, dating from around 1530, into modern Norwegian. Wahl writes that this directive “is drawn up in the language used by the mercenaries of the time – a curious mixture of Northern Eu- ropean languages, in particular Low German”. (Translated from the Norwegian original.) 14 Lennart Elmevik and Ernst Håkon Jahr

Wahl gives no indication of the origins of the source, nor does he provide any example of this “curious mixture”, as he calls it. But Wahl’s short note has subsequently been referred to frequently (initially by Brattegard 1932, p. 303) as evidence that such a pidgin language existed in Bergen in the 16th century. Simensen (1989) finally established that Wahl’s source must have been Charter No. 566 in Diplomatarium Norvegicum, vol. 13. However, this charter is written in the Low German of the period, and there is nothing in the text which merits the term ‘mixed language’. It must be concluded that, to date, no one has been able to provide any actual proof that a pidgin or mixed language derived from Low German and the Scan- dinavian languages existed during the Hanseatic period. The most probable reason for the lack of such a pidgin-like mixed Scandinavian–German idiom is that, at the time, Scandinavian and Low German were, as it seems, mutually intelligible (cf. Jahr 1997, 1998b, 1998c).

References Brattegard, Olav, 1932: Über die Organisation und die Urkunden des hansischen Kon- tors zu Bergen bis 1580. In: Skrifter utgivne av Bergens historiske Forening 38. Pp. 237–303. Braunmüller, Kurt, 1994: Semikommunikation og lingvistiske simulationsmodeller (om sprogkontakt i hansetiden). In: Ulla-Britt Kotsinas & John Helgander (eds.), Dialektkontakt, språkkontakt och språkförändring i Norden. Föredrag från ett fors- karsymposium. Stockholm. (Meddelanden från Institutionen för nordiska språk vid Stockholms universitet 40.) Pp. 92–97. Braunmüller, Kurt (ed.), 1995: Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Braunmüller, Kurt & Diercks, Willy (eds.), 1993: Niederdeutsch und die Skandinavi- schen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Cederschiöld, Wilhelm, 1913: Studier över genusväxlingen i fornvästnordiska och fornsvenska. Gothenburg. (Göteborgs Kungl. Vetenskaps- och Vitterhets-Samhälles Handlingar, Fourth series. Vol. 14, No. 2.) Dalen, Arnold, 1994: The Influence of Low German on the Norwegian Language. In: Volker Henn & Arnved Nedkvitne (eds.), Norwegen und die Hanse. Wirtschaftli- che und kulturelle Aspekte im europäischen Vergleich. Frankfurt am Main. Pp. 31–39. Diplomatarium Norvegicum 1889, Vol. 13. Christiania [Oslo]. Elmevik, Lennart, 1977: Prosjektet ‘Det tyska inflytandet på svenska språket under medeltiden’. In: Meddelanden från Institutionen för nordiska språk vid Stockholms universitet 1. Stockholm. Pp. 5–11. Elmevik, Lennart, 1979: Svenskt och lågtyskt i medeltidens Stockholm – kring ett tvåspråkighetsproblem. In: A. Stedje & P. af Trampe (eds.), Tvåspråkighet. Föredrag vid det andra Nordiska tvåspråkighetssymposiet. Stockholm. Pp. 226–232. Elmevik, Lennart & Schöndorf, Kurt Erich (eds.), 1992: Niederdeutsch in Skandinavien III. Akten des 3. Nordischen Symposions “Niederdeutsch in Skandinavien” in Sig- tuna 1989. Berlin. Holst, Clara, 1903: Studier over middelnedertyske laaneord i dansk i det 14. og 15. aar- hundrede. Kristiania [Oslo]. Contact between Low German and Scandinavian in the Late Middle Ages 15

Jahr, Ernst Håkon (ed.), 1995: Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Jahr, Ernst Håkon, 1997: New perspectives on the language contact between Middle Low German and mainland Scandinavian in the Late Middle Ages, and about a foot- note on mixed languages which gave rise to a “detective story”. In: Multilingua 16. Pp. 325–337. Jahr, Ernst Håkon (ed.), 1998a: Språkkontakt i Norden i middelalderen, særlig i hansatiden. Nord 1998/4. København. (Skrift nr. 1 fra prosjektet Språkhistoriske prinsipp for lånord i nordiske språk.) Jahr, Ernst Håkon, 1998b: Sociolinguistics in historical language contact: the Scandina- vian languges and Low German during the Hanseatic period. In: Ernst Håkon Jahr (ed.), Language change: advances in historical sociolinguistics. Berlin/New York. Pp. 119–140. Jahr, Ernst Håkon (ed.), 1998c: Language change: advances in historical sociolinguistics. Berlin/New York. (Trends in Linguistics, Studies and Monographs 114.) Jahr, Ernst Håkon (ed.), 2000: Språkkontakt – Innverknaden frå nedertysk på andre nordeuropeiske språk. Nord 2000/19. København. (Skrift nr. 2 frå prosjektet Språk- historiske prinsipp for lånord i nordiske språk.) Jahr, Ernst Håkon, 2001: Historical sociolinguistics: the role of Low German language contact in the Scandinavian typological shift of the Late Middle Ages. In: Lingua Posnaniensis 43. Pp. 95–104. Jahr, Ernst Håkon, 2006: Clara Holst – kvinnelig pionér i akademia i Norge. Oslo. (Skrift nr. 5 fra prosjektet Språkhistoriske prinsipp for lånord i nordiske språk.) Moberg, Lena, 1989: Lågtyskt och svenskt i Stockholms medeltida tänkeböcker. Upp- sala. (Acta Academiae Regiae Gustavi Adolphi, 58.) Nesse, Agnete, 2002: Språkkontakt mellom norsk og tysk i hansatidens Bergen. Oslo. (Skrift nr. 4 fra prosjektet Språkhistoriske prinsipp for lånord i nordiske språk.) (Det Norske Videnskaps-Akademi. II. Hist.-Filos. Klasse. Skrifter og avhandlinger nr. 2.) Rambø, Gro-Renée, 2009: Historiske og sosiale betingelser for språkkontakt mellom nedertysk og skandinavisk i seinmiddelalderen – et bidrag til historisk språksosiolo- gi. Oslo. (Skrift nr. 6 fra prosjektet Språkhistoriske prinsipp for lånord i nordiske språk.) Ronge, Hans, 1967: Nedertysk sprog. Sverige. In: Kulturhistorisk leksikon for nordisk middelalder 12. Pp. 275–276. Simensen, Erik, 1989: Zum Kontakt zwischen Niederdeutsch und Nordisch im Mittel- alter unter besonderer Berücksichtigung der Frage nach einer niederdeutsch-nordischen Mischsprache. In: K. Hyldgaard-Jensen, V. Winge & B. Christiansen (eds.), Niederdeutsch in Skandinavien II. Berlin. Pp. 65–77. Skancke, Kaja, 2001: Mellomnedertysk påvirkning på nordnorske dialekter. Nord 2001/ 26. København. (Skrift nr. 3 fra prosjektet Språkhistoriske prinsipp for lånord i nordiske språk.) Skard, Vemund, 1973: Norsk språkhistorie. Bind 1 – til 1523. Oslo–Bergen–Tromsø. Tegnér, Esaias, 1889: Tyska inflytelser på svenskan. In: Arkiv för nordisk filologi 5. Pp. 155–166, 303–344. Ureland, Per Sture (ed.), 1987: Sprachkontakt in der Hanse. Aspekte des Sprachaus- gleichs im Ostsee- und Nordseeraum. Akten des 7. Internationalen Symposions über Sprachkontakt in Europa, Lübeck 1986. Tübingen. Wahl, Johan Olaus, 1927: Esge Bilde. Lensherre paa Bergenhus 1529–37. In: Skrifter utgivne av Bergen historiske Forening 33. Pp. 211–266. Wessén, Elias, [1929] 1954: Om det tyska inflytandet på svenskt språk under medeltiden. Stockholm. 16 Lennart Elmevik and Ernst Håkon Jahr

Winge, Vibeke, 1992: Dänische Deutsche – deutsche Dänen. Geschichte der deutschen Sprache in Dänemark 1300–1800 mit einem Ausblick auf das 19. Jahrhundert. Hei- delberg. A post-national perspective on the German Hansa in Scandinavia 17 A post-national perspective on the German Hansa in Scandinavia

Arnved Nedkvitne, former professor, University of Oslo

German cultural and economic influence has been strong in all of Eastern Europe and Scandinavia since the 12th century, and until 1945 many Germans lived there in areas with a mixed German and Slav population. German historians saw it as part of their responsibility to describe and explain the history of this German diaspora to the east. They felt Hansa history to be important, and in the century between 1870 to c. 1970 a series of excellent monographs of high scientific quality were written. The names of Dietrich Schäfer, Wilhelm Koppe, Friedrich Bruns and Walter Vogel were well known and much referred to among Scandinavian historians. In the Soviet period (1945–1989) this Ger- man culture in the east was eradicated. With it also disappeared the most important ideological basis for traditional Hansa history, and interest in research into Hansa history declined. But some historians and philologists in both Ger- many and Scandinavia have maintained the tradition, attempting to liberate it from its nationalist past. Their central themes remain the same, but the concepts they employ are often different, as are even the conclusions sometimes. For this article I have chosen three basic themes which run through general Hansa history and are highly relevant in a discussion of the Hansa in Scandina- via: – Should the Hansa be seen as transmitters of more advanced West Euro- pean technologies and social practices to Scandinavia and Eastern Europe? – How was it possible to keep the international Hansa organisation in existence for several centuries without military power, compulsory courts of law and a central bureaucracy? – Scandinavian towns with German settlers existed as multiethnic societies.

The Hansa as transmitters of West European economic and social practices to Scandinavia In 1945 and 1946, 14 million Germans fled or were evicted from Eastern Europe, many of them from old Hansa towns like Reval, Riga, Königsberg, Danzig and Stettin. Their ancestors had emigrated eastward over a period of 18 Arnved Nedkvitne

700–800 years and were so numerous that they had kept their German language and culture.1 There was German immigration to Scandinavia during the Hansa period on a more modest scale. What made them emigrate and form this large diaspora? Did they embody a more advanced West European technology and social organisation? This question was interesting for nationalist historiography, and still is today. In 2000, the German historian Carsten Jahnke published his thesis on the herring trade in the Baltic region during the period 1100–1600. The first commercial fisheries in the Baltic were located on Rügen, and they became the pro- totype for the organisation of herring fisheries in the region. The merchants played the central role; they brought salt, organised the preservation process for the herring, and transported it to the markets in the German interior. These merchants were Germans, from Bardowiek on the Elbe around the year 1130, later on mainly from Lübeck and Stralsund (Jahnke 2000, pp. 21–23, cf. pp. 37–38). The much more important Scanian fisheries catered only for local needs up to the middle of the 12th century. In the second half of the century these fisheries were drawn into long-distance trade, with the German interior as an important market. Lübeck merchants are the first who are mentioned as curing and selling this herring, and even though Jahnke does not say so explicitly, he presents the founding of Lübeck in 1159 and the sale of Scanian herring to distant markets as connected developments (Jahnke 2000, pp. 63–64, 69, 47–48, 352). In my own PhD dissertation on the Bergen trade (Nedkvitne 1983), I discussed the origin of stockfish exports from Bergen more explicitly. Stockfish had been produced along the Norwegian north and west coasts for local consumption centuries before the Middle Ages. Drying cod without salt is simple to organise, and peasant fishermen needed no help from merchants to produce the stockfish, but merchants were needed to transport it to foreign markets. In the last decades of the 11th century this stockfish appeared on the English market, and in the 12th century its export became economically important for the town of Bergen and for the coastal population. The driving force for the first phase of this long-distance trade came from England, and possibly from West German ports on the Rhine estuary. The exporting merchants were Norwegian, English and West German. Around 1250, Lübeck and its neighbouring Baltic towns took the lead; their merchants had a wide northern European trade net- work, which the early merchants lacked. They imported large quantities of Bal- tic grain, which encouraged peasant fishermen to produce more stockfish to trade for it. Archaeological excavations confirm that around 1250 there was a substantial growth of fishing villages in northern Norway. Lübeck and the other Wendish towns (Hansa towns east of Lübeck) exported the fish mainly to England and to continental North Sea ports between Bruges and Bremen (Ned- kvitne 1983, chapter I). The Germans were able to organise their stockfish trade more effectively

1 See a special issue of Der Spiegel in January 2011 entitled “Die Deutschen im Osten”. A post-national perspective on the German Hansa in Scandinavia 19 and quickly because they were professionals, while the Scandinavians more often were engaged in it to supplement their main income from farming or ad- ministering landed property. The Hansa trade had a more permanent character, and this permitted production of the goods for long-distance trade to become more entrenched as well. Hansa merchants accumulated more capital, which enabled each merchant to ship larger quantities and offer credit to their customers. They also had more extensive trading contacts, which enabled them to combine exports of high-quality cloth from England and Flanders, fish from Bergen and grain from the Baltic, giving their customers a wider choice of products. Carrying out trade in many towns at the same time was possible through the use of written correspondence and accounts. A literate merchant could buy and sell for a partner living in another town, and this partner could exercise a certain amount of control through the accounts he received and the letters he sent and received (Nedkvitne 1983, chapter IV). The main exports from Stockholm were copper and iron. Göran Dahlbäck has been the most active historian of medieval Stockholm in recent decades. He has never discussed explicitly whether the Hansa created the Swedish metal industry, but his work leads the reader to think this was the case. Hansa merchants sailed into Lake Mälaren to buy metals from the end of the 12th century (Dahlbäck 2002, p. 18). Stockholm was founded just before 1250, and one of the main motives behind its establishment was to protect and control international trade. Copper and iron were the town’s main exports throughout the medieval period, and the exporters were almost exclusively from Lübeck, later also from Danzig. Many of them chose to become citizens of Stockholm. At the end of the 15th century about one third of the taxpayers in Stockholm were ethnic Germans, but there were fewer of them in the lower classes, who did not pay taxes (Dahlbäck 1988, pp. 16, 52, 71–72, 80; Dahlbäck 2002, p. 40). Did Hansa merchants create the Swedish mining industry? The question has been of interest in Swedish and German historiography during the national as well as post-national period. Research over the last decades has confirmed traditional views about this (Gustafsson 2006, p. 23). Through the Hansa, Scandinavian producers were assured a permanent market at predictable prices and a variety of desirable goods in return. This encouraged specialised market production. Few would dispute that this was to the advantage of Scandinavian fishermen and other producers. The impetus came from outside and would no doubt have reached Scandinavia even without the Hansa, but the process would then have been slower, and more transmitters of the new technologies would have been native Scandinavians. While Scandina- vian nationalist historiography focused mainly on the slower development of a Scandinavian merchant class, modern historians concentrated more on the ad- vantages to the producers. This change of focus is due to a shift of interest among social scientists from nationalism to social differences within the nation, and from the bourgeoisie to the common people. But the description of the social realities has not changed radically. 20 Arnved Nedkvitne

In other towns like Kalmar, Oslo and Tønsberg, the Hansa did not develop new products. They bought butter, hides, furs, bog iron and other goods which the peasants had produced from time immemorial. These products came as rent payments to the landowners, who in turn sold the goods to Hansa merchants. The landowners received from the Hansa prestigious West European products like quality cloth, objects of copper and bronze, weapons, wine and wheat. The Hansa trade made it desirable for landowners to demand higher rent from ten- ants. The result was harsher economic oppression, but also the creation of a West European-type elite class (Blomkvist 1979, pp. 167–278, 187–188, 264, 276; Nedkvitne & Norseng 1991, pp. 192–193). Originally all Scandinavian towns were governed by a royal bailiff in co- operation with the town’s assembly. At the end of the 13th century this arrange- ment was marginalised, to be replaced by a town council whose members were prosperous merchants. In Norway and Denmark, the consensus is that the state was behind this change. In Sweden, the influential Swedish urban historian Adolf Schück voiced this opinion as early as 1926 (Schück 1926). His German contemporary Fritz Rörig, on the other hand, maintained that Hansa merchants who had settled in Swedish towns and became members of the local town councils were the motivating force. They brought the idea of town councils manned by the urban elite with them from their Hanseatic home towns (Rörig 1928). German settlers have also been seen as having an influence on urban legislation and ordinances regulating trade in general. A modern Swedish voice is Sofia Gustafsson (2006). She concluded that both royal officials and town councillors with German origins were behind new urban legislation. They derived their ideas mainly from German and English towns, but she rejects the view that the laws of Lübeck were taken as the main model. The legislators knew about the norms in North European towns and adopted what was most relevant to their situation. Scandinavians imported general West European ideas on urban law, with sections formulated to fit their particular circumstances (Gus- tafsson 2006, pp. 31–34, 209–211). In Nedkvitne (1983) I concluded that Norwegian laws and ordinances from the period 1280–1380 on the rights of foreign merchants in Norway were modelled on those from England and Germany, possibly also from the Low Countries. For each regulation or paragraph, it was possible to find parallels in these countries. Norwegian authorities did not import their legislation from one particular town or district (Nedkvitne 1983, p. 220). This is the same conclusion as Gustafsson’s for Sweden, but in Norway it is less con- troversial. About 1100–1350, Scandinavia experienced a “commercial revolution”, influenced by Western Europe and not by Germany alone. In Western Norway trade was initially conducted mainly with England, but later the Hansa towns became more important. The Danish isles, Scania and Sweden received their economic influence from Germany from the very start. Legislation accompanying this was taken from German and English towns. With this modification, A post-national perspective on the German Hansa in Scandinavia 21 there is consensus today that the Hansa introduced more advanced commercial organisational systems and technologies into Scandinavia.

The ties which united the Hansa What kept the Hansa together as an organisation for several centuries? For Ger- man historians in the nationalist tradition, it was the national identity of the Hansa merchants. According to Rörig, their identity in the 13th century was focused on the Reichsidee or Reichsgedanken. The Reich is to be understood here as the German kingdom or empire. In the second half of the 13th century this empire broke down as a political and administrative structure. What then kept the Hansa together was exclusively their ethnicity as Germans (Rörig [1940] 1971, p. 539). Modern historians distinguish between “ethnic identity”, which is an identity among people who speak the same language and have a common history, and “nationalism”, which is the idea that people with a common ethnicity should have a state of their own. Rörig states that Hansa merchants had a common ethnic identity which in the early period was nationalist. For German and Scandinavian historians writing in recent decades, the strength of ethnic and national identities has been less self-evident. But the fact remains that the Hansa Kontor2 in Bergen, the town councils of Hansa towns and the Hansa diets (Hansetage) frequently referred to their German ethnicity, and often added the adjective erbarnn, ‘honourable’, to their titles. The Hansa diets are time and again called dudischenn erbarnn anse steder (Norges Gamle Love 1934, p. 674). The Kontor called citizens of Bergen ‘the Norwegians’ (de Norman), while the Norwegian authorities and the town council of Bergen called the Hansa merchants ‘Germans’ or de Kopman, which shows that they too thought in ethnic categories (Norges Gamle Love 1934, pp. 245–248, 251– 257). German ethnicity created ties between Hansa merchants travelling and living in Scandinavia and around the Baltic, but historians in recent decades no longer see this as the main force that kept the Hansa together. Today historians discuss large organisational units like the state less in terms of ideas and more in sociological terms. A fruitful point of departure here is Max Weber’s definition of the pre-modern state as an organisation which has the monopoly on legitimate violence within a certain area, has superior control over the courts of law, and has a bureaucracy at its disposal. The Hansa existed on two organisational levels. On the lowest level, it was an association between individual merchants who visited the same market at the same time, the most important of these in Scandinavia being in Bergen and Scania. On a higher level, it was a league between Hansa towns, represented by their town councils. The Hansa towns were politically independent, and each exercised state power as defined by Weber.

2 One of four major German trading settlements within the Hanseatic area. 22 Arnved Nedkvitne

The central Hansa institution was the Hansa diet, where representatives of the member towns met to discuss subjects of common interest. Attendance at the diet was voluntary, and if a debate took a turn which was not in a particular town’s interest, its representatives could simply leave, or neglect to implement the decisions when they returned home. The diet could wage war against an external enemy, but a member town was not obliged to put its soldiers or warships at the disposal of the Hansa for that purpose. The Hansa diets did not have courts of law or a bureaucracy of their own, but individual towns put their re- sources at the disposal of the Hansa for particular purposes. State power belonged to individual member towns, and not to the Hansa diets. In Bergen there was a Hanseatic settlement which in winter numbered about 1000 merchants including their servants, and in summer had more than double that number, sailors and visiting merchants included. In 1366 they organised themselves in a Kontor subordinated to the Hansa diets. They elected aldermen, who exercised extensive state powers over the German settlement; the Hansa merchants had exterritorial rights. In the period between the establishment of the Kontor in 1366 up to 1559, the Hansa merchants dominated Bergen mili- tarily. They were armed and could mobilise between 1000 and 2000 men. In contrast, the royal castle Bergenhus had only about 50–70 armed men. In 1455 the aldermen’s militia killed 60 Norwegians, including the town’s bishop and the commander of the king’s castle. The aldermen also acted as judges in internal disputes between Germans; when there were disputes between Germans and Norwegians, the Germans were often smuggled out of the country to avoid facing a Norwegian court. The six aldermen, 18 assistants and the Kontor’s secretary constituted a small part-time bureaucracy which, among other things, corresponded with the councils in the Hansa towns. Judgements by the aldermen could be appealed to Lübeck (Nedkvitne 1983, pp. 260–264, 267; Ned- kvitne 2011, pp. 140–142). Justyna Wubs-Mrozewicz (2005) has analysed how the aldermen used their powers. Their main purpose was to organise Hansa trade with Norwegian customers to the benefit of all Germans in the Kontor, and prevent Norwegian officials from interfering to the detriment of Hansa interests. The majority of the merchants and almost all the aldermen were from Lübeck, and they sought to organise trade and shipping in a way which was advantageous to winter residents, particularly those from Lübeck, and discriminated against merchants from North Sea towns, particularly those from Kampen, Deventer and Zwolle in today’s Netherlands. The aldermen used strong-arm tactics against Hansa merchants who collaborated with outsiders like the Hollanders, such as public humiliation, large fines and expulsion (Wubs-Mrozewicz 2005). The fair in Scania in Denmark was different because no single Hansa town dominated, and because the Hansa merchants were present there only 3–4 months a year. Each Hansa town was given its own limited area called a Vitte, and the town sent a bailiff (Vogt) to administer it. He passed judgement in conflicts which originated in the Vitte, here the town had exterritorial rights. Inside A post-national perspective on the German Hansa in Scandinavia 23 its Vitte, the Hansa town was at liberty to organise trade as they pleased. Indi- vidual citizens were given plots where they could erect private booths or market stalls, selling goods and preparing salted herring without restrictions (Jahnke 2000, pp. 193, 207). Customs duties and other fees were sometimes decided by the town in question (Jahnke 2000, p. 199). Hansa towns had great problems cooperating in the settlements they visited abroad because each town pursued its own interests, and the feeling of communality was weak. A settlement was administratively strong only when one town dominated and in practice organised it alone. Urban identity seems to have been stronger than Hansa identity. There were numerous guilds in all Hansa towns, and a merchant could be member of several. Their main purpose was to create social ties, exchange information, put pressure on the town councils in matters of common interest, and organise religious rituals to the benefit of guild members. Several of these guilds organised merchants who traded at certain markets (Wernicke 1999). There were Bergenfahrer guilds in Lübeck, Wismar, Stralsund, Greifswald and Deventer, and in Rostock there was a Wiekfahrer guild for merchants who traded in Oslo and Tønsberg (Brück 1999). In even more towns there were guilds for merchants trading at the market in Scania. All these guilds took members from one town only. This was partly for practical reasons; social gatherings and feasts would be difficult to organise if the members lived in different towns. But politically these guilds lobbied for certain policies, and the most effective channels for this were the separate town councils. In 1350–1470 the price of stockfish compared to other goods was particularly favourable (Nedkvitne 1983, chapter VI), and during this period several prosperous Bergenfahrer from Lübeck were members of the town council and the most prestigious guilds of the town (Burkhardt 2009, pp. 153–167). The guilds strengthened and encouraged town identity and not Hansa identity, and the flowering of such guilds is one of several indications that urban identity was stronger than Hansa identity. The overall pattern was that state power belonged to the individual Hansa towns, a merchant’s identity was primarily tied to a specific town, and Hansa political power functioned best when one town took responsibility and the others followed its lead. But despite this fact, the Hansa community existed as an organisation to be taken seriously for 300 years (1250–1550), and in some form for 500 years (1159–1669). What were the ties that kept the Hansa together? The Hansa towns and their merchants followed laws which were largely identical, or at least similar. In the 13th century numerous new towns were founded along the Baltic Sea and in Eastern Europe, and they copied each others’ laws. These were supplemented by ordinances passed by the individual town councils (Theuerkauf 2006). In Bergen and other Hansa settlements, merchants from several towns adopted statutes which were enforced by elected aldermen. In 1494 the Kontor in Bergen had statutes (Wilkor) comprising 100 24 Arnved Nedkvitne paragraphs (Norges Gamle Love 1934, document number 416). This created a community of law between merchants from several towns. It was common for two or three merchants to pool their capital and work in partnership. This was the rule in the Bergen trade, where one of the partners would live in Lübeck and usually own most of the capital, and the other would live in Bergen and do most of the work. In such economic partnerships, both merchants would usually come from the same town. Merchants living in different towns could also order goods from each other by letter and send them by ship without travelling themselves and without being partners in an economic sense. In the Bergen trade, the person who organised all of this lived at the Han- sa Kontor in Bergen, and he received grain and beer from Lübeck, Wismar, Rostock, Stralsund and Danzig, and cloth and other western goods from Boston in England, Deventer or Bruges (Nedkvitne 1983, pp. 314–316; Cordes 1999, pp. 70, 72–76; Burkhardt 2009, pp. 188–190). These buying and selling networks created ties between merchants living in different Hansa towns, Kontore and other settlements abroad. But the strongest ties between Hansa merchants were created through their joint efforts to obtain trading privileges in foreign countries. Scandinavia had laws which were valid for everybody living in a certain region, but they were mainly meant for peasant societies. In addition, there were urban laws for problems which were likely to arise in towns and marketplaces. In the 13th century the foreign merchants grew more numerous at the same time that state power was being strengthened. State officials wanted special laws for foreign merchants, and the latter demanded to be exempted from duties which the state imposed on native merchants. Conflicts concerning the legal position of foreign merchants erupted all over northern Europe in the second half of the 13th century. The conflicts were settled through privileges which were negotiated between the state and groups of foreign merchants. Hansa towns often opted for individual privileges instead of those which were common to all Hansa merchants. The first known privilege related to the market in Scania dates from 1251, and it gave merchants from Rostock the right to be judged by their own bailiff according to Rostock’s laws in disputes among themselves (Jahnke 2000, p. 69). During the period 1319–1340 there was a civil war in Denmark, and the warring parties sought military support from neighbouring Hansa towns by granting them almost all the privileges they asked for in Scania. Individual towns obtained privileges by exerting political pressure, this favoured the most valuable political allies (Jahnke 2000, pp. 81– 90). The first privileges granted in Norway were negotiated in 1278 by two councillors from Lübeck, acting on behalf of merchants from unnamed German maritime towns. Several privileges were granted by successive kings until 1343. During this early period, Lübeck and its Baltic neighbours jointly negotiated their privileges, while North Sea towns like Bremen, Hamburg and Kampen negotiated individually (Nedkvitne 1983, pp. 31, 198–205). The military situation in Bergen 1366–1559 described above made it poss- A post-national perspective on the German Hansa in Scandinavia 25 ible for the Kontor to introduce its own trading practices without worrying too much about the opinion of Norwegian authorities. The Kontor community adopted its own statutes, which were kept secret from Norwegian authorities in Bergen, and they enforced them by mobilising the Kontor militia if necessary. A Norwegian who owed money to a Hansa merchant had to trade with this merchant until the whole debt was paid back, and Hansa merchants stayed in Bergen as long as they wished. Neither the Kontor’s organisation nor its practices described above were legitimised through privileges; they were defended through the use of military force. Such use of a mobilised militia demanded a high degree of solidarity among all merchants who were present (Nedkvitne 1983, pp. 270–278; Nedkvitne 2011, pp. 142–149). The strong Bergen Kontor was exceptional in the Hansa organisation. A weak central authority made the Hansa organisation flexible. When a problem arose, the towns which felt their interests to be affected could take action. The town which had the strongest authority and interest in the matter took the lead. The Hansa could mobilise to tackle large problems, but also small ones where only a few towns had an interest. Nationalist historiography tended to present the Hansa as a state-like organisation, and focused on its conflicts and strong-arm tactics in Bergen and Scania. Modern historiography sees its loose organisation as more interesting. And it is challenging to try to understand how the Hansa was able to defend its merchants without support from a strong state. The Hansa in relation to its member towns was in a similar position as the EU is today in relation to its member states.

Scandinavian towns with German settlers as multiethnic societies In the last half century there has been a growing interest in local history all over Scandinavia, and a large portion of the published work has been authored by academically trained historians. Today most Scandinavian research into Hansa history is written as part of the history of a particular town. German minorities are seen as contributing to the town’s identity, and their economic, political and social role is discussed in relation to the town as a whole. Problems of integration arouse great interest in contemporary society and are discussed in the town histories of Copenhagen (Kjersgaard 1980), Aalborg (Knudsen & Kock 1992), Malmö (Tomner 1971), Stockholm (Dahlbäck 2002, 1998), Kalmar (Blom- kvist 1979), Bergen (Helle 1982) and Oslo (Nedkvitne & Norseng 1991). In the German nationalist tradition, most clearly represented by Rörig, the first step in German emigration to the Baltic was that merchants from Western Germany created corporations with clear ideas about how their new settlements should be organised. Then they sailed east and founded new towns as planned. 26 Arnved Nedkvitne

From the start the German merchants comprised a ruling class who exercised power through the urban councils (Rörig 1928). Scandinavian and East European nationalist historians, on the other hand, took as their main focus the process of building the state in their respective countries. Some of them saw the German merchant settlers as an instrument in the kings’ efforts at state-building, as was clearly the case in Sweden. Others saw the German settlers as over-powerful subjects who were harmful to the project of creating a state, on the same level as powerful feudal lords, as was the case in Norway. Since state-building was of paramount importance to na- tionalists, Swedish historians had a more positive attitude to the Hansa than their Norwegian colleagues. It was part of the nationalist way of thinking that when ethnic groups met, one national culture should dominate the other, and in our case German and Scandinavian historians held different opinions about which ethnic group dominated in particular towns. Post-nationalist historiography has been more open to processes of accommodation and integration. This was discussed in Nils Blomkvist’s 1979 history of Kalmar and in my own 1991 history of medieval Oslo. Political and administrative power in these and other towns has to be seen as the result of interaction between the king’s officials, who often resided in a fortified castle, and the town council, which could be dominated by Germans. Economic power was a compromise between local landowners and tax collec- tors, on the one hand, and Hansa merchants on the other. Social integration between ethnic groups has to be examined from the background of economic, political and other factors in each town (Blomkvist 1979, pp. 228, 250; Nedkvitne & Norseng 1991, chapters 8 and 16). Hansa merchants started to settle in Scandinavian towns after about 1250, and they grew numerous after about 1370. In the following discussion it is important to distinguish between Hanseatic ‘summer guests’ who only visited for a few weeks or months, ‘winter residents’ or liggere who were citizens of a Hansa town but settled in a Scandinavian town for a limited number of years, and German merchants and craftsmen who settled permanently in Scandinavia with their families. The question of integration only arose for the two last groups. Integration practices largely depended on state legislation, which was different in Denmark, Sweden and Norway.

Denmark The Danish king had German-speaking subjects who did not come from Hansa towns. From c. 1200, nobles, townsmen and peasants emigrated from the Ger- man County of Holstein to the Danish province of Schleswig/Sønderjylland, where the inhabitants were both German and Danish (Winge 1991, p. 89; Winge 1992, pp. 33–35). At the beginning of the 14th century most of Denmark was pledged to the Count of Holstein. During this period German-speaking nobles won prominent positions in the state administration, at court and among A post-national perspective on the German Hansa in Scandinavia 27 merchants and craftsmen (Poulsen 2008, pp. 130, 134; cf. Braunmüller 1993, p. 16). In 1460 the inhabitants of Holstein became the Danish king’s subjects and were free to settle wherever they wanted in his kingdom. After 1699, 20% of the inhabitants of Copenhagen had German as their mother tongue, and 20% of the king’s subjects were Germans, mainly from Schleswig-Holstein (Winge 1991, pp. 91, 108). Many medieval Danish kings had Low German as their mother tongue, and Christian IV, whose reign began in 1588, was the first king to be able to write Danish correctly (KLNM Nedertysk sprog). Hansa merchants were integrated into Danish towns more easily because of the numerous Germans from Schleswig-Holstein who were there already. In the 14th century formal citizenship was introduced in Danish towns, and immigrants from Hansa towns received citizenship and were elected to the town council (KLNM Borgare). Hansa merchants were also admitted as winter residents against the payment of certain taxes or charges (Knudsen & Kock 1992, p. 378; Tomner 1971, pp. 228–230; KLNM Vinterliggare). The Germans were free to create their own guilds or join the guilds of Danish merchants. As the market in Scania declined after 1370, Hansa merchants settled in Copenhagen, Malmö and other towns in Scania and Zealand as citizens or winter residents. In 1382 they founded a German guild (Tyske Kompagni) in Copenhagen (Kjersgaard 1980, pp. 111, 145). The oldest guild in Malmö was founded 1329 and was for Germans only: if a member married a Danish woman, he was expelled. Its aldermen assisted Germans in conflicts with Danish authorities. In 1475 the king abolished all German guilds in Den- mark (Tomner 1971, p. 230; cf. Christensen 1957, p. 90). In Jutland there were fewer Germans, and they found it to their advantage to join Danish guilds. The members of the Corpus Christi guild in Aalborg were merchants and state officials from both ethnic groups; about half of them had names which indicate a German origin (Knudsen & Koch 1992, p. 378; Christensen 1957, p. 90). In Odense, a German guild was founded in 1435, but a few years later it was merged with the Danish merchant guild (Christiensen 1957, pp. 90–91). Social intercourse and feasting in combined guilds must have strengthened integration. There were tensions between Hansa merchants and Danes, but they were not due to integration problems. Denmark had no general urban body of laws with special sections about foreigners. The Hansa had to negotiate privileges which could apply to only one town or market or to all Denmark, and these had to be renewed with each new king (Jahnke 2000, pp. 81–87; KLNM Stadsrett. Dan- mark). Another source of conflict was the repeated attempts by Lübeck and its allies to intervene in Danish civil wars, and in wars between Denmark and Sweden, in the hope of obtaining better privileges. There is no evidence that Hansa merchants who became Danish citizens or stayed in Danish towns for a limited number of years had problems integrating or met opposition from the authorities. 28 Arnved Nedkvitne

Sweden In Sweden, Kalmar seems to have been the first town to receive Hansa settlers, starting just after 1200. The newcomers may have applied Lübeck’s legal system in conflicts between Germans, since this was what they had done in other towns along the Baltic (Blomkvist 1979, pp. 172, 174, 258; Gustafsson 2006, p. 21). Around the year 1250, the Swedish regent Birger Jarl founded Stock- holm. He saw it as important to keep immigrants from Lübeck and other Hansa towns under Stockholm’s urban law. Shortly afterwards he ordained that Lübeck’s ‘summer guests’ were to be free from taxes and customs duties, but if they wanted to remain in Sweden and live there, the king ordered his officials to treat them according to Swedish law, and they were eligible for the ruling town council, which also was the law court of the town (Dahlbäck 1988, p. 20; Brandt 1953, p. 210). The Swedish state used its military ascendancy to retain control of urban jurisdiction and at the same time clear the way for immigration and a smooth integration of Hansa merchants. Like other princes along Baltic coast in the 13th century, Swedish kings wanted strong urban communities on their territory and achieved this by making it attractive for German merchants to settle there. The policy was a success, and the Germans became the core of urban merchant communities in Swedish towns (Dahlbäck 1988, pp. 17, 53; Gustafsson 2006, pp. 29–30). In 1350 the state issued a body of urban laws which were valid for all Swe- dish towns. Several paragraphs legislated on the rights of German ‘summer guests’, ‘winter residents’ and Germans who had become citizens in Swedish towns.3 Swedish citizens with German origins were given half of the seats on the town councils; the king could do this without endangering his control of the town because both Stockholm and Kalmar had powerful royal castles. The law established the rights of Hansa merchants without a time limit, and saved the urban community from future conflicts about Hansa privileges. Medieval Stockholm had about 20 guilds, and two of them were for Germans only, the guilds of St. Gertrud and Our Lady. Both were religious and social, but they played no political or economic role (Dahlbäck 1988, p. 169). Lena Moberg claims that Swedes and Germans could speak their own language and be understood by the other group, like modern Swedes and Norwe- gians. Middle Low German and Middle Swedish were more similar than today’s German and Swedish. Many people were also bilingual, and they could write as well as speak the other language. The Swedish language went through fundamental changes under German influence, which suggests that communication between the two groups must have been extensive (Moberg 2004, pp. 417–431). There were sources of conflict between the two ethnic groups. The Swedish

3 Magnus Erikssons stadslag, Konungsbalken, chapter XIX, and Köpmålabalken, chapters XXX– XXXIV; Birger Jarl’s first privilege granted for Lübeck included ‘winter residents’ and ‘summer guests’ in one category, called ‘guests’. A post-national perspective on the German Hansa in Scandinavia 29 state must have seen it as a problem that Swedish citizens with German family origins often returned to Lübeck with their capital. This may have delayed the formation of a group of native patricians who could have helped in the crown’s administration of urban communities. From the 1430s there were repeated in- surrections in Sweden against Nordic union. This was primarily directed against Danish dominance, but also against the Danish royal families with Ger- man origins (Pommerania, Bavaria, Oldenburg). The conflict did not have its origin in urban communities but had consequences for Germans living there (Blomkvist 1979, p. 253). In the years between 1350 and 1470, an increasing number of merchant families with German origins came to consider themselves as Swedish, and at the end of the period Stockholm and other towns had a strong native merchant class. They came to resent the paragraph in the town law which said that half of the members of town councils should be ‘Germans’ if possible (Magnus Erikssons stadslag, Konungsbalken, chapter II). These negative attitudes to Germans and Danes merged, and in 1471 the town law was changed and only ‘Swedes’ could be elected to the councils. The law’s definition of a ‘German’ and a ‘Swede’ is not clear. Swedish historians have considered it likely that a German family who had lived for two or three generations in Sweden was considered Swedish, which means that a citizen whose father was born in Sweden was classified as Swedish (Dahlbäck 1988, pp. 17, 53; Gustafsson 2006, pp. 29–30). German immigrants had now outlived their usefulness and could be marginalised. But there are no signs that these political changes led to social tensions or a reversal of the integration of immigrants. German members of the town council in Stockholm were drawn into national politics even earlier, in 1389, when the German-born Swedish king Albrecht of Mecklenburg was defeated by his Danish-born challenger Margareta. Ger- man urban councillors of Stockholm then organised the arrest and execution of many Swedish citizens in order to help Albrecht. The actual circumstances are only vaguely described in the sources (Lamberg 2000, pp. 64–65), but the origin of the conflict was not internal problems between ethnic groups in Stock- holm. Swedish historians in recent decades have described the integration of the Germans as unproblematic, or at least they have found no indications to the contrary (Dahlbäck 1988, pp. 52–53, 80; 2002, pp. 32–33; Svanberg 2004, pp. 381, 391; Gustafsson 2006, pp. 28–31; Blomkvist 1979, p. 258).

Norway Hansa merchants in the two East Norwegian towns of Oslo and Tønsberg had a local organisation under the protection of the Hansa diets to defend their interests. In 1378 a Hansa diet gave the two local aldermen authority to function as a court of justice for all Hansa merchants who were present in the towns. From 1420 at the latest, the town council in Rostock controlled the settlements and was their court of appeal, since most of the merchants were Rostock citi- 30 Arnved Nedkvitne zens. The Rostock council passed statutes for how Hansa trade in the two towns should be conducted (Norges Gamle Love 1912, document number 403) and regulated “merchants who lie (licht) in Oslo and Tønsberg making use of the privileges of the Hansa towns”. ‘Lie’ clearly includes both ‘winter residents’ and ‘summer guests’. Numerous citizens and town councillors of the two towns had German names in the Late Middle Ages. Some definitely came from Hansa towns, while others may have come from Schleswig-Holstein. The obstacles to Hansa merchants becoming Norwegian citizens were put in place by the Hansa. Merchants born in a Hansa town and who later became citizens of Oslo or Tønsberg initially were allowed to make use of Hansa privileges abroad in towns like Bruges in Flanders, but from about 1400 the Hansa no longer accepted this and expelled them from the Hansa. Hansa merchants who, despite this, wanted to immigrate to Oslo and Tønsberg were welcomed by local authorities, just as they were in Sweden and Denmark; they did not experience problems integrating there (Nedkvitne & Norseng 1991, pp. 362–366). In Bergen, however, no integration took place during the medieval period, which is exceptional in a Scandinavian context. How should this be explained? In the period 1247–1299, Norwegian authorities gave foreigners who owned or hired a house for at least one year in a Norwegian town the same legal status as Norwegians. Foreigners who hired a house for 6–12 months had some taxes and charges waived, but these ‘winter residents’ seem to have had the same economic rights as Norwegians. ‘Summer guests’ were exempted from more taxes and charges but had to pay customs duties (Nedkvitne 1983, pp. 198– 205). During this period, the Norwegian authorities conducted a policy of integration which was similar to that in Sweden and Denmark. In the following period, 1299–1380, Norwegian authorities narrowed the conditions for obtaining a legal status as a Norwegian, so that it applied only to a person who was born in Norway, or was married to a Norwegian woman, or had fled from his country of origin with his whole family (Norges Gamle Love 1849, document number 70). No easy access to citizenship in Bergen was offered. During the reign of King Håkon V (1299–1319), foreigners experienced severe restrictions; they could only conduct wholesale trade in towns, and not with other foreigners. In periods winter residence was made illegal. After 1319 Norway became part of a Nordic union, and the king reduced his military and administrative presence in Bergen. In 1366 the Hansa merchants organised themselves in the Bergen Kontor, which meant that their military, judicial and administrative presence was strengthened (Nedkvitne 1983, pp. 205–230). The 1000 winter residents came to feel that their success depended on a strong Kon- tor and not on favourable privileges from the state. It goes without saying that this situation was not advantageous to the integration of Hansa merchants. The Kontor prohibited its members from visiting the same guilds as Norwegians, and only two guilds exclusively for Germans were acceptable. A post-national perspective on the German Hansa in Scandinavia 31

Post-nationalist historians have wanted to see the Hansa as promoters of European integration, bridging ethnic differences. The main promoter of this view has been a French historian from Strasbourg, Philippe Dollinger, in his 1964 survey La Hanse. The situation in Bergen during the period 1366–1559 has been difficult to fit into this general picture. Dollinger included an incident in 1455 when the Hansa merchants in the Kontor killed the king’s bailiff, the Bishop of Bergen and 60 others, and burnt down the monastery where they had sought refuge. But he is at pains to explain that the episode was an exception and should not influence our general understanding of the Hansa (Dollinger 1989, pp. 383–384). Knut Helle, in his history of Bergen in the Middle Ages, tries to argue for the same general view, but has problems doing so. On several occasions he em- phasises that he wants to distance himself from the traditional negative attitude to the Hansa found in Norwegian historiography (Helle 1982, pp. 762, 768, 770–773). But he also relates several incidents where Hansa merchants during the period 1350–1370 committed acts of violence without answering for them in court. In the following paragraph he excuses the Hansa leadership by explaining that Hansa diets were unable to control the Kontor aldermen, who in turn did not control the violence of individual Kontor members. Despite this, the Kontor’s leadership “sometimes felt strong enough to promote its members’ interests through arrogance and violence” (Helle 1982, p. 734). He also recognises that the Kontor had a policy of promoting segregation between Ger- mans and Norwegians (Helle 1982, pp. 761–762). Facts presented in this un- systematic way nullify each other; what is lacking is an analysis of what the Hansa’s use of violence and disregard for Norwegian jurisdiction meant for their integration into Bergen society. Is the best supported conclusion that Bergen was an exception to the general picture given by Dollinger? In a collection of articles entitled ‘Die Hanse – Lebenswirklichkeit und Mythos’, which has been reprinted several times, Carsten Müller-Boysen from Kiel wrote about the Kontor in Bergen. His empirical material is taken from Helle’s book, but he makes no effort to present the Hansa in Bergen as civilised and integrated. The manners of this masculine society were rough and rude, he sees their rituals and games (spill) as the expression of a violent mentality. They exploited their power “without pity” to promote economic as well as judicial interests (Müller-Boysen 2006, pp. 230–232). Perhaps Müller-Boysen can distance himself more easily from the general understanding of the Hansa on this point because he is German? Marko Lamberg compared the town councils of Bergen, Stockholm and Malmö (which was Danish until 1658). In all three towns, the council spoke for the local citizens, but its voice was weaker in Bergen because of the Hansa presence. The Kontor had powers similar to those of a state, and only the commander of Bergenhus castle was able to oppose it, with mixed results (Lamberg 2000, pp. 59–60). Lamberg sees the Hansa merchants’ attack on competing merchants on the night of November 8–9, 1523 as partly directed against the 32 Arnved Nedkvitne town council – two of the victims were councillors, and the attackers broke into the town hall and stole the chest which held documents setting out the privileges given to the town. The town council could not negotiate with the Kontor aldermen on its own; they depended on support from the state and from common townsmen. Bergen councillors never became an urban elite with power and prestige independent from that of state and commoners (Lamberg 2000, pp. 236–237, 246, 249). It was not tempting for individual Hansa merchants to leave the Kontor and become citizens of Bergen (Lamberg 2000, p. 246). Dag Lindström studied the craft guilds in Stockholm, Malmö and Bergen for the period 1350–1622. Craft guilds with statutes grew common in Denmark and Bergen around 1350, in Sweden somewhat later, about 1450. Until about 1550 these guilds enjoyed a certain amount of independence, but they were all controlled by an external power; in Sweden and Denmark this was the town council, which was in turn supervised by the state. Bergen was special because all the craftsmen were Germans without Bergen citizenship, and external control was exercised by the Kontor, which in turn was supervised by Lübeck (Lindström 1991, pp. 254–262). Lamberg states several times that the Kontor held powers normally exercised by the state, and this influenced the character of Bergen society (Lindström 1991, pp. 256–257). Müller-Boysen, Lamberg, Lindström and myself all agree that Bergen during the Hansa period, 1366–1559, was different compared to the way the Hansa normally functioned in other Scandinavian and North European urban communities. Helle did not hold this view because he did not compare Bergen to other towns visited by the Hansa. The Kontor exercised violence, issued statutes and organised courts of law just like the king’s representatives could do, and they took political action and administered craft guilds like the urban councils did. The Bergen Kontor merchants were not integrated into urban society and its institutions, and this created severe problems for Norwegian officials, townsmen and other foreigners in the town. But Bergen was an exception. In Swedish and Danish towns, and Oslo and Tønsberg in Eastern Norway, integration went smoothly. Hansa merchants were close to Scandinavians in language and culture, and the authorities welcomed them because they were seen as representing a more advanced West European commercial organisation.

Hansa or state as transmitters of Low German in Norway? Of all the cultural impulses which Germans transmitted to Scandinavia during this period, one of the most important was the Low German language. The present publication is part of a linguistic project which has as its aim to analyse how the three Scandinavian languages changed under the influence of Mittel- niederdeutsch. A post-national perspective on the German Hansa in Scandinavia 33

It seems to be a premise in these projects that the basic changes in all the Scandinavian languages were due to oral communication with speakers of Mit- telniederdeutsch. In Denmark the arguments for this are convincing, but for Norway an alternative hypothesis exists: the linguistic changes came through written Danish. The Norwegian linguist Arnold Dalen points out that it is extremely difficult to distinguish between the consequences of oral communication between German Hansa merchants and Norwegian townsmen, and the result of the situation after about 1400 whereby those who received an education in reading and writing in Norway increasingly were taught Danish (Dalen 1994, pp. 31–32). Gro-Renée Rambø also mentions this alternative hypothesis, referring to Dalen (Rambø 2008, pp. 361–363). But neither of them analyses the problem, nor do they discuss which empirical evidence supports or weak- ens each hypothesis or draw conclusions as to their relevance. I can only present the historical aspects of such an analysis and leave the linguistic side of it to others. First, let us look at the influence of the oral language used by Hansa merchants in Bergen, Oslo and Tønsberg. In the period between the founding of the Bergen Kontor in 1366 and 1559, the aldermen enforced statutes which were aimed at economic and social segregation. Hansa merchants were forbidden to visit Bierstuben (ale houses) outside the “Bryggen” quarter of the town (“the wharf”) where most Germans lived. They were not permitted to stay outside Bryggen after 9 pm in winter and 10 pm in the summer, or be a member of the same guild as Norwegians. They were obliged to attend social occasions at Bryggen. They were forbidden to have close relations with honourable Norwe- gian women, but nothing was said about prostitutes. There was a group of personal mistresses who fell into a grey area between the two, but relations with individual women is not necessarily evidence for generally close social relations between Germans and Norwegians. There is no evidence of extensive social contacts between the two ethnic groups in the period 1366–1559 (Ned- kvitne 2011, pp. 149–157). In the years after the Reformation, the Kontor gradually was forced by a stronger state to abandon its policy of economic segregation, and we must assume that they also loosened their policy of social segregation. In the 1560s a priest in Bergen called Absalon Pederssøn Beyer wrote a diary where, among other things, he described in detail relations between Germans and Norwe- gians. At that time Norwegians often seem to have visited the wine cellars and Bierstuben at Bryggen, probably because the beer was considered to be better there.4 It was more rare that Hansa merchants visited Norwegian Bierstuben. The aldermen and eighteen other members of the Kontor leadership were invited to banquets or parties by the Danish commander of Bergenhus castle and vice versa. But Absalon’s diary does not give an impression of extensive social ties and communication between Germans and Norwegians.

4 For example, in Absalons Pederssøn’s diary dated 19th April 1566, pp. 114–115, and 23rd Janu- ary 1567, pp. 126–127. 34 Arnved Nedkvitne

The evidence indicates that oral communication between the two ethnic groups before about 1560 mainly concerned business. I have difficulty understanding how influence from spoken Mittelniederdeutsh could change the structure of the language spoken in Bergen; it is more reasonable to think that it may have influenced parts of the vocabulary. Oral communication between Germans and Norwegians in Bergen was most limited before the Reformation, but that was the period when changes in the Norwegian written language were greatest. In Oslo and Tønsberg there were fewer Hansa merchants, but they were more integrated into urban society; several of them became Norwegian citizens and even members of the town council. Closer social interaction in Oslo and Tønsberg promoted linguistic influence, and what limited this was the fact that the Hansa merchants were few in number there (Nedkvitne & Norseng 1991, pp. 365–366). It is even more difficult to reconcile with the social realities the idea that important linguistic changes should find their way from the 6000 Norwegian citizens of Bergen to the 150,000–200,000 inhabitants of the rest of Norway. Agnete Nesse has shown that the influence from Hansa merchants’ oral language created dialect peculiarities in Bergen. This is perfectly reconcilable with the social realities. But she has not claimed that spoken communication in Bergen was the source of structural or other changes in the language spoken throughout Norway (Nesse 2002, pp. 95–100). From a historian’s point of view, it is natural to see the larger changes as caused by Danish influence through administrative writing and schooling, to a smaller degree also orally through Danish state and church officials residing in Norway. The strongest influence from Low German on Norwegian took place in the Late Middle Ages, a period when almost all written communication in Norway was for administrative or other pragmatic purposes (Nedkvitne 2004, chapter 3). The Danish and Danish/German kings of the Nordic union increasingly put Danes, Germans and Swedes into Norwegian church and state offices. In 1370 a Dane was made Bishop of Bergen for the first time (Helle 1982, p. 879), and in 1407 the first Danish Bishop of Oslo was appointed (Bull 1922, p. 284). For- eigners became parish priests; in 1432 a German Dominican was parish priest on Røst in Lofoten (Querini 1583, fol. 204r). Secular power was increasingly concentrated in the two royal castles of Akershus and Bergenhus. The first foreign commander of Bergenhus was a Swede, appointed in 1453 (Helle 1982, p. 827). The first known foreign commander of Akershus was a German, who took office in 1424 (Nedkvitne & Norseng 1991, p. 413), and during 40 of the following 50 years the commanders of Akershus – the most important castle in Norway (Nedkvitne & Norseng 1991, p. 415) – were foreigners. Erik of Pomer- ania was the Norwegian king from 1389–1439; from this period we know the names of 26 Norwegian and 16 foreign state officials in Norway (Taranger 1915, p. 278). The officials were Norwegians, Danes and Germans mainly A post-national perspective on the German Hansa in Scandinavia 35 from Schleswig-Holstein, but it was pragmatic to have one administrative language, and that was Danish. After c. 1400 Danish must have replaced Norwe- gian as the language taught and probably spoken in the schools. An emphasis on state and elite power gives a different picture of the social mechanisms at work. Language change is not only determined by voluntary oral imitation of prestigious models, but also by power relations.

References

Absalon Pederssøns dagbok 1552–1572. Ed. Ragnvald Iversen. 1963. Bergen. Blomkvist, Nils, 1979: Kalmars uppkomst och äldsta tid. In: Kalmars stads historia volume 1, Ed. Ingrid Hammarström. Kalmar. Pp. 167–308. Brandt, Ahasverus von, 1953: De äldsta urkunderna rörande tysk-svenska förbindelser. In: Historisk tidskrift (Sweden). Pp. 209–230. Braunmüller, Kurt, 1993: Entwicklung des niederdeutsch-skandinavischen Sprachkon- takts. In: Niederdeutsch und die skandinavischen Sprachen. Heidelberg. Pp. 9–40. Braunmüller, Kurt, 1998: Sprogkontakt i hansetiden – en sammenfattende oversigt over Hamborg-projektet. In: Språkkontakt i Norden i middelalderen, særlig i hansatida. Copenhagen. Pp. 17–31. Brück, Thomas, 1999: Korporationen der Bergenfahrer in den wendischen Hansestäd- ten. In: Genossenschaftliche Strukturen in der Hanse. Köln. Pp. 123–133. Bull, Edvard, 1922: Oslos historie. Kristianias historie volume 1. Oslo. Burkhardt, Mike, 2009: Der hansische Bergenhandel im Spätmittelalter. Handel – Kauf- leute – Netzwerke. Köln. (Quellen und Darstellungen zur Hansischen Geschichte. Neue Folge LX.) Cordes, Albrecht, 1999: Die Anfänge des Gesellschaftshandel im Hanseraum bis zur Mitte des 14. Jahrhunderts. In: Genossenschaftliche Strukturen in der Hanse. Ed. Nils Jörn et al. Köln. Pp. 65–78. Christensen, Aksel E., 1957: Danmark. In: Hansestæderne og Norden. Århus. (Det nordiske historikermøde i Århus 7–9 august 1957.) Pp. 55–96. Dahlbäck, Göran, 1988: I medeltidens Stockholm. Stockholm. (Monografier utgivna av Stockholms stad no. 81.) Dahlbäck, Göran, 1998: Invandring – särskilt tysk – till Sverige under medeltiden. In: Invandrarna och lokalsamhället. Historiska aspekter på integrationen av invandrare i nordiska lokalsamhällen. Stockholm. Pp. 11–30. Dahlbäck, Göran, 2002: Stockholm blir stad 1250–1560. In: Staden på vattnet, volume 1. Stockholm. Pp. 17–72. Dalen, Arnold, 1994: The influence of Low German on the Norwegian language. In: Norwegen und die Hanse. Frankfurt am Main. Pp. 31–40. Dollinger, Philippe, 1989: Die Hanse. Stuttgart. (Originally La Hanse. 1964.) Gustafsson, Sofia, 2006: Svenska städer i medeltidens Europa. Stockholm. (Stockholm studies in history 86.) Helle, Knut, 1982: Bergen bys historie 1. Bergen. Hyldgaard-Jensen, Karl, 1993: Unterschiede in der Aufnahme mittelniederdeutscher Rechtstermini in die altdänische und altschwedische Rechtsprache. In: Nieder- deutsch in Skandinavien IV. Berlin. Pp. 102–117. Jahnke, Carsten, 2000: Das Silber des Meeres. Fang und Vertrieb von Ostseehering zwischen Norwegen und Italien (12–16 Jahrhundert). Köln. 36 Arnved Nedkvitne

Jahr, Ernst Håkon, 2008: Historical sociolinguistics: the role of Low German language contact in the Scandinavian typological shift of the Late Middle Ages. In: Språk- historie og språkkontakt. Festskrift til Ernst Håkon Jahr på 60-årsdagen. Oslo. Pp. 246–257. Kjersgaard, Erik, 1980: Byen og borgen Havn. Københavns historie 1. Copenhagen. KLNM = Kulturhistorisk leksikon for nordisk middelalder, vols. 1–21. 1956–1976. Co- penhagen. Knudsen, Bodil Møller & Kock, Jan, 1992: Fra Aalborgs fødsel til Grevens Fejde 1534. In: Aalborgs historie 1. Ed. Erik Johansen et al. Aalborg. Lamberg, Marko, 2000: Dannemän i stadens råd. Rådmanskretsen i nordiska köpstäder under senmedeltiden. Stockholm. Lindström, Dag, 1991: Skrå, stad och stat. Stockholm, Malmö och Bergen ca 1350– 1622. Uppsala. Magnus Erikssons stadslag i nusvensk tolkning. 1966. Translation by Åke Holmbäck and Elias Wessén. Stockholm. Moberg, Lena, 2004: Det tyska inflytandet på svenskan. In: Signums svenska kulturhis- toria. Medeltiden. Lund. Pp. 417–431. Müller-Boysen, Carsten, 2006: Die ‚Deutsche Brücke’ in Bergen und die Niederlas- sungen in Tønsberg und Oslo. In: Die Hanse – Lebenswirklichkeit und Mythos. Lübeck. Pp. 223–233. Nedkvitne, Arnved, 1983: Utenrikshandelen fra det vestafjelske Norge 1100–1600. (Unpublished PhD in history, University of Bergen.) Nedkvitne, Arnved & Norseng, Per G., 1991: Byen under Eikaberg. Oslo bys historie 1. Oslo. (Reprint with new title: Middelalderbyen ved Bjørvika. Oslo 2000.) Nedkvitne, Arnved, 2004: The social consequences of literacy in medieval Scandinavia. Turnhout. Nedkvitne, Arnved, 2011: Ære, lov og religion i Norge gjennom tusen år. Oslo. Nesse, Agnete, 2002: Språkkontakt mellom norsk og tysk i hansatidens Bergen. Oslo. Norges Gamle Love, 1849: First series volume 3. Christiania. Norges Gamle Love, 1912: Second series, volume 1. Kristiania. Norges Gamle Love, 1934: Second series, volume 2. Oslo. Poulsen, Bjørn, 2008: Hertugdømmets dannelse 700–1544. In: Sønderjyllands historie 1. Aabenraa. Pp. 41–186. Querini, Pietro, 1583: Viaggio del magnifico messer Piero Quirino, gentilhuomo vini- tiano. In: Delle navigationi et viaggi 2. Ed. Giovanni Battista Ramusio. Venezia. Fol. 199d–206r. Rambø, Gro-Renée, 2008: Historiske og sosiale betingelser for språkkontakt mellom nedertysk og skandinavisk i seinmiddelalderen. Et bidrag til historisk språksosiolo- gi. (PhD, University of Agder.) Kristiansand. Rörig, Fritz, 1928: Der Markt von Lübeck. Topographisch-statistische Untersuchungen zur deutschen Sozial- und Wirtschaftsgeschichte. Reprinted in: Wirtschaftskräfte im Mittelalter. 1971. Köln. Pp. 36–133. Rörig, Fritz, 1940: Reichssymbolik auf Gotland. Reprinted in: Wirtschaftskräfte im Mittelalter. 1971. Köln. Pp. 490–539. Schück, Adolf, 1926: Studier rörande det svenska stadsväsendets uppkomst och äldsta utveckling. Stockholm. Svanberg, Ingvar, 2004: Främlingar i medeltidens Sverige. In: Signums svenska kul- turhistoria. Medeltiden. Lund. Pp. 379–399. Taranger, Absalon, 1915: Norges historie 3.1. Kristiania/Oslo. Theuerkauf, Gerhard, 2006: Recht, Rechtsaufzeichnung, Gerichtsbarkeit. In: Die Hanse – Lebenswirklichkeit und Mythos. Lübeck. Pp. 497–502. A post-national perspective on the German Hansa in Scandinavia 37

Tomner, Lennart, 1971: Malmö stads historia före 1500. In: Malmö stads historia 1. Malmö. Pp. 173–252. Wernicke, Horst, 1999: Die Fahrtrichtungsgenossenschaften in den Hansestädten. In: Genossenschaftliche Strukturen in der Hanse. Köln. Pp. 135–163. Winge, Vibeke, 1991: Dansk og tysk i 1700-tallet. In: Dansk identitetshistorie 1. Co- penhagen. Pp. 89–110. Winge, Vibeke, 1992: Dänische Deutsche – deutsche Dänen. Geschichte der deutschen Sprache in Dänemark 1300–1800. Heidelberg. Wubs-Mrozewicz, Justyna, 2005: Traders, ties and tensions. The interaction of Lübeck- ers, Overijsslers and Hollanders in Late Medieval Bergen. Hilversum. 38 Arnved Nedkvitne Language contact, communication and change 39 Language contact, communication and change

Gro-Renée Rambø, University of Agder

Introduction Modern Scandinavian language varieties clearly demonstrate effects of contact with and influence from other languages. Today, English is the most influential source of language interference on the Scandinavian languages, but historically Middle Low German is still reckoned to have been the most influential source of borrowing. In addition to numerous word loans, the Scandinavian languages also took different kinds of grammatical structures (such as word-formation categories) from Low German during their intense and lengthy language contact period. Because of this, there is a long tradition within Scandinavian linguistics of studying language contact between Middle Low German and the Scandinavian languages in the Late Middle Ages, mainly the 14th and 15th centuries, when linguistic interference from Middle Low German was at its strongest. This was first and foremost due to the impact of the Hanseatic League in northern Europe at the time, but also a result of more general cultural influence from the North Germanic area. In this article I will discuss some specific aspects of this language contact period. First, I will present some important information about the history and background of research on language contact between Middle Low German and the Scandinavian languages, discuss different focus areas within language contact studies, and then place my own research within the theoretical distinctions I draw. Then I will present and discuss my research design, as well as some findings and insights from my doctoral thesis (Rambø 2009). This will include examples of how we can gain important knowledge and insights about historical language contact periods by applying the results of research carried out within the frameworks of modern language contact theory and sociolinguistics/ sociology of language to this traditional field of study. Finally, I will further investigate some problems and issues which arose while working on my PhD, focusing more specifically on the language users. I will then discuss what I find to be important problems and considerations for further research on this period of language contact. 40 Gro-Renée Rambø

History and background Scandinavia and the Hanseatic League There has been a very long tradition of trade between Scandinavia and the low- lands in the northern parts of Germany – in fact, we can trace this back to pre- historic times. The most important economic and cultural contacts developed about 1200–1550, starting with minor trading contacts on Gotland (Visby) from 1164 onward, and ending up with a vast area of economic influence – the Hanseatic League. The Hanseatic League was an alliance of trading cities and their guilds that established and maintained a powerful trade organization along the coast of northern Europe, from the Baltic to the North Sea and inland, during the Late Middle Ages and early modern period (c. 13th–17th centuries). The Hanseatic cities had their own legal system and provided their own protection and mutual aid. At its height, the league included some 70 cities and 100 to 130 smaller towns located in an area that is spread over several European countries today – from the Zuidersee (Netherlands) in the west to Estonia in the east, and from Visby (Sweden) in the north to a line running through Cologne-Erfurt-Breslau- Krakow in the south. The Hanseatic merchants supplied western and central Europe with luxury goods, food and raw materials from northern and eastern Europe – for example furs, wax, cereal, and fish, as well as flax, hemp, wood, and construction products such as pitch, tar and potash. In return, they imported finished products such as cloth, metal goods (especially weapons), and spices from the west and south to the north and east. The Hansa must be considered an important mediating factor for different kinds of cultural contacts between the various regions and countries in the vast trade area that they controlled. The effects of contact with their language, Middle Low German, gives witness to such cultural contacts. Middle Low German was the lingua franca used in different trading situations, as well as other communication situations, throughout the trading area. There were intimate and long-lasting contacts between North German traders and craftsmen and the Scandinavians living in urban centers (in particular) due to the social, economic and cultural position of the Germans at the time. As Dalen (1994) has pointed out, during any historical period, some nation takes a leading role, and its culture and language are imitated by other nations. Thus, France had a dominant position in northern Europe in the 17th and 18th centuries, and French culture and the French language were then fashionable all over Europe. The culture and the language of the Hansa were in a similar position in northern Europe during the Late Middle Ages. Due to this, the economic and cultural domination of northern Germany during the Hanseatic period is an important explanatory factor when exploring the contact situations and effects of this contact. Language contact, communication and change 41

Historical and social preconditions It is quite common within linguistics to differentiate between internal and external factors when describing and explaining linguistic changes. Whereas research dealing with internal factors has its main focus on the linguistic structures involved, research dealing with external factors focuses on issues such as historical elements (for instance power relations, laws, inventions), the social environment, cultural conditions, and considerations linked to the language users themselves, including psychological factors, in order to explore how such factors affect language use and linguistic choices. My PhD thesis was an attempt to describe and discuss specific aspects of this language contact period which have not received significant attention in earlier research – the historical and social preconditions within which language contact took place. Even though it is absolutely clear that more work has to be done at the intra-linguistic level, I have chosen to focus on specific kinds of external linguistic factors connected with language contact, excluding internal factors (although of course I do not dismiss such factors as unimportant). I have not tried to analyse specific results of linguistic contact, but have instead concentrated on the situations which defined the frames within which contact leading to linguistic changes took place. In my thesis, the focus is on language contact between speakers of different languages and varieties, but I concentrated on the group level, not the individual. I tried to show how external factors governing different groups of speakers’ possibilities and restrictions for language use choices have an impact on how both the specific contact situations and the results of language contact may have varied across Scandinavian countries and cities in the Late Middle Ages. In doing this, I have drawn on insights from modern language contact theories, as well as theories within the sociology of language and sociolinguistics. As Jahr (1997) has pointed out, a renewal of interest in contact linguistics can be seen in several of the more recent contributions to the field (see for instance Ureland 1987; contributions in the four volumes of Niederdeutsch in Skandinavien, edited by Schöndorf & Westergaard 1987; Hyldgaard-Jensen et al. 1989; Elmevik & Schöndorf 1992; Menke & Schöndorf 1992; Moberg 1989; Winge 1992; Braunmüller & Diercks 1993; Braunmüller 1995; Jahr 1995, 1997, 2000). This newer research is innovative in focusing specifically on the social aspects involved in language contact situations, as well as on multilingualism, which raises questions linked to what language competence involves. Both directly and indirectly, these studies demonstrate clearly that contact linguistics is a highly multidisciplinary field. Post-war research in language contact and sociolinguistics/sociology of language have found that a linguistic community is never fully homogeneous and self-contained. On the contrary, all languages are in some respects what we could call mixed languages. Language contact (short term or long term) is quite common. How we define language contact has importance for what questions we ask, and, as a consequence, for what answers we obtain. 42 Gro-Renée Rambø

Language contact – what is it? It is not languages that are in contact with each other, but rather the speakers of different languages and varieties. Milroy points out: What a theory of language change must account for, therefore, are the conditions in which speaker innovations are successful in penetrating the system, and this implies of course that we should also try to understand the conditions in which they are not successful. There appears to be no alternative but to suppose that the conditions involved must be social in a broad sense (including interactional and communicative aspects of speaker-behaviour), because candidates for change are sometimes accepted and sometimes rejected. Thus, a purely intra-linguistic account will usually fail to explain why a particular change took place at the time and place that it did and not at some other time and place when the intra-linguistic conditions were similar. Furthermore, it will not explain why potential changes that were always inherently possible did not take place for long periods of time during which the appropriate linguistic conditions existed. (Milroy 1997, p. 313f.) According to Milroy, then, an intra-linguistic focus will not be sufficient when studying language contact or linguistic change as a result of contact. Language contact studies deal with situations in which people using different linguistic varieties meet and communicate. Communicative possibilities are the crucial point. If there are several individuals present at a given time in a given place, and if these individuals use different linguistic varieties, no language contact takes place unless the individuals actually communicate with one another in some way or other (through writing or speaking) (see for instance Thomason 2001, p. 1f). This way of defining language contact calls for special attention to be directed at two domains. First, there are the individuals themselves. What can we find out about their linguistic competence(s), both in their society and at an individual level? Second, what can we discover about restrictions and possibilities governing the communication situations? What are the conditions for actual bilingual meetings to take place? The first question calls for a description and an understanding of the psychological, social and linguistic competence of the individual, or groups of individuals, involved in the contact situation. The second question calls for an understanding of all kinds of conditions defining the structures within which the contact takes place, such as social structures within community (or communities), historical frames, general cultural contacts, and so on.

Language contact and language change All natural languages change through time, and variation at different linguistic levels is the normal situation. When individuals using different linguistic varieties (languages, dialects) meet and communicate on a long term level, the conditions for language change is strong, if we consider linguistic varieties to be dynamic entities. What intra- and extra-linguistic conditions are present when Language contact, communication and change 43 linguistic changes occur as a result of contact and communication between people using different linguistic varieties, and what determines whether some changes are short term and others long term? What changes which occur in the actual communication situation cause long-lasting changes in a linguistic variety? We necessarily have to consult different kinds of sources to answer these kinds of questions. We can categorize the factors defining the linguistic outcomes of contact situations in different ways. One way of organizing the different factors is to divide them into the following three categories:

INTERNAL FACTORS 1) Linguistic – typological distance between languages – the degree of congruence between elements – markedness – other constraints and factors, e.g. the existence of gaps (first and foremost on the semantic level, but also on the morphological and phonological level); frequency (in the use of a specific linguistic item)

EXTERNAL FACTORS 2) Social, political and cultural – the length and intensity of the contact (degree of contact, societal bilingualism) – patterns of interaction; social networks – settlement structures; social meeting arenas – power and prestige relations; social mobility – demography (for instance the size of the different groups involved) – education – migration 3) Psychological and individual – attitudes and identity – motivation to use a particular linguistic variety – the individual’s strategies in second language acquisition – awareness of language norms (including informal norms) – other cognitive processes and mechanisms (some of these insights are from Killie 2007).

All of these factors are relevant and important, but we need to clarify how we conceive the relationships between the different factors, and how they influence each other. We take as a starting point what choices individuals have in a contact situation using a language other than their own. Erik Simensen (1994, p. 33) claims that, in principle, there are six different communication strategies which are possible in an oral language contact situation: each individual uses his or her own dialect or language, a regional or national standard, the variety of the interlocutor, a third language (a lingua franca), a “mixed” language, or an interpreter. 44 Gro-Renée Rambø

Which solution is chosen depends on different sets of conditions. Simensen claims the most important conditions to be: 1) the communication situation itself, 2) personal, social and cultural preconditions of the interlocutors (for instance, motivation and experience with other linguistic varieties), and 3) the relationship between the mother tongues of the interlocutors, that is, the distance between the varieties linguistically as well as geographically. Simensen claims that the last set of conditions, the distance between the varieties or languages, is fundamental: the more two languages or varieties have in common, the easier it will be for the individual to understand and learn the other variety (Simensen 1994, p. 34). Observable similarities between language varieties could be due to one or more of the following: 1) historical (or so-called genetic) kinship, 2) structural (or typological) similarities, and/or 3) loans (either by one language from the other, or by both from the same language(s)). However, one could also claim that linguistic factors are not the most important, depending on what the research question is, and depending on how we define linguistic and communicative competence. In their famous and often- cited book on language contact, Thomason & Kaufman state: The starting point for our theory of linguistic interference is this: it is the sociolinguistic history of the speakers, and not the structure of their language, that is the primary determinant of the linguistic outcome of language contact. Purely linguistic considerations are relevant but strictly secondary overall. (Thomason & Kaufman 1988.) This approach has been one of the most important guidelines for my own research. I have focused on language contact factors listed under category 2) above – social, political and cultural factors.

Historical and social preconditions for language contact between Middle Low German and Scandinavian in the Late Middle Ages Different places, different people, different varieties – different results? In my PhD thesis, I attempted to shed new light on the language contact which occurred between Middle Low German and Swedish, Danish and Norwegian in the late Middle Ages (14th and 15th centuries) by answering the central question: What can we find out about different sociolinguistic conditions which might have been determining factors for language contact in the most important cities in Scan- dinavia during the Hanseatic period, and what different contact conditions and results can be observed as consequences of the different sociolinguistic situations? (Rambø 2009, p. 9; my translation.) Language contact, communication and change 45

Contact between the Hanseatic League and the various countries and cities in Scandinavia differed in several ways. I have therefore tried to answer the above questions by looking at specific cities in Denmark, Sweden and Norway more closely with respect to a range of extra-linguistic factors. The cities studied and compared are Copenhagen and Ribe in Denmark, Stockholm and Kalmar in Sweden, Bergen and Oslo/Tønsberg1 in Norway, and Visby on Gotland (the first Scandinavian city in which the Hanseatic League achieved a central position). To enable us to say more about how language contact processes may have worked in these different Scandinavian cities during the Late Middle Ages, I have considered it necessary to take a highly intra- and interdisciplinary approach – insights are drawn from different fields within linguistics, as well as from general history, medieval archaeology, sociology, social anthropology, and interactional psychology. The most important theoretical frameworks used here are historical sociolinguistics (Romaine 1982; Jahr 1999), the sociology of language (Fishman 1968, 1976a, 1976b), modern language contact theory, and insights from interactional sociolinguistics, including accommodation theory (for instance, Giles & Smith 1979) and the Acts of Identity model (Le Page & Tabouret-Keller 1985). In addition, findings from social anthropology and sociology concerning identity and negotiating identity in a social context, and from general history concerning political conditions, economic structures, cultural contacts, settlement structures, demographical conditions, etc., have been used to the extent that they can say something of importance about the linguistic communities dealt with here. In my work I describe and discuss language contact between Middle Low German and the different Scandinavian languages during the Late Middle Ages. It is important to state that none of these languages were standardized in the modern sense of the term; this fact effects researchers’ assumptions and conclusions about possibilities for language/dialect contact, language mixing, code-switching, and bilingualism. It is also worth mentioning that Latin was an important lingua franca in northern Europe during the Late Middle Ages, and probably the most common one in Scandinavia (at least in certain environments and domains) during this period, indicating that multilingualism was probably quite common in various circles. This differs from the representations we find

1 Oslo and Tønsberg are usually considered together in literature concerning the presence and situation of the hanseats in the eastern part of Norway in the Late Middle Ages. Mainly this has to do with important similarities concerning different kinds of political and demographical matters at the time in these two cities, as well as obvious geographical matters. For instance, they were both relatively small cities at the time, the hanseats had a “faktorei” in both cities, and the settlement conditions were the same for the German hanseats in these cities, and the time period for hanseatic presence in the two cities were similar. Privileges given to the German hanseats were at several occasions directed towards the hanseats in Oslo and Tønsberg together. Historians usually point out the differences between the western parts of Norway on one hand, represented by Bergen, and the eastern parts of Norway, represented by the cities of Oslo and Tønsberg on the other, when it comes to the hansa period. Both Oslo and Tønsberg were part of the greater area referred to as the ‘Viken-area’. 46 Gro-Renée Rambø in several earlier works, which tended to directly or indirectly present Scandi- navian communities in the Middle Ages as monolingual. Braunmüller (2000) points out that, paradoxically, almost nobody has mentioned the very obvious point that large parts of this population during the Late Middle Ages must have been bi- or trilingual (in Middle Low German, Latin, and the Scandinavian language of their city) to be able to participate in urban daily life. This means that earlier research has, to a large extent, concentrated on making monolingual descriptions which are abstractions of the actual linguistic reality, a form of linguistic reductionism (Braunmüller 2000, p. 2). A realistic theory of linguistic competence must take as a starting point the abilities of the particular individual to communicate adequately in different situations (Braunmüller 2000, p. 2f). Therefore, if we want to say anything about the language contact processes in the Late Middle Ages, we cannot do so without considering the social, historical and cultural preconditions for achieving bi- or multilingualism at the individual level. Since language is indissolubly linked with the human mind, it is a social and psychological phenomenon. Languages exist and change because they are used by people. Language change comes about through actual communicative interactions which take place within a specific linguistic community, in a specific situation, involving specific people. The totality of the language use context therefore has to be taken into account when we, as linguists, try to explore and understand how and why language contact situations lead to language change at different linguistic levels. We cannot assume that the contact conditions were exactly the same throughout the whole Scandinavian area, and throughout the entire contact period of several hundred years. Therefore, we need to try to discover what differences and similarities there were in these various contexts and explore how important these differences and similarities may or may not have been for the contact situations, and in turn for the outcome of this contact.

Societal extra-linguistic factors – similarities and differences In my PhD thesis (Rambø 2009), each of the seven Scandinavian cities mentioned above were studied and compared to try to account for differences and similarities with regard to language contact and its consequences. The description for each city is divided into three subsections: language sociology and historical preconditions, conditions for language contact, and language use in the specific city under discussion. The first subsection describes the political and historical basis for the position of the Hanseatic traders in the city. This demarcates the timeframe for the Hanseatic period in the specific city, which may be at variance with what is usually considered to be the Han- seatic era for the country as a whole, or even for Scandinavia. This is an important point, since what we call the Hanseatic period – when contact with the German traders was at its most intense – varies between the different countries in Scandinavia, as well as between the different towns in each Language contact, communication and change 47 country. Language contact is seen as part of this general cultural, economic and political contact. I also consider demographic relations, like settlement structures and the constitution of population groups. These kinds of structure tell us about prestige and/or power relationships between different population groups, as well as what the conditions were for social interaction, including language use interaction, in different arenas. The second subsection takes a closer look at the linguistic communities, or societies, and specific language use situations. I make an effort to use the information provided in the first subsection to illuminate our understanding of the language society or societies in each city, as well as how the specific language use situations have come about and have worked in practice. The third subsection compares the information given in the other two in order to investigate more closely what we can know or assume about the actual language use in the specific city during the Late Middle Ages. This allowed me to make well-founded assumptions about the totality of the sociolinguistic conditions, and say more about contact between Middle Low German and Danish, Swedish or Norwegian during the Hanseatic period. I also compared the findings for each of the cities. I will illustrate this below with a few examples from the Norwegian cities. In my thesis I also discussed the use of the term ‘semi- communication’ (see Braunmüller 1993, 1994) and concluded that it is not sufficiently enlightening. Instead, it seems reasonable to use the term ‘receptive bilingualism’. When reporting what assumptions can be made about the possibilities for communication and language use through this study of the sociolinguistic conditions, I therefore use the terms ‘active bilingualism’ and ‘receptive bilingualism’.

Two Norwegian cities – an example The Hanseatic period must be defined somewhat differently for Bergen on one hand and Oslo and Tønsberg on the other. If we take as a point of departure the question of when these German traders were in a dominant position (in some way) in each city, then we can define the Hanseatic period as running from about 1250 to 1650 in Bergen, whereas it lasted from about 1250 up until about 1500 in Oslo/Tønsberg (Nedkvitne 1994, p. 9f). There was a higher percentage of Germans living in Bergen than in Oslo and Tønsberg during this period – some researchers mention that they made up 15– 20% of the total population in Bergen, whereas the proportion was much lower in Oslo and Tønsberg. In both areas the German traders had their own formal organisation, but they were different with regard to both size and importance. In addition, the Germans staying in Bergen and those in Oslo and Tønsberg did not come primarily from the same German areas. The patterns of integration also differed between the two urban areas: whereas the Germans settled in their own colony in a specific part of Bergen, the Bryggen, the Germans living in Oslo and Tønsberg were scattered around the area. These factors hold impor- 48 Gro-Renée Rambø tance for what we can assume about the conditions for language contact situations to occur. There were some similarities between all the cities I studied. The Ger- mans, regardless of what city they lived in, used their own language (Low German) in several different arenas. In all the cities they were considered “strangers” in some sense, even in those where they were allowed to take up Scandinavian citizenship alongside their German citizenship (for instance in Stockholm). The differences are nevertheless important. First of all, the stability of the German population in each city differed. This is important, because it tells us something about the foundations for bilingualism on the individual and societal level. It is more likely that a person living in an area for a long period of time would develop some degree of bilingualism than a person just “visiting” for a shorter period of time. In Bergen there was a rather large semi-permanent German population (traders and craftsmen), plus a large more unstable group (“visitors”). In Bergen the Germans consisted of only single males living together in their colony; their families were left behind in Germany. The Ger- mans brought their families with them to Oslo and Tønsberg, which makes it probable that there were children in these cities growing up as true bilinguals (which is often regarded as very important in contact situations; see for instance Trudgill 1986). These conditions could tell us something about restrictions and possibilities for actual language use in contact situations. It seems reasonable to assume that the Germans living in Oslo and Tønsberg must have had some active competence in Norwegian to manage daily life there because of the way the social structures were organized in the area. They were not a large enough group to form their own separate German institutions alongside the Norwegian ones (unlike in Bergen; see Nesse 2002), and to be able to participate in social activities with the majority of Norwegians, they would have to have gained some competence in Norwegian. For the Norwegian population, on the other hand, the social conditions and structures probably did not create a need for widespread bilingualism (neither active, passive, or receptive) in order to communicate satisfactorily in all necessary social arenas. They were not exposed to German to the same extent as the population in Bergen was. In Oslo and Tønsberg, the people using Middle Low German were probably only the Germans themselves (on a group level), whereas in Bergen some Norwegians probably also used Middle Low German (to some extent) in certain communication arenas, as they could gain socially or economically from having some competence in the language. We can assume that many (maybe even most) Germans who settled in Oslo and Tønsberg would develop bilingual competence of some kind after a while, while the Germans living in Bergen probably did not. It is important to empha- size that we are talking about probabilities here, based on what different kinds of sources and different theoretical insights give us reason to assume (for more examples, see Rambø 2009). Language contact, communication and change 49

What new insights has the focus on sociolinguistic conditions provided? It seems quite clear from the source material that we cannot talk about language contact as if it were identical in the different Scandinavian countries and cities. I have shown that the social, historical and cultural preconditions for speakers of different varieties to come into contact were not the same throughout Scan- dinavia – different people met in different places, communicated differently, and passed on different results of language contact within the various communities they were part of (see more on this in Rambø 2009).

The individual language user – reflections and considerations for further research In my PhD, I included insights from accommodation theory, the Acts of Iden- tity model, and Goffman’s theory of the presentation of self in everyday life, but other considerations should be included on the micro level of language contact. In the following I draw on insights from Joshua Fishman’s approach to language sociology, Einar Haugen’s theory of language ecology, and Dell Hymes’s model of communicative competence to discuss what competence in a language (or linguistic variety) means at the individual level, and what the interplay is between the individual and group level (the societal structures) in a contact situation.

Linguistic varieties and the individual language user Within modern language contact research, it has recently been thoroughly and (partly) convincingly argued that distinguishing between language contact and dialect contact can yield fruitful insights into the consequences of different kinds of language contact. The crucial point is that different languages are not mutually intelligible, whereas dialects of the same language are. Given this fact, it is claimed that the possibilities for communication and understanding differ profoundly in these different contexts (see e.g. Trudgill 1986). For the type of language contact in question here, it has been argued that perhaps Middle Low German and the Scandinavian languages were structurally similar enough to consider them dialects of the same language, and not different languages. Middle Low German and Scandinavian varieties would then have been mutually intelligible to some extent, and this would in part explain how the influence of Middle Low German could be so profound. But beyond this, we need to consider how we conceive bilingualism and communicative competence. Trudgill has himself moderated his model (as it was first presented) by stating that there is no clear-cut division between language contact and dialect con- 50 Gro-Renée Rambø tact – there are all sorts of intermediate cases, and it is therefore more fruitful to talk about a continuum. It is also a critical point that intelligibility to a large extent is dependent on exposure – even if the linguistic varieties involved are fairly different structurally, the ability to understand the variety strengthens with intense exposure at the individual level. The degree of intelligibility within a speaker-based approach is not an absolute; the results and preconditions depend on the ways in which the conversation is negotiated between the speakers of different languages or varieties in a specific situation and arena (Trudgill 2000).

Language, language user and language use When studying a language contact situation, it is necessary to stress that the relationship between language, language user and language use is the key to understanding what happens linguistically. Fishman points out that the sociology of language is concerned with examining the interaction between two aspects of human behavior: use of language, and the social organization of behavior. He further on states: Briefly put, the sociology of language focuses upon the entire gamut of topics related to the social organization of language behavior, including not only language usage per se but also language attitudes, overt behaviors toward language and toward language users. (Fishman 1976b, p. 217.) Similar thoughts are found in Einar Haugen’s work within the discipline he himself calls ‘language ecology’. Haugen points out that it is the psychological and social aspects which define language ecology, here also language change. The psychological and the social components are complementary – they depend upon each other. The social structures of a society are composed by individuals, and language is part of the human mind. But what happens in the human mind also has to be related to the social surroundings of the individuals. Haugen emphasizes that what regulates the ecology of a given language, by ne- cessity, is the language users. He says: The true environment of a language is the society that uses it as one of its codes. Lan- guage exists only in the minds of its users, and it only functions in relating these users to one another and to nature, i.e. their social and natural environment. Part of its ecology is therefore psychological: its interaction with other languages in the minds of bi- and multilingual speakers. Another part of its ecology is sociological: its interaction with the society in which it functions as a medium of communication. The ecology of a language is determined primarily by the people who learn it, use it, and transmit it to others. (Haugen 1972, p. 325.) Middle Low German’s extensive influence on the Scandinavian languages during the Late Middle Ages has to be described and explained using a model comprising an understanding of both the linguistic varieties involved, as well as historical, sociological and cultural relations at the time the contact took place. Language contact, communication and change 51

These latter factors were important for defining how people actually involved themselves in different kinds of communication situations in different arenas, which is again of importance at the individual psychological level. Trudgill says: The languages that are in contact with each other socially may become changed linguistically, as a result of also being in contact psychologically, in the competences of individual speakers. (Trudgill 1986, p. 1.) Different terms have been used to define the linguistic competence of the speakers involved in communicative situations in Late Medieval Scandina- vian cities where Low German was in contact with Scandinavian varieties; terms like ‘mixed languages’ (Winge 1992; Moberg 1989), ‘semi-communication’ (e.g. Braunmüller 1993), ‘active and passive bilingualism’ (Nesse 2002), and ‘receptive bilingualism’ (Rambø 2009). It is central what terms we end up using, since they guide the way we understand and explain what we observe. The descriptions, or assumptions, we make on the individual and societal level govern our understanding of the communicative situations at crucial points.

Language contact and bilingual (multilingual) societies and individuals We cannot thoroughly examine a contact situation without focusing on the actual individuals involved in it. In my PhD, I placed more focus on social, historical and cultural factors governing societal structures, and less on the individuals – partly because of the way my research questions were formulated, and partly because of all the difficulties and constraints there are when we try to find enough “evidence” at the individual level from a long-term diachronic perspective. I will now briefly discuss some issues and difficulties that need to be dealt with when taking the individual level thoroughly into account. I take as a starting point the bilingual individual. The psychological component, as Haugen defines and understands it, has above all been linked to bi- or multilingualism in communication situations where there is more than one linguistic variety involved. In language contact situations – whether these are defined as language contact, or dialect contact, or somewhere along a continuum – bilingualism or multilingualism is crucial. However, we can talk about bilingualism in different ways. On one hand, we can differentiate theoretically between individual and societal bilingualism (Appel & Muysken 1987, p. 1f). We could say that societal bilingualism is found when two or more languages are used (spoken) in a given society. Such a definition necessarily includes many different kinds of bilingual societies, and we could therefore further divide societal bilingualism into three different types (Appel & Muysken 1987, p. 2): different languages are used by different groups of the population in a society, and each group is mono- 52 Gro-Renée Rambø lingual; all the individuals in a society are bilingual; or one group in a society is monolingual, whereas another group is bilingual. On the other hand, most linguistic changes which are the consequence of language contact – as discussed thoroughly by Weinreich (1953), among others – appear above all to be the result of individual bilingualism. If the people in a contact situation do not speak mutually intelligible varieties, one or both/all parties involved necessarily would have to have some sort of bilingual competence in order to communicate effectively. This bilingual competence at the individual level (repertoire) in the long term leads to influence and interaction, or what many researchers call ‘interference’ (Trudgill 1986, p. 1). Bilingual individuals could, of course, be found in any sort of linguistic society. This means that also a linguistic society considered to be monolingual, could encompass bilingual individuals. Trudgill (1986) points out that it is the language competence of the individual which determines what happens linguistically in a contact situation. How- ever, it is not always easy to describe what kind of competence this is. What level of competence must be reached to call an individual bilingual? The question is even more complex when dealing with contact situations far back in time. What is the actual difference between active bilingualism and passive, or receptive, bilingualism? If we take Hymes’ theory of communicative competence as a starting point, we must admit that linguistic competence is only one of several components – other factors such as sociolinguistic competence, strategic competence, discourse competence, etc. are equally important. Within modern language-contact studies, it has been considered sufficient to define the distinction between a language and a dialect by referring to the presence or absence of structural intelligibility. But what happens when intelligibility in a contact situation is also linked with discourse strategies, gestures, strategic competence, etc.? And when is such competence ‘active’, and when is it ‘receptive’? It seems absolutely clear that we cannot talk about active or receptive competence as if these were opposites – they must be considered different stages on a continuum, which must be defined at the individual level. It seems unreasonable to assume that a person should, for instance, be able to understand certain linguistic items (have ‘passive’ competence) in a foreign linguistic variety without at the same time being able to actively produce the words herself, if necessary and/or wanted (unless there are severe problems causing restrictions for active competence on the phonetic level). In other words, the question is how much ‘receptive’ competence in a linguistic variety the individual actually needs in order for making it an ‘active’ competence if needed? We could claim that ‘receptive’ competence is of course also an ‘active’ competence (psychologically) – but to a different degree. Language contact, communication and change 53

Conclusion If we want to understand language contact processes, whether in present-day or historical studies, it is both interesting as well as necessary to include different intra- as well as inter-disciplinary ideas, approaches to understanding, perspectives, methods and theories. We cannot understand language contact processes without including knowledge linked to the individual language users and the human mind. We need to explore how to define communicative competence. What complicates the study of language, language variation and language change is that most of the terms we use are in themselves relational. This goes for central terms like language and dialect, as well as for terms like culture, ethnicity, mother tongue, (language) society, bilingualism, etc. All these are social constructs, which in part means that they are to some extent floating concepts; they are dynamic and change according to time, place, situation, etc. Whereas these ideas have been rather successfully integrated in several newer works within the academic fields of modern language-contact research and modern sociolinguistics or the sociology of language, this approach is still less visible and less integrated in diachronic studies, for instance within historical linguistics and historical sociolinguistics. An understanding of the phenomenon of bilingualism has been essential in much of the language contact research, but Bilingual or bidialectal phenomena have been the main focus of the interest that has been shown. Yet bilingualism is not in itself an adequate basis of a model or theory of the interaction of language and social life. From the standpoint of such a model or theory, bilingualism is neither a unitary phenomenon nor autonomous. The fact that two languages are present in a community or are part of a person’s communicative competence, is compatible with a variety of underlying functional (social) relationships. Conversely, distinct languages need not be present for the underlying relationships to find expression. Bilingualism par excellence … is a salient, special case of the general phenomenon of linguistic repertoire. No normal person, and no normal community, is limited to a single way of speaking, to an unchanging monotony that would preclude indication of respect, insolence, mock seriousness, humor, role distance, and intimacy by switching from one mode of speech to another. (Hymes 2003.) Language contact phenomena and situations represent an enormously complex subject because there are so many important factors involved, and because these factors are so interwoven. There are even more profound and special problems involved when the object of study lies far back in time, since we will also be faced with problems of finding relevant, reliable and/or representative sources – or determining exactly what they are representative of. The problems in defining the communicative competence(s) of language users are even more evident. Nevertheless, these research constraints do not make this kind of work less interesting – rather, it becomes more so. By shining a light on the difficulties we are faced with when carrying out diachronic language contact research, we also illuminate potential problems in synchronic studies which are perhaps not always that conspicuous, due to much better access to relevant sources. 54 Gro-Renée Rambø

References Appel, René and Muysken, Pieter, 1987: Language contact and bilingualism. London. Braunmüller, Kurt & Diercks, Willy (eds.), 1993: Niederdeutsch und die Skandinavi- schen Sprachen I. Heidelberg. Braunmüller, Kurt, 1993: Nye hypoteser om sprogkontakt mellem nedertysk og dansk i Hansetiden. En projektrapport. In: Mette Kunø & Erik Vive Larsen (eds.), 4. Møde om Udforskning af Dansk Sprog. Århus. Pp. 46–55. Braunmüller, Kurt, 1994: Semikommunikation og lingvistiske simulationsmodeller (om sprogkontakt i hansetiden). In: Ulla-Britt Kotsinas & John Helgander (eds.), Dialektkontakt, språkkontakt och språkförändring i Norden. Föredrag från ett fors- karsymposium. (Meddelanden från Institutionen för nordiska språk vid Stockholms universitet, MINS 40.) Pp. 92–97. Braunmüller, Kurt (ed.), 1995: Niederdeutsch und die Skandinavischen Sprachen II. Heidelberg. Braunmüller, Kurt, 2000: Historisk syntax og flersprogethed. In: Nordlyd 28. Pp. 1–24. Dalen, Arnold, 1994: The influence of Low German on the Norwegian language. In: Volker Henn & Arnved Nedkvitne (eds.), Norwegen und die Hanse. Wirtschaftliche und kulturelle Aspekt im europäischen Vergleich. Frankfurt am Main. Pp. 31–39. Elmevik, Lennart & Schöndorf, Kurt Erich (eds.), 1992: Niederdeutsch in Skandinavien III. Akten des 3. Nordischen Symposions „Niederdeutsch in Skandinavien“ in Sig- tuna, 1989. Berlin. Fishman, Joshua Aaron (ed.), 1968: Readings in the sociology of language. The Hague. Fishman, Joshua Aaron, 1976a [1971]: Advances in the sociology of language I. Basic concepts, theories and problems: alternative approaches. The Hague/Paris. Fishman, Joshua Aaron, 1976b: The sociology of language. An interdisciplinary social science approach to language in society. In: Joshua A. Fishman (ed.), Advances in the sociology of language I. Basic concepts, theories and problems: alternative approaches. The Hague/Paris. Pp. 217–405. Giles, Howard & Smith, Philip, 1979: Accommodation theory: optimal levels of convergence. In: H. Giles & R. N. St. Clair (eds.), Language and social psychology. Ox- ford. Pp. 45–65. Haugen, Einar, 1972: The ecology of language. In: Anwar S. Dil (ed.), The ecology of language. Essays by Einar Haugen. California. Hyldgaard-Jensen, Kristian, Winge, Vibeke & Christensen, Birgit (eds.), 1989: Nieder- deutsch in Skandinavien II. Berlin. Hymes, Dell, 2003: Models of the interaction of language and social life. In: C. B. Paulston & G. R. Tucker (eds.), Sociolinguistics: the essential readings. Blackwell. Pp. 30–47. Jahr, Ernst Håkon, 1995: Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Jahr, Ernst Håkon, 1997: New perspectives on the language contact between Middle Low German and mainland Scandinavian in the Late Middle Ages, and about a foot- note on mixed languages which gave rise to a “detective story”. In: Multilingua 16. Pp. 325–337. Jahr, Ernst Håkon (ed.), 1999: Language change: advances in historical sociolinguistics. Berlin. Jahr, Ernst Håkon (ed.), 2000: Språkkontakt – Innverknaden frå nedertysk på andre nordeuropeiske språk. København. (Skrift nr. 2 frå prosjektet Språkhistoriske prinsipp for lånord i nordiske språk. Nord 2000:19.) Killie, Kristin, 2007: Language contact in a historical perspective. (Unpublished hand- out. HiSon seminar, Lesbos.) Language contact, communication and change 55

Le Page, Robert L. & Tabouret-Keller, Andrée, 1985: Acts of identity. Cambridge. Menke, Hubertus & Schöndorf, Kurt Erich (ed.), 1993: Niederdeutsch in Skandinavien IV. Berlin. Milroy, James, 1997: Internal vs external motivations for linguistic change. In: Multi- lingua 16. Pp. 311–323. Moberg, Lena, 1989: Lågtyskt och svenskt i Stockholms medeltida tänkeböcker. Upp- sala. Nedkvitne, Arnved, 1994: How important was Hansa trade for the Norwegian economy? In: Volker Henn & Arnved Nedkvitne (eds.), Norwegen und die Hanse. Wirtschaftliche und kulturelle aspekte im europäischen vergleich. Frankfurt am Main. Pp. 9–18. Nesse, Agnete, 2002: Språkkontakt mellom norsk og tysk i hansatidens Bergen. Oslo. Rambø, Gro-Renée, 2009: Historiske og sosiale betingelser for språkkontakt mellom nedertysk og skandinavisk i seinmiddelalderen – et bidrag til historisk språksosiolo- gi. Oslo. Romaine, Suzanne, 1982: Socio-historical linguistics: its status and methodology. Cam- bridge. Schöndorf, Kurt Erich & Westergaard, Kai-Erik (eds.), 1987: Niederdeutsch in Skandi- navien I. Berlin. Simensen, Erik, 1994: Språkkontakt over landegrenser. Munnleg kommunikasjon i mellomalderen. In: Collegium medievale 7.1. Oslo. Pp. 33–50. Thomason, Sarah Grey & Kaufman, Terrence, 1988: Language contact, creolization and genetic linguistics. Oxford. Thomason, Sarah, 2001: Language contact: an introduction. Edinburgh. Trudgill, Peter, 1986: Dialects in contact. Oxford. Trudgill, Peter, 2000: On locating the boundary between language contact and dialect contact – Low German and continental Scandinavian. In: Ernst Håkon Jahr (ed.), Språkkontakt – Innverknaden frå nedertysk på andre nordeuropeiske språk. Køben- havn. (Skrift nr. 2 frå prosjektet Språkhistoriske prinsipp for lånord i nordiske språk. Nord 2000.19.) Pp. 71–87. Ureland, Per Sture (ed.), 1987: Sprachkontakt in der Hanse. Aspekte des sprachausgleichs im Ostsee- und Nordseeraum. Tübingen. (Akten des 7. Internationalen Sym- posions über Sprachkontakt in Europa, Lübeck 1986.) Weinreich, Uriel, 1953: Languages in contact: findings and problems. New York. Winge, Vibeke, 1992: Dänische deutsche – deutsche Dänen. Geschichte der deutschen sprache in Dänemark 1300–1800 mit einem ausblick auf das 19. Jahrhundert. Hei- delberg. 56 Gro-Renée Rambø Gender reduction in Bergen Norwegian: a North-Sea perspective 57 Gender reduction in Bergen Norwegian: a North-Sea perspective1

Peter Trudgill, University of Agder

The Norwegian of the city of Bergen holds a special place in discussions of the influence of Low German on the Scandinavian languages. Indeed, it also holds a special place in the study of Norwegian dialectology, being generally considered to be aberrant in a number of respects in the context of regional varieties of Norwegian. One respect in which it is highly unusual has to do with the unique position which the grammatical gender system of the Bergen dialect has amongst Norwegian dialects. In Bergen Norwegian the distinct feminine definite and indefinite articles have disappeared, and the markers of the new common gender are all original masculines, e.g. bok-en cf. bok-a ‘the book’, en bok cf. ei bok ‘a book’. This reduction of three genders to two is a clear example of simplification; and a number of writers have ascribed this to language contact. Ernst Håkon Jahr has developed a contact-based explanation for this loss of the feminine in Bergen (Jahr 1995, 1998, 2001), in which he argues that this development in the town was the result of the “heavy influence of language contact between Norwegian and Low German” (2001, p. 100) during the Hansa period. Nesse (2002) makes the same kind of point in more detail; and my own writings are amongst a number of others supporting this view. As Jahr says: Siden voksne som lærer nye språk, kjennetegnes ved å være “imperfect adult learners” i forhold til hva barn er, vil vi vente å finne flere og tydligere språkkontaktre- sultater i Bergen enn i Oslo og Tønsberg.2 Jahr’s explanation, that is, is couched in terms of the simplification which is well known to be associated with the imperfect language learning abilities of adults in contact situations (Trudgill 2011). Other writers have discussed whether this should be considered to be dialect contact rather than language contact (Trudgill 2000), supposing that Low Ger- man might have been rather mutually intelligible with Norwegian at the time

1 I am very grateful for their help with this paper to Torben Arboe, Gunther De Vogelaer, Martin Durrell, Jarich Hoekstra, Ernst Håkon Jahr, Agnete Nesse, Hans Frede Nielsen, Inge Lise Peder- sen, Karen Margrethe Pedersen, Gro-Renée Rambø, Gertrud Reershemius, Peter Siemund, Philip Stickler, Wim Vandenbussche, Alastair Walker, and Roland Willemyns. 2 ‘Compared to children, adults learning new languages are characterised as being “imperfect adult learners”, and we would therefore expect to find more – and more obvious – consequences of language contact in Bergen than in Oslo or Tønsberg.’ 58 Peter Trudgill in question, and noting that the gender reduction was partial, from three to two, rather than the total loss of gender altogether. And yet other writers have queried the role of any type of contact, Perridon (2003) for example suggesting that we need no other explanation for what happened than simple phonological change (see also Heide 2003). In this paper, however, I show that the Bergen Norwegian gender system is not at all unusual when viewed in a wider geographical context. And, given that this is the case, I look at the possibility that there might be another, complementary explanatory factor that has been at work in producing this “aberrant” Bergen gender system. After all, it is often a good idea to look at the possibility of multi-causality in linguistic change: there are so many developments which could have taken place that when some particular change has actually taken place this suggests there is likely to have been more than one favourable, predisposing factor at work.

The geography I begin by noting the geographical location of the city of Bergen on the eastern shores of the northern North Sea. Across this sea from Bergen, on the western shores, lies another area where the original three grammatical genders of Ger- manic have also not been retained, namely Britain. Bergen is about 300 nautical miles from the Scottish coast at Aberdeen, and we know that contact around and across the North Sea has been of considerable historical and cultural importance over the centuries. Can there be a connection between the loss of gender in the two places? One’s first answer has to be that it is not immediately obvious that this is anything other than pure coincidence. But notice, then, that Bergen Norwegian was not the only continental Germanic variety to experience a reduction of genders from three to two. Further south, again on the east coast of the North Sea, and at more or less exactly the same time as in Bergen (see below), the metro- politan Dutch varieties of the northern Netherlandic-speaking area also merged the feminine with the masculine (Geerts 1966). It is perhaps not quite so easy to shrug this off as simply being a coincidence. The claim that it is simply a coincidence becomes perhaps even more difficult to maintain when we observe the following. The dialects of the Flemish south of the Low Franconian Netherlandic-speaking area, as well as neighbouring areas, did not undergo this change, and retained the three original genders (De Vogelaer 2009), as they still do to this day. The boundary between the innovating northern two-gender dialects and the southern three-gender Dutch dialects is conventionally drawn by Dutch dialectologists as running from west to east across the Netherlands, in a rather straight line from the Rhine delta on the North Sea coast. At the eastern end of this line, it takes a brief turn to the north and then another back east, just before it meets the Netherlands-Germany border to the east of Hardenberg. Gender reduction in Bergen Norwegian: a North-Sea perspective 59

Of course we conventionally divide the dialects of the West Germanic dialect continuum into “German” and “Dutch”, but this is a politico-sociolinguistic matter since there is no linguistic reason for the division of the continuum into two3 to be made at the border between Germany and The Netherlands. So surely this west-east isogloss will not just stop at the national border. My investigations show that of course it does not: the west-east two-gender /three-gender isogloss does indeed extend from the Netherlands across into Germany, into the County of Bentheim in Lower Saxony. So the question then arises as to how much further it extends into the Low German-speaking area of this part of western Germany. My major source in tackling this question has been Wahrig-Burfeind (1989). Her work, and her maps, give a strong indication that all of the Low German (Plattdeutsch/Niederdeutsch) dialects spoken in western Lower Saxony have only the two genders. In the Low German spoken in Emsland, Lower Saxony, the “Synkretismus von Maskulinum und Femininum dominiert” (1989, p. 57). And further to the north, the loss of the feminine has also been totally carried through in East Frisia: Wahrig-Burfeind (1989, pp. 82–83) discusses the East Frisian Low German [EFLG] merger using data from Wiesenhann (1977); but the fact of this merger is also confirmed in the book-length study of East Frisian Low German by Matras & Reers- hemius (2003, p. 18), who write: EFLG, like other Low German varieties, has a two-gender system: common (or non-neuter) and neuter. As in Dutch and the Scandinavian languages, common nouns tend to continue historical masculines and feminines, while neuter nouns generally continue historical neuters. Wahrig-Burfeind shows, too, that the Low German of the Oldenburg area also lacks a distinct feminine (1989, p. 57) – which takes us geographically as far east, more or less, as the banks of the River Weser. I am therefore also assum- ing, though I have not actually had this confirmed, that all the intervening areas of Lower Saxony between East Frisia/Emsland and Oldenburg, including at least Cloppenburg, Ammerland and Wesermarsch, also have the merger; it would be rather remarkable if they did not. Goltz & Walker (1990) divide the Low Saxon dialects of Low German in Germany into six major sub-divisions; according to my assumption, three of them, Ostfriesisch, Oldenburgisch and Emsländisch, have the merger. This assumption is illustrated in Map 1. All three of Ostfriesisch, Oldenburgisch and Emsländisch are spoken on former Frisian-speaking territory. In view of this configuration of the two-gender dialects, it is no surprise to find that, back west on the other side of the border, the Low Saxon-speaking areas of the Netherlands around Groningen and in Overijssel also have the syncretism. To complete the geographical jigsaw, it is also rather predictable that we should find that West Frisian, the language spoken between the northern Low Franconian Dutch-speaking area and Groningen Low Saxon, is no differ-

3 Or three, if we include Luxemburgisch. 60 Peter Trudgill ent. The West Frisian language follows northern Dutch and East Frisian Low German in having exactly the same masculine–feminine syncretism. The two-gender system of Modern West Frisian is due, that is, to the loss of the feminine, just as in Dutch and in Bergen Norwegian (Tiersma 1985, p. 47). So it is the whole of the north of the Netherlands, regardless of which of the three languages is spoken there, that has been affected by this Bergen-style simplification; together with a large neighbouring area of northwestern Germa- ny up to the River Weser. These two-gender West Germanic varieties lie in a single contiguous geographical area which comprises the northern dialects of Dutch, nearly all the dialects of West Frisian, and the northwestern dialects of Low German. The fact that a long contiguous stretch of the eastern coastline of the North Sea, and adjoining inland areas, from the Rhine delta in the southern Netherlands to the mouth of the Weser in northern Germany, is occupied by varieties of three different languages which have all lost the masculine–feminine distinction, while southern Dutch and inland Low German have not, looks to be a geographical pattern of some significance. But what is the significance? Or is this still a coincidence? Staying on the eastern shores of the North Sea, we can notice something else which perhaps causes us to think a little more. It is true that if we travel from the mouth of the River Weser eastwards along the German coast, we come first into more conservative territory: the coastal dialects of Low Saxon Low Ger- man spoken from Bremerhaven and up into western Holstein do still retain the three genders. But it is remarkable that one does not have to travel much further along the coast, as the shoreline swings to a northerly direction, before one returns to two-gender territory. This territory stretches from the mouth of the River Eider onwards. According to Jørgensen (1954), as reported in Wahrig-Burfeind (1989, p. 50), the three-gender/two-gender line runs east to west from south of the Eckenförde Bay – so just to the north of Kiel – and then via Haby and along the River Eider as far as mouth of the river on the North Sea at Tönning. Wah- rig-Burfeind (1989, pp. 75–76), using data from Bock (1933), shows that north of this line the masculine/feminine merger has been totally carried through in the Low Saxon of Schleswig. Schleswig thus represents the fourth of Goltz and Walker’s six Low Saxon divisions to have the merger, leaving only Hol- steinisch and Nord-Hannoverisch with three genders. In the island dialects of North Frisian, the feminine has also been lost, though on the islands of Föhr and Amrum it has been lost through merger not with the masculine but with the neuter (Hoekstra 1996). If we continue north- wards along the coast from North Friesland into Denmark, we come next to the Danish dialects of the Jutland peninsula which, with the exception of the far north and east, have also undergone gender reduction (Arboe 2008). Do we, then, still want to maintain that this is all a coincidence? The spatial pattern of feminine loss in modern Germanic does not look to be random. There is a large area of Europe stretching from the Rhine delta along the North Sea coast and up to and including most of Jutland, with a small break only between Gender reduction in Bergen Norwegian: a North-Sea perspective 61

Lower Saxony and Schleswig, where the local dialects of each Germanic language all have the two-gender innovation. Indeed, if we consider Bergen from a more international perspective, it is a remarkable fact that from around 1600 or so (see below) it has been part of a larger Germanic-speaking area of coastal western Europe around the North Sea where the original Germanic three-gender system has not been retained. As is illustrated in Map 1, gender-loss of various kinds has occurred in seven different North Sea Germanic languages in a unified geographical area: Bergen Norwegian, English, and varieties of Dutch, West Frisian, Low German, North Frisian, and Danish. Again, perhaps significantly, this development did not take place in any areas away from the North Sea (except for Stockholm and Copenhagen – see below). What exactly are we to make of this?

Map 1: The geographical gender-reduction zone c. 1700. To the west and north of the line, Ger- manic dialects of English, Dutch, West Frisian, Low German, North Frisian, Danish, and Bergen Norwegian have all lost the original three-gender system. 62 Peter Trudgill

The linguistics Although we are presented with this large contiguous area where the three Ger- manic genders have been reduced, we have to concede that the actual linguistic outcomes do vary considerably from dialect to dialect, to the extent that it could be argued there is actually no need to consider whether or not this is a coincidence. However, according to Wahrig-Burfeind (1989), there is a regular linguistic change-path which Germanic gender loss may pass along. Thus the outcomes, albeit rather different, simply represent different chronological stages, and/or different branches, of the same gender-loss process. The path starts from the baseline of the original Germanic three-gender situation, in which pronouns corresponding to ‘he’, ‘she’ and ‘it’ are employed purely on the basis of the grammatical gender of the nominal referent, as still happens in conservative Germanic varieties such as High German and Nor- wegian. In the latter, we still find forms such as:

vegg-en … han ‘wall-the … he’ bok-a ... ho ‘book-the … she’ ku-a … ho ‘cow-the … she’ jent-a … ho ‘girl-the … she’

The first stage of feminine loss along the path occurred when, as in 16th-century Bergen Norwegian, feminine nominals were no longer formally distinct from the masculine, as a result e.g. of the loss of the distinction between masculine and feminine articles, but ‘she’ was still used to pronominalise grammatical as well as natural feminines. Using the equivalent modern forms, this would be:

vegg-en … han ‘wall-the … he’ bok-en ... hon ‘book-the … she’ ku-en … hon ‘cow-the … she’ jent-en … hon ‘book-the … she’

In the next chronological stage, ‘he’ takes over in the case of former non-animate feminines; this is the crucial stage for grammatical gender merger. This is what happened during the 1600s in Bergen; and it is also what has occurred in northern Dutch. So modern Standard Dutch is like Bergen Norwegian in that it has the definite article forms de (common gender – original masculine and feminine), while the southern dialects have den (masc.) and de (fem.), with corresponding differences in adjectival agreement (Taeldeman 2005, p. 62):

‘the thin farmer’ ‘the thin farmer’s wife’ Standard N. Dutch: de dunne boer de dunne boerin Flemish: den dunnen boer de dunne boerin Gender reduction in Bergen Norwegian: a North-Sea perspective 63

Importantly, the effects of this merger can be seen in grammatical as well as natural gender:

‘The table … it’ Standard N. Dutch: De tafel … hij (= ‘he’) Flemish: De tafel … ze (= ‘she’)

The noun tafel was originally feminine in all varieties of Dutch, and is therefore pronominalised by ‘she’ in varieties which preserve the feminine. But in Stand- ard Dutch, where the feminine has been lost, all common gender nouns, regardless of historical gender status, are now pronominalised by ‘he’, except for animate feminines such as vrouw ‘woman’ (De Vogelaer & De Sutter 2011). So this gives us a system as follows, where boom ‘tree’ is an original masculine:

de boom… hij ‘the tree … he’ de tafel … hij ‘the table … he’ (originally fem.) de koe … zij ‘the cow … she’ de meisje … zij ‘the girl … she’

Remarkably, however, Bergen has gone one stage further than this. In basilectal Bergen, not only are originally feminine nouns pronominalised with the masculine pronoun han ‘he’, but so are non-human female animates: one says ku-en … han ‘cow-the … he’. Thus in the modern basilectal Bergen dialect we have (Nesse 2005):

vegg-en … han ‘wall-the … he’ bok-en ... han ‘book-the … he’ (originally fem.) kuen … han ‘cow-the … he’ (naturally fem.) jent-en … hon ‘girl-the … she’

Hon ‘she’ in Bergen is entirely restricted to women and girls.4 There is another Germanic development which is illustrated in the modern Bergen dialect. Although han ‘he’ is still the pronoun used for all common gender inanimate nouns (and indeed as we have just seen for all animates except humans), it is subject to a grammatical constraint such that den, originally ‘that’, has to be used instead when the pronoun is fronted or stressed (Nesse 2005). This is also what has happened in northern Dutch (Shetter 1994, p. 68): although hij is used for all common gender inanimates, die, also originally ‘that’, has to be used with fronting or stress. Here we can see the beginnings of another stage of development, as represented in the current Standard Danish and Standard Swedish (Skrzypek 2010)

4 This gives the Bergen dialect a “feminine personal” gender equivalent to the Polish “masculine personal” or “virile” gender, a gender class which is confined to male human beings (and in the plural only; Sadowska 2011). 64 Peter Trudgill systems. In these two-gender Scandinavian varieties, feminine has merged with masculine; but there is no pronominalisation of the type boken … han. Rather, these varieties have developed a newer system in which han and hun/ hon are both reserved for animates, while inanimate common gender nouns (including original feminines) are pronominalised with the new pronoun den ‘it’, derived historically from the masculine distal demonstrative den ‘that’. In these systems, han ‘he’ has now dropped out of use altogether except for male animates, while den ‘that’ has taken over and is used for all originally masculine and feminine inanimates in all contexts, not just when fronted or stressed. This change is said to have been completed in Standard Swedish during the 1700s (Davidson 1990). In modern Danish, these forms are:

vegg-en … den ‘wall-the … it’ bog-en ... den ‘book-the … it’ koen … den/hun ‘cow-the … it/she’ pigen … hun ‘girl-the … she’

The situation in East Frisian Low German is similar. East Frisian Low German, as we saw earlier, has a two-gender system: common (or non-neuter) and neuter. As in Dutch and the Scandinavian languages, common nouns derive from historical masculines and feminines. Importantly also, the linguistic details do indicate a truly total gender merger, à la Bergen. According to Matras & Reers- hemius (2003), in East Frisian Low German the pronoun zäi ‘she’ is reserved for female animates only, as in Dutch. Wiesenhann (1977), as reported in Wah- rig-Burfeind (1989), agrees that East Frisian Low German no longer uses the feminine pronoun for non-animate referents. But as in Standard Danish and Swedish – and unlike in Bergen and Dutch – inanimate common gender nouns in East Frisian Low German are pronominalised by an originally demonstrative pronoun däi. However, däi can also be used for animates in certain discourse contexts (see Matras & Reershemius 2003, p. 23 for details), which is not dissimilar to the usage in stressed position of den in Bergen and die in Dutch. The Low German of Schleswig also has a pronoun system which operates with ‘he’ and ‘she’ for animates but has a different common gender pronoun for non-animates (Bock 1933). A further step along this path may ensue, which can lead to the loss of grammatical gender altogether. This involves the loss of a distinct neuter grammatical gender. According to De Vogelaer & De Sutter (2011), such a process is currently taking place in modern Dutch in that non-animate masculines and feminines are increasingly being treated as neuters, while animates are not, something which Curzan (2003) says also happened in Old English/Middle English, where grammatical gender eventually disappeared altogether, to be replaced by a purely natural gender system. The Danish dialects of West Jutland have progressed even further along this gender-loss path. Above, I cited Jutland as an area which was almost entirely Gender reduction in Bergen Norwegian: a North-Sea perspective 65 two-gender territory. However, there is a significant difference between the dialects of western Jutland and those of the east. West Jutland has developed an innovative system where gender has been lost and replaced, as in English, by a system which distinguishes between male animates, female animates, and inanimates. However, inanimate nominals are further divided in West Jutland into the two semantically/grammatically-determined categories of count vs. mass noun (Arboe 2008). The pronominal forms den and det, as in Standard Danish, have taken over from ‘he’ and ‘she’ for inanimates. But the original merged common masculine-plus-feminine gender marked by den has come to be used as the nominal class for all inanimate count nouns, while mass nouns behave like original neuters, taking det. Haugen (1976, p. 61) cites the following examples:

Æ ægetræ i vor have, den er stor ‘The oak-tree in our garden, it’s big’ BUT Ægetræ, det er best til møbler ‘Oak (wood), it’s (the) best for furni- ture’

It can be seen that original common gender den is used to pronominalise the inanimate count noun, regardless of historical gender, while det is used for the mass noun. Some of the neighbouring North Frisian dialects (and certain Low German dialects in the same area) have a similar system. Löfstedt (1968, p. 12) tells us that in modern North Frisian, certain original inanimate masculines and feminines have become neuter, while others have become neuter only when used as mass nouns and retain their original gender when used as count nouns. Alastair Walker (p.c.) kindly supplies an example: di kafee (‘the coffee’, masculine) re- fers to a particular cup of coffee, whereas dåt kafee (neuter) is coffee as a sub- stance.

Diffusion and independent development There is, then, some diversity in the linguistic outcomes of gender simplification in the different Germanic languages around the North Sea. But it is possible to see these outcomes as simply representing the same outcome at different stages of development, making the “coincidence” hypothesis less likely. Nevertheless, there is a puzzle here that arises rather frequently in the historical human sciences. If an innovation event takes place in one location at a certain time, and then in another location at a later time, the question is: has the innovation diffused from the first location to the second, or is it just an independent development? 66 Peter Trudgill

Separate explanations for each area It is perfectly possible to claim that the picture presented in 1 is the result of the coincidence of a number of independent developments, since for each particular dialect explanations can be advanced which work rather well in that individual case. Gender-loss in English is often ascribed, for instance, to contact with Old Norse and/or Late British (see Trudgill 2010). Gender reduction in Dutch can be ascribed to rapid urbanisation and growth during the Golden Age, particularly with the movement of large numbers of refugees fleeing from Spanish persecution in the south to the north of the Dutch-speaking area. Similar urbanisation explanations have been advanced for Stockholm and Copenhagen, where the feminine gender was also lost (and so also in Standard Swedish and Danish). At the end of the 15th century the population in Stockholm was about 6,000, but this grew to around 60,000 by 1700 as the Swedish empire expanded. Similarly, at around this same time Copenhagen was experiencing a five-fold increase in the number of its inhabitants. In each of these three cases, the mechanism would have been the simplification that typically accompanies dialect mixture (Trudgill 1986). According to another theory, gender simplification in Dutch and West Fri- sian may also, or instead, be due to Dutch/Frisian gender mismatches: Kusmen- ko (1996, 2000) suggests that simplification may be due to large-scale confusion on the part of speakers about masculine/feminine/neuter grammatical gender membership (which is, after all, arbitrary), with the confusion resulting from contact between, and bilingualism in, Dutch and West Frisian – it is certainly the case that “there are a good many Frisian nouns whose gender differs from that of their Dutch cognates” (Visser 2011, p. 32). Kusmenko also makes a similar point about simplification in Jutish Danish, North Frisian and Schleswig Low German, where nouns also belonged to different genders in the different languages.5 Language shift may also have been a factor in some areas: during the rapid language shift which occurred as speakers of South Jutish abandoned their North Germanic mother tongue and shifted to Low German in the 19th century as a result of political developments and border changes, the two-gender system was simply transferred intact from South Jutish to Low German – although this still begs the question as to why South Jutish had only two genders. And the western Low Saxon dialects of Low German – the Groningen and Overijs- sel variants in the Netherlands; and Ostfriesisch, Oldenburgisch and Emslän- disch in Germnay – are also spoken in an area which was Frisian speaking (in East Frisia until the 1400s, and as a minority language until the 1500s; Matras and Reershemius 2003). So these varieties of Low German could also be regarded as substrate-influenced dialects resulting from contact-induced lan-

5 Haugen (1976) makes a similar point about Stockholm and Copenhagen, suggesting that the entry of new loanwords from Low German into the dialects of the big cities might have caused problems of gender assignment which contributed to the merger of masculine and feminine. Gender reduction in Bergen Norwegian: a North-Sea perspective 67 guage shift. This could be important because, as Wahrig-Burfeind says, there are indications from rather sparse records that some varieties of East Frisian also had the masculine–feminine syncretism. It would therefore be tempting to suppose that gender reduction in northern Dutch and northwestern Low Ger- man might be the result of substratal Frisian influence, but in fact written Old Frisian had the original three Germanic genders until at least the 1400s (Brem- mer 2009). There is also, finally, and as we saw above, the well-known argument that gender reduction in Bergen (as well as in Copenhagen and Stockholm) was due to language/dialect contact involving Norwegian (Danish, Swedish) and Low German, with resultant imperfect adult language learning.

Drift In diachronic linguistics, however, there is also a further complication with independent developments. The problem concerns how truly “independent” such developments might be. This is because of what is often called, following Sa- pir, drift (see also Lakoff 1972). Sapir writes (1921, p. 150) that “language moves down time in a current of its own making. It has a drift.” Importantly, he discusses inherent or inherited tendencies in languages and language families: The momentum of … drift is often such that languages long disconnected will pass through the same or strikingly similar phases … The English type of plural represented by foot: feet, mouse: mice is strictly parallel to the German Fuss: Füsse, Maus: Mäuse. Documentary evidence shows conclusively that there could have been no plurals of this type in Primitive Germanic … There was evidently some general tendency or group of tendencies in early Germanic, long before English and German had developed as such, that eventually drove both of these dialects along closely parallel paths (p. 172). Sapir’s argument is essentially that language varieties may resemble one another because, having derived from some common source, they continue to evolve linguistically in similar directions by undergoing similar linguistic changes. In many cases of drift it is possible to argue that varieties which are derived from a common source may have inherited shared tendencies or propensities which can subsequently lead to the development of similar but new changes and hence similar or identical but new characteristics, even after separation, as with Sapir’s umlaut plurals example. In these cases, we have to say that drift involves propensities to linguistic changes resulting from structural properties which varieties inherit. The inheritance is of structural conditions which at a later date may, but need not, lead to the development of new but identical linguistic changes in different varieties with a common ancestry. The structural preconditions in our case here would be a factor or factors which favour the loss of the morphological distinctiveness of Germanic masculines and feminines. 68 Peter Trudgill

Diffusion Such favourable predisposing factors might well not be impossible to find. But then we would be presented with the problem of why they were operative only in the areas concerned: there is no escaping the picture portrayed in Map 1. The geography of Germanic gender reduction that has occurred around the North Sea makes it look very much as if it is the result of the geographical diffusion of an innovation out of one or a number of centres. This follows the familiar pattern of spatial diffusion known to dialectologists and geolinguists, having at least one outlying exclave to which the innovation has jumped, namely Bergen (Trudgill 1973; Chambers & Trudgill 1998). We also recall that Bergen, like the other possible exclaves, Copenhagen and Stockholm, has had a long history of being a highly important maritime trading port. Wahrig-Burfeind (1989) points to the two geographical areas which have total gender loss as being kernel areas which seem to her to have led the way in this process: English, where grammatical gender has disappeared altogether and been replaced by a semantic natural gender system; and West Jutish, where the same thing has happened but where an additional semantic distinction has also been introduced. But was it any more than these areas just “leading the way”? If we are to be able to decide whether the geographical distribution of gender loss in Germanic dialects is something more than a coincidence, we need also to have some over- sight into the chronology of gender loss in the contiguous West Germanic area, including the developments in English and the three major Scandinavian cities. It is clear, of course, that gender loss started in English before it began any- where else – there is plenty of information about the timing of the loss of gender which shows that it began earlier than on the other side of the North Sea. According to Jones (1988), it was an ongoing process for more than 300 years which began to be evident in texts from the 900s in the north of the Old Eng- lish-speaking area. The loss then gradually spread south, and appears to have been complete by 1300 everywhere except in Kent in the far southeast of Eng- land, where it lingered on somewhat longer. This loss, as noted above, is often ascribed to language contact with Brittonic Celtic and/or Old Norse. The next development in terms of chronology would appear to be the one in West Jutland. Skautrup (1944, p. 270) says, in connection with the development of this “whole new type of gender”,6 that the writer of a section of the Jyske Lov (Jutland Law) written about 1325 “certainly” had this system because “he has difficult in distinguishing neuter words from common-gender words”.7 As is well known, post hoc is not necessarily propter hoc. But importantly for our discussion of the possibility of geographical diffusion as an explanatory factor, Perridon (1997, p. 360) explores the interesting angle that these two

6 “helt ny genustype”. 7 “han har vanskelighed ved at skelne neutrumsord fra fælleskønsord”. Gender reduction in Bergen Norwegian: a North-Sea perspective 69 gender-simplification events, in Britain and in West Jutland, might be linked, in that the former may in some sense have exported it to the latter. Russ (1982) has made the same suggestion – the chronology makes it clear that, if there is a link, it would have to be that the change occurred in England or southern Scotland first and then spread across the North Sea to West Jutland. Perridon writes that, while it is well known that Old Norse had a major impact on Eng- lish, it has not been usual to suggest that “the dialect contact in the Danelaw might have had some consequences for the invaders as well”. But there are a number of similarities between West Jutlandic and English which might have this origin, and he specifically mentions the fact that they have both lost grammatical gender. He believes that it is possible that this change, amongst others, was “the result of extensive dialect contact in the Danelaw”. What then remains to be explained is why West Jutlandic was more influenced by English – or, more likely, by anglicised Old Norse – than other forms of Danish. Perridon does not mention the geographical proximity which is illustrated in Map 1: it is 334 nautical miles from Newcastle, England, to Es- bjerg, Jutland. But he says that any “explanation has to start with the fact that West Jutland mainly consists of barren moorlands, which in the past could hardly sustain its population” (Perridon 1997, p. 261). This led to a greater proportion of people from this area emigrating to Britain than from elsewhere, and “it is not unreasonable that those who returned to West Jutland after having accumulated some wealth in England” could have been, as successful re-mi- grants, rather influential. Nielsen (2007, p. 102) objects to Perridon’s thesis on the grounds that the linguistic outcomes were not the same in England and West Jutland. As I have noted above, however, they are not so very different because, in both cases, grammatical gender has been entirely lost and replaced by a noun-class system based on semantics. The difference simply lies in the fact that, like in English, the West Jutland classes are male, female and neuter – but the neuter class is additionally subdivided into mass and count categories. This is especially so since a not totally dissimilar development has actually also occurred in certain English dialects (Siemund 2008). Traditional dialects in the English southwest have a pronominal category difference between count and mass nouns, such that inanimate count nouns are pronominalised with he but mass nouns with it:

Pass the loaf – he’s over there. vs. I likes this bread – it be very tasty.

Nielsen also objects that it could just as well be the case that “the linguistic parallels between English and the Jutland dialect may have arisen independently”, “not least because the two idioms have a shared Germanic heritage which, to some extent, enables them to move in the same direction without being in direct contact” (2007, p. 102) – a clear reference to drift, as discussed above. 70 Peter Trudgill

But, again, how likely is it that we can explain the remarkably coherent geographical configuration of this phenomenon simply through drift? That drift was operative in all and only this contiguous area would in itself also seem to be too much of a coincidence. On the other hand, a spread from England to West Jutland, if this is what happened, and from there east and south to most of the rest of Jutland and down into Schleswig, would account for all the Danish, Low German and North Fri- sian gender loss areas indicated on the map. Then the fact that the two different innovating continental masculine–feminine syncretism areas – Schleswig Low German/Jutland Danish/North Frisian; and Low Saxon Low German/West Frisian/Dutch – are separated from one another by the small conservative three-gender Weser-Eider zone involving the Low German dialects of the Bremerhaven and Cuxhaven areas and Holstein opens up the possibility that the northern limit of this intermediate zone might represent the current southern limit of the spatial diffusion of gender loss out of West Jutland – and that perhaps conceivably as a result of earlier diffusion from Britain. Then the southern edge of the intermediate zone might equally well represent the northern limit of diffusion out of some focal point in the Hol- land/Friesland area, which is perhaps most likely to be the urban area of the northern Netherlands. As far as this Dutch/West Frisian/Lower Saxony area is concerned, moreover, there is an interesting convergence of chronology with Bergen and the other major Scandinavian cities. In all these cases the loss seems to have occurred in the period 1450–1650. Pedersen (1999) writes that in Copenhagen the masculine–feminine distinction was totally lost during the period 1450–1600. For Bergen, the timing appears to have been much the same. Nesse (2002) shows that, although there is an awareness of the distinctness of the two genders in the writings of Mester Absalon8 (1528–1575) in that he shows etymo- logically-correct usage of the pronouns hon and han for inanimates, he has already merged all other originally distinct forms. It was somewhat later that hon ‘she’ stopped being used to pronominalise original inanimate feminines and was replaced by han ‘he’. The simplification also began in Stockholm at around the same time. In the northern Netherlands the merger can also be seen to have been under way in the 1500s and to have become consolidated in the early 1600s (Geerts 1966). While the date for the first available texts showing the merger in West Frisian is not until the 1600s (Hoekstra 2001), this is almost certainly because there are no extant texts from the late 1500s. I have not been able to find any information about gender loss in the northwestern dialects of Low German mentioned above, but EFLG is spoken in an area which was Frisian speaking until the 1400s, when the Hanseatic League became dominant, and Frisian remained as a minority language there until the

8 Absalon Pederssøn Beyer, “Mester [Master] Absalon”, was a Bergen priest, teacher, and writer whose diaries from 1552–1572 provide invaluable historical data (Skard 1972, p. 32f). Gender reduction in Bergen Norwegian: a North-Sea perspective 71

1500s (Matras & Reershemius 2003). Many of the other relevant areas were also originally Frisian speaking – East Frisian was formerly spoken as far east as the river Elbe.

Conclusion The fact that all the relevant gender-loss varieties are located in more or less coastal areas suggests that what may be relevant here in differentiating between, say, coastal Dutch/Frisian/Low German which have the merger, and inland Low German/High German areas which do not, is the factor of differential accessibility to the geographical diffusion of linguistic innovations (Trudgill 1973). Wahrig-Burfeind talks of “das Phänomen der arealen Gerichtetheit des Genussynkretismus” (1989, p. 295) – there is a spatial directionality to the spread of the gender change. The implication is that the loss of gender is an innovation which has spread from one area to another even across language boundaries. It is important here to recognise that spatial diffusion takes two rather different forms. Long-term contact between geographically neighbouring communities can lead, through face-to-face interaction and accommodation (Trudgill 1986), to the diffusion of linguistic innovations gradually outwards from areas where they have already become established until they cover a wider and wider area. But at the same time, long-distance contact between urban centres also occurs as a result of communication and trading patterns, with movements of speakers from one place to another leading to the jumping of innovations from one urban area to another and the formation of exclaves such as Bergen (Chambers & Trudgill 1998). In one example of long-distance contact, we know that the Hanseatic trade brought large numbers of Low German speakers to Bergen over a long period of time. It is also known (Rambø 2008) that traders from Lübeck played an especially important role. Lübeck itself is situated in the Holstein three-gender area, but the two-gender Schleswig and Jutish-speaking areas were not far away – the distance from Lübeck to the Eckenförde-Tönning line is only about 50 miles/80 kilometres – and so it is not difficult to imagine that ships travelling out from Lübeck might well have been carrying a number of two-gender speakers. We also know that there were very many Danes in Bergen – Nesse (2002) cites figures showing that in the 200 years from 1550, more than a quarter of the foreigners resident in Bergen were from Denmark – and especially in the earlier years the West Jutland town of Ribe played an important part in North Sea trading activity. Rambø (2008, p. 329) also has it that, while the Germans were the largest group of foreigners in Bergen, there were other sizeable groups of foreigners, notably English, Scots and Dutch; and the considerable and long-term presence 72 Peter Trudgill of a Dutch population in Bergen up until the late 1500s is made clear in Wubs-Mrozewicz (2008). She discusses the role in Bergen of the Dutch “Hol- landers”, mostly from Amsterdam; and of “Overijsselers”, mostly from Kampen, Zwolle and Deventer (on former Frisan territory). Notice that the former group spoke northern Low Franconian and the latter group spoke Low Saxon – both groups thus bringing a two-gender system to Bergen with them. The “Dutch” presence also probably included actual Frisian speakers. The role of Groningen in Bergen also merits some careful consideration. As a semi-independent entity, it was the only city from what is now the Netherlands to be a member of the Hanseatic League; and it is situated in an originally Frisian- speaking and subsequently Low Saxon-speaking two-gender area. It would seem to be quite possible, in any case, that Bergen, as a highly important maritime trading city, acquired a two-gender system not only because simplification was brought about by imperfect adult second language and/or second dialect learning; and not only because gender mismatches and multilingualism brought about confusion about nominal gender membership, but also because of the geographical diffusion from continental North Sea Europe of gender loss as an innovation. This diffusion resulted from contact, in Bergen itself, between Bergen Norwegian speakers, on the one hand, and speakers of other Germanic varieties which already had two-gender systems, on the other. During the crucial decades when gender loss happened, there may well have been, walking around the streets of Bergen, enough northern Dutch, Frisian, Low Saxon and Jutish speakers – men who had actually brought their own two-gender systems with them – to enable us to consider that the geographical diffusion factor should take its place alongside the other better-known explanations for why Bergen only has two grammatical genders.

References

Arboe, Torben, 2008: Pronominal repræsentation i danske dialekter. In: Inger Schoon- derbeek Hansen & Peter Widell (eds.), 12. Møde om udforskningen af dansk sprog. Århus. Pp. 29–38. Bock, Karl Nielsen, 1933: Niederdeutsch auf dänischem Substrat. Studien zur Dialekt- geographie Südostschleswigs. Marburg. Bremmer, Rolf, 2009: An introduction to Old Frisian. Amsterdam. Chambers, J. K. & Peter Trudgill, 1998: Dialectology. Cambridge. Curzan, Anne, 2003: Gender shifts in the history of English. Cambridge. Davidson, Herbert, 1990: Han, hon, den: genusutvecklingen i svenskan under nysvensk tid. Lund. De Vogelaer, Gunther, 2009: Changing pronominal gender in Dutch: transmission or diffusion? In: S. Tsipaklou, M. Karyolemou & P. Pavlou (eds.), Language variation: European perspectives II. Amsterdam. Pp. 71–80. De Vogelaer, Gunther & Gert De Sutter, 2011: The geography of gender change. Pro- nominal and adnominal gender in Flemish dialects of Dutch. In: Language sciences 33. Pp. 192–205. Gender reduction in Bergen Norwegian: a North-Sea perspective 73

Geerts, Guido, 1966: Genus en geslacht in de Gouden Eeuw. Brussels. Goltz, Reinhard & Alastair Walker, 1990: North Saxon. In: Charles Russ (ed.), The dialects of Modern German. London. Pp. 31–58. Haugen, Einar, 1976: The Scandinavian languages. An introduction to their history. London. Heide, Eldar, 2003: Felleskjønnet i bergensk: resultat av mellomnedertysk kvantitetspå- verknad? In: Maal og minne. Pp. 23–35. Hoekstra, Jarich, 1996: Transitive pronouns and gender syncretism in Fering-Öömrang. In: NOWELE 27. Pp. 45–66. Hoekstra, Jarich, 2001: Comparative aspects of Frisian morphology and syntax. In: Horst Haider Munske et al. (eds.), Handbook of Frisian studies. Tübingen. Pp. 775– 786. Jahr, Ernst Håkon, 1995: Nedertysk og nordisk. Språksamfunn og språkkontakt i Han- satida. In: E. H. Jahr (ed.), Nordisk og nedertysk: språkkontakt og språkutvilking i Norden i seinmellomalderen. Oslo. Pp. 9–28. Jahr, Ernst Håkon, 1998: Sociolinguistics in historical language contact. The Scandina- vian languages and Low German during the Hanseatic period. In: E. H. Jahr (ed.), Language change: advances in historical sociolinguistics. Berlin. Pp. 119–133. Jahr, Ernst Håkon, 2001: Historical sociolinguistics. The role of Low German language contact in the Scandinavian typological split of the late Middle Ages. In: Lingua posnaniensis 43. Pp. 95–104. Jones, Charles, 1988: Grammatical gender in English 950–1250. New York. Jørgensen, Peter, 1954: Zum Schleswiger Niederdeutsch: Kritik und Forschung. Copen- hagen. Kusmenko, Jurij, 1996: Sprachkontakte und Sprachveränderungen in der süddä- nischnordfriesisch-niederdeutschen Kontaktzone. In: Walter Baumgartner & Hans Fix (eds.), Arbeiten zur Skandinavistik XII. Vienna. Pp. 65–83. Kusmenko, Jurij, 2000: Entwicklung des Genussystems in den skandinavischen Sprachen. In: Fritz Paul (ed.), Arbeiten zur Skandinavistik XIII. Frankfurt. Pp. 469– 478. Lakoff, Robin, 1972: Another look at drift. In: Robert Stockwell & Ronald Macaulay (eds.), Linguistic change and generative theory. Bloomington. Löfstedt, Ernst, 1968: Beiträge zu einer nordfriesischen Grammatik. I. Das Substantiv und das Adjektiv, das Zahlwort und der bestimmte Artikel. Uppsala. Matras, Yaron & Gertrud Reershemius, 2003: Low German: East Frisian dialect. Mu- nich. Nesse, Agnete, 2002: Språkkontakt mellom norsk og tysk i hansatidens Bergen. Oslo. Nesse, Agnete, 2005: Boken – han og kua – den. Om endringer i norske genussystem. In: Maal og minne. Pp. 136–145. Nielsen, Hans Frede, 2007: English and the Jutland dialect; or, the demise of a romantic notion. In: A. Wawn et al. (eds.), Constructing nations, reconstructing myth: essays in honour of T. A. Shippey. Turnhout. Pp. 97–108. Pedersen, Karen Margrethe, 1999: Genusforenkling i københavnsk. In: Danske folke- mål 41. Pp. 79–106. Perridon, Harry, 1997: Is the definite article in Jutlandic a borrowing from Low Ger- man? In: Multilingua 16. Pp. 351–363. Perridon, Harry, 2003: Language contact and grammatical change: the case of Bergen. In: Tijdschrift voor skandinavistiek 24. Pp. 235–256. Rambø, Gro-Renée, 2008: Historiske og sosiale betingelser for språkkontakt mellom nedertysk og de skandinaviske språkene i seinmiddelalderen – et bidrag til historisk språksosiologi. Kristiansand. 74 Peter Trudgill

Russ, C. V. J., 1982: Die Vereinfachung der Nominalflexion im Englischen und Ju- tisch-Dänischen. Ein Fall der gegenseitigen Beeinflussung? In: Orbis 29. Pp. 120– 129. Sadowska, Iwona, 2011: Polish: a comprehensive grammar. London. Sapir, Edward, 1921: Language. New York. Shetter, William, 1994: Dutch: an essential grammar. London. Siemund, Peter, 2008: Pronominal gender in English. A study of English varieties from a cross-linguistic perspective. London. Skard, Vemund, 1972: Norsk språkhistorie 1523–1814. Oslo. Skautrup, Peter, 1944: Det danske sprogs historie, I. Copenhagen. Skrzypek, Dominika, 2010: Interactions of grammatical categories: how three genders became two. In: Skandinavistik: European journal of Scandinavian studies 38. Pp. 91–106. Taeldeman, Johan, 2005: Oost-Vlaams. Tielt. Tiersma, Pieter Meijes, 1985: Frisian reference grammar. Dordrecht. Trudgill, Peter, 1973: Linguistic change and diffusion: description and explanation in sociolinguistic dialect geography. In: Language in society 3. Pp. 215–246. Trudgill, Peter, 1986: Dialects in contact. Oxford. Trudgill, Peter, 2000: On locating the boundary between language contact and dialect contact: Low German and continental Scandinavian. In: E. H. Jahr (ed.), Språkkon- takt: innverknaden frå nedertysk på andre nordeuropeiske språk. Copenhagen. Pp. 71–86. Trudgill, Peter, 2010: Investigations in sociohistorical linguistics. Stories of colonisa- tion and contact. Cambridge. Trudgill, Peter, 2011: Sociolinguistic typology. Social determinants of linguistic complexity. Oxford. Visser, Willem, 2011: Historical gender change in West Frisian. In: Morphology 21. Pp. 31–56. Wahrig-Burfeind, Renate, 1989: Nominales und pronominales Genus im südlichen Nordseegebiet: eine areallinguistische Untersuchung. Bremen. Wiesenhann, Tjabe, 1977: Einführung in das ostfriesische Niederdeutsch. Leer. Wubs-Mrozewicz, Justyna, 2008: Traders, ties and tensions. The interactions of Lübeckers, Overijsslers and Hollanders in Late Medieval Bergen. Hilversum. Norwegian and German in Bergen 75 Norwegian and German in Bergen

Agnete Nesse, University of Bergen

Bergen was a bilingual city for more than four hundred years due to the large number of Germans who lived there from the 13th century onwards. This bilingualism can be explored from many different angles. My aim in this article is to illustrate some of these possible approaches and the results of different studies. The research fields involved include dialectology, textual work, social history and language contact studies, all within a sociolinguistic frame.

Dialectology When studying the local vernacular of Bergen, one point always springs to mind, namely that in a Norwegian context, this dialect is an exception. Its combination of dialect features that are regarded as typically West Norwegian (WN) with others that are regarded as typically East Norwegian (EN) is interesting; even more interesting are the different simplifications found only in this dialect. For the sociolinguistically very salient words ‘I’ and ‘not’, Bergen has the WN forms /e:g/ and /içe/, whereas EN has /je/ or /jei/ and /ike/ or /ite/; the written standard used in Bergen (Bokmål) has and . The tendency to have a short vowel and long consonant in syllables that were short in Old Norwegian is regarded as an EN and northern feature: Bergen has /spik:eçøt:/ whereas WN /has spe:keçø:t/ (‘dried and salted mutton’). The fact that the short vowel is higher in Bergen (/i/ vs /e/) is also seen in the now obsolete dialect form /vik:e/ for WN /ve:ka/ (‘week’). In this case, the Bergen dialect now uses the EN form /u:ke/. The loss of the feminine gender in the Bergen vernacular represents a simplification. For example, ‘cow’ (fem. in all other varieties of Norwegian) and ‘ox’ (masc.) are treated the same grammatically: en ku – kuen ‘a cow – the cow’ and en okse – oksen ‘an ox – the ox’. Another simplification is the use of the singular adjectival form for plural predicate adjectives: whereas other Norwe- gian dialects use plural -e, as in di e store ‘they are big’, the Bergen vernacular has the singular form, di e stor. In Nesse (2002), five Bergen dialect features were chosen for closer scruti- ny. The questions asked were: how can these features be traced in the written sources over time, and could language contact with Low German account for the development of these features? 76 Agnete Nesse

Loss of the feminine gender Both Standard Danish and Standard Swedish merge the masculine and feminine gender, and since Norwegian Bokmål is based partly on Standard Danish and partly on (east) Norwegian dialects, the use of only two genders is also frequent in written Norwegian. Therefore this Bergen dialect feature is often regarded as standard, and its sociolinguistic associations are very different from other dialect features. There is an important difference between the Danish/ Bokmål and Bergen mergers, however, if, like Corbett (1991), we choose to look not only at phrase-internal but also phrase-external matters concerning gender. In both Danish and Bokmål, the merger has resulted in a common gender, and inanimate1 nouns are referred to by the common pronoun den, as opposed to the normal Norwegian dialect system where feminine nouns take ho (‘she’) and masculine nouns han (‘he’). In Bergen, however, all inanimate nouns that are non-neuter are referred to by han. Only when the pronoun is stressed is den obligatory. Therefore, the best way to analyse the merger between masculine and feminine gender in the different varieties is to say that the Bergen system has neuter and masculine gender, whereas the Danish/Bokmål system has neuter and common gender (Nesse 2005 gives a broader view of the different Nor- wegian gender systems). The differences between Bergen and Danish have to do with the dialect situation at the time of the merger. In Zealand, den was already replacing ho/han as the pronoun of reference (Pedersen 1999), while in West Norwegian the system using ho and han was still common when the merger took place. In other WN dialects, that still has three grammatical genders, ho and han are still used, even though den has become more frequent during the 20th century. The loss of the feminine gender in Scandinavian varieties involves several factors. One is that masculine is the default gender, another is that neuter is the most marked gender, so it is less likely that the neuter would be merged with the masculine. Today, it is held that the best explanation for the loss of the feminine gender is that even if language-internal factors explain the form the merger took, the fact that a merger took place in certain Scandinavian varieties is due to language contact with Low German from the Late Middle Ages onwards (Pedersen 1999; Jahr 1999; Nesse 2002, 2005, Enger 2011). Language contact situations where one or more of the varieties involved has grammatical gender will in most cases lead to changes in the gender systems – or at least in the realisation of the gender system (Weinreich 1953; Unterbeck et al. 2000). The reason behind this is obvious: gender is to a large extent redundant, and in most cases communication will not be hampered if a speaker uses the wrong gender for a noun. In an adult language learning situation, energy will be spent on other, less redundant language features. The different situations where a Scandinavian variety has been in contact

1 Most animals also belong to the inanimate group in these varieties. Norwegian and German in Bergen 77 with another language or languages – with the immigrant languages in the USA, the Finno-Ugric languages in Norway, Sweden and Finland, or with Ger- man – have all led to alterations in the gender system (Haugen 1953; Jahr 1984; Bull et al. 1986; Sandström 2000). The actual outcome depends on the varieties in contact, but the fact that these changes happen is a result of contact as such. The Low German that was used in Bergen from the 14th until the 18th century had three grammatical genders, but the realisations were not as distinct from one another as was the case with Middle Low German in the 13th century. For example, the masculine definite article der and the feminine definite article die had merged to de, and the masculine and neuter case endings were practically identical, which was also the situation in the Middle Low German described by Lasch (1914). In non-nominative cases, -en was a frequent ending in Low German, and the fact that -en was also frequent in Norwegian and used for definite singular masculines may also have contributed to the loss of the feminine gender in the Bergen vernacular. In the contact situation in Bergen, where mutual accommodation was a dominant feature, any surface structure that was the same in the two varieties in contact was favoured (see also the periphrastic genitive).

Past tense of weak verbs in -et Weak verbs of the first declension, like kaste (‘throw’) and hoppe (‘jump’), take -et in the past tense in the Bergen vernacular, both in the preterite and in participle: kastet and hoppet. This dialect feature is sociolinguistically marked, like the merger between feminine and masculine gender. Most Norwegian dialects have the ending -a in the past for first declension verbs, but -et is the most common ending in Bokmål (-a is possible but not frequently used), and thus -et is regarded as ‘standard’ and not ‘dialect’. As Jahr (1999, p. 127) has shown, the fact that Bokmål and the Bergen dialect have identical endings for this class of verbs is the result of two separate processes that accidentally led to the same result: in Bergen because of contact between Norwegian and Low German, and in Bokmål because of contact between Norwegian and Danish. There is no variety of Norwegian in which the preterite and past participle endings for this verb group are different, in contrast to in Danish, where the preterite is -ede and the past participle is -et. During the period when Danish was used in writing in Norway, failing to master this distinction was one of the most frequent mistakes made by Norwegian writers. Low German had a system identical to the Danish one, with different endings for the preterite and past participle. There are exceptions to this, since dialects of Low German both on the Danish and the German side of the border have apocope, and in some of these varieties -ede is reduced to -ed or -et as a result of “Auslautverhärtung” (Braun- müller 2003, p. 211). In the latter case, the Low German system was identical to the one emerging in Bergen. The important factor is that while Norwegian has had a system where the 78 Agnete Nesse two forms are identical, most Danish and Low German dialects have had a system with two different endings. It can be questioned, however, whether the Norwegian system with one ending had come into use during the Hansa contact period in Bergen; earlier, Norwegian had a system similar to that in Germany and Denmark, with -aði and -it as frequent realisations of the preterite and the past participle, respectively. The fact still remains that the only two varieties of Norwegian where the endings -et evolved are the only ones that are the result of contact between Norwegian and Danish/Low German. In the matter of verbs, it might be expected that language contact in Bergen would lead to an increase in regular verbs at the expense of irregular verbs. As will be shown in the section on the social conditions in Bergen, this contact was not very intense, even though it lasted for a long time. A more intense contact situation might have led to more changes to the verb system. One feature can be mentioned, even though it only concerns very few verbs.These are verbs that have two syntactic forms: one that takes a direct object and one that does not. Other Norwegian varieties distinguish between sitte (without an object) and sette (with an object), and between ligge (without an object) and legge (with an object): Jeg sitter (‘I sit’) / Jeg setter meg ned (‘I sit down’, literally ‘I sit myself down’). In the Bergen vernacular, the system is simplified: sitte and ligge are used in both functions – eg sittar meg ner (‘I sit down’), eg liggar ongene klokken syv (‘I put the children to bed at seven’). Whether this simplification in the system of the vernacular is due to contact with Low German or whether it is due to dialect contact between different Norwegian varieties in the city has not yet been examined, but both explanations are theoretically possible.

Periphrastic genitive The periphrastic genitive, which uses the reflexive pronouns sin (masc.), sitt (neutr.) and sine (plural), is, as far as I know, the only feature from the Bergen vernacular that has spread to most of Norway. In contrast to the features discussed above, it has not been considered proper in writing, but today we see an increased use of it in Nynorsk as well as in Bokmål, probably due to the fact that during the 20th century this feature spread to the prestigious East Norwe- gian dialects. Before that, this genitive was mostly used in West and North Nor- wegian, areas where Bergen – and the Bergen dialect – had an influence. Its origin in Bergen has not been disputed, nor has the fact that the construction was influenced by Low German. But what recent research has discovered is that the Norwegian periphrastic genitive is not a loan translation, but rather can be seen as a regrammaticalisation based on homophony, and that during the first centuries of its existence in the Bergen dialect, this variant had to compete with a loan translation which is now obsolete in Norwegian. The German periphrastic genitive construction often includes a dative, but this will not be discussed here, since it is of less importance to the Norwegian development, and since the written sources from Bergen show that this was not Norwegian and German in Bergen 79 obligatory in Bergen German. What is important is the fact that the Low Ger- man variant uses the personal pronouns sin (masc.), ehr (fem.) and ehr (plural). It is also important that the choice of pronoun is dependent on the possessor, and the pronoun ending is dependent on the possessed noun: Jochum is Clamer sin borge (‘Jochum is Clamer’s witness). Here, the masculine sin is used because Jocum is a man. When this construction was translated into Norwegian, the result was expressions like de theydske derris fremganng2 (‘the Germans’ prosperity’), where the personal pronoun deres (plural) agrees with the plural possessor. This construction was also used in Danish, and we can regard its occurrence in Norwegian texts as an influence from the Danish written language. The construction that survived in spoken Norwegian however, is less often found in the written sources. An early example is from a diary from 1562: Jacob Christiernson sine vidnesbyrd (‘Jacob Christiernson’s confessions) (Iversen 1963, p. 16). Here we see that instead of the personal pronoun hans, the author used the reflexive pronoun sin. In the Norwegian construction, the choice of pronoun is dependent on the possessed noun, yielding sentences like 1–3. 1) Jacob Christiernson (masc.) sin (masc.) bror (masc.) (‘J. C.’s brother’) 2) Jacob Christiernson (masc.) sitt (neutr.) hus (neutr.) (‘J. C.’s house’) 3) Jacob Christiernson (masc.) sine (plural) brødre (plural) (‘J. C.’s broth- ers’).

When trying to explain why this construction survived instead of the “correct” translated construction, we must look at the sociolinguistic conditions in Bergen at the time when both constructions were in use. The vast majority of those speaking German in Bergen were men, and when they talked about who owned what and who backed whom, they were talking about men (Nesse 1998, 2002). The Low German construction was therefore usually used in its masculine form, involving the personal pronoun sin. This pronoun was homophonous with the Norwegian reflexive (masc.) pronoun sin. It is therefore likely that the people of Bergen, who heard Low German more often than they read Danish (since very few of them could read), adapted the word sin to form a new genitive pattern. This construction provides good evidence of how a “change from below” (Elspass et al. 2007) can succeed at the expense of a “change from above” (the loan translation).

Declension of names In Norwegian dialects in most parts of the country, there is a way of signalling whether a person is known or unknown to the interlocutors. The most common

2 From Bergen Fundas (unknown author), written during the second half of the 16th century in Bergen (Sørlie 1957, p. 47). 80 Agnete Nesse way is to do this by placing a personal pronoun in front of the name. In example 4, at least one of the interlocutors is not familiar with who Grete is, whereas in example 5, Grete is known to all: 4) en som heter Grete blir med (‘one called Grete comes along’). 5) a Grete blir med (East Norwegian) / ho Grete blir med (North Norwe- gian) (‘she Grete comes along’).

The form a which is used in EN stems from an older accusative form hana (‘her’), while the North Norwegian ho (‘she’) stems from an older nominative form hon which is now used in all syntactic functions. In Bergen, however, the second sentence would be realised as 6: 6) Greten blir med.

The name Grete (fem.) is declined here in the same way as all masculine nouns and takes the ending -(e)n in the definite singular: bil – bilen (‘car’ – ‘the car’), flaske – flasken (‘bottle’ – ‘the bottle’). Both female and male first names can be declined in this way, but when it comes to surnames, usually only men are referred to with their surname in its definite form. Thus for sib- lings called Grete Friele and Hans Friele, both Greten and Hansen would be used, but only Hans would be called Frielen.3 This may have changed, however, with women being more likely to keep the same surname throughout their lives. On the other hand, surnames are used less than was common some generations ago. An investigation into the sociolinguistic distribution of the current use of declined names in Bergen is called for, concerning both who is more likely to use this feature and who is more likely to be referred to in this way. Names are also usually declined in German, as in most Norwegian dialects, by placing a pronoun in front of the name: die Grete, der Hans. So the exceptional way of declining names in Bergen is not due to direct influence from Low German and must be explained differently. The Low German diminutive suffix -ken has been investigated as an explanation, and if it had not been for the sociolinguistic circumstances in bilingual Bergen, this might have proved fruitful. But since German-speaking families were not common in Bergen (Nesse 2007), there were neither children nor wives to address with the diminutive. One could say that the idea that the diminutive form of first names was used among the Hanseatic merchants at Bryggen is as strange as the thought that these forms were common at a military camp. Therefore, it seems more likely that the way names are declined in the Bergen vernacular must be seen as an example of linguistic simplification. In- stead of dividing singular nouns into two groups as is done in other Norwegian

3 Pronounced /fʁi:ḷ/ as opposed to indefinite /fʁi:lə/. Norwegian and German in Bergen 81 dialects, distinguishing between bil – bilen and Grete – ho Grete, the system in Bergen contains only one noun group: bil – bilen and Grete – Greten. This result is very typical of language contact situations. However, this is not a language feature that is often found in the written sources, so it is not easy to discover when forms like Greten and Hansen were first used in Bergen. There are, however, many examples of occupational titles used instead of a person’s name, and these nouns are of course used in the definite as well as in the indefinite form. There is a possibility that the habit of calling a man Glassmakeren (‘the glass maker’) and Snekkeren (‘the carpen- ter’) instead of or in addition to their actual names eased the transfer to forms like Hansen, and later also Greten.

The infinitive marker te The dialect features discussed above are all still in everyday use in the Bergen dialect, even though, as mentioned above, not all Bergen speakers will use the definite form of names. The last feature discussed in this section, however, is on the verge of becoming obsolete, but its use in earlier stages of the dialect is known from the very thorough description of the dialect of Bergen from 1911– 1912 (Larsen & Stoltz). In several Germanic languages, the infinitive marker is homophonous with a preposition: English has to, German has zu and the Low German used in Bergen had to. In the Scandinavian languages, the system was different: the preposition til (often pronounced /te/) is distinct from the infinitive marker at/å. In Bergen, however, a system used to exist that was identical to that in the West Germanic languages, with te used both as the infinitive marker and a preposition. This can be analysed in a similar way to the past tense verb system above: the system in the dialect comes from one of the varieties in contact, while the form used comes from the other. In this case, the system of having a homophonous infinitive marker and preposition came from Low German, whereas the realisation – te – comes from Norwegian.

The sociolinguistic conditions The written sources that were used to find examples of the different dialect features through time have also been studied for their content in order to see what they can tell us about the society and the linguistic situation in Bergen. The external factors that serve as a frame for the linguistic situation can be very briefly summed up as follows (Nesse 2000a, 2002, 2003): From the 13th century onwards, Bergen became increasingly important to German merchants, and by 1360 a German settlement was established by the Hanseatic League in the area 82 Agnete Nesse called Bryggen in the centre of Bergen. The Norwegian king granted the merchants privileges, but their way of organising their lives in the city was to a large degree monitored not by Norwegian but by Hanseatic authorities. The policy was that of non-integration: Hanseatic merchants were not citizens of Bergen and could not take part in the administration of the city, nor would the local court handle Hanseatic matters. Only if there was a conflict between a Norwegian and a German merchant would the local court be involved; otherwise, the Hanseatic court at Bryggen dealt with internal conflicts, and also some trade-related conflicts involving Germans and Norwegians (Ersland 2011, pp. 47–56). From 1408 until 1766, there were one to three German churches in Bergen, paid for by the German merchants, and sermons were de- livered in German. Last but not least: the Hanseatic merchants in Bergen were required to be bachelors. Only after returning to Germany were they allowed to marry, so Bryggen in Bergen was a society of single men, the youngest of them only 14 years old. The Hanseatic settlement declined from the 17th century onwards, and by about 1750 it was no longer a reality. But the sociolinguistic patterns that developed during the height of the Hanseatic League lasted as long as there were German immigrants in Bergen. Only after 1800 did their numbers fall rapidly, with German apprentices being outnumbered by Norwegian at Bryggen. The increase in the Norwegian population after 1800 meant that immigration no longer was necessary, and for both Norwegian and German fortune seekers, new possibilities opened up in the USA. Due to the relative isolation of the Hanseatic merchants, primarily imposed by the Hanseatic League but also confirmed through the privileges granted by the Norwegian king, language contact in Bergen was never very intense. The effect that Low German had on the local vernacular must be due to the fact that contact lasted so long, and because there was contact between the groups in informal spheres. It seems clear that there was no need for the Germans to learn to speak or write Norwegian – all their matters were conducted in German. It also seems clear that the Norwegians did not have to speak German, since they never were forced to have anything to do with the Germans. Still, we know that need is not the sole reason for people adopting the language of another group. Power and prestige are key concepts here; if one group is far more powerful and/or prestigious than the other, the weaker group may have an interest in learning to use the stronger groups’ language in order to achieve the same power and prestige. In Bergen, for the most of the Hanseatic period the Germans had the strongest economy. On the other hand, they had no political power apart from what automatically comes with money. More importantly, they had no access to marrying into the local elite, which at times was very strong both economically and culturally, and was limited to a certain number of families. The few Han- seatic merchants who broke the rules of the Hanseatic League and got involved Norwegian and German in Bergen 83 in the local society were quick to start to use Norwegian in some domains, even if they continued using German as their working language. I have called this situation one of power balance in order to explain why we do not find sources written in the “other” language: those Norwegians who could write used Norwegian (later Danish), and the Germans, by far the most literate group, wrote in Low German (later in High German). But there is evidence that the members of both groups could read the other language. Corre- spondence between the Germans and Norwegians was bilingual, in that each wrote his own language. It is also striking to see that in “neighbours’ books” from Bryggen (presented below), the different authors wrote in the language they felt most comfortable with, obviously in the knowledge that choice of language did not matter, since both languages were understood by all and regarded as equal within this specific context. The sociolinguistic system in Bergen thus involved receptive bilingualism, where both Norwegian and German were understood by all people in the city. One might wonder if a mixed code developed; the answer to this is more a matter of definition than of linguistic facts. There were great numbers of loanwords in each direction (Nesse 2011; Brattegard 1934), and there were lexical items that must be regarded as hybrids. In translations at the time, we see that the similarity of the two languages is made use of extensively in order to render the translation as accurately as possible. I have not wished to call the language ‘mixed’, however, first of all because there are no instances where there is any doubt whether the author is writing Norwegian/Danish or a variety of German. Second, I have found no instances of code switching in the stricter sense of the word; that is, there are no instances where we see a change of language during a text or during a sentence. Within Meyer Scottons’ (2006) framework, however, where the term ‘code switching’ is also used when a loan word is “em- bedded” in the dominant language, it would be possible to argue that code switching is found in the written material from Bergen. If we look at the relationship between Norwegian and Danish on the one hand and between Low and High German on the other, the situation is very different. The Danish written in Norway was always to some extent marked by the Norwegian dialect of the writer, and Low German interference is common in the sources written in High German. So one could certainly find examples of mixed code from Bergen, which is not peculiar. Whereas our sources tell us that the inhabitants of Bergen indeed viewed Norwegian and German as two separate languages, and that which language you used determined which group you belonged to and thus which privileges you held, the differences between varieties of Norwegian and Danish, and between Low and High German, were not considered important. The term ‘Danish’ was used for both Norwegian and Danish, and the different German varieties were simply called ‘German’ (see Nesse 2002, pp. 96–104 for examples of what the different varieties were called). 84 Agnete Nesse

The textual approach Introduction The texts in Norwegian, Danish, Low German and High German from the period 1350–1936 that I have been working with are not only valuable sources for linguistic examples and an understanding of the society where these languages were used. They are also historical documents that give insights into text types, literacy and writing skills for the period in question. Most of these texts are unpublished. In addition, most of them only exist in one copy, either as an original or a copy. Most of the texts are dated, and the author is known. The authors are, with extremely few exceptions, men, and most of them are merchants, though texts by priests and other scholars also do occur. Today these manuscripts are situated in four different archives: the city archive of Lübeck, the Royal Danish Library, the library of the University of Bergen, and the city archive of Bergen. Those that have been published were for the most part published between 1850 and 1900 by Norwegian historians. This means that attention must be paid not only to the texts themselves, but also to the way they were edited and published.

Translations Unsurprisingly, since Bergen was a bilingual city, not many translations between Norwegian/Danish and German were carried out. There were, of course, numerous translations from German into Danish made in Denmark, and these were also disseminated in Norway, but it is not known whether these translations were used in Bergen, and they will therefore not be discussed here. I will concentrate on the two major works of translation that were carried out in Bergen. Why were these translations made, and what can they tell us about the languages in contact? The first is a group of real estate documents that were translated from Nor- wegian into Low German in the 1550s (Nesse 2008, 2009, 2010). The original papers have been lost, so what we have are copies of the Norwegian documents, written by the same hand as the translations. From the handwriting it is possible to suggest that the writer (and translator?) was the secretary of the Hanseatic settlement in Bergen (Bruns 1939, pp. 54–55 gives handwriting samples for all the secretaries). At the time, there was a threat of increased payments for property that the Germans rented in Bergen. There may thus be two reasons why the translations were needed under these circumstances. One is that some of the original documents were written in an older form of Norwe- gian, and we have evidence from other sources that this was difficult for the Germans to read. More important was the fact that the increased payment might have led to a major conflict, and the Hanseatic authorities in Lübeck, who did not necessarily read Norwegian, might have become involved. Thus there were juridical reasons to have these translations. Norwegian and German in Bergen 85

We suppose, then, that the translator was German. One interesting fact is that the Norwegian documents have many more abbreviations than the Ger- man. The translator may have felt uncertain about Norwegian grammar, especially if it was an older stage of the language than the one he knew from the streets of Bergen and contemporary writing, and this may be why he chose to leave out parts of the Norwegian words. Even more extraordinary is the likeness in syntax between the two versions. In two of the documents, the word order has not been changed at all! This was possible since both languages had more flexible syntax than their modern versions have, especially when it comes to the position of the auxiliary. Still, it is very clear that the translations were kept as close to the original as possible to avoid any dispute due to deviant for- mulations. In contrast to this is the other major translation, which was of Bergen Fun- das, the first historical record of Bergen. It was written in Dano-Norwegian in the 1550s in Bergen, and the first translations were made before 1600. Bergen Fundas exists in 22 different manuscripts, half in Dano-Norwegian and half in German. There are two Dano-Norwegian editions (Nicolaysen 1858 and Sørlie 1958) and one edition of the German translation (appendix to Nesse 2002). The aim of the German translation was not to ensure legal accuracy, and the differences between the language of the Dano-Norwegian and German versions are much greater than in the translated real estate documents discussed above. All but one of the German manuscripts are now in Denmark, and that may not be a coincidence: the translations may have been made not for the Hanseatic merchants in Bergen but for the German- speaking elite of Denmark. All the Dano-Norwegian manuscripts of Bergen Fundas end with the dispute between the Hanseatic merchants and Christoffer Valckendorff, the King’s representative in Bergen between 1556–1560. The German manuscripts, however, continued to be updated after this; the last copy must have been made as late as 1670, because the history continues up to that date. It was quite common for scribes to feel free to add important facts up until their own time, but it is not easy to explain why only the German scribes did this. As we shall see below, this has led to quite curious editing of Bergen Fundas.

Texts in German My studies of Norwegian and German in Bergen have all been carried out within the frame of Norwegian language history. That is to say, the German language as such was not investigated apart from what it could say about the language contact situation (loanwords, translations, some grammatical constructions). The two German texts presented here, are very different from one another, and the way they have been edited and published show us how differently they have been regarded as historical documents. 86 Agnete Nesse

A dialogue from the 15th century A dialogue often referred to as “de dudesche unde de norman” is structured as a conversation between a German merchant and a Norwegian in Bergen. The text is based on a charter issued by King Kristoffer in 1444 (Taranger 1912, p. 251), but where the charter simply states what the Germans may and may not do, the dialogue is constructed so that the Norwegian makes accusations against the merchant, and the merchant defends himself. The background for the charter lies in several letters of complaint by both German merchants and Norwegians regarding their co-existence in Bergen. The dialogue is believed to have been written in Lübeck between 1512 and 1538 (Bruns 1901, p. 142). It stands in a long tradition of using dialogues as didactic and exemplar texts. But in spite of what has been claimed, for example by Nielsen (1877, p. 34), this text is more balanced than other such texts: it is not evident whether it is the merchant or the Norwegian who deserves the sym- pathy of the reader. Linguistically, it is interesting to see which Norwegian loan words are used. There are quite a few of them, and all are of legal importance in this specific context, like oltap (‘beer tap’). There was, of course, a German word corresponding to this,4 but since the selling of beer was regulated in Bergen and there was a dispute surrounding it, the Norwegian technical term was chosen. This text exists in only one manuscript, but it was published three times between 1877 and 1912. The editions are quite similar to one another, and for the most part are accurate. A thorough comparison between the charter and the dialogue could probably shed new light not only on the conflicts in Bergen at the time, but also on the accommodation between Norwegian and Low Ger- man. A Norwegian translation has been carried out, but is not yet published. In time it will provide interesting information, especially to historians.

Politics, religion and astronomy The other German text worth mentioning here is about one hundred years younger than the dialogue, written at the end of the 16th century in Bergen. It exists in several manuscripts, but has not yet been published in full. An abbre- viated version translated into Danish was published by Nicolaysen (1868), but it fails to give a full picture of this very remarkable text. It is called Die norsche saw (‘the Norwegian sow’), even though the pig is not central in the book. But the sow serves to depict the decadence that, according to the unknown author, was typical of Bergen at the time. In addition, a Norwegian poem from 1583 about a monster pig born in Akershus near Oslo is quoted in the text. Politically, the text is interesting in its harsh criticism of the exploitation of Norwegian fishermen and farmers. Because of this treatment, it is claimed,

4 In fact, King Kristoffer’s charter uses the term bior tapp. This means that the author of the dialogue either had more sources than this charter for the dialogue, or that he knew the Bergen Ger- man variety well and knew that oltap was the word commonly used in the city. Norwegian and German in Bergen 87 they will never become free to use their potential, and the country will remain poor. As well as Norwegian and Danish officials, the German merchants are criticised. There is not much religious tolerance evident in the text; it expresses an intense hatred towards non-Lutherans, especially Catholics and anabaptists. The decadence of Bergen society is criticised, and the immoral behaviour of several “good” citizens is mentioned, but their names are written in code. As was quite common at the time, astronomy was closely tied to religious beliefs. Sky formations and the like were interpreted as God’s signs to men, and many pages of the text are filled with such speculations. Dialogue is also used in this text to make certain issues clearer. But the di- dactical form that dominates the text is one often used in religious texts, namely a parable followed by an explanation. A more suitable title for the text would be “the Norwegian hen-basket”, since the Norwegians are compared to hens trapped in a basket instead of being free-range. The author used different breeds of chicken to illustrate the different social classes of the society. Linguistically, the text is interesting mainly because of the type of German it employs, which can be described as High German with many Low German elements in it. For a study of the history of Norwegian, the most interesting aspect is the poem about the monster pig, which is quoted in Norwegian and not translated into German. Thus the author must have written the book for bilingual readers. This poem is, as far as I know, the only example we have of written Norwegian with heavy German interference. The writer must have known the poem orally and wrote it down using his – German – spelling rules, varied as they were. Furthermore, the different scribes who copied the text chose different solutions when spelling the Norwegian words of the poem, giving us an idea of how they perceived spoken Norwegian (Nesse 2009, pp. 124–126). Finally, one can ask why a text that is so critical of the Hanseatic League came to be written by a German in Bergen. Wouldn’t the Germans stick together? Probably not. Only those belonging to the Hanseatic League had the special privileges – and limitations – of Bryggen society bestowed on them. Germans who were not part of this were treated as other foreigners were, Eng- lishmen, the Dutch, etc. Furthermore, Hanseatic merchants who chose to leave the League were punished severely, and this of course might create bad feelings towards Hanseatic society. This means that in this mixed city, it was possible to be a Norwegian patriot in the German language.

Texts in Norwegian Most of the texts that form the basis for the different studies of Norwegian and German in Bergen are written in Norwegian or Danish, or any conceivable mixture of the two varieties. From the oldest period we have mostly charters, while later on we find letters, laws, books, book keeping accounts and other kinds of texts. There are four texts that should be mentioned here, since they have been used more extensively than the others. 88 Agnete Nesse

History Bergen Fundas, mentioned above, is the first history of a city that was written in Norway. It has been seen as a product of the so-called Bergen humanists, who were a group gathered around the king’s representative at the castle of Bergen. But it might also be seen as evidence that a self-conscious bourgeoisie was evolving in Bergen, a group of people who had their identity tied to their trade and to their city, as much as to their king and their church. It is written in the Dano-Norwegian so typical of the 16th century, a variety that is as Norwegian as it is Danish, and marked by an extensive variation in spelling. The history of Bergen is described from the time when the area that later became the city was pasture land, and starts with a magic tale: One day the shepherds heard strange sounds, as if many people were speaking, and when they told their master of this, he prophesised that a rich and mighty city would arise on the pasture land. It is not mentioned that this was a message from God, but there is a clear resemblance to texts like die norsche saw and Ab- salon’s diary.

Absalon’s diary The diary of the priest and teacher Absalon Pederssøn is one of the best-known Norwegian texts from the 16th century. The language resembles that of Bergen Fundas, but large parts of the diary are written in Latin, whereas Bergen Fun- das does not contain Latin. The use of Latin in Bergen during the Hanseatic era has not been investigated, apart from some observations of code shifting in letters written by Absalon’s mentor, the bishop Geble Pederssøn (Nesse 2009). Analysing the entries in order to find out why Absalon chose which language would probably yield insights about how Latin was perceived. It would be interesting to see how the bilingual priest used the two languages, if his bilingualism was complementary in that some of the expressions related to religion only existed for him in Latin. Since the diary to some degree is concentrated around Absalon’s life, it is a very valuable source of knowledge about social patterns in informal life in Bergen. He lists the different people who were invited to baptisms and wed- dings, and he is the most cited source when it comes to contact between the Hanseatic merchants and the Norwegians, at least for the upper social layers of both groups.

Court protocols from the 1590s and the 1660s The two oldest court protocols from Norwegian Bergen were written in the 1590s and 1660s. When reading them, the separation between Norwegian and German society in Bergen really springs to mind. The pages are as good as empty of German merchants; if one read only these books, one would think this was a pure Norwegian city. The reverse case holds for the court protocols from Bryggen, which will not be mentioned here. Norwegian and German in Bergen 89

The content of the two Norwegian protocols is similar, apart from the great number of witch trials in the older protocol, which is absent from the younger one. Linguistically, we can see how much the written language changed from the 16th to the 17th century. In the latter, there is still much variation, but it is within a narrower range, is less influenced by Norwegian dialect, and is more in line with written Danish; it is still not codified as a fixed norm, but it is on its way. In these protocols, all kinds of people appear, with their names, occupations and their disputes listed. Women are present to a much larger degree than in other sources; they dominate the witch trials, both as the accused and as witnesses, but also in other cases they stand up as witnesses, bypassers or accusers. And since it was important to state exactly what people said in court, the protocols have many quotations that, even though they are not written in anything like pure dialect, at least when it comes to vocabulary and partly when it comes to syntax, give an idea of what the spoken language of the people was like.

Texts in both languages When studying language contact, texts that display examples of the different varieties in context are of great importance for understanding the kind of bi- or multilingualism in the community in question. From Bryggen in Bergen we have a large corpus of “neighbours’ books” (Nesse 2012a), consisting of about 20 books from 10 different housing complexes. In these books, everything related to life in the housing complex was recorded – from the very practical division of payment for the upkeep of the buildings, fire protection and taxes, to rules for behaviour (when and where smoking was allowed, church duties), and last but not least records of employment. The oldest of these books starts in 1529, and the last ends in 1936. During this period, the language of the books changes first from Low German to High German, then from High German to Danish, and finally from Danish to Norwegian. All the writers are merchants with varying levels of education, and the writing skills range from very basic to rather elaborate. The three language shifts that these books display show us that the gradual shift from one variety to another took place in different ways. The first shift, from Low to High German, occurred during the 17th century, and can be described as a gradual process where more and more High German elements were introduced, until around 1700 we find that there are few Low German elements left in the written language. The secretaries of the Hanseatic settlement in Bergen, most of whom held a law degree from a German university, wrote High German in all external correspondence from 1580 onward. However, in texts meant for internal use the same secretaries wrote Low German long after that and they distinguished between the two written varieties. The common merchants did not seem to have this awareness. Instead, they mixed the two: Low German lexical items like ik and dat are replaced by High German forms 90 Agnete Nesse ich and dass, but the grammar remains Low German. For example, the Low German merger of der (masc. nom.) and die (fem. nom.) to de results in the writing der (nom.) in front of feminine nouns. The change from High German to Danish took place around 1800. The first Danish entries in the neighbours’ books dates from the 1770s, and the last Ger- man entry is from 1820. This language shift does not involve a mixing of the two varieties; rather, it came about because an increasing number of merchants – after 1750 they were no longer subjects of the Hanseatic League but of Bergen – had Norwegian as their first language and Danish as their written code. The immigrants from Germany, however, continued to write in German – at least in some domains – because of the tradition at Bryggen. They were not themselves responsible for Bergen bilingualism, but they came to a city where this bilingualism was well established and profited from it. This established pattern may, of course, have been one of the reasons immigration lasted so long after the Hanseatic League faded away, since it made the move easier for both young apprentices and adult merchants. But even though the German immigrants of the 18th and 19th centuries wrote in German in the neighbours’ books and in their private diaries, they must have learned Norwegian/Danish in order to join the various social and formal or- ganisations of the city. In these settings, all writing was in Danish only. An example is provided in the minutes from the shooting association of Bergen, which had many German merchants as members, where everything was written in Danish. The final change from Danish to Norwegian occurred around 1900. Like the switch from Low to High German, it happened through mixing, but with one large difference: while those who altered their written code from Low to High German still spoke Low German and thus started writing in a variety different from their spoken dialect, those who switched their written code from Danish to Norwegian all spoke Norwegian and thus began writing in a variety more similar to the one they spoke. The neighbours’ books can be regarded as semi-private writing; they were not written for publication or archives, but rather were intended to be read by a group of ordinary people who shared responsibility for a group of buildings. Whether these people were subjects of the Hanseatic League or the Danish, Swedish or Norwegian king did not really make a difference to the way this material was written. The texts remained remarkably stable through the centuries, which makes them excellent sources for diachronic research, whether on linguistics or history. The reason why the books ceased to be produced in the 1930s is partly due to the invention of the typewriter, but most of all because of a change in the way the ownership of such buildings was organised. If only one man owned the whole housing complex, the need for a book that each owner wrote in and signed no longer existed. Norwegian and German in Bergen 91

Editorial practices The texts that form the basis for all the different studies, apart from one, were not printed in their time. Those that have been edited and published were worked on several centuries after they were written. The others remain as hand-written manuscripts. This has significance, of course, for the choice of methodology, and most likely for the outcome of the studies. Working with hand-written material is time-consuming, but on the other hand, what you read is what was written, and not the interpretation of an editor. Both during attempts to edit hand-written material (Nesse 2002, 2008, 2009) and while working with editions produced by others, the question of how editing should be done has arisen. There is a large literature on this (Nesse 2012b), and different scholars have different views. Linguists have different needs from those of historians, even though historians may also end up reading a document in the wrong way if a comma is put in the wrong place. Ersland (2011, p. 18) shows, for example, that an editor’s insertion of a comma has led to a false understanding of where the market place in Bergen was situated in the 1500s. When the language of the original text, as in the case of Low German, is not known to the potential readers – Norwegian historians and linguists – the challenge is of a different type than questions concerning punctuation and capital letters. The issue of translation comes up, and whether the translation should be published on its own or along with the original version. A change in research questions leads to different requirements from the editing process. Solutions that seemed correct 150 years ago can now create more misunderstandings than the editor could imagine. The editions of Bergen Fundas by Nicolaysen (1858, 1868) can serve as a brief example. The Norwegian version of Bergen Fundas was published in 1858. As mentioned above, all the Norwegian manuscripts end in the 1550s, and so does the edited version. The German translation was not considered important enough to publish at the time, apart from the expansions from 1550–1670. These existed only in the German versions, and since they contained information about Bergen society that could be found only in these German texts, they needed to be published. But how? The solution was to translate them into Danish and give them a separate name: Mikjel Hofnagels optegnelser (‘the records of Mikjel Hofnagel’). Modern readers who do not pay special attention to the footnotes will believe that this was a text written in Danish by a man called Mikjel Hofnagel, and indeed, many writing about the history of Bergen have done so, and quoted the imaginary Hofnagel in pure Danish. This is unfortunate, but the choice made in 1868 can be understood in light of the importance of keeping the original Bergen Fundas “clean” – only the core text deserved to go by that name. Still, the editions from the 1800s are valuable to us in many ways, even if they need to be used with care. 92 Agnete Nesse

Concluding remarks Applying theoretical and methodological tools from different disciplines has been fruitful in investigating the bilingual language history of Bergen. At the same time, it will be evident that the wide scope that has given a very good overall view of the linguistic situation is not the only possible approach. To choose one specific corpus and have it digitalised in order to carry out quantitative analyses is a possibility that clearly would make important data accessible to further interpretation. I have done this only to a very limited degree (Nesse 2002, pp. 206–210), and more of this kind of work would be welcomed. At the other end of the methodological scale, it is clear that more editions of the existing material are needed in order to involve other researchers in the field of the linguistic and textual history of Norway. The texts from Bergen influenced other places in Norway, especially in the west and the north, and these texts should be made available to those who do not have the time or the skills to work with the handwritten manuscripts in the different language varieties used in the city. The research field of language contact in Bergen over the ages is wide, and there are many tasks still to attend to. Most critical in my view is to reopen the investigations of German that were started in the works of Olav Brattegard in the 1930s and 1940s. A sociolinguistic approach to the German language of Bergen, compared to the language in the northern part of Germany, could give further insights not only into the Bergen German language as such, but also into language contact between Norwegian and German.

References

Brattegard, Olav, 1934: Einige norwegische wörter in mittelniederdeutschen han- seatischen texten. In: Norsk tidsskrift for sprogvidenskap VII. Pp. 278–285. Braunmüller, Kurt, 2003: Agnete Nesse: Språkkontakt mellom norsk og tysk i hansatidens Bergen. Dr.art.-disputas, Universitetet i Tromsø, Det humanistiske fakultet, 11. mai 2002. In: Norsk lingvistisk tidsskrift 21. Pp. 195–215. Bruns, Friedrich, 1901: Norweger und Deutsche zu Bergen. In: Hansische Geschichts- blätter. Pp. 142–152. Bruns, Friedrich, 1939: Die sekretäre des deutschen kontors zu Bergen. Bergen. Bull, Tove, Junttila, Jorid Hjulstad & Pedersen, Aud Kirsti, 1986: Nominalfrasen i ski- botnmålet i Troms. In: Norsk lingvistisk tidsskrift 1−2. Pp. 60−71. Corbett, Greville, 1991: Gender. Cambridge. Elspass, Stephan, Langer, Nils, Scharloth, Joachim & Vandenbussche, Wim (eds.), 2007: Germanic language histories ‘from below’ (1700–2000). Berlin/New York. Enger, Hans Olav, 2011: Gender and contact – a Natural Morphology Perspective on Scandinavian Examples. In: P. Siemund (ed.), Linguistic Universals and Language Variation. Trends in Linguistics Studies and Monographs 231. Berlin. Pp. 171–203. Ersland, Geir Atle, 2011: Das Kaufmannshaus. Det hanseatiske kontorets rettslokale og administrasjonshus i Bergen. Bergen. Norwegian and German in Bergen 93

Haugen, Einar, 1953: The Norwegian language in America: a study in bilingual behavior. Philadelphia. Iversen, Ragnvald, 1963: Absalon Pederssøn. Dagbok og oration om mester Geble. Tekstbind. Oslo/Bergen. Jahr, Ernst Håkon, 1984: Language contact in northern Norway. Adstratum and substratum in the Norwegian, Sami and Finnish of northern Norway. In: Acta borealia 1. Pp. 103–112. Jahr, Ernst Håkon, 1999: Sociolinguistics in historical language contact: the Scandina- vian languages and Low German during the Hanseatic period. In: E. H. Jahr (ed.), Language change. Advances in historical sociolinguistics. Berlin/New York. Pp. 119–140. Larsen, Amund B. & Stoltz, Gerhard, 1911–1912: Bergens bymål. Kristiania. Lasch, Agathe, 1914: Mittelniederdeutsche grammatik. Halle. Myers-Scotton, Carol, 2006: Multiple voices. An introduction to bilingualism. Malden. Nesse, Agnete, 1998: Mellom lån og hjemlig utvikling. Den såkalte garpegenitiven som språkkontaktresultat. In: E. H. Jahr (ed.), Språkkontakt i Norden i middelalderen, særlig i hansatida. København. (Nord 4.) Pp. 121–138. Nesse, Agnete, 2000a: Språksosiale forhold i Bergen i hansatiden. In E. H. Jahr (ed.), Språkkontakt – Innverknaden frå nedertysk på andre nordeuropeiske språk. Køben- havn. (Nord 19.) Pp. 203–216. Nesse, Agnete, 2000b: Er det mulig å kombinere sosiolingvistikk og filologi? In: Mot- skrift 1. Pp. 108–121. Nesse, Agnete, 2002: Språkkontakt mellom norsk og tysk i hansatidens Bergen. Oslo. Nesse, Agnete, 2003: Written and spoken languages in Bergen in the Hansa era. In: K. Braunmüller & G. Ferraresi (eds.), Aspects of multilingualism in European language history. Amsterdam/Philadelphia. Pp. 61–84. Nesse, Agnete, 2005: Boken – han og kua – den. Om endringer i norske genussystem. In: Maal og minne. Pp. 136–146. Nesse, Agnete, 2007: 1750–1850: The disappearance of German from Bergen, Norway. In: S. Elspass, N. Langer, J. Scharloth & W. Vandenbussche (eds.), Germanic language histories ‘from below’ (1700–2000). Berlin/New York. Pp. 423–436. Nesse, Agnete, 2008: Bilingual texts from a bilingual city. In: G. A. Ersland & M. Treb- bi (eds.), Neue studien zum archiv und zur sprache der Hanseaten. Bergen. Pp. 47– 64. Nesse, Agnete, 2009: Flerspråklige kilder fra Bergen. In: M. Lethi-Eklund, C. Wide, G. Harling-Kranck, S. Tiisala & M. Lamberg (eds.), Folkmålsstudier XXXXVII. Pp. 109–132. Nesse, Agnete, 2010: Norwegisch-niederdeutsche sprachkontakte. Zweisprachige quellen aus dem Hansekontor Bergen. In: Niederdeutsches jahrbuch 133. Pp. 87– 104. Nesse, Agnete, 2011: Hanseatene og språket. In: G. A. Ersland (ed.), Das Kaufmanns- haus. Det hanseatiske kontorets rettslokale og administrasjonshus i Bergen. Bergen. Pp. 41–46. Nesse, Agnete, 2012a: Four languages, one text type. The neighbours’ books of Bryggen 1529–1936. In: M. Stenroos, M. Mäkinen & I. Særheim (eds.), Language contact and development around the North Sea. Amsterdam/Philadelphia. Pp. 81– 98. Nesse, Agnete, 2012b: Editorial practices and language choice: ‘Low German language monuments’ in Norway. In: N. Langer, S. Davies & W. Vandenbussche (eds.), Lan- guage and history, linguistics and historiography. Inderdisciplinary Approaches. Oxford. Pp. 111–126. Nicolaysen, Nicolay, 1858–1868: Norske magasin I–II. Christiania. 94 Agnete Nesse

Nielsen, Yngvar, 1877: De dudesche kopman unde de Norman. In: Forhandlinger i Vi- denskabs-Selskabet i Christiania Aar 1876. 8. Christiania. Pedersen, Karen Margrethe, 1999: Genusforenkling i københavnsk. In: Danske folke- mål 41. Pp. 79–105. Sandström, Caroline, 2000: The changing system of grammatical gender in the Swedish dialects of Nyland, Finland. In: B. Unterbeck, M. Rissanen, T. Nevalainen & M. Saari (eds.), Gender in grammar and cognition. Berlin/New York. Pp. 793–806. Sørlie, Mikjel (ed.), 1957: Bergens fundas. Bergen. Sørlie, Mikjel (ed.), 1962: Mattis Størssøn. Den norske krønike. Oslo/Bergen. Taranger, Absalon (ed.), 1912: Norges gamle love, 2. rekke, 1388–1604 a Tekst. Chris- tiania [Oslo]. Unterbeck, Barbara, Rissanen, Matti, Nevalainen, Terttu & Saari, Mirja (eds.), 2000: Gender in grammar and cognition. Berlin/New York. Weinreich, Uriel, 1953: Languages in contact. Findings and problems. New York/The Hague. Semi-communication and beyond 95 Semi-communication and beyond: some results of the Hamburg Hanseatic Project (1990–1995)

Kurt Braunmüller, University of Hamburg

Introduction At the beginning of the 1980s, researchers once again became more interested in how the Mainland Scandinavian languages were influenced by the impact of (Middle) Low German, the language of the Hanseatic League. But the focus of interest in language contact was still primarily on the lexical loans from Low German and sometimes on the borrowed word-formation patterns as well (see Ureland (ed.) 1987). Both of these aspects were quite obvious and could hardly be overlooked (cf. the survey in Törnqvist 1977 for Swedish). Structural, grammatical or typological investigations in general were, however, rare or absent, though they are crucial for any deeper understanding of the development and typological restructuring of the Mainland Scandinavian languages. In Copenhagen in 1987, I presented for the first time an outline of my ideas, which were published two years later in the proceedings (Braunmüller 1989). In the section entitled “Semikommunikation in Skandinavien und im Ostsee- raum zur Zeit der Hanse?” [Semi-communication in Scandinavia and in the Baltic during the time of the Hansa?], I suggested a new research project, the main purpose of which was to investigate the hypothesis that there was a form of communication in the late Middle Ages similar to the way people in the Mainland Scandinavian countries communicate with each other today, a phenomenon which has been called ‘semi-communication’ (see Haugen 1966). This term is quite misleading, however, at least for native speakers of English, since it supposes that one is only able to understand half of what has been said. In fact, it actually means ‘receptive multilingualism’, and can be traced back to a short dictum in Hockett’s popular introduction to structural linguistics in the 1950s and 1960s (retrieved again by Börestam Uhlmann 1997, p. 240f.). The misleading term ‘semi-communication’ should therefore, in principle, be replaced by the more appropriate term ‘semi-bilingualism’: Among educated Danes and Norwegians, however, communication is quite unimpeded: each speaks his own personal variety of his own language, but has learned by experience to understand the speech pattern of the others. The result may be called semi-bilingualism: receptive bilingualism accompanying productive monolingual- ism. (Hockett 1958, p. 327.) 96 Kurt Braunmüller

Haugen (1966) himself, however, speaks rather vaguely of some ‘code noise’ for the divergence between speakers of different varieties. But the term ‘code noise’ is not at all clear, nor does it cover the fact that the linguistic codes involved must be genetically more or less closely related, which is indispensable for any proper treatment of the linguistic situation in Scandinavia. Attitudes and other social factors did not get the attention they deserved either. One may also view this kind of direct communication between speakers of genetically related codes as an interdialectal form of communication which is quite common between speakers of varieties within the same linguistic ‘diasystem’ (a term coined by Uriel Weinreich in 1954). If one takes this point of view, one has to disregard the fact that the northern and western Baltic Sea languages, Old Danish, Old Swedish and Middle Low German, were considered Ausbau languages – which they actually are, to use Heinz Kloss’ term (cf. Kloss 1975) – but rather consider them mutually intelligible varieties that are part of an over-arching (pan-Baltic) diasystem, and not as national languages in their own right. Old Norwegian and its dialects belong to this scenario as well, but cannot be considered languages/varieties of the Baltic area in the strict sense of this term. The hypothesis that ‘semi-communication’ was possible between the pan- Baltic varieties and beyond (in Norway) in the Middle Ages later became the starting point for a smaller research project, financed by the German Research Foundation (DFG) in the years 1990–1995. The three researchers were Dr Wil- ly Diercks (full-time), a specialist in the history of Low German (and a native speaker of this language himself), Gabriela Niepage (part-time), who had worked before with Middle Low German texts from a philological point of view at the University of Hamburg, and myself (part-time) as the principal in- vestigator with expertise in the history of the Scandinavian languages, plus as- sistance from students in the Scandinavian Department. Unlike other on-going research projects in Scandinavia and Germany, our main focus was on sociolinguistics and contact linguistics, and not on the philological aspects of text analyses or on lexical borrowing from Low German as such. Moreover, our project was the first one which used the new personal computer technology wherever possible, and a (former) student, Per Warter, developed a special morphological parsing programme called MANAL (see the bibliography for further references). With this programme it was, at least in part, possible to analyse morphological items of the target language using an algorithm (originally) developed for the source language, e.g. using the Low German version of MANAL for an analysis of Danish or Swedish word structures, simulating parts of the receptive morphological decoding process of this parent language and thus making semi-communication understandable. I still consider this idea a hypothesis, which became very plausible through our research in this field but which never can be proved with absolute certainty due to the lack of meta-linguistic data, e.g. comments on how people communicated with each other in the (early) times of the Hanseatic Semi-communication and beyond 97

League. We only know for certain that Latin texts were translated, if necessary, for those who were not able to understand this language (primarily un- educated people). But as far as I know, there are no notes or references attest- ing that utterances or texts in the Germanic languages were translated in Han- seatic cities during the Middle Ages and beyond. Evidence has been found of translations at much later times, however, when High German replaced Low German as a medium of documentation. But this did not happen before the second half of the 16th, and especially in the 17th century, e.g. in Bergen (see Nesse 2010, pp. 91, 97). The reasons for this linguistic change in practice were the substantial divergent developments of the varieties involved, both in the dialect of the city of Bergen and in late Middle Low German, as well as in early New High German. The most plausible explanations for direct receptive communication in the Baltic in the early days of the Hansa are, in my view, (1) multilingualism, including receptive multilingualism, as a default before the age of nationalism, (2) the omnipresence of numerous linguistic varieties (or the lack of standardisation) in any type of communication, and (3) the absolute desire to understand each other, especially in trading situations, because only the results – economic success – counted, and not linguistic correctness or stylistic appro- priateness. This hypothesis was well received within a very short time and is sometimes mentioned without any reference, a fact which I take as a kind of a compliment that my research hypothesis has become part of the received knowledge when dealing with Scandinavian language history. For example, Teleman (2002, p. 29) says about the role of the German language in Sweden: Här kan det ha funnits en tämligen utbredd tvåspråkighet eller möjligen semikommunikation mellan svenska och tyska borgare. [Here one could find some rather widespread bilingualism or possibly semi-communication between Swedish and German citizens; my emphasis.] This view has been referred to again recently and more explicitly (see Tele- man 2010, p. 333), now with reference to Braunmüller (1993c), where the structural/grammatical similarities between Middle Low German and the (Mainland) Scandinavian languages have been pointed out in detail. Teleman claims, however, that divergences in the vocabularies of the contact languages have to be considered the real hindrance to unimpeded semi-communication in the Baltic. Moreover, the spread of Low German lexical loans, earlier mostly confined to the varieties used in trading stations and towns, has to be seen instead as a result of the nation-building process and integration of the dialects into the standard language over the course of the 19th century (Teleman 2010, p. 339). 98 Kurt Braunmüller

Sources and methods Our project primarily used translations of different types of texts. The (direct) sources for these translations were, in most cases, of Middle Low German origin but were in some cases also written in Early High German. The main motivation for proceeding in this way was to obtain valid results within a very limited period of time (two years of funding each for a maximum of two periods). Moreover, the working capacity of the two researchers who would do most of the work was also limited. But why did we restrict our project to translations? Preliminary research in the archive of Lübeck did not give us hope of finding translations, code-switches or even code-mixing in the official documents we looked at. We got the impression that all the texts in the archive were monolingual, either in Latin or in Low German, and obviously did not contain traces or paragraphs written in any Scandinavian contact language. This fact, though not directly encouraging at first glance, fitted, however, the initial hypothesis presented in Copenhagen in 1987, that there was no need for interpreting in pan-Baltic communication since all the languages involved in contact were immediately and directly mutually intelligible, at least for travelling professionals such as the Hanseatic tradesmen (if not for everybody). The translated texts selected for the projects were: A) popular reading texts (so-called Volksbücher), all of them of southern or central European origin, among others from the Italian Decamerone, in their Low German versions; B) selected Bible texts from both the Old and the New Testament. The first text investigated was a Low German Bible translation from the year 1480, and the last ones were the first comprehensive national Bible translations, the Gustav Vasa Bible for Sweden (1541) and Christian III’s Bible for Denmark (1550). Both translations were dependent on preliminary versions of Luther’s (High German) Bible translation published in 1545; C) law texts, especially the town laws of Flensburg and Schleswig (Flens- borg/Slesvig stadsret, 1284 and 1200, respectively), and The Jutish Law Codex (Den jyske lov), but the research carried out on these law texts was not very intense due to the lack of time; D) another semi-religious type of text which became wide-spread in the Bal- tic, the so-called ‘danses macabres’ (Totentänze: Des Dodes Danz [Low Germ. source text 1489/1496] and Den gamle danske dødedans [Danish translation 1536]).

Since the texts of (A) and (B) formed the bulk of our data, we will give a more detailed survey of those, mentioning the Low German source texts: (A1) Broder Rusche, (A2) Eyne schone historie van twen kopluden vnde eyner thuchtigen framen frawen, (A3) Griseldis, and (A4) Reynke de Vos; (B1) a se- Semi-communication and beyond 99 lection of Genesis I, (B2) the Swedish Pentateuch paraphrasis (compiled from the five books of Genesis), and (B3) a selection of the four gospels; B1 and B3 are in parallel translations into Middle Low German (functioning as a model), Danish and Swedish. Often it was quite obvious that the Danish version was the first one in Scandinavia and subsequently became translated into Swedish, though these two languages did not differ considerably from each other. They were translated anyway because they played a central role in the church service and should therefore be fully idiomatic. All texts were made available both in WORD and ASCII/DOS files and analysed by rather simple word-crunching and parsing programmes written in FORTRAN (Octaword and Octoword, developed by Ralf Thiele in Kiel), some of them especially designed and/or modified for this Hansa project. One has to keep in mind that, at that time, personal computer technology was still in its in- fancy, and the most advanced computers were IBM AT machines (with 8-bit technology, a very slow processing capacity and equipped with a hard disc maximum capacity of 100 MB, or at best 200 MB). One might suppose that most of the texts analysed were from rather late periods of contact when receptive multilingualism had been superceded by active and proficient bilingualism (which certainly was the case for many people in Scandinavian towns). Comparisons with one of the oldest (religious) texts available, the Swedish Pentateuch paraphrasis from the middle of the 14th century, showed, however, no significant structural or lexical differences compared with later translations. Moreover, the number of loanwords was not the most important factor as far as mutual and direct intelligibility is concerned: most similarities were either inherited due to the close genetic relationship or had their sources in replicated word-formation patterns, which made use of common lexical items and borrowed prefixes and suffixes.

The historical and linguistic background As we argued for in our first outline of the Hamburg Hanseatic Project (see Diercks & Braunmüller 1993), there was a very favourable linguistic situation for language contact in the Baltic in the Middle Ages when the merchants of the Hanseatic League took over the established Visby trade across the Baltic Sea in the middle of the 12th century. Among other factors, we mentioned the big socio-economic and technological differences between the European con- tinent and Scandinavia. This fact facilitated various kinds of transfer from the south to the north of Europe, involving such things as Christianity, Latin-based education and literacy in general, administration and the banking system and, last but not least, book printing (cf. the export of book(block)s from Mohnkopf in Lübeck to Stockholm; Menke 1987). In addition, the Jutish peninsula always functioned as a pivotal link between the North, West, and especially North Sea, Germanic dialects. Common (or wide-spread) grammatical features are: 100 Kurt Braunmüller

1) the lack of morphosyntactic gender markers and the emergence of instances of semantic gender marking; 2) the pre-posed definite article, in some West Germanic dialects sometimes rendered in two different forms with divergent referential functions;1 3) morphological simplification and reduction, such as lack of inflexion due to apocope; 4) only one possessive-reflexive pronoun in the 3rd person (unlike in North Germanic, which has two); cf. Eng. his/her, Dut. zijn/haar and Jutish dialects hans/hendes [± reflexive] vs. (Standard) Dan. both sin/sit [+ reflexive] and hans/hendes (in principle [– reflexive] or [± reflexive], due

to certain syntactic restrictions, such as in NPs like (here) Sw. Peri och

hansi /*sini bror [P.i and hisi brother]; and 5) common (sentence) intonation patterns (rapid fall in intonation at the end of a sentence vs. a continuous fall; no rising pitch on unstressed vowels as in Insular Danish varieties, esp. eastern ones). In the most southern Jutish dialects, we also see some German word-order patterns (the so-called ‘Satzklammer’, i.e. the use of discontinuous complex verb constructions, or the tendency to place the finite verb at the very end in subordinate clauses).

Seen from a more general perspective, the Baltic Sea area can be considered a typical instance of a dialect continuum due to close genetic relationships. This fact facilitates many forms of convergence, and grammatical and lexical transfer. When language contact occurs, there are, in principle, at least four possibilities for coming to grips with this situation: a) A subjected population can be forced to adopt the conquerors’ language, the result of which will be a foreign (mostly syntactic) substratum in the restructured local language. Lexical borrowing of parts of the vocabulary will lead to an expansion of the inherited vocabulary, a phenomenon which normally is called a superstratum. b) Speakers can use a third or intermediate language as a lingua franca, often only for specific purposes (e.g. in trading relations or for writing documents). Latin and, later, Middle Low German functioned as such a neutral, transnational third language. c) Becoming bi/multilingual is another option when immigrating to another country. This can, in the long run, result in (i) the simultaneous acquisition of two mother tongues from birth (2 L1s), (ii) early sequen- tial bilingualism, or (iii) second (or adult) language acquisition, which

1 Cf. the dialect of Amern in the Lower Rhine area, referred to in Heinrichs (1954), or the North Frisian variety Fering, spoken on the island of Föhr, as described in Ebert (1971). One of the two forms there might be due to language contact with a Jutish (or a lost North Sea Germanic) dialect. Semi-communication and beyond 101

comprises all forms of language acquisition which take place beyond puberty and which hardly every result in native speaker competence (L1 + L2 + …). d) Speakers can make use of the (more or less) close genetic relationship between the languages or dialects involved, which is still the default case in (Mainland) Scandinavia today when communicating with other fellow Scandinavians. In this case speakers either take advantage of the various relationships and similarities (including historical and stylistic) between the varieties involved, thus expanding their potential knowledge of the diasystem of their mother tongue, or sooner or later they become, at least receptively, bilingual with respect to the neighbouring languages due to learning by communicating. As already mentioned above, we have called this semi-communication or, more generally and more precisely, receptive (or ‘passive’) multilingualism.

Towards a typology of language contact during the era of the Hanseatic League (1150–1550), seen from the Hamburg perspective I would now like to give a short typological outline of language contact between the people who came into contact with each other and spoke either Low German or a (Mainland) Scandinavian variety. Written communication was far more restricted than today and was used only for certain purposes, especially for official documentation, such as municipal records (tänkeböcker – Stadtbücher), juridical statements of any kind, treaties, etc., but also for bookkeeping. Latin was the default language for documentation and later was replaced by Low German, but not before the 1370s as far as Hanseatic League’s own documents were concerned. This development coincided with a change in the social status of Low German, which became more and more a lingua franca in the Baltic and, later, a prestigious language as well. Oral communication made use of the narrow distance between the North Saxon variety of Low German on the one hand and the Scandinavian languages and dialects on the other. Even before the intense period of contact in the Late Middle Ages, we can observe many lexical and structural correspondences between these languages which facilitated face-to-face communication considerably, primarily in traditional situations, where the context, the terms of trade and the merchandise were well known to everybody working in this field. In face-to-face situations one can also point to items whose names are unknown, or use gestures. In short, during the early times of the Hanseatic League and beyond, one can say that communication involved far more than only verbal ex- changes. 102 Kurt Braunmüller

A very important point was that the languages involved in contact were not yet standardised, nor were there certain (written) norms which had to be followed. So people needed to be open to linguistic variation to a very high extent but could, in return, also make use of any diasystematic variation they were familiar with. In our project we thus talked about “diffuse code intersection” and “fuzzy rules” (cf. Braunmüller 1995d), and investigated language contact phenomena in terms of family resemblances and gestalt perception mechanisms, since codified grammatical forms did not yet exist at that time. Moreover, accommodation was the key technique in oral communication: speakers could adjust to their trading partners’ or neighbours’ variety and could immediately see which convergence strategy gave the best results. In addition, one will earn more money in trading contexts when one’s accommodation strategy (here: the convergence towards the dialect of the addressee) is effective. The very positive effects of extrinsic (here: commercial) motivation and strategic accommodation can hardly be overestimated! Another important point about this intense language contact is the fact that we observed many polycentric points of contact, especially along the coastal areas in Scandinavia. In contrast to later forms of language contact, which were bound to courts or economically dominant centres, the Hanseatic contacts were not tied to the upper classes nor to the merchants or citizens alone, but reached even fishermen and craftsmen (cf. Andersen 1995, where it is demonstrated that the core vocabulary of Jutish dialects was deeply influenced by Low Ger- man but also includes some Frisian loan words). In these coastal towns and trading stations we find (a) receptive multilingualism between genetically closely related varieties (Low German and a Scandinavian vernacular), (b) diglossia (many domains had specific languages or at least vocabularies of their own), (c) simplified Latin (also called ‘dog Latin’) as a kind of lingua franca (e.g. for travelling purposes), (d) L2 varieties and inter-languages as intermediate stages in second language acquisition, and finally (e) partial (active) competence in various languages for specific purposes. Besides these forms of multilingualism, we observe not only the (inevitable) interference and transfer from the L1 varieties into the target languages due to incomplete second language acquisition by adults, but also forms of creative code-mixing, resulting in new word-formation patterns which were well received and often succeeded as productive innovations, since linguistic standardisation was unknown in the Middle Ages and in Early Modern Times. Typical of Hanseatic language contact is the typological restructuring of the Mainland Scandinavian languages, which adopted word-formation patterns previously only found in West Germanic varieties, such as verbal prefixes (an-, be-, for-/för-, miss-) and nominal derivational suffixes (mostly -he(i)t, -else). Along with Heine & Kuteva (2005) we call this process ‘linguistic replication’. The transfer of these patterns was facilitated by the close genetic relationship of the languages involved in contact, speakers’ mutual linguistic flexibility and willingness to directly understand their neighbours Semi-communication and beyond 103 and, last but not least, the lack of standardisation and absence of norms, at least in oral communication. The lexical impact and transfer from Middle Low German has often been overestimated (cf. Rosenthal 1987). One of the main problems for getting a realistic picture is the close genetic relatedness between the languages in contact. Often it is hard to determine whether a word has been borrowed due to intense language contact, became reactivated, or was inherited and could be found in one of the varieties of the target language. Rosenthal (1987, p. 186ff.) mentions as realistic figures about 17% loan words from Low German and about 27% from High German into the Mainland Scandinavian languages. In texts containing languages for specific purposes (Fachsprachen), one will find between 23% and up to 30% loans from Low German. These figures, based on random samples, were supported by our project as well (cf. Zeevaert 1995b). It has to be observed, however, that the basic vocabulary has generally not been affected by this contact: the only instance that could be documented was a new modified reading of a conjunction replicated from Low German (but not the form of the conjunction itself) in Danish and Swedish (Old Scand. þa ‘then’ [temporal conjunction; cf. Germ. dann] > Dan. da – Swed. då ‘since’ [temporal/causal] < Low Germ. dô [temporal/causal, introducing clauses]). Unlike in Old English, no pronouns, verbs or nouns in the basic vocabulary were borrowed into the Scandinavian languages to a larger extent. The bulk of transfers involved new derivations which could be documented in detail when we compared translations from Low German into Danish or Swedish, where we observed more tokens than types of word-formation elements in the target texts (Diercks 1993b). It is possible that a new type of word formation led to the re- placement of an inherited noun, probably helped by an increase in transparency (cf. Old Swed. sûtari > Swed. sko-makare ‘shoe maker’). Other domains of grammar did not show significant divergence between the varieties involved: the differences in the phonological system and in pho- notactics were rather marginal and did not impede mutual understanding or prevent language acquisition. As far as inflexional morphology is concerned, the picture is more ambiguous. While Ringgaard (1986) argued for an independent, language-internal development, at least for texts written in Jutland, can the loss of inflexional markers also be traced back to language contact due to widespread imperfect second language acquisition? The predomi- nance of oral communication in the early days of the Hansa certainly plays a considerable role in this process as well: Low German and Jutish varieties had already lost their weakly stressed vowels (apocope) before the intense period of contact began. Moreover, the mastering (or recognition) of inflexional morphemes is not so important when decoding any Germanic dialect since the lexical roots and the derivational suffixes contain the crucial semantic information, unlike the Finnic languages, where e.g. the case or possessive morphemes following the lexical morpheme cannot be neglected when decoding an utterance. 104 Kurt Braunmüller

Since the word-order patterns in the languages of this contact scenario were not yet fixed, much more variation in word order was possible than would have been the case in modern times. This provided a good chance for imperfect (adult) bilinguals to predominantly use parallel syntactic patterns in both of their languages, ignoring e.g. less frequent syntactic structures or word-order patterns not found in their L1. The ordering of constituents in utterances could thus follow communicative needs and preferences, since it was not (yet) bound by strict syntactic norms. Moreover, the syntactic structures of the contact languages already showed significant correspondence at many crucial points,, such as (a) V2 placement in main clauses, (b) agreement in the position of the subject, the finite verb and the sentential adverb in the so-called nexus field, and free variation in the content field (to use two terms from Paul Diderichsen’s topological model), and (c) rather free variation in the position of the finite verb in (dependent) clauses, since the absolute final position of the finite verb was far from being mandatory in Middle Low German or in any Scandinavian variety (for more details, see Braunmüller 1998b, pp. 330ff.). To sum up, the following were crucial factors for contact between Low Ger- man and the Scandinavian languages in the Middle Ages and beyond: 1) at the beginning of the contact period, receptive multilingualism (or semi-communication) in direct oral communication, like in Scandinavia today; 2) the dominance of oral communication in the vernaculars, which also led to the use of non-standard varieties and forms that were not yet conventionalised. Oral communication thus gave rise to linguistic creativity and unorthodox word-formation and word-order patterns. Some of them, however, were well received and integrated into the target language within a very short period of time. 3) inter-dialectal accommodation and the occurrence of spontaneous inter- language varieties, comparable to strategies in modern inter-Scandina- vian communication, when people who accommodate say that they are speaking ‘Scandinavian’; 4) the absence of established norms in writing and the lack of standardisation. This gives rise to parallel constructions, frequently used by language learners and many bilinguals as well, unimpeded transfer and grammatical innovation, especially from the prestigious language Low German into the Scandinavian vernaculars.

Conclusion The Hamburg Hanseatic Project and the follow-up SFB project H 3 “Scandina- vian Syntax in a Multilingual Setting”, conducted within the Collaborative Re- search Centre (SFB 538) on Multilingualism from 2005 onwards, investigated Semi-communication and beyond 105 not only Low German and the Mainland Scandinavian languages, but also considered the subsequent contacts with High German and the parallel contact with Latin as the dominant written language in the Middle Ages and beyond (cf. Braunmüller 2000a, 2004 and Höder 2010). We not only analysed the different grammatical layers and domains of the languages involved but introduced, for the first time and to a larger extent, sociolinguistic issues, issues of language contact and bilingual communication, based on linguistic variation. We thus changed the paradigm from a predominantly philological approach, still based on neogrammarian principles, to modern sociolinguistics and contact linguistics, including investigations into the amount and types of vocabulary borrowed into the Scandinavian languages. Moreover, our project was the first to make use of the new personal computer technology, which changed the data base for empirical projects completely: now it was possible to analyse all phenomena in a text and carry out quantitative analyses as well. The change in the research paradigm was complete: new methods (a preference for structural and typological features of the languages involved), a new focus of interest (a sociolinguistic and contact linguistic approach, combined with research on variation and multilingualism), and new technology (word-crunching programmes) were applied.

References (cf. also the bibliography of the Hamburg Hanseatic Project)

Andersen, Torben Arboe, 1995  bibliography. Börestam Uhlmann, Ulla, 1997: Utblick för inblick – några aspekter på grannspråks- kommunikation i teori och praktik. In: Nordisk tidskrift för vetenskap, konst och in- dustri 3. Pp. 239–254. Braunmüller, Kurt, 1989  bibliography. Braunmüller, Kurt, 1993c  bibliography. Braunmüller, Kurt, 1995d  bibliography. Braunmüller, Kurt, 1998b  bibliography. Braunmüller, Kurt, 2000a  bibliography. Braunmüller, Kurt, 2004  bibliography. Diercks, Willy, 1993b  bibliography. Diercks, Willy & Braunmüller, Kurt, 1993  bibliography. Ebert, Karen H., 1971: Referenz, Sprechsituation und die bestimmten Artikel in einem nordfriesischen Dialekt (Fering). Bredstedt. (Studien und Materialien 4.) Haugen, Einar, 1966: Semicommunication: the language gap in Scandinavia. In: So- ciological inquiry 36. Pp. 280–297. Heine, Bernd & Kuteva, Tania, 2005: Language contact and grammatical change. Cam- bridge. (Cambridge approaches to language contact.) Heinrichs, Heinrich Matthias, 1954: Studien zum bestimmten Artikel in den germanischen Sprachen. Gießen. (Beiträge zur deutschen Philologie 1.) Hockett, Charles F., [1958] 1965: A course in modern linguistics. 9th edition, New York. 106 Kurt Braunmüller

Höder, Steffen, 2010: Sprachausbau im Sprachkontakt. Syntaktischer Wandel im Altschwedischen. Heidelberg. (Germanistische Bibliothek 35.) Kloss, Heinz, 1975: Die Entwicklung neuer germanischer Kultursprachen seit 1800. Düsseldorf. (Sprache der Gegenwart, Schriften des IdS 37.) Menke, Hubertus, 1987: Na dem Holme 1 vat mit boken. Zum spätmittelalterlichen Buchvertrieb Lübecks in den Ostseeraum. In: Kurt Erich Schöndorf & Kai-Erik Westergaard (eds.), Niederdeutsch in Skandinavien. Berlin. (Beihefte zur Zeitschrift für Deutsche Philologie 4.) Pp. 147–157. Nesse, Agnete, 2010: Norwegisch-niederdeutsche sprachkontakte. Zweisprachige Quellen aus dem Hansekontor Bergen. In: Niederdeutsches Jahrbuch 133. Pp. 87– 104. Ringgaard, K[ristian], 1986: Flektionssystemets forenkling og middelnedertysk. In: Arkiv för nordisk filologi 101. Pp. 173–183. Rosenthal, Dieter, 1987: Über den Umfang des niederdeutschen Einflusses auf das Schwedische. In: P. Sture Ureland (ed.), Sprachkontakt in der Hanse. Aspekte des Sprachausgleichs im Ostsee- und Nordseeraum. Tübingen. (Linguistische Arbeiten 191.) Pp. 170–202. Teleman, Ulf, 2002: Ära, rikedom & reda. Svensk språkvård och språkpolitik under äldre nyare tid. Stockholm. Teleman, Ulf, 2010: Lågtyska lånord. Om stad och land i svensk språkhistoria. In: Gu- nilla Byrman, Anna Gustafsson & Henrik Rahm (eds.), Svensson och svenskan. Med sinnen känsliga för språk. Lund. Pp. 331–341. Törnqvist, Nils, 1977: Das niederdeutsche und niederländische Lehngut im schwedischen Wortschatz. Neumünster. (Sprache und Schrifttum, NF B 10.) Ureland, P. Sture (ed.), 1987: Sprachkontakt in der Hanse. Aspekte des Sprachaus- gleichs im Ostsee- und Nordseeraum. Tübingen. (Linguistische Arbeiten 191.) Weinreich, Uriel, 1954: Is a structural dialectology possible? In: Word 10. Pp. 388–400. Zeevaert, Ludger, 1995b  bibliography.

Bibliography of the Hamburg Hanseatic Project and related research in Hamburg Andersen, Torben Arboe, 1995: Niederdeutsche Lehnwörter in jütischen (west- dänischen) Dialekten. In: Kurt Braunmüller (ed.), Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Pp. 181–223 [guest researcher]. Braunmüller, Kurt, 1989: Voraussetzungen für die Übernahme mittelniederdeutscher Sprachstrukturen in die skandinavischen Sprachen. In: Karl Hyldgaard-Jensen et al. (eds.), Niederdeutsch in Skandinavien II. Berlin. (Beihefte zur Zeitschrift für Deutsche Philologie 5.) Pp. 9–29. Braunmüller, Kurt, 1993a: Syntaktisch-typologische Divergenzen zwischen dem Mit- telniederdeutschen und dem älteren Dänischen (am Beispiel des Volksbuchs ‚Bruder Rusche‘). In: Hubertus Menke & Kurt Erich Schöndorf (eds.), Niederdeutsch in Skandinavien IV (…). (Zeitschrift für Deutsche Philologie, Beihefte 7.) Pp. 44–54. Braunmüller, Kurt, 1993b: Syntaktische Divergenzen und Transferenzen – Ein struk- tureller Vergleich anhand mittelniederdeutscher und altskandinavischer Volks- bücher. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 231–270. Braunmüller, Kurt, 1993c: Voraussetzungen für die Übernahme mittelniederdeutscher Sprachstrukturen in die skandinavischen Sprachen. In: Kurt Braunmüller & Willy Semi-communication and beyond 107

Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3). Pp. 137–160. Braunmüller, Kurt, 1993d: Nye hypoteser om sprogkontakt mellem nedertysk og dansk i Hansetiden – en projektrapport. In: Mette Kunøe & Erik Vive Larsen (eds.), 4. Møde om Udforskningen af Dansk Sprog. (...). Århus. Pp. 46–55. Braunmüller, Kurt, 1994a: Semikommunikation og lingvistiske simulationsmodeller (Om sprogkontakt i Hansetiden). In: Ulla-Britt Kotsinas & John Helgander (eds.), Dialektkontakt, språkkontakt och språkförändring i Norden (...). Stockholm. (MINS 40.) Pp. 92–97. Braunmüller, Kurt, 1994b: Der Einfluß des Mittelniederdeutschen auf die altskandi- navischen Sprachen in neuer Sicht. In: Niederdeutsches Jahrbuch 117. Pp. 93–108. Braunmüller, Kurt, 1995a: Forudsætninger for at overtage middelnedertyske sprog- strukturer i de skandinaviske sprog. In: Ernst Håkon Jahr (ed.), Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Pp. 29–54. Braunmüller, Kurt, 1995b: Syntaktiske divergenser og transferenser. En strukturel sam- menligning ud fra middelnedertyske og ældre skandinaviske folkebøger. In: Ernst Håkon Jahr (ed.), Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Pp. 106–146. Braunmüller, Kurt, 1995c: Formen des Sprachkontakts und der Mehrsprachigkeit zur Hansezeit. Eine einführende Übersicht. In: Kurt Braunmüller (ed.), Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Pp. 9–33. Braunmüller, Kurt, 1995d: Semikommunikation und semiotische Strategien: Bausteine zu einem Modell für die Verständigung im Norden zur Zeit der Hanse. In: Kurt Braunmüller (ed.), Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Pp. 35–70. Braunmüller, Kurt, 1996: Forms of language contact in the area of the Hanseatic League: dialect contact phenomena and semicommunication. In: Nordic Journal of Linguistics 19. Pp. 141–154. Braunmüller, Kurt, 1997a: Kontaktlinguistische Probleme im Ostseeraum zur Zeit der Hanse. In: Wolfgang W. Moelleken & Peter J. Weber (eds.), Neue Forschungsar- beiten zur Kontaktlinguistik. Bonn. (Plurilingua 19.) Pp. 81–88. Braunmüller, Kurt 1997b: Communication strategies in the area of the Hanseatic League: the approach by semicommunication. In: Multilingua 16. Pp. 365–373. Braunmüller, Kurt, l998a: Sprogkontakt i Hansetiden – en sammenfattende oversigt over Hamborg-projektet. In: Ernst Håkon Jahr (ed.), Språkkontakt i Norden i middelalderen, særlig i hansatiden. Forskningsprogrammet Norden och Europa. Copen- hagen. (Nord 1998. 4.) Pp. 17–31. Braunmüller, Kurt, 1998b: Wortstellungstypologische Untersuchungen zu den Kon- taktsprachen der Hansezeit (Mittelniederdeutsch, Dänisch, Schwedisch). In: John Ole Askedal (ed.), Historische germanische und deutsche Syntax. Frankfurt/M. etc. (Osloer Beiträge zur Germanistik, NF 21.) Pp. 315–334. Braunmüller, Kurt, 2000a: Voraussetzungen für die Übernahme hochdeutscher Sprachstrukturen in die skandinavischen Sprachen. In: Hans-Peter Naumann & Sil- via Müller (eds.), Hochdeutsch in Skandinavien. Internationales Symposium, Zürich, 14.–16. Mai 1998. Tübingen/Basel. (Beiträge zur Nordischen Philologie 28.) Pp. 1–18. Braunmüller, Kurt, 2000b: Højtysk som ’naturlig’ fortsættelse af den nedertyske sprogkontakt i Norden i 1500-tallet? In: Ernst Håkon Jahr (ed.), Språkkontakt – Inn- verknaden frå nedertysk på andre nordeuropeiske språk. Copenhagen. (Nord 2000: 19.) Pp. 277–288. Braunmüller, Kurt, 2000c: On types of multilingualism in Northern Europe in the late Middle Ages: language mixing and semicommunication. In: Gudrún Thórhallsdóttir 108 Kurt Braunmüller

(ed.), The Nordic Languages and Modern Linguistics 10. Proceedings of The Tenth International Conference of Nordic and General Linguistics. University of Iceland. June 6–8, 1998. Reykjavík. Pp. 61–70. Braunmüller, Kurt, 2002a: Semicommunication and accommodation: observations from the linguistic situation in Scandinavia. In: International Journal of Applied Linguistics 12. Pp. 1–23. Braunmüller, Kurt, 2002b: Om flerspråkighet och språkförändring. In: Folkmålsstudier (Helsinki) 41. Pp. 11–33. Braunmüller, Kurt, 2002c: Language contacts during the Old Nordic period I: with the British Isles, Frisia and the Hanseatic League. In: Oskar Bandle et al. (eds.), The Nordic languages. An international handbook of the history of the North Germanic languages. Volume 1. Berlin/New York. (Handbücher zur Sprach- und Kommuni- kationswissenschaft 22.1.) Pp. 1028–1039. Braunmüller, Kurt, 2002d: Semikommunikation, ackommodation och interdialektal kommunikation. Tre centrala begrepp för att beskriva språksituationen under Hansa- och reformationstiden. In: Svante Lagman, Stig Örjan Ohlsson & Viivika Voodla (eds.), Svenska språkets historia i Östersjöområdet. Tartu. (Nordistica Tartuensia 7; Studier i svensk språkhistoria 7.) Pp. 39–48. Braunmüller, Kurt, 2004: Niederdeutsch und Hochdeutsch im Kontakt mit den skandinavischen Sprachen. Eine Übersicht. In: Horst Haider Munske (ed.), Deutsch im Kontakt mit germanischen Sprachen. Tübingen. (Reihe Germanistische Linguistik 248.) Pp. 1–30. Braunmüller, Kurt, 2005: Language contacts in the Late Middle Ages and in Early Modern Times. In: Oskar Bandle et al. (eds.), The Nordic languages. An international handbook of the history of the North Germanic languages Volume 2. Berlin/ New York. (Handbücher zur Sprach- und Kommunikationswissenschaft; 22.2.) Pp. 1222–1233. Braunmüller, Kurt, 2006a: Left periphery of C – a vulnerable domain in language contact situations? Studies in older Danish and Swedish syntax and discourse structure. In: Wilfried Kürschner & Reinhard Rapp (eds.), Linguistik International. Festschrift für Heinrich Weber. Lengerich etc. Pp. 41–49. Braunmüller, Kurt, 2006b: Wortstellung und Sprachkontakt: Untersuchungen zum Vor- feld und Nebensatz im älteren Dänischen und Schwedischen. In: Amsterdamer Bei- träge zur älteren Germanistik 62. Pp. 207–241. Braunmüller, Kurt, 2007: Receptive multilingualism in Northern Europe in the Middle Ages. A description of a scenario. In: Jan D. ten Thije & Ludger Zeevaert (eds.), Re- ceptive multilingualism. Linguistic analyses, language policies and didactic concepts. Amsterdam/Philadelphia. (Hamburg Studies on Multilingualism 6.) Pp. 25– 47. Braunmüller, Kurt, 2008: Zu den sprachlichen Verhältnissen und Kommunikationsfor- men in Nordeuropa im späten Mittelalter und in der frühen Neuzeit. In: Geir Atle Ersland & Marco Trebbi (eds.), Neue Studien zum Archiv und zur Sprache der Han- seaten. Bergen. (Det Hanseatiske Museums skrifter 28.) Pp. 127–141. Braunmüller, Kurt & Steffen Höder, 2011: The history of complex verbs in Scandina- vian languages revisited: only influence due to contact with Low German? [in this volume]. Braunmüller, Kurt (to appear in 2012?): How Middle Low German entered the Main- land Scandinavian languages. In: Lars Bisgaard & Lars Bøje Mortensen (eds.), Town, guilds, and cultural transmission in the North (c. 1300–1500). Odense. Braunmüller, Kurt (ed.), 1995: Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Semi-communication and beyond 109

Braunmüller & Willy Diercks (eds.), 1993: Niederdeutsch und die Sprachen I. Heidel- berg. (Sprachgeschichte 3.) Diercks, Willy, 1993a: Niederdeutsch-dänisch-schwedische Übersetzungsliteratur. Ein linguistischer Strukturvergleich anhand der Schwankgeschichte vom Bruder Rausch. In: Hubertus Menke & Kurt Erich Schöndorf (eds.), Niederdeutsch in Skan- dinavien IV. Berlin. (Zeitschrift für Deutsche Philologie, Beihefte 7.) Pp. 68–90. Diercks, Willy, 1993b: Zur Verwendung prä- und postmodifizierender Morpheme im Mittelniederdeutschen und in den skandinavischen Sprachen. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidel- berg. (Sprachgeschichte 3.) Pp. 161–194. Diercks, Willy, 1995: Om anvendelsen af præ- og postmodificerende morfemer i middelnedertysk og i de skandinaviske sprog. In: Ernst Håkon Jahr (ed.), Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Pp. 147–176. Diercks, Willy & Kurt Braunmüller, 1993: Entwicklung des niederdeutsch-skandinavischen Sprachkontakts. Untersuchungen zur Transferenz anhand von volks- sprachlichen Texten des 15., 16. und 17. Jahrhunderts – eine Projektübersicht. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 9–40. Engelbrecht, Michael, 1993: Mitteleuropäisch-skandinavischer Kontakt zwischen 800 und 600 aus historischem Blickwinkel. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 41–49 [guest researcher]. Höder, Steffen, 2010: Sprachausbau im Sprachkontakt: syntaktischer Wandel im Altschwedischen. Heidelberg. [PhD dissertation, Hamburg 2009.] Höder, Steffen, 2011: Dialect convergence across language boundaries. A challenge for areal linguistics. In: Frans Gregersen, Jeffrey K. Parrott & Pia Quist (eds.), Lan- guage variation – European perspectives III. Selected papers from the 5th Interna- tional Conference on Language Variation in Europe (ICLaVE 5), Copenhagen, June 2009. Amsterdam/Philadelphia. (Studies in language variation 7.) Pp. 173–184 Höder, Steffen & Ludger Zeevaert, 2008: Verb-late word order in Old Swedish subordinate clauses. Loan, Ausbau phenomenon, or both? In: Peter Siemund & Naomi Kintana (eds.), Language contact and contact languages. Amsterdam/Philadelphia. (Hamburg studies on multilingualism 6.) Pp. 163–184. Jahr, Ernst Håkon, 1995: Niederdeutsch, Norwegisch und Nordisch. Sprachgemein- schaft und Sprachkontakt in der Hansezeit. In: Kurt Braunmüller (ed.), Nieder- deutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Pp. 125–144 [guest researcher]. Niepage, Gabriela, 1993: Rezeptionsbedingungen volkssprachlicher Erzählstoffe im spätmittelalterlichen Skandinavien. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 51–86. Warter, Per, 1993: MANAL. Automatische Lemmatisierung zur Computersimulation des Sprachkontakts. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 195– 229. Warter, Per, 1995a: Computersimulation von Wortverstehen am Beispiel mittelniederdeutsch-skandinavischer Sprachkontakte. Hamburg: Magisterarbeit (Institut für Germanistik: Skandinavistik) [137 pages]. Warter, Per, 1995b: Computersimulation von Wortverstehen am Beispiel mittelniederdeutsch-skandinavischer Sprachkontakte. In: Kurt Braunmüller (ed.), Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Pp. 71–123 [a shortened version of Warter 1995a]. 110 Kurt Braunmüller

Warter, Per, 1995c: Automatisk lemmatisering och datorsimulering av den lågtysk- skandinaviska språkkontakten. In: Ernst Håkon Jahr (ed.), Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Pp. 177–198. Zeevaert, Ludger, 1992: Sprachkontaktforschung anhand von Übersetzungsliteratur: Volkssprachliche skandinavische Erzähltexte des 16. und 17. Jahrhunderts. Ham- burg: Magisterarbeit (Institut für Germanistik: Skandinavistik) [119 pages]. Zeevaert, Ludger, 1993: Sprachkontaktforschung anhand von Übersetzungsliteratur. Volkssprachliche skandinavische Erzähltexte des 16. und 17. Jahrhunderts. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 87–136. Zeevaert, Ludger, 1995a: Språkkontakt och litteraturkontakt. Översättningstexter som material för språkkontaktundersökning. In: Ernst Håkon Jahr (ed.), Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen. Oslo. Pp. 81–105. Zeevaert, Ludger, 1995b: Wie intensiv war der mittelniederdeutsch-skandinavische Sprachkontakt wirklich? Untersuchungen von Transferenzen im Grundwortschatz. In: Kurt Braunmüller (ed.), Niederdeutsch und die skandinavischen Sprachen II. Heidelberg. (Sprachgeschichte 4.) Pp. 145–179. Zeevaert, Ludger, 2002: Multilingualismus und Sprachwandel. Zum möglichen Einfluß deutsch-schwedischer Mehrsprachigkeit auf die Gustav-Vasa-Bibel. In: John Ole Askedal & Hans-Peter Naumann (eds.), Hochdeutsch in Skandinavien. II. Interna- tionales Symposium, Oslo, 19.–20. Mai 2000. Frankfurt am Main etc. Pp. 101–125. Zeevaert, Ludger & Chrystalla Thoma, 2006: Klitische Pronomina im Griechischen und Schwedischen. Eine vergleichende Untersuchung zu synchroner Funktion und dia- chroner Entwicklung klitischer Pronomina in griechischen und schwedischen narra- tiven Texten des 15. bis 18. Jahrhunderts. In: Arbeiten zur Mehrsprachigkeit B/70. Hamburg: Sonderforschungsbereich 538 Mehrsprachigkeit. Zeevaert, Ludger, 2004a: Rezension von: Veturliði Óskarsson, Middelnedertyske låne- ord i islandsk diplomsprog frem til år 1500, 2003. In: Skandinavistik 34. Pp. 195– 196. Zeevaert, Ludger, 2004b: Rezension von: Agnete Nesse, Språkkontakt mellom norsk og tysk i hansatidens Bergen, 2002. In: NOWELE 45. Pp.125–132. Zeevaert, Ludger, 2005: Språkkontakt, syntaktisk variation och syntaktisk förändring. Kan det påvisas ett sammanhang i fornsvenska texter? In: Cecilia Falk & Lars-Olof Delsing (eds.), Svensk språkhistoria 8. (...) Lund. Pp. 333–342. Zeevaert, Ludger, 2006: Variation und kontaktinduzierter Wandel im Altschwedischen. In: Arbeiten zur Mehrsprachigkeit B/74. Hamburg: Sonderforschungsbereich 538 Mehrsprachigkeit. Zeevaert, Ludger, 2007: Receptive multilingualism and inter-Scandinavien semicommunication. In: Jan D. ten Thije & Ludger Zeevaert (eds.), Receptive multilingualism. Linguistic analyses, language policies and didactic concepts. Amsterdam/Phila- delphia. (Hamburg Studies on Multilingualism 6.) Pp. 103–135. Zeevaert, Ludger, 2008: The development in subordinate clauses in Old Swedish: An example for contact-induced language change? In: Anna Zanchi (ed.), Skálda- mjöðurinn. Selected proceedings of the UCL graduate symposia in Old Norse literature and philology, 2005–2006. London. Pp. 181–225. Zeevaert, Ludger, 2009: Deutscher Einfluss und syntaktischer Wandel im Schwedi- schen. In: Lars Wollin, Dagmar Neuendorf & Michael Szurawitzki (eds.), Deutsch im Norden. Akten der nordisch-germanistischen Tagung zu Åbo/Turku, Finnland, 18.–19. Mai 2007. Frankfurt/M etc. (Nordeuropäische Beiträge aus den Human- und Gesellschaftswissenschaften 28.) Pp. 279–306. Semi-communication and beyond 111

Unpublished lectures Braunmüller, Kurt: Sprogkontakt som dialektkontakt (Om sprogforandring i Hansetiden). (Gothenburg, August 1993.) Nedertysk og dansk i kontakt: forskningshypoteser og metoder. (Flensburg, September 1993.) Neue Perspektiven zu den niederdeutsch-skandinavischen Sprachkontakten in der Hansezeit. (Berne, Zurich, Fribourg, June 1994 and Frankfurt/M., July 1995.) Nye perspektiver for den nedertysk-nordiske sprogkontakt i Hanse-tiden. (Tromsø, Oc- tober 1994.) Semikommunikation: teori och praktik i dag och under Hansatiden. (Uppsala, January 1996.) Språkkontakt och flerspråkighet under Hansatiden. (Uppsala, January 1996.) Formen des Sprachkontakts und der Mehrsprachigkeit in Skandinavien zur Hansezeit. (Kiel, April 1997.) Sprachkontakte zur Hansezeit: Studien zum Sprachkontakt zwischen genetisch ver- wandten Sprachen. (Munich [LIPP], February 2010.)

Zeevaert, Ludger: Mittelniederdeutsch-schwedischer Sprachkontakt. Untersuchungen anhand des Grund- wortschatzes in Bibeltexten. (Hamburg: Workshop: ‚Sprachkontakte in der Hanse- zeit‘, October 1993.) 112 Kurt Braunmüller Low German texts from Late Medieval Sweden 113 Low German texts from Late Medieval Sweden

Stefan Mähl, Uppsala University

Introduction In medieval times, Sweden was a multilingual area where three languages were used in written communication: Latin, Low German and Swedish. The impact of Low German is an important part of Swedish language history. Not only were a substantial number of words transferred from Middle Low German into Old Swedish, but it has also been argued that Middle Low German influence played an important role in the loss of inflectional morphology and in some changes in the syntactic field (cf. Braunmüller & Diercks 1993; Braunmüller 1995). However, during the 20th century Swedish and German scholars paid little attention to the Low German texts that were written in Late Medieval Sweden. The aim of this paper is threefold: to provide some background information on the multilingual situation in Sweden, and especially in Stockholm, during the Late Middle Ages; to describe the origin of the use of Low German in documents written in Sweden; and to give a brief summary of a project which investigates and describes the Middle Low German variety that was used in Stockholm during the 14th, 15th and 16th centuries.

The emigration from Germany to Sweden during the Late Middle Ages The impact of German emigration and of the growing trade between North Germany and Sweden has been investigated in several studies by German and Swedish historians (e.g. Ahnlund 1953; Dahlbäck 1995, 1998; Koppe 1933; Lamberg 2001; Schück 1940). This section is based upon the research carried out in these studies. During the 14th and 15th centuries, cities like Arboga, Kalmar and Stockholm grew rapidly. This period is characterized by increasing trade and a better economic climate than previously. A close investigation of city records shows that a great many inhabitants were German immigrants. Most of them were craftsmen and merchants (cf. Sundqvist 1957). The origin of Stockholm pre-dates its written history, and how and why the city evolved in the mid-13th century is still disputed. However, a number of sur- 114 Stefan Mähl viving documents seem to indicate that Stockholm grew rapidly and became the largest city in Sweden towards the end of the 13th century. Researchers still argue about whether this development was a planned process initiated by the king and German merchants or a more spontaneous type of growth. It is reasonable to assume that Birger Jarl’s primary aim when founding Stockholm was not only to strengthen his own domestic and international position, but also to achieve general economic growth. German immigrants are frequently attested in documents from the late 13th century, so we can assume that German immigration started much earlier, before the reign of Birger Jarl (cf. Maillefer 2007). Around 1250 Birger Jarl renewed an agreement with the city of Lübeck in which he invited Germans to settle in Sweden as Swedish citizens and prom- ised the immigrants generous benefits (Wessén 1992, p. 4): Om åter några från eder stad vilja uppehålla sig hos oss och bo i vårt rike, så vilja vi, att de skola bruka vårt lands lagar och styras efter dem och i övrigt kallas svenskar. [If some people from your city wish to reside in Sweden and live in our kingdom again, we would like them to use our country’s law, be governed by them and be called Swedes in all other respects.] During the reign of Magnus Birgersson (1275–1290), relations with North Ger- man cities were strengthened. In 1276, Magnus married Helvig von Holstein, and through the marriage he strengthened his ties to the North German elite. During the 14th and 15th centuries German immigration flourished. According to some calculations, Stockholm had a population of approximately 5,000 by the end of the 15th century, which made it the largest city in Sweden. German immigration probably reached its peak during the reign of the German Albrecht von Mecklenburg (1364–1389). German noblemen, knights, servants and war- riors emigrated to Sweden in order to increase their own wealth. The Swedish king Albrecht often selected German knights for certain administrative positions, for instance Raven van Barnekow (bailiff at the royal castle of Nyköping) and Vicke van Vitzen (bailiff at the royal castle of Kalmar). He was deposed after his defeat at the battle of Åsle/Falköping. As a result, Germans temporar- ily lost their political influence, but their cultural and linguistic influence was still massive. Many of the German immigrants came from cities in the Baltic region, for instance Lübeck, Stralsund and Wismar. But a large proportion also came from cities in Westphalia or in the eastern part of Holland. Their surnames are often derived from their cities of origin: van Ossenbrügge (Osnabrück), van Unna (Unna, near Dortmund), van Warendorpe (Warendorf, near Münster) (cf. Sundqvist 1957).

Middle Low German – a definition The term Middle Low German is used here to refer to the compound of dialects (Märkisch, Nordniederdeutsch, Ostfälisch, Westfälisch) spoken in the northern Low German texts from Late Medieval Sweden 115 part of Germany during the period 1200–1650. The exact time span covered by the term Middle Low German sometimes varies between studies. The dialects in question were spoken north of the so-called Benrather Linie, the boundary line of the zweite Lautverschiebung (the High German Sound Shift). In the Low German area the consonants p, t, k are intact, whereas in High German they have undergone affrication:

HIGH GERMAN LOW GERMAN SWEDISH ENGLISH Apfel appel äpple apple Zunge tunge tunga tongue Lauch lôk lök leek

Middle Low German, like any label of its kind, is only a convenient abstrac- tion. At no time was there ever one uniform language, but what existed was a continuum of many different dialects (cf. Mähl 2008, pp. 42–45).

The use of Low German in Sweden during the Late Middle Ages Within the last five decades, scholarly interest in the study of Low German texts from Denmark, Finland and Norway has increased considerably (cf. Kan- tola 1987; Schöndorf 1987; Tiisala 2004). Unfortunately, most of the studies are based on very little empirical evidence. However, Low German texts from Sweden have hardly received any attention at all. An important exception is the study of Low German documents from Visby carried out by Artur Gabrielsson (1971/1972; see also Ahlsson 1989 on Everhard von Wampen). The first Middle Low German letters from Sweden date from 1360. Issuers are King Magnus of Sweden and his son King Håkon of Norway (Diplomata- rium Suecanum 6300–6302). The table below lists the first Middle Low Ger- man documents from Swedish cities during the period 1360–1375: Date Place Issuer(s) Diplomatarium Suecanum (DS) 28 June 1360 Helsingborg King Magnus and King Håkon DS 6300–6302 20 September 1360 Helsingborg Duke Albrecht von Mecklenburg DS 6348 28 September 1362 Söderköping King Magnus and King Håkon DS 6676 6 July 1364 Söderköping Duke Albrecht von Mecklenburg DS 7024 26 July 1364 Jönköping King Albrecht von Mecklenburg DS 7030 14 November 1364 Stockholm Karl Ulfsson DS 7095 5 February 1367 Kalmar King Albrecht von Mecklenburg DS 7498 6 June 1367 Kalmar Vicke van Vitzen DS 7537a 16 November 1367 Borgholm King Albrecht von Mecklenburg DS 7605 5 August 1369 Stockholm Mayor and Council DS 7935 19 November 1370 Stäkeholm Johann and Vicke Ummereise DS 8235 19 November 1370 Stäkeholm Johann and Vicke Ummereise DS 8236 116 Stefan Mähl

24 February 1371 Stockholm Bo Jonsson DS X 28 24 February 1371 Stockholm Bo Jonsson DS X 29 3 March 1371 Stockholm Minter Hans Kolnare DS X 35 15 May 1371 Stockholm Mayor and Council DS X 53 8 November 1374 Stockholm King Albrecht von Mecklenburg and DS 8675 Duke Albrecht von Mecklenburg 11 January 1375 Stockholm Vicke van Vitzen DS 8707 2 February 1375 Stockholm Raven van Barnekow DS 8719 2 February 1375 Stockholm Raven van Barnekow DS 8720 8 May 1375 Stockholm Duke Albrecht von Mecklenburg DS 8776 6 July 1375 Stockholm King Albrecht von Mecklenburg DS 8811 8 September 1375 Västerås Bo Jonsson DS 9170

In the 15th century, Low German became the dominant language in correspondence between cities in the Baltic region. Latin slowly lost its leading position owing to the breakthrough of the vernacular into written texts, a process which, in Germany, started in the 13th century and reached its conclusion in the 15th century (cf. Gärtner 1995; Mähl 2008, pp. 36–45). Swedish was seldom used in correspondence with foreign cities. One example of it is the letter from Kris- ter Niklasson to the city council in Reval dated 1436 where, at the beginning of the letter, he points out: Ok bidher jac idher kærliga, at i thet ey til mistycke takin, at jak idher scrifwer wppa swensco, thy jak nw enghen tyschen scrifware ner mik hawer. [And I sincerely ask you not to be offended by my writing to you in Swedish, for I do not have a German scribe with me.] (Finlands medeltidsurkunder 3, no. 2168; see also Tiisala 1996, p. 278.) Although Latin slowly lost its dominance in correspondence, it continued to be used in certain domains, such as the church and universities. It was also used among merchants, by whom it was still regarded as a very important language (cf. Tiisala 1996, p. 282).

Low German documents from Swedish cities – the case of Stockholm German presence in Stockholm during the Late Middle Ages From its origins, the city of Stockholm has been a location of German settlement. In several ways, the Stockholm records tell us about the close connections between Stockholm and the northern part of Germany during the Late Middle Ages. Right from the beginning Germans formed an essential part of the population of Stockholm, roughly estimated at 30–40%. German immigrants played such an important role in the city administration that it was necessary to regulate the number of German members of the city council so that no more than 50% of the posts could be held by Germans (cf. Mähl 2008, p. 24), in Swedish called the tudelningsprincipen. There is a great deal of evidence indicating that German immigrants formed an upper class of wealthy merchant families who were able to Low German texts from Late Medieval Sweden 117 maintain their German identity and their Low German language for generations after their emigration to Sweden. In Stockholm they visited the same church and the same guilds. They also married within the same circles. Furthermore, German immigrants maintained close contacts with their business friends and their rela- tives in the Baltic region. Because of this immigration, Stockholm – like many cities in the Baltic region – was a heterogeneous city, where both Swedish and Low German were written and spoken. From the outset, multilingualism was a feature of medieval Stockholm. As mentioned above, Germans played an essential role in the city administration, which led to the regulation of the number of German members of the city council manifested in Magnus Eriksson’s city law around 1350. The city records, however, indicate that the nationality of the council members did not always play a central role. An example is Lambert Westfal, who in 1438 was a German member of the council and in 1444 a Swedish mayor in Stockholm. After the battle of Brunkeberg in 1471, the tudelningsprincipen was cancelled immediately, and no German could have a seat in the city council. The political power of the German immigrants was now reduced.

City administration and scribes in Late Medieval Stockholm In Late Medieval Stockholm there were at least two administrative organiza- tions: the city administration and the royal administration in the castle of Stock- holm. Unfortunately, there are very few studies of Swedish administrations before King Gustav I Vasa. In this section, I will describe some aspects of the organization of the city administration in Stockholm. According to the city law issued by Magnus Eriksson, the city scribe had to be a Swede. Furthermore, the law states that the city documents must be written in Swedish, which indicates that (Low) German had a dominant position in the city. Through this law, the use of Swedish in official documents was promoted. The city scribe was responsible for a great many documents, for instance city (court) records, letters and other legal documents. The city scribes of Stock- holm in the 14th century are unknown, but during the period 1420–1533 the following scribes have been identified in the records of Stockholm (Mähl 2008, p. 50): City scribe Period Laurens Sonason –1440 Conradus Stadsskrivare 1440–1444 Mickel Skrivare 1444–1452 Engelke Hanson 1453–1459 Nicolaus Hanson 1459–1472 Ingevald Stadsskrivare 1472–1487 Helmik van Nörden 1487–1511 Andreas Anderson 1511–1524 Olaus Petri 1524–1533 118 Stefan Mähl

It is reasonable to assume that the city scribes knew three languages: Swe- dish, Latin and Low German. However, it is clear that there were differences between the scribes listed above. Lena Moberg, who has carried out a detailed investigation of two of the scribes, Ingevald Stadsskrivare and Helmik van Nörden, claims that Ingevald had limited knowledge of Middle Low Ger- man: Ingevalds anslutning till en mera allmänt omfattad fsv. norm i många av de beskriv- na fallen medger slutsatsen att denne skrivare varit huvudsakligen enspråkig och haft endast mer eller mindre ytlig kännedom om medellågtyskan. [Ingevald’s adherence to a more general Old Swedish norm in many of the cases described indicates that the scribe was chiefly monolingual and had only a more or less superficial knowledge of Middle Low German.] (Moberg 1989, p. 253.)

Low German documents from Stockholm – an analysis of the variety In the first half of the 14th century, Latin and Swedish were used in letters sent from the Stockholm city council to other cities. Latin was used in correspondence with foreign cities, and Swedish was used within Sweden. In the second half of the 14th century, scribes began producing letters in Middle Low Ger- man, mostly in correspondence with cities in Northern Germany, such as Lübeck. Within Low German philology, the importance of the city of Lübeck for the development of the Middle Low German language in the Baltic region has always been stressed. Because of Lübeck’s dominant position in the German Hansa, researchers have assumed that the Middle Low German variety used by the city administration in Lübeck (the Lübecker Norm) was a model for the Middle Low German language in the Baltic region, but also for the other Middle Low German dialects. In handbooks, it is often claimed that only one Middle Low German language existed. This theory, defended in the 20th century by, for instance, Conrad Borchling and William Foerste (cf. Foerste 1949), has never been verified by sufficient empirical evidence. Recent studies show that the Middle Low German variety used by the city administration in Lübeck in the 14th and 15th centuries was not homogeneous at all. For instance, it was less homogeneous than the variety used by the city administration in Hamburg. In Westphalian cities such as Herford, Münster and Osnabrück, very little influence of the Lübeck variety has been attested. Therefore the German researcher Robert Peters labels the Lübecker Norm a myth (Peters 1995). An interdisciplinary project funded by the Foundation for Baltic and East European Studies gave me the opportunity to investigate the Middle Low Ger- man variety that was used for internal and external correspondence in Stock- holm (Mähl 2008). The main aim of my study was threefold: – to analyse the Middle Low German variety used in Stockholm during the Late Middle Ages; Low German texts from Late Medieval Sweden 119

– to compare the development of the Middle Low German variety in Stock- holm with the development of Middle Low German varieties in northern Germany; – to analyse the effects of the multilingual environment in Stockholm. Is there any transfer from Swedish in the Middle Low German documents?

In my project I investigated 111 Middle Low German documents that were written in Stockholm during the period 1364–1508. These documents are mostly original texts that are kept in different European archives. In general, three types of documents were investigated: – letters from the city mayor(s) and the city council to other cities in the Bal- tic region, for instance Danzig, Lübeck and Reval; – deeds; – real-estate transactions that were registered in the city books.

An example of one of the document types investigated is given below. It is dated 20 September 1420, and the issuer is Jens from Svartnö, who was selling a cabbage plot at Stockholms norra malm to Erik Skinnare (Mähl 2008, pp. 33, 162): Vor alle den genen de dessen breff seen vnde høren lezen bekenne ik jønis i swar- tenøø dat ik myt wol vordachten mode rechtliken vnde redeliken vorkoft hebbe erik schynner mynen kolhoff vppe deme norderen malme beleghen welke vorscreuen kolhoffes rwm bret is xj ele vnde is lank xxv ele to water wart dyt heuet my de vor- benomede erik to danke wol betalet hir vmme late ik myt mynen eruen vnde antwar- de deme vorbenomeden erike vnde synen eruen dat vorscreuen rwm to brukende vnde to hebbende to ewigen tiden sunder ansprake myner vnde mynen eruen vnde des to tughe vnde hoghere bewarynge zo hebbe ik vorbenomen jønis i swartenøø ghebeden her clauwes braschen ratman to deme stocholme vnde tideken vp der goten dat ze ere ingesegele to eyn witlicheit henghen an dessen breff wante ik seluen nen ingesegel en hadde de ghegheuen vnde screuen is to deme stocholme na der bort vn- ses heren xiiijc iar in deme xx iare in vigilia mathei apostoli et ewangeliste. My investigations show that the Middle Low German variety used in Stock- holm during the period 1364–1508 was influenced not only by the Lübeck variety, but also by the Middle Low German used in the Bremen–Hamburg– Lübeck–Rostock area (North Low Saxon) (Mähl 2008, pp. 124–130). A smaller investigation of Middle Low German letters from other Swedish cities yielded a similar result (Mähl 2004). My study also shows that there is West- phalian influence in these Middle Low German texts, mostly in the city-internal documents. Typical Westphalian features are the forms vrent ‘friend’, nîn ‘nobody’, and sal ‘shall/will’, where the cities Bremen, Hamburg, Lübeck and Rostock instead use vrünt, nên and schal (Mähl 2008, pp. 130–132). It is not surprising that both North Low Saxon and Westphalian features are attested in Middle Low German documents from Stockholm, as a large proportion of the German immigrants came from these two areas. Quite often, leading members of the city council were born in the western part of the Low German area (cf. 120 Stefan Mähl

Sundqvist 1957), and their native dialect of course would have influenced their language. My findings indicate that the description of the Middle Low German variety used in Swedish cities, which assumes that only the city administration in Lübeck influenced the Baltic region linguistically during the Late Middle Ages, must be revised, especially for Stockholm but also for other Swedish cities. Surprisingly, there are very few Swedish elements attested in Middle Low German documents from Stockholm. Swedish words sometimes occur in the texts, such as gatebode ‘street shop’, hustru ‘wife’, and tompte/tumpte ‘plot of land’. Most of the words that were not translated into Middle Low German belong to Swedish city culture. It is reasonable to assume that legal issues were sometimes the reason words were not translated into Middle Low German. It is also clear that these words were transparent to the city’s populace, so a translation was not always necessary (Mähl 2008, pp. 132–134).

Concluding remarks German immigration was important for the development of Swedish society and the Swedish language during the Late Middle Ages. An understanding of contact between Swedes and Germans in Swedish cities in the Late Middle Ages is fundamental to the interpretation of the development of Swedish cities. In this paper, I have briefly described one of these interesting fields; my study of course needs to be followed up by further research, since many questions are still unanswered. My research shows that essential knowledge about the use of language, in this case Latin, Low German and Swedish, can be obtained through philological methods. It is clear that Middle Low German played an important role in Swedish cities’ correspondence with other cities in the Baltic region during the Late Middle Ages. But I have also shown that Middle Low German was used for internal communication in the city of Stockholm, which was not known before. More research is needed in this field, which is of central importance to Swedish and German medieval studies.

References Ahlsson, Lars-Erik, 1989: Everhard von Wampen: Spiegel der Natur. Ein in Schweden verfasstes mnd. Lehrgedicht. In: Karl Hyldgaard-Jensen, Vibeke Winge & Birgit Christensen (eds.) in conjunction with Kurt Erich Schöndorf, Niederdeutsch in Skandinavien 2. Akten des 2. nordischen Symposions ‘Niederdeutsch in Skandina- vien’ in Kopenhagen 18.–20. Mai 1987. Berlin. (Beihefte zur Zeitschrift für deutsche Philologie 5.) Pp. 154–172. Ahnlund, Nils, 1953: Stockholms historia före Gustav Vasa. Stockholm. (Monografier utg. av Stockholms kommunalfullmäktige 15.) Low German texts from Late Medieval Sweden 121

Braunmüller, Kurt, 1995: Niederdeutsch und die skandinavischen Sprachen 2. Heidel- berg. (Sprachgeschichte 4.) Braunmüller, Kurt & Diercks, Willy, 1993: Niederdeutsch und die skandinavischen Sprachen 1. Heidelberg. (Sprachgeschichte 3.) Dahlbäck, Göran, 1995: I medeltidens Stockholm. 2nd edition. Stockholm. Dahlbäck, Göran, 1998: Gast oder Bürger? Zur rechtlichen Stellung des deutschen Kaufmanns im mittelalterlichen Schweden. In: Horst Wernicke & Nils Jörn (eds.), Beiträge zur hansischen Kultur-, Verfassungs- und Schiffahrtsgeschichte. Weimar. (Hansische Studien 10.) Pp. 309–314. Diplomatarium Suecanum. Riksarkivet (ed.). 1829ff. Stockholm. Finlands Medeltidsurkunder 3, 1431–1450. Reinhold Hausen (ed.). 1921. Helsingfors. Foerste, William, 1949: Über plattdeutsche Sprache und Dichtung. Münster. Gabrielsson, Artur, 1971/1972: Zur Geschichte der mittelniederdeutschen Schrift- sprache auf Gotland. 1. Teil. In: Niederdeutsches Jahrbuch 94. Pp. 41–82. Fortset- zung. Die Entwicklung von 1350 bis 1500. In: Niederdeutsches Jahrbuch 95. Pp. 7– 65. Gärtner, Kurt, 1995: Zur Erforschung der westmitteldeutschen Urkundensprachen im 13. Jahrhundert. In: Gotthard Lerchner et al. (eds.), Chronologische, areale und situative Varietäten des Deutschen in der Sprachhistoriographie. Festschrift für Ru- dolf Große. Frankfurt am Main. (Leipziger Arbeiten zur Sprach- und Kommunika- tionsgeschichte 2.) Pp. 263–272. Kantola, Markku, 1987: Zur Sprache mittelniederdeutscher Urkunden aus Finnland. In: Kurt Erich Schöndorf & Kai-Erik Westergaard (eds.), Niederdeutsch in Skandina- vien. Akten des 1. nordischen Symposions ‘Niederdeutsch in Skandinavien’ in Oslo 27.2–1.3 1985. Berlin. (Beihefte zur Zeitschrift für deutsche Philologie 4.) Pp. 67– 73. Koppe, Wilhelm, 1933: Lübeck-Stockholmer Handelsgeschichte im 14. Jahrhundert. Neumünster. (Abhandlungen zur Handels- und Seegeschichte. Neue Folge 2.) Koppe, Wilhelm, 1934: När Stockholm var Hansestad. Några notiser ur Lübecks stats- arkiv. In: Samfundet S:t Eriks årsbok 1934. Pp. 73–94. Lamberg, Marko, 2001: Dannemännen i stadens råd. Rådmanskretsen i nordiska köp- städer under senmedeltiden. [Jyväskylä dissertation.] Stockholm. (Monografier utgivna av Stockholms stad 155.) Mähl, Stefan, 2004: Mittelniederdeutsche Urkunden aus Schweden (1360–1375). In: Robert Damme & Norbert Nagel (eds.), westfeles vnde sassesch. Festgabe für Robert Peters zum 60. Geburtstag. Bielefeld. Pp. 101–113. Mähl, Stefan, 2008: Geven vnde screven tho deme holme. Variablenlinguistische Un- tersuchungen zur mittelniederdeutschen Schreibsprache in Stockholm. Uppsala. (Acta Academiae Regiae Gustavi Adolphi 99.) Maillefer, Jean-Marie, 2007: När folkungarna anlitade tyska adelsmän för sin utrikes- politik. Stockholm. (Sällskapet Runica et Mediævalia, Lectiones 7.) Moberg, Lena, 1989: Lågtyskt och svenskt i Stockholms medeltida tänkeböcker. [Stockholm dissertation.] Uppsala. (Acta Academiae Regiae Gustavi Adolphi 58.) Peters, Robert, 1995: Die angebliche Geltung der sog. mittelniederdeutschen Schrift- sprache in Westfalen. Zur Geschichte eines Mythos. In: José Cajot et al. (eds.), Lingua Theodisca. Beiträge zur Sprach- und Literaturwissenschaft. Jan Goossens zum 65. Geburtstag. Münster & Hamburg. Pp. 199–213. Schöndorf, Kurt Erich, 1987: Mittelniederdeutsche Urkunden aus Norwegen. In: Kurt Erich Schöndorf & Kai-Erik Westergaard (eds.), Niederdeutsch in Skandinavien. Akten des 1. nordischen Symposions ‘Niederdeutsch in Skandinavien’ in Oslo 27.2– 1.3 1985. Berlin. (Beihefte zur Zeitschrift für deutsche Philologie 4.) Pp. 46–66. 122 Stefan Mähl

Schück, Herman, 1940: Stockholm vid 1400-talets slut. Stockholm. (Kungl. Vitterhets Historie och Antikvitets Akademiens Handlingar 48.) Sundqvist, Birger, 1957: Deutsche und niederländische Personenbeinamen in Schwe- den bis 1420. [Uppsala dissertation.] Stockholm. (Anthroponymica Suecana 3.) Tiisala, Seija, 1996: Mellan latin och lågtyska. Svenskans ställning i hansatidens Sverige-Finland. In: Ann-Marie Ivars et al. (eds.), Svenskans beskrivning 21. Förhandlingar vid Tjugoförsta sammankomsten för svenskans beskrivning. Helsing- fors den 11–12 maj 1995. Lund. Pp. 278–283. Tiisala, Seija, 2004: Power and politeness. Language and salutation formulas in correspondence between Sweden and the German Hanse. In: Journal of Historical Prag- matics 5.2. Pp. 193–206. Wessén, Elias, 1992: Om det tyska inflytandet på svenskt språk under medeltiden. Stockholm. (Nytryck i Nordiska språk och svenska 10.) A survey of Low German loanwords in Danish 123 A survey of Low German loanwords in Danish in the medieval period and the transition from Low German to High German as the written language in Tønder in the 17th century1

Birgit Christensen, Vanløse

Low German loanwords in Danish medieval charters My motive for selecting Low German loan words as the topic for my Master’s thesis was the fact that the subject has been rather neglected. I was familiar with Peter Skautrup’s discussion of the topic in Det Danske Sprogs Historie (vol. I, 1944, pp. 298–302 and vol. II, 1947, pp. 102–111), as well as with Marius Kris- tensen’s doctoral thesis (1906) and the works of Clara Holst (1903) and Ida Marquardsen (1908), together with more recent articles such as Anders Bjer- rum’s (1969) on the Oath of Allegiance from 1387, Magda Nyberg’s (1971) on loan words in a North Schleswig dialect, and H. Bach’s (1977) on Modern Ger- man influence on the Danish standard language. I wanted to find out whether there was a difference between the borrowing of Low German words between various parts of the country and at differing social levels, and I therefore decided to work with texts that were both dated and localised, namely charters. From among a collection of Old Danish charters, duplicated for the use of the Dictionary of Old Danish (Gammeldansk Ordbog), I selected a number of charter groups: the “Halland group”, consisting of a deed, two mortgage deeds and a receipt from noblemen (1378–1387); an “Ol- der chancellery group” (1397–1400); the “State papers group” (1387–1397), including inter alia the Act of Union from 1397; a “Younger chancellery group” (1433–1435); a group of letters from the Copenhagen city court (byting) (1430–1432); a similar group from Køge (1431–1433); and three groups of depositions from the district courts (herredstingene) of Sjælland (1433–1435), Fyn (1433–1435), and Jylland (1432–1435). The loanwords were divided into different types, e.g. simple words and composita for which only one of the components is a loanword, and the frequency of their occurrence was calculat- ed as the percentage of loanwords occurring in the total number of running words counted. These figures were then compared for the different groups of

1 This paper was translated into English by associate Professor Dr. Phil. Gillian Fellows-Jensen. 124 Birgit Christensen charters. The result was that Low German influence was not as strong in the “Halland group” (6.39) as in the “Older chancellery group” (9.88), the “State papers group” (10.0), or the “Younger chancellery group” (14.22). Thus the Low German influence was weakest in the “Halland group”. In the 1430s, Low German influence was strongest in the “Younger chancellery group” and weakest in the charters from the district courts from Jylland (9.38), Sjælland (9.13) and Fyn (7.95). The documents from the city court (byting) in Copenhagen are on a middling level (11.44) and have most in common with those from the chancellery in terms of percentage of loan words, while the documents from Køge (9.23) are on a level with those from Jylland. It is hardly surprising that there are more loanwords in texts from a higher social level, the chancellery, than in the texts from the lower social levels, since the scribes who worked in the higher social levels probably had a more comprehensive education than the other scribes; they may even have studied at the North German universities. With respect to the depositions from the district courts, I would like to mention one factor which I did not touch upon in my earlier work: it is thought-pro- voking that the frequency of occurrence of the loanwords in the 1430s is weakest in Fyn. This raises the question as to whether the presence of Low German loan words in Sjælland reflects an influence from Copenhagen or whether the influence in Jylland reflects a spread from south to north. My material is unfortunately too limited to enable me to draw a reliable conclusion.

The suffixes -else, -inge and -nisse In 1908 Marquardsen demonstrated for the first time that Danish words with the suffix -else often correspond to Low German words with the suffix -inge. It had previously been thought that they were renderings of Low German words with the suffix -nisse. Marquardsen’s discovery was not, however, generally accepted. Thus it was in Seip’s opinion (Seip 1952, pp. 53–54) probably incor- rect to say that the Low German suffix -nisse lies behind the great productivity of the Danish suffix. Rather, we should say that when Low German verbs were borrowed, they received the native suffix -else instead of the suffix -nisse that they had had in Low German. Now that I had a large collection of relevant material, it suddenly became clear that Marquardsen had been right and that -else did not as a rule correspond with -nisse but with -inge. It seems that in the Middle Ages the idea had spread that when a word ends in -inge in Low Ger- man, there must be a corresponding word in Danish that ends in -else, and vice versa. In addition, there are some Danish words in -else which correspond both with a Low German word in -inge and one in -nisse; the latter, however, is West Low German, specifically Westphalian. It is on analogy with these that the Low German words in -nisse which do not have a corresponding form in -inge receive a Danish suffix -else (Christensen 1982, pp. 57–61). This was a surprise; no one could have anticipated that this would prove to be the case. A survey of Low German loanwords in Danish 125

Low German loanwords in letters from Køge Later I studied the frequency of occurrence of loan words in some documents from the Køge town archive from the period 1488–1494 (Christensen 1989); the frequency for this subgroup was 12.11. A comparison was then made with the older documents from Køge dating from the 1430s, where an increase in the frequency of loanwords can be observed. However, this result has to be taken with certain reservations, since the documents from the 1430s were all written by one scribe, while those from 1488–1494 were written by four or five different scribes, and their personal preferences may have made themselves felt. The documents in question are letters of patent, that is, proclamations to the general public, so I thought it advisable also to examine some private letters sent to a specified recipient. I selected three letters from the private archive of the nobleman Bent Bille: two drafts written by him, and a letter in Danish from the merchant Oluff Lar- ber in Stralsund. One letter from Bent Bille was to another nobleman, Iver Axelsøn Thott, dealing with a legal dispute about a mortgage deed. It is slightly formal, although not sufficiently so to prevent the inclusion of a small amount of direct speech. The frequency of loanwords for this letter is 11.30. The second letter is a complaint to the Bishop of Roskilde that the Bishop’s half-brother had stolen four of Bent Bille’s best pigs, which he would like to have returned, together with compensation. The frequency of loan words for this letter is 9.70. Thus, in the more formal letter connected with the legal dispute there are more loanwords, while in the angry letter about the theft of the pigs there are fewer loanwords. The Danish letter from Stralsund occupies an intermediate position.

Low German loanwords in private letters Over the course of the last 30–40 years, it has become more and more usual within research in the humanities to take an increasing interest in people’s daily life, their private and intimate concerns. Inspired by Peter Skautrup’s claim that it is in private correspondence that one can find a style approaching that of the spoken word (Skautrup 1944–1970, vol. II, p. 73), I studied the frequency of occurrence of loanwords in some private letters exchanged between close relations and married couples (Christensen 1992). Such letters, which are less formal, might also be referred to as family letters. The family in question was the nobleman Jørgen Rud of Vedby, who was a rigsråd (state counsellor, 1435–1440) and lensmand (lord lieutenant, 1504–1505), his wife Kristine Rosenkrantz (dates of birth and death unknown), their daughter Anne Rud (date of birth unknown; she died in 1533), and his son-in-law Hen- rik Krummedige (c. 1465–1530), a Norwegian and Danish counsellor and knight; he was born in Norway but the family came originally from Holstein and had also lived in Schleswig. Their letters, which are from the years 1497– 126 Birgit Christensen

1502, are presumably all written in their own hands. The frequency of occurrence of loan words was greatest in Henrik Krummedige’s two letters, at 9.52 and 12.60, but not as great as in the chancellery letters from the 1430s. The frequency in Anne Rud’s letter is 12.59. The older couple had the lowest frequency of loanwords, which was not even as high as that in the depositions from the district courts in Fyn in the 1430s. Kristine Rosenkrantz’ letter has a frequency of 5.54, and Jørgen Rud’s 4.72. In spite of the fact that Jørgen Rud was a very learned man, the letter written to his daughter contains few loan words.

The transition from a Low German written language to a High German one in Tønder My motive for taking up the topic of the written language in Tønder was once again the fact that this topic has been rather neglected. The reason why I selected Tønder from among the four market towns in North Schleswig was that its archive, which is kept in the North Schleswig provincial archives, was the most well preserved, the largest, and the one whose documents go furthest back in time. This last fact was important, since it was impossible to determine in ad- vance exactly when the change in written language from Low German to High German took place. I had imagined that it had taken place much earlier than proved to be the case. It turned out, in fact, that this change happened at a very late date. Emden is the only other town where the change of language took place equally late (Hahn 1912), and nowhere did it happen later. It must be em- phasised that it is the written language with which I am concerned. I cannot, however, avoid touching on the spoken language from time to time, but here I rely on the findings of Anders Bjerrum, who stated that in about 1600 everyone, whether of high or low status, spoke Sønderjysk (the Danish dialect of North Schleswig), and that in addition most adults could speak Low German (Bjerrum 1943, p. 443; 1973, p. 63).

A survey of the source material It was first necessary to clarify the chronology of the language change. In the town archives there are 75 bundles of documents with material antedating 1700, as well as a number of well-preserved court records. It is possible to work through the court records and find out when the language changes, but this cannot be done with the bundled material, for those documents are not kept in an exact chronological order. Random sampling would not be satisfactory, since there might be material in the unopened bundles that could cause the chronology to be seriously disrupted. All the bundles were therefore examined. For each year, I compiled a survey of the material and divided it up into groups ac- A survey of Low German loanwords in Danish 127 cording to the issuer of the documents: the Duke, members of the municipal government, persons outside the municipal government etc. This survey is too extensive to publish here, but it has proved to be an essential tool for me. Work- ing in this way, I could not only produce a frame for all future work but also fulfil my most important aim, namely to have access to the individual human beings who lived and worked in the town under the conditions prevailing at the time so that I could see what written language they employed. And the chronology was a vital factor not least because of the correlation with political history. Reports of the actual use of language are rare. The scribes themselves never said why they wrote in Low German or High German. Nor did I find any directive from one person to another that this was the occasion to write in High Ger- man. And I only found a single reference to knowledge about the language of another person (see below). Historians distinguish between a relic and a report (Erslev [1926] 1987, p. 5). I treat the sources primarily as relics.

Division into phases It is possible to say that all the material contained in the town archive belongs to one domain: administration. This is not an entirely satisfactory label, however. Administration consists of many branches, and it is therefore advisable to divide it into sub-domains. Here it is important to distinguish between whether the texts were for external consumption, in practice in most cases correspondence with the Duke, or for internal use, which would in practice consist of accounts and other texts concerned with financial matters. In addition, information can sometimes turn up about the use of language in other domains. I oper- ated not only with the concept of domain but also with the concept of diglossia (Christensen 2006a, pp. 138–151; cf. Fishman 1967, Ferguson 1973, Trudgill 1995). I divided the sequence of events into five phases (Christensen 2006a, pp. 141–151). The first phase extends from 1436 – the date of the oldest document in the town archive, a confirmation by Duke Adolf of the town’s privileges – up to 1536. All the documents in the town archive from this phase are written in Low German. I include this phase, defined by the exclusive use of Low Ger- man, in anticipation of future research. This will make it possible to examine whether there is evidence of a gradually increasing influence on the actual language in this period. The second phase extends from 1537 to 1604. This phase begins with the oldest surviving High German document, a charter issued by King Christian III, who had introduced the Reformation to the Kingdom of Denmark the preceding year and who had already introduced it to the Duchy of Schleswig in 1528 (Gregersen 1986, 2000). The first High German document from the Duke, who was the King’s brother, is from 1545. From then on the Dukes regularly corresponded with Tønder in High German with one exception, a 128 Birgit Christensen proclamation from 1570 which concerns the surrounding area. The correspondence from Tønder to the Duke, on the other hand, consisting of petitions, continued to be in Low German. For the short period 1600–1603, the court records were entered in High German by town clerks who came from outside, but subsequently they returned to being written in Low German. During this phase High German was beginning to make its entry from outside. The third phase extends from 1605 to 1651. The first petition to the Duke in High German dates from 1605. It was written by the town clerk Jurgen Thim- sen, who must have taken up his post in 1603. He died in 1651 or at the beginning of 1652. He mastered both of the written languages. In the course of that period High German increasingly came into use, and this development seems to have been particularly swift in the 1630s. During this phase High German was the language of the Duke, while Low German was the language of financial matters and accounts, as well as of the church – with the exception of the Danish morning service (matins) – and of municipal schools. The fourth phase extends from 1652 to 1672. In 1652 Dean Stephan Kenckel from Flensburg (Danish Flensborg) took up his post. High German then became the language of the church – with the exception of the Danish matins – and presumably High German also became the language of the borger- skole (the municipal school), although I have not been able to find information about this. Low German was still the language used for accounts. The last treasurer’s accounts in Low German are from 1672. The fifth phase begins in 1673. It would seem that the change of language was by then more or less complete, although we still find various financial accounts in Low German, both from a deputy mayor and from craftsmen. I stopped searching at the year 1700, and I cannot therefore guarantee that there was no Low German used thereafter. If there was, it cannot be expected to have been very much. The historian H. V. Gregersen, according to whom High German as a written language did not become significant until late in the whole of the Duchy, is probably right to think that this was because Low German was easier to learn than High German if one had the North Schleswig Danish dialect as one’s home language (Gregersen 1974, p. 247; Christensen 2006a, p. 152). If one accepts Fishman’s definition of diglossia as an alternation between two languages – which Jan Goossens does in his works on the situation around the western boundary for Low German (see particularly Goossens 1984) – one comes up with the following simple description of the situation in Tønder: phase 1– diglossia, involving the Danish dialect and Low German; phases 2 and 3– increasing triglossia, with High German as the high language, the Danish dialect as the low language, and Low German as an intermediate language; phases 4 and 5– decreasing triglossia (Christensen 2006a, p. 151). In addition, there were traces of a knowledge of the Danish written language in Tønder as early as in the first half of the 17th century, a matter to which I shall return. How widespread this knowledge was is impossible to determine. Per- A survey of Low German loanwords in Danish 129 haps one ought to speak of quadroglossia instead of triglossia, and triglossia instead of diglossia? At least the situation ended up with not just a new diglossic situation but a quadroglossic situation.

The transition from a Low German to a High German written language in various social and socio-functional groups The division into phases is supplemented with a survey of the change of written language in the various social and socio-functional groups (Christensen 2007a) and a number of articles which go deeper into the question. I shall mention the most important results from these. Decrees promulgated by the prefects (amtmænd) in Tønderhus, Dietrich Blome (1594–1608), Hans von der Wisch (1608–1624) and Wolf Blome (1624–1664), continued from the beginning of the 17th century to alternate between Low German and High German, but most of them were written in High German. The last Low German document from a prefect in Tønderhus that I have found in the Tønder town archive dates from 1652 and was issued by Wolf Blome (Christensen 2007a, p. 100). As already mentioned, the court records were kept in High German during the short period 1600–1603. The men who did this had come from the south. The town clerk Johann Bouke, who was probably from Ostfriesland, kept the records in High German in 1600 and the town clerk Johannes Faust, who came from Flensburg (Danish Flensborg) but whose family originated from West- phalia, continued keeping the records in High German until 1603 (Christensen 2000a, pp. 124–126). He was succeeded by the town clerk Jurgen Thimsen, who in 1612, probably in connection with a visit by the Duke to the city, went over to using High German in the records (Christensen 2000a, p. 136). The other town clerks both before and after this short period wrote in Low German. Peter Jacobsen, who was town clerk in Tønder from, at the latest, 1578 to 1594, was born in Tønder, educated in Rostock, and wrote in Low Ger- man. At the university in Rostock the transition to High German had been made in 1550 – although Low German had not entirely disappeared in the second half of the 16th century – but this did not mean that he wrote in High German during his employment in Tønder (Christensen 2000a, p. 120–121). Jurgen Thimsen was the town clerk who was particularly affected by the language change in the Tønder administration. It is not known where Thimsen was educated. He took over his post in 1603. He mastered both written languages and his use of these was systematic: he wrote in High German from 1605 to the Duke and from 1612 in the court records. He retired as town clerk in 1638 but continued to be employed in the town administration, and he wrote texts dealing with financial matters for internal use in the town until his death in 1651 or 1652 (Christensen 2000a, pp. 126–136; 2005, pp. 112–130). The system began to disintegrate with his son, Frederich Thim, who also worked for the town administration and was town clerk for the short period 130 Birgit Christensen

1648–1649 but had to resign his post in favour of a man from outside whom the Duke wished to be employed (Christensen 2005, pp. 131–137; 2007a, p. 103). Frederich Thim wrote in High German to the Duke and the King and in the court records; in daily life he mostly wrote in Low German, particularly in connection with money matters, but he also occasionally wrote in High German in this context (Christensen 2005, p. 137; 2000a, pp. 116–118). From Henricus Schallichius onwards, who came from Westphalia in 1638 (Christensen 1999, p. 35; 2006a, p. 147; 2007a, p. 103), the written language of the town clerk for the administration was High German. The town clerks were too well-educated to be considered typical representatives of the population of Tønder. More typical were the mayors, treasurers and craftsmen. The mayors, who were merchants and only left behind a few texts, from the middle of the 17th century wrote mainly in High German for administrative af- fairs (Christensen 2006a, p. 148; 2008). The treasurers, who were also merchants, continued to use Low German for a long time (Christensen 1999; 2000b; 2007, p. 104). The last accounts kept by treasurer Hinrich Meysahl in Low German date from 1672. It should also be mentioned that treasurer Jacob Jensen Roost also used Low German for his accounts in 1675 (Christensen 2006a, p. 149), but these were probably not as formal as the official treasurer’s accounts. The treasurers’ texts were for internal use in the town. Just as in the German towns, there is a clear difference here: High German was used for correspondence going out of the town, which would normally be to the Duke, and Low German was for internal use within the town. This becomes very clear in connection with the administration of hospitalet, which was not a hospital but an institution for the poor. The town ran a kind of bank associated with this institution, and the accounts for this have survived. Jurgen Thimsen kept these accounts in Low German from at least 1615 till his death in 1651 or 1652 (Christensen 2005, p. 125–126; 2009, p. 103). In 1652 they were kept for a short time by the mayor, Thomas Andersen, and then by the Dean, Stephan Kenckel, in both cases in High German (Christensen 2009, p. 103). Bonds were issued, largely written in Low German by Jurgen Thimsen and in both Low German and High German by Frederich Thim. The interesting fact about these bonds was that the citizens themselves had signed them, sometimes adding a brief accompanying text saying that they were signing in their own hand. Among these bond holders were a few women; one of them, Agete Dalleres/Daleres, wrote the whole bond herself in Low German in 1639 (Chris- tensen 2009, p. 110). The last bond written in Low German was issued in 1662 to Christian Amdersen, a citizen of Tønder. The identity of the scribe is unknown; Frederich Thim had died the preceding year (Christensen 2011). Women were not involved in the administration of the town, but there is a certain amount of evidence to show that they could write – for example, they signed bonds, as mentioned above. Laureta Preuss, who was married to Peter Preuss, the mayor of Tønder in 1675–1680, and who after his death continued A survey of Low German loanwords in Danish 131 to run his merchant business, added a note in Low German to a letter from 1646 somewhat later, but in 1696 as an old woman she wrote a receipt in High Ger- man (Christensen 2011). It was the craftsmen who continued using Low German longest. The material they left behind can all be said to fall within the domain of money matters, particularly receipts for their work for the town. Their use of Low German first began to decrease in the period 1688–1699, particularly after 1690. The last Low German text is an invoice from the locksmith Christian Petersen in 1699 (Christensen 2006a, p. 150). Irmtraud Rösler also confirms that the craftsmen in Mecklenburg went over to using High German much later than did the ad- ministrators of the town (Christensen 2006b, p. 85; Rösler 2000, p. 43). At the lowest level of the social scale, we see that in 1696 there was an executioner in Tønder who could write an invoice to the town, and this was in High German. He was one of a family of executioners called Asthusen, from Hamburg (Christensen 2007a, p. 109). His invoice was not for an execution but for the cleaning of “the secret room”, that is the toilet, in Rådsvinkælderen (the council’s wine cellar), a hostelry in the Town Hall.

School and church It was a disappointment that I did not find more material about the municipal school, which was the normal form of schooling available to citizens. Such documents might have revealed which language or languages were taught and employed for teaching. It is assumed, as already mentioned, that the change of language in the school took place at the same time as the transition to High Ger- man in 1652 for preaching in the church (Christensen 2006a, p. 148). It was a surprise that names of private schoolmasters turned up in the records, but unfortunately these were found in the tax censuses, which provide no information about the language of eduation (Christensen 2006c). It was an equally great disappointment that there was so little information about the language used in church, as this would certainly have been of great significance. The use of High German as the language of preaching would have lent this language authority. One priest made an attempt to employ High German as the preaching language in 1631, but on this occasion the citizens protested (Christensen 2006a, p. 146; 2007a, p. 106). High German was first used at the morning service in 1652, when Stephan Kenckel took office as Dean (Christensen 2006a, pp. 147–148; 2007a, p. 106). In addition, throughout the 17th century there was an early morning service in Danish, although German hymns were sung during the service (Christensen 2007a, p. 106). It is conceivable that it was easier to get hold of German hymn-books at that time. 132 Birgit Christensen

Danish in Tønder Finally, I shall make a brief mention of one element in this project that was quite unexpected: there were a few texts in Danish, which demonstrated that there were people who could read and write Danish in Tønder. It was already known that during the years 1614–16 there had been a municipal school that taught in Danish in Tønder (Christensen 2006c, p. 219). The oldest Danish text that I have discovered in the town archive is from 1602, but this was written by a plumber (a specialist on lead roofs) in Møgeltønder who belonged to the Danish enclaves. The oldest text in Danish that was written by a citizen of Tønder is the text by a signature on a hospital bond from 1632 (Christensen 2006c, p. 220). In addition, there were several indirect indications of knowledge of Danish as a written language in Tønder, for example correspondence from 1685–1696 between some people in Odense who wrote in Danish to Tønder, with responses in German from Tønder (Chris- tensen 2006c, pp. 223–224). Most significantly, I found a reference to actual language use: the nobleman Benedikt Rantzau, former provincial governor of Møgeltønder, had his scribe write in 1609 that “de Herrn Burgermeister vnnd Rath der denischen jo so wol kundlich dann der duttschen sprache” (‘the Lord Mayor and the council know both the Danish and the German language’) (Christensen 2006c, p. 219). In this connection, it was also interesting that rather unexpectedly a letter turned up dated 1634 from the priest Petrus Jacobi von Alsleff (Alslev, a vil- lage in Højst parish), who had studied in Germany, spent some time in Copen- hagen and been offered a living in Tønder. However, he refused to return home until he had become more proficient in the Danish standard language; he later ended up in Vodder parish, slightly north of Tønder (Christensen 2007b, 2007c, pp. 10–12).

Summary and concluding remarks In conclusion, it can be said that the High German written language was brought to Tønder by the Duke and his chancellery at the beginning of the 17th century, but only spread slowly throughout the administration. From the middle of the 17th century, when the town clerk Jurgen Thimsen died and the Dean Stephan Kenckel arrived, High German became more commonly used in the town, although the treasurers still kept to Low German in the 1670s for internal administration, while some craftsmen in the town still were writing Low German in the 1690s. At the same time, there were also instances of Danish as a written language in the 17th century. I would not claim to have brought the Tønder project to completion. It would be interesting to move up to a higher cultural level where the changes of language could be observed, to go more deeply into the matter of how Low A survey of Low German loanwords in Danish 133

German and High German were written, and to continue in the same manner with the other market towns, Sønderborg, Aabenraa and Haderslev. I believe that the chronological division into phases will give a firmer basis for further research.

Literature Low German loanwords Bach, H., 1977: Der niederdeutsche Einfluss auf die dänische Standardsprache. Sprach- liche Interferenz. Festschrift für Werner Betz zum 65. Geburtstag. Hrsg. von Herbert Kolb und Hartmut Lauffer. Tübingen. Pp. 526–531. Bjerrum, Anders, 1969: Sproget i hyldingsbrevene fra 1387. Kopenhagener Germanis- tische Studien 1. Pp. 256–254. Christensen, Birgit, [1982]: MA Thesis: De middelnedertyske låneord i et udvalg af danske diplomer fra 1378–1435. Accessible at Det Humanistiske Fakultets speciale- bibliotek, Afdeling for Navneforskning, University of Copenhagen, Institutionen för nordiska språk and Seminariet för ortnamnsforskning, Uppsala University. Christensen, Birgit, 1982: Die mittelniederdeutschen Lehnwörter in dänischen Urkun- den aus dem Zeitraum 1378–1435. Kopenhagener Beiträge zur Germanistischen Linguistik 20. Pp. 1–66. Christensen, Birgit, 1989: Niederdeutsche Lehnwörter in Quellen aus der Gegend von Køge aus der zweiten Hälfte des 15. Jahrhunderts. Zeitschrift für deutsche Philolo- gie, Beiheft 5: Akten des 2. nordischen Symposions “Niederdeutsch in Skandina- vien” in Kopenhagen 18.–20.5. 1987. Unter Mitwirkung von Kurt Erich Schöndorf herausgegeben von Karl Hyldgaard-Jensen, Vibeke Winge und Birgit Christensen. Berlin. Pp. 173–187. Christensen, Birgit, 1992: Niederdeutsche Lehnwörter in einigen dänischen Privatbrie- fen aus den Jahren 1497 und 1502. Zeitschrift für deutsche Philologie. Beiheft 6. Ak- ten des 3. nordischen Symposions “Niederdeutsch in Skandinavien” in Sigtuna 17.– 20.8. 1989. Herausgegeben von Lennart Elmevik und Kurt Erich Schöndorf. Berlin. Pp. 37–48. Holst, Clara, 1903: Studier over middelnedertyske laaneord i dansk i det 14. og 15. Aar- hundrede. Kristiania. Kristensen, Marius, 1906: Fremmedordene i det ældste danske skriftsprog omtr. 1300. København. Marquardsen, Ida, 1908: Der Einfluss des Mnd. auf das Dänische im 15. Jahrhundert. Beiträge zur Geschichte der deutschen Sprache und Literatur 33. Halle. Pp. 405– 458. Reprinted New York/Graz 1969. Nyberg, Magda, 1971: Om nogle tyske låneord i M. B. Ottsens ordbog. Studier i dansk dialektologi og sproghistorie tilegnede Poul Andersen på halvfjerdsårsdagen d. 8. juni 1971. København. Pp. 255–261. Seip, Didrik Arup, 1952: Omstridde spørgsmål i norsk språkutvikling. Oslo. Skautrup, Peter, 1970: Det Danske Sprogs Historie I–IV. København 1944–1968 Index volume V. 134 Birgit Christensen

The transition from Low German to High German written language in Tønder in the 17th century Bjerrum, Anders, [1943] 1973: Folkesproget i Tønder gennem Tiderne. I. Tønder gennem Tiderne. Skrevet af danske forfattere. In: M. Mackeprang (ed.), Skrifter, udgivne af Historisk Samfund for Sønderjylland 3.2. Pp. 440–464. English translation: The language of the people of Tønder through the ages. In: Anders Bjerrum. Lin- guistic papers. Published on the occasion of Anders Bjerrum’s 70th birthday. Sel- skab for nordisk Filologi, Copenhagen. Pp. 51–74. Christensen, Birgit, 1999: Einige Handwerkerbezeichnungen in den Rechnungen aus Tønder (Tondern) aus dem 17. Jahrhundert. In: Germanistische Schlaglichter 4. Festschrift für Märta Åsdahl Holmberg zu ihrem 80. Geburtstag. Pp. 34–55. Christensen, Birgit, 2000a: Die Stadtschreiber und der Wechsel von niederdeutscher zu hochdeutscher Schriftsprache in den Gerichtsprotokollen der Stadt Tondern. In: Bei- träge zur nordischen Philologie 28. Pp. 117–141. Christensen, Birgit, 2000b: Ortsnamen in den Kämmereirechnungen der Stadt Tondern beim Wechsel von Niederdeutscher zu Hochdeutcher Schriftsprache im 17. Jahrhun- dert. In: Niederdeutsches Wort 40. Pp. 105–118. Christensen, Birgit, 2005: Stadtschreiber in Tondern während des 17. Jahrhunderts und der bei ihnen zu beobachtende Schriftsprachenwechsel vom Niederdeutschen zum Hochdeutschen. In: Lennart Elmevik, Stefan Mähl & Kurt Erich Schöndorf (eds.) in conjunction with John Ole Askedal & Otto Erlend Nordgreen, Niederdeutsch in Skandinavien V–VI. (Osloer Beiträge zur Germanistik 36.) Pp. 105–140. Christensen, Birgit, 2006a: Schriftsprachenwechsel Nd.–Hd. in der Stadt Tønder (deutsch Tondern) im deutsch-dänischen Grenzgebiet. Eine Phaseneinteilung. In: Stuttgarter Arbeiten zur Germanistik 434. Pp. 135–155. Christensen, Birgit, 2006b: Zum Schriftsprachwechsel Nd.-Hd. der Handwerker in der Stadt Tondern. In: Stuttgarter Arbeiten zur Germanistik 436. Pp. 59–87. Christensen, Birgit, 2006c: Private skolemestre i Tønder i 1600-tallet. Underviste de i at læse og skrive dansk? (With summary in English.) In: Harald Gustafsson & Hanne Sanders (eds.), Vid gränsen. Integration och identiteter i det förnationella Norden. Gothenburg. Pp. 204–229. Christensen, Birgit, 2007a: Hd./nd./dän. Schriftsprachgebrauch der Vertreter verschie- dener sozialer bzw. soziofunktionaler Gruppen in Tønder/Tondern. Ein Überblick. In: Stuttgarter Arbeiten zur Germanistik 439. Pp. 97–112. Christensen, Birgit, 2007b: Petrus Jacobi von Alßlef – Lidt om nedertysk, højtysk og især dansk i Tønder i 1600-tallet. In: Peter Widell & Ulf Dalvad Berthelsen (eds.), Møde om Udforskningen af Dansk Sprog 11. Pp. 76–90. Christensen, Birgit, 2007c: Dansk i Tønder i 1600-tallet. In: Ord og Sag (edited by Peter Skautrup Centret for Jysk Dialektforskning, Aarhus Universitet), December. Pp. 4–18. Christensen, Birgit, 2008: Die Schriftsprachkenntnisse der Bürgermeister der Stadt Tønder/Tondern zur Zeit des Schriftsprachwechsels Niederdeutsch-Hochdeutsch im 17. Jahrhundert. In: Beiträge zur Kanzleisprachenforschung 4. Pp. 141–158. Christensen, Birgit, 2009: Der Schriftsprachwechsel von Niederdeutsch zu Hoch- deutsch in der Verwaltung des Hospitals zu Tønder im 17. Jahrhundert. In: Beiträge zur Kanzleisprachenforschung 6. Pp. 97–115. Christensen, Birgit, 2011: Ein neuer Blick auf die Sprachverhältnisse in Tønder und Umgebung: Die Schreib- und Sprachkenntnisse der Frauen im 17. Jahrhundert. In: Stuttgarter Arbeiten zur Germanistik 454. Pp. 159–172. Erslev, Kr., [1926] 1987: Historisk Teknik. Den historiske undersøgelse fremstillet i sine grundlinier. 10th printing, with an afterword by Kai Hørby and Hans Vammen. Copenhagen. A survey of Low German loanwords in Danish 135

Ferguson, Charles A., 1973: Diglossia. In: Word 15. Pp. 325–340. Fishman, Joshua A., 1967: Bilingualism with and without diglossia; diglossia with and without bilingualism. Journal of social issues 32. Pp. 29–38. Hahn, Louis, 1912: Die Ausbreitung der neuhochdeutschen Schriftsprache in Ostfries- land. Teutonia: Arbeiten zur germanischen Philologie 24. Goossens, Jan, 1984: Die Herausbildung der deutsch-niederländischen Sprachgrenze. Ergebnisse und Desiderata der Forschung. In: W. Besch, K. Hufeland, V. Schupp, P. Wiehl (eds.), Festschrift für S. Grosse. Göppingen. Pp. 23–44. Gregersen, H. V., 1974: Plattysk i Sønderjylland. En undersøgelse af fortyskningens historie indtil 1600-årene. Deutsche Zusammenfassung. Odense. Gregersen, H. V., 1986: Reformationen i Sønderjylland. Skrifter, udgivne af Historisk Samfund for Sønderjylland 63. Aabenraa. Gregersen, H. V., 2000: Reformationen i hertugdømmet. Sønderjylland år 0–2000: V. In: Sønderjysk Månedsskrift, June. Pp. 159–166. Rösler, Irmtraud, 2000: Mecklenburgische Handwerkerrechnungen und -quittungen 16.–18. Jahrhundert. (Stuttgarter Arbeiten zur Germanistik 372.) Pp. 37–52. Trudgill, Peter, 1995: Sociolingiustics. An introduction to language and society. 3rd edition. Harmondsworth, UK. 136 Birgit Christensen MLG loanwords in medieval charters issued in Närke 137 Middle Low German loanwords in medieval charters issued in the Swedish province of Närke

Helena Wistrand, Stockholm University

Introduction This paper constitutes an effort to comprehensively describe the range, distribution, and use of words imported from Middle Low German (MLG) in medieval charters originating from the Swedish province of Närke. A more extensive account of the study which this paper summarizes, and its results, can be found in my dissertation (Wistrand 2006). The aim of that study was primarily descriptive from a quantitative perspective, although some qualitative aspects are discussed here, mainly in the sections on lexical variants and regional provenance. The charters studied were all written between the years 1350 and 1520. They have been divided according to three different aspects – social provenance, regional provenance and chronological provenance – and the vocabulary of MLG origin used in them has been studied accordingly. The main focus here is on regional provenance. Three lesser parts of the full study deal with foreign words (Germ. Fremdwörter), MLG affixes, and loan words used in the different textual formulas of the charters. Of these, only the issue of foreign words is briefly accounted for in this paper.

The province of Närke during the Middle Ages The fairly small and geographically limited area that forms the medieval province of Närke contains a series of culturally and socially interesting environments. For example, the town of Örebro dates back to the end of the 13th century, although at that time it was already an important crossroad between routes from the provinces of Värmland, Öster- and Västergötland, the mining districts of Central Sweden and the region of Lake Mälaren. During earlier medieval times, the Eriksgata (i.e. the journey traditionally taken by newly elected kings to have the elections acknowledged by the rural courts and assemblies) led here from the southwest (Lovén 1999, p. 142; Jonasson 1984, p. 23f.). Also situated in Närke are various mining districts, the Cistercian convent of Riseberga, the 138 Helena Wistrand castles of Örebro and Göksholm, and not least the rural court assemblies of the province. Some of the hundreds (small administrative districts) of medieval Närke were loosely united in larger judicial districts called treding (appr. ‘third’), which perhaps was a remnant of an older division of the province into three parts (Tunberg 1911, p. 48). The two tredings remaining in the Middle Ages were Västra tredingen (‘the Western third’) and Östra tredingen (‘the Eastern third’); there was one rural court judge per treding, while the hundreds of Glanshammar, Sundbo and Örebro, and also the mining district of Noraskogen were not part of any larger judicial district of this kind, and subsequently each had their own rural court judges (Styffe 1911, p. 302). However, at the end of the 15th century, all of Närke seems to only have had one rural court judge prac- tising in the entire province (Almquist 1955, pp. 50, 104).

Material The selection of charters for the source material of the study was based on the following criteria: the charters are original versions issued between the years 1350 and 1520 by people who can be assumed to have permanently resided in the province. The source material for the study consists of 270 charters, of which the majority has not yet been published. This means that they have had to be read from photocopies or from the actual originals. Most of the charters constitute official texts of a judicial nature, such as deeds of transfer of different kinds (i.e. bills of sale, deeds of gifts, wills, etc.) or judgments. There are also some letters of a non-official nature which mainly concern administrative matters. These texts were for the most part written by members of the Sture family, who were also members of the council of the realm, at the end of the 15th and the beginning of the 16th centuries, and who resided at Örebro castle (Wiktorsson 1983, p. 19). MLG loan words were located in the charters by referring to different dictionaries. Initially, every lemma with a MLG equivalence was extracted from Ordbok öfver svenska medeltids-språket (Söderwall 1884–1918; Söderwall et al. 1925–1973), resulting in a glossary of possible MLG loanwords in Old Swedish (OSw). This glossary was used to locate loanwords in the charters, while comparisons were also made with etymological information in other dictionaries (SAOB 1893 ff.; Hellquist 1948; NEO 1995–96), resulting in a total of 258 lemmas (107 nouns, 104 verbs, 33 adjectives, 13 adverbs and 1 conjunction) of MLG origin being identified in the source material. These 258 lemmas occur in a total of 2,100 separate instances in the source material, and all are accounted for in chapter 3 of Wistrand (2006). All MLG forms referred to in this paper and in the study which it is based upon have also been drawn from dictionaries (Lasch & Borchling 1928–; Schiller & Lübben 1875–1881). MLG loanwords in medieval charters issued in Närke 139

Lexical variants In a section concerning graphonomical and phonological integration of MLG loanwords in Lena Moberg’s dissertation (1989, p. 54f.), she states that it is primarily the form of the word in the source language that constitutes its base in the target language. She writes that one should thus assume that the form of a MLG loanword in OSw reflects its form in the source language, rather than seeking to explain differences between them as being a result of changes in OSw. In this context, she points out the problem that dictionaries often do not give a fair representation of the variations in form that MLG vocabulary items most likely had. Moberg also writes that a person who was familiar with MLG ought to have produced words which were more similar to corresponding forms in the source language than someone with a lesser knowledge of MLG. Unfamiliar pronunciation, accent, inflection, word formation and orthography are all characteristics of foreign words. The orthography may be atypical in different ways: it could be unfamiliar in the target language, or it could be familiar but correspond to an unexpected pronunciation (Dahlstedt 1969, p. 18ff.). The study discussed in this paper reveals that some examples of imported words display foreign traits even though in many cases these loanwords had existed in OSw for quite some time in more orthographically and morphologically integrated forms. Such forms are thus assumed to have been produced by scribes with a relatively good knowledge of MLG due to their position in society and/or level of education. Seven forms which have been extracted from charters issued in the 14th century can be considered to be quite early. These forms are geographically distributed over three locations. In charters from the town of Örebro we find the forms ‹macht› (in charter no. 11942, see Sources and Literature below)1 with MLG ‹cht› for /kt/, and ‹qwytan› (14176), which indicates a long stem-vowel as in MLG quit. From charters issued at the convent of Riseberga, I have extracted the forms ‹iuncfru› (13031), with /k/ preserved in the first element of the compound, like in the MLG principal form, and ‹acht› (13979). In the area south of Lake Hjälmaren, charters contain the forms ‹ganzlica› (12079) and ‹ganslika› (12848), clearly lacking /k/ in the root morpheme like the MLG form, and ‹macht› (13816).2 During the 15th century, words which may be characterized as foreign are more scarce and geographically scattered, and they will not be mentioned further here. Foreign forms are most frequent in charters issued in the early 16th century, primarily from the office of Örebro castle. Two of the scribes of the

1 References to specific charters are given here, since specific written forms are discussed in this section. 2 MLG macht, OSw. makt ‘power’; MLG quit, OSw. qvitter ‘quits, free’; MLG juncvrûw(e), OSw. iungfru ‘maiden’; MLG acht, OSw. akt ‘intention’; MLG ganslĩk, gantz-, -lĩke(n), OSw. gansklika ‘entirely’. 140 Helena Wistrand charters in question have been identified as Erik skrivare (‘Erik the scribe’) (Sjödin in Pers 1932, p. 691), and Kettil skrivare Påvelsson (‘Kettil the scribe Påvelsson’) (Utterström 1968, p. 19). Erik uses the forms ‹befollandes› (34517b, 34680), which seems to be derived from the MLG alternative bevolen, and ‹legx› (34517b), which reflects the MLG monosyllabic subsidiary form leysch. Kettil Påvelsson uses the forms ‹Rögtid›/‹Rökcthä› (34826), ‹röcthe› (35512/ 35695), probably from the MLG parallel form röchte. He also writes ‹Joffrws› (34803), ‹Joffrvner› (34822), ‹Joffrv› (34864) and ‹joffrv› (35808), where the first element of the compound is perhaps derived from the MLG western subsidiary form juffer; and ‹ffor vetten› (35512) which, with its short vowel in the root morpheme, obviously is influenced by the MLG alternative form vorwetten. Furthermore, he writes ‹Rittärss› (34863) and ‹Rittherss› (34865), which reflects the MLG subsidiary form ritter.3 The early instances of forms characterized as foreign here can perhaps partially be explained as being the result of their fairly young age in relation to their time of appearance in OSw, i.e. they occur chronologically relatively close to the actual borrowing of the words. The fact that relatively young MLG loanwords are written in ways which closely reflect their forms in the source language may not be considered strange, but it nevertheless reveals that the scribes must have had some knowledge of the MLG forms. Perhaps the later instances, i.e. those found in charters issued in Örebro in the first two decades of the 16th century, are more noteworthy, not least since they are attributed to known scribes. It is particularly interesting when forms can be attributed to a specific MLG dialectal area, as is the case with Kettil Påvels- son’s way of writing iungfru. A further study of the charters written by this particular scribe might perhaps reveal more about MLG influence on his text production.

MLG loanwords according to social, regional and chronological provenance The frequency of the imported words was also studied from social, regional and chronological perspectives. The charters issued by the burghers of Örebro and the miners of Noraskogen were found to be richest in imported words, followed by texts issued by members of the nobility. Regarding the regional comparison, the charters originating from the northern parts of Närke, where the town of Örebro and the Noraskogen mining district are located, and from the south-east

3 MLG bevālen, bevēlen, bevolen, OSw. befala ‘command’, ‘entrust’; MLG leydisch, leitsch, leysch, OSw. leidhisker ‘cloth from Leiden’; MLG rüchte, röchte, OSw. rykte ‘piece of news’; MLG juncvrûw(e), jon-, juffer, OSw. iungfru ‘maiden’; MLG vorvēten, OSw. forvita ‘know of’; MLG ridder, riddere, ritter, OSw. riddare ‘knight’. MLG loanwords in medieval charters issued in Närke 141 of the province, contain the largest number of imported words. The chronological comparison shows that both the number and the frequency of the imported words increase over time.

Social provenance Comparison on a social basis is somewhat problematic, starting with the cate- gorization of the source material. One drawback is that in most cases we do not know who actually wrote the charters, or to what extent the person issuing a charter influenced in the wording of the text, i.e. how the responsibility for formulating the text was distributed between the scribe and the person issuing a charter. The comparison according to social provenance was thus made in pairs: chief judges and rural judges, knights and squires, members of convents/ monasteries and priests, burghers of Örebro and members of mining communities (see Table 1 below). Looking at the first two pairs, the group of people with the highest social standing – i.e. chief judges and knights – have the greatest variation in their use of MLG loanwords Counting all parts of speech together, there are a larger number of loanwords in charters issued by chief judges than by rural judges, with regard to both lemmas (types) (on average 2,0 types of MLG loanwords per text in charters issued by chief judges (54/27), compared to 1,5 in those issued by rural judges (65/42)) and separate instances (tokens) of loanwords (on average 6,6 tokens of MLG loanwords in charters issued by chief judges (179/ 27) compared to 4,7 tokens in charters issued by rural judges (196/42)).4 How- ever, a slightly different pattern emerges for charters issued by knights and squires: knights’ charters have on average the largest number of types of MLG origin (on average 3,3 types of MLG loanwords per text in charters issued by knights (30/9), compared to 2,4 in those issued by squires (65/27)), while the squires’ have the largest number of tokens (on average 8,0 tokens of MLG loanwords in charters issued by squires (216/27) compared to 6,6 in charters issued by knights (59/9)). Although the squires used a smaller number of types than the knights, they used these types all the more frequently. Texts issued by members of convents and monasteries do not differ much from those issued by priests concerning the number of loanwords. MLG nouns and especially verbs are somewhat more frequent in charters issued at convents

4 Concerning all given average numbers in this paper, it must be considered that they are not related to the length of the charter texts respectively, i.e. the number of words in each text. When writing my dissertation, I refrained from making such calculations. In OSw it is not easy to clearly state what is a clearly-defined lexical unit, and what is not. What we in modern Swedish consider solid compounds, is in OSw sometimes written as two words and sometimes as one, often varying within each text. Even Söderwall (1884–1918, 1925–73) do not give consistent directions in this matter, since some compounds are lemmatized as one word, others as two. The reliability of such calculations simply would not correspond to the amount of work they would have led to. There- fore, the calculations of average numbers of MLG loanwords in the texts are simply based on the number of loanwords and the number of charters in each group. 142 Helena Wistrand

Table 1: Social, regional and chronological variation. Frequency of types and tokens of MLG loanwords in absolute numbers. Number Adj./ In Nouns Verbs of adv. total charters types tokens types tokens types tokens types tokens Chief judges 27 28 91 16 60 10 28 54 179 Rural judges 42 25 90 22 52 18 54 65 196 Knights 9 12 29 9 15 9 15 30 59 Squires 27 27 82 22 74 16 60 65 216 Members of convents/ 39 36 105 36 107 16 43 88 255 monasteries Priests 41 30 93 24 80 18 52 72 225 Burghers 33 35 149 37 84 19 65 91 298 Members of mining 4 10 18 11 16 4 6 25 40 communities

Social variation In total 221 203 657 177 488 110 323 490 1468 Northern Närke 24 22 40 15 37 7 20 44 97 -The mining district 1 1 1 0 0 11 2 2 of Noraskogen -The hundred of 19 20 35 15 32 513 40 77 Örebro -The hundred of 4 4 4 3 5 6 6 13 15 Glanshammar Western third 20 21 50 12 20 14 27 47 97 Eastern third 31 25 82 23 62 17 50 65 194

Regional variation In total 75 68 172 50 119 38 97 156 388 1358–1399 52 24 89 14 66 18 71 56 226 1400–1432 63 38 209 36 138 21 100 95 447 1433–1466 56 48 202 32 134 17 69 97 405 1467–1499 50 38 148 35 123 17 65 90 336 1500–1520 49 62 263 69 247 28 181 159 691

Chronological variation In total 270 210 911 186 708 101 486 497 2105 and monasteries, which probably is due to the fact that verbs generally occur more often in a couple of specific types of charters, namely testaments and deeds of will. This group of charters contains a number of verbs of MLG origin which are not found in any other charters in the data base, e.g. begripa ‘agree, decide’, bevara ‘protect, keep’, forbarma ‘have mercy’, kosta ‘defray’, qvitta ‘reduce, settle, quit’, stikta ‘found, establish’, tøva ‘delay, put off’, and æra ‘honour, venerate’. The group of charters which has been found to contain the largest number of MLG loanwords was issued by the burghers of Örebro and by members of MLG loanwords in medieval charters issued in Närke 143 the mining communities. Of these two groups, the first impression is that the latter (the average number of MLG loanword types in charters issued by members of the mining communities is 6,3 (25/4), and the average number of tokens is 10 (40/4)) seems to contain a somewhat larger number of loanwords than the former (the average number of MLG loanword types in charters issued by burghers of Örebro is 2,8 (91/33), and the average number of tokens is 9 (298/ 33)). However, the charters issued by members of mining communities are fewer and occur very late; the four texts which constitute this group were all issued at the end of the 15th century or later. So, excluding charters issued by members of the mining communities, the charters issued by the burghers of Örebro on average contain the largest number of MLG loanwords, followed by those issued by members of convents and monasteries (e.g. mainly from the convent of Riseberga, but a few texts originate from the monastery of Örebro).

Regional provenance For the regional comparison, two categories of charters have been selected: those issued by rural judges and those issued by parish priests. The reason for this is that rural judges and parish priests can be linked to specific places within the province. Since a rural judge in late medieval Närke did not exclusively practise in one particular hundred but rather in a whole third (unless he worked in one of the northern hundreds), the division into thirds has formed the base of the regional comparison. It is true that the ecclesiastical administration did not make use of the same divisions, but charters issued by parish priests have nevertheless been categorized according to the same principle in this study. The hundreds which were not part of these judicial districts, i.e. those of Glanshammar, Sundbo and Örebro, and the mining district of Noraskogen, have been combined into a regional group of their own, whose the size better corresponds to the size of Västra tredingen (‘the Western third’) and Östra tredingen (‘the Eastern third’), both concerning geographical area and number of charters. This group of charters is called Norra Närke (‘Northern Närke’) in the study. Thus we have three regional groups of charters: Northern Närke containing 24 charters, the Western third containing 20, and the Eastern third containing 31 (see Table 1). Average years of issue of the charters in these groups are 1446 (Northern Närke), 1412 (the Western third), and 1417 (the Eastern third); so the charters issued in Northern Närke differ from the others in being on average younger. Although only charters issued by rural judges and parish priests were singled out in this part of the study, one should keep in mind that the centre for issuing charters in the Western third was the castle of Göksholm, with its resident noblemen; in the Eastern third it was Riseberga, with its convent; and in Northern Närke it was the town of Örebro, with its burghers. Thus charters from these environments might still to some extent be subject to influence from other social classes, although more or less indirectly. 144 Helena Wistrand

Figure 1: The geographical distribution of MLG loanwords within the province. (N=nouns, V=verbs, A=adjectives and adverbs combined.)5

5 N: bevarning ‘certainty, confirmation’; borghare ‘burgher’; borghmæstare ‘chief magistrate’; byte ‘exchange, barter’; damber ‘dam, mill-dam’; del ‘part’, ‘share, lot’; foghate ‘sheriff, bailiff’; forordh ‘reservation, condition’; gava ‘gift’; hovudsman ‘captain’; hælft ‘half’; iungfru ‘maiden’; klensmidher ‘locksmith, smith who repairs and manufactures small metal objects’; kompan ‘com- panion’; liverne ‘life, way of life’; makt ‘power’, ‘importance’; mata ‘degree, extent’; nødhthorft ‘necessities (of life)’; panter ‘pledge, pawn’; plata ‘(metal) plate’, ‘cuirass’; provaster ‘(ecclesiastical) dean’; redha ‘account’; riddare ‘knight’; rænta ‘income, levy’; rættoghet ‘right, claim’; sinne ‘mind, reason’; skomakare ‘shoemaker’; skræddare ‘tailor’; stadher ‘town’; thanke ‘thought’; vidhergiæld ‘compensation, payment’; vikt ‘weight’; vilkor ‘condition, stipulation’; væghna ‘behalf’; vælmakt ‘power, health, well-being’; væpnare ‘squire’; værkmæstare ‘foreman, superintendent of a craft guild’; æra ‘honour’. V: akta ‘intend’, ‘notice’, ‘respect, consider’; anama ‘receive, accept’; befala ‘order’, ‘entrust’; behalda ‘keep’; bekænna ‘confess’; beradha ‘confer, consider’; beredha ‘pay’; bestanda ‘be in possession of’; betala ‘pay’; bevara (sik) ‘concern (oneself)’; bevisa ‘prove, establish’; bliva ‘remain’, ‘become’; byta ‘change, exchange’; børa ‘ought to’, ‘be fitting for’; fata ‘stow (in drum or barrel)’; forbiudha ‘forbid’; forfylghia ‘claim, exact’; forlika ‘reconcile’; formagha ‘be capable of’; forvara ‘protect’, ‘keep’, ‘make secure’; forvissa ‘ensure, guarantee, promise’; fulbordha ‘fulfil’, ‘confirm’; føgha ‘ordain’; hantera ‘handle’, ‘discuss, consinder’; hindra ‘prevent, re- strain’; hopa ‘hope’; ivirsea ‘consider’; ludha ‘run, read’; rænta ‘yield, bring in’; tilstanda ‘con- MLG loanwords in medieval charters issued in Närke 145

Figure 1 constitutes an attempt to illustrate a somewhat more qualitative analysis of the MLG loanwords occurring in the source material. The words which are common to all three regional groups are also fairly common in the general charter vocabulary; many of them are part of formulas, such as til mere visso ok høghre bevarning (‘to make doubly sure and further protect’), medh akt ok forordh (‘with intent and conditions’), full makt ok vald (‘complete power and might’), i sva mato (‘to that extent’), a nokors væghna (‘on some- body’s behalf’), bekænnis medh thesso øpno brefe (‘avow with this public letter’), medh beradhno modhe (‘with aforethought’), beskedheliker man (‘sensible man’), ærliker man (‘respected man’). These words are all frequently used in the source material as a whole and occur throughout the research period 1375–1520. Except for borghmæstare, the nouns which are only found in charters issued by rural judges and parish priests in Northern Närke occur only once each. Skomakare, skræddare and værkmæstare are used as bynames, and epithets like borghare and borghmästare are both likely to be found here, since the town of Örebro is situated in the area. The verbs unique to this part of the province are found only once each in the material, and the one unique adjective occurs twice. This group of charters has the smallest number of types of MLG loanwords per text (see Table 1: on average 1,8 types per charter (44/24) compared to 2,4 (47/20) in the Western third and 2,1 (65/31) in the Eastern third)). This may seem strange, since the town of Örebro is situated in this part of the province, and one might expect a fairly extensive use of MLG loanwords in the towns compared to rural areas, not least since this is also the youngest group of charters in the regional comparison. However, looking at social provenance, the group of charters with the highest number of loanwords per text was issued by the burghers of Örebro, and these charters are not included in the comparison by regional provenance. A possible interpretation of these results is that lexical influence between different social classes in the area, or maybe more correctly between their respective scriveners, was not very extensive. The nouns which are only used in the charters of the Western third occur once each, of which klensmidher is used as a byname. The verbs unique to this group of charters are also only used once each, as are the adjectives and adverbs, except for yterlika, which is found twice. However, when looking at all the MLG loanwords occurring in the charters in this group, one finds that on fess, admit’; tygha ‘certify, attest’; upantvardha ‘transfer, convey, assign’; vilkora ‘bind, pledge’. A: beskedheliker ‘sensible, deliberate’; frir ‘free, unimpeded’; gansker ‘whole, entire’; gansklika ‘entirely, completely’; kranker ‘frail, infirm, ill’; lidhugher ‘free, free of liability’; likervis ‘as if’; lødhogher ‘stirling’; myndogher ‘authoritative, empowered’; mæktogher ‘authorized, empowered’; opinbar ‘public, official’; opinbara ‘publicly, officially’; pliktogher ‘bound, obliged’; qvitter ‘quits, free’; redhelika ‘clearly, plainly, lucidly’; redho ‘ready (money)’, ‘reliable, honest, scrupulous’; saligh ‘blessed, delighted, blissful’; strænger ‘strong, powerful, competent, prominent’; sunder ‘sound, healthy’; sunderlika ‘particularly, above all’; yterlika ‘further, additionally, in greater detail’; ælænde ‘miserable, unfortunate, in distress’; ærliker ‘respected, esteemed, distinguished’. 146 Helena Wistrand average they actually contain the highest number of types of MLG origin per text (see Table 1: on average 2,4 tokens (47/20) of MLG loanwords per charter). This is interesting, since this group is also the oldest of the three. One possible reason for this might be some influence from the convent of Riseberga, as- suming that the nuns there were mainly members of the nobility. The comparison according to social provenance, which is briefly accounted for above, shows that charters issued by members of the convent (combined with a few charters issued by members of the monastery of Örebro) on average contain a larger number of MLG loanwords than those issued by rural judges and priests. Also, the texts of the charters issued at the convent of Riseberga tell us that rural judges and priests from time to time were involved in issuing charters in the convent; that there had to be some contact and influence between these priests and the members of the convent is scarcely questionable. The group of charters issued by rural judges and parish priests of the Eastern third contains the highest number of separate instances of MLG loanwords. The explanation for this is, of course, that it also comprises the largest group of charters in the regional comparison. The nouns which are unique to the charters issued by rural judges and parish priests in the Eastern third are all only found once, except for plata, which is used twice as a byname. This is also the case concerning all of the verbs found only in this set of data. Of the adjectives and adverbs unique to this group, gansklika and pliktogher occurs three times each, while the rest occur only once. The number of lemmas which are unique to this group of charters is considerably higher than it is for either of the other groups. The Eastern third contains the castle of Göksholm. In the late Middle Ages, important members of the Natt och Dag family, which belonged to the high nobility of medieval Sweden, resided there. In the 15th century the knights Bengt Stensson and his son Magnus Bengtsson lived there, and both in turns held the office of chief judge of Närke. The son married a woman of German birth, namely Ermegård Fickesdotter Bülow, who also issued a few charters which appear in the source material. Compared with the circumstances mentioned in Northern Närke, this might be interpreted to mean that lexical influence between social classes was more extensive in the Eastern third than in Northern Närke.

Chronological provenance All charters in the source material issued in the 14th century (i.e. 1375–1399) form the oldest group of charters, and all charters issued in the 16th century (i.e. 1500–1520) form the youngest group. Those issued in the 15th century have been divided into three periods: 1400–1432, 1433–1466, and 1467–1499. The five chronological groups are adequate for seeing a change over time, and since every group contains a large enough number of charters, the results of the comparison are reasonably reliable. As one might expect, the average number of MLG loanwords increases MLG loanwords in medieval charters issued in Närke 147 steadily over time, except for the combined number of adjectives and adverbs, whose frequency is fairly constant throughout the four first periods; they even decrease in number to some extent. Between the fourth and fifth periods, a significantly larger increase in the number of nouns and verbs of MLG origin appears. During this time the adjectives/adverbs also increase in use somewhat. This increase in the use of MLG loanwords at the turn of the 16th century is partly explained by the nature of the charters issued in the latest period. For some reason, the number of preserved deeds of transfer from the end of the 15th century is not as extensive as it is from the 14th century and the bulk of the 15th century. The majority of charters from the period 1500–1520 consist of texts of a non-official nature – private letters concerning administrative matters – and it is natural that these are more varied in their wording than the deeds of transfer, since the latter are more dependent on formulas and model texts. Neverthe- less, the texts from the youngest group of charters are lexically influenced by MLG to a remarkable extent, as well as morphologically, e.g. the MLG prefix be- often has the form bi-, and the suffix -are often has the form -er(e).

Conclusion There are many problems connected to the completion of a study like this. One drawback is that in most cases we do not yet know who wrote the Swedish medieval charters, or to what extent the scribe relied on formulas and model texts or drafts. Another problem has to do with the survival and conservation of source material; in this study, the lack of preserved charters of a regular official nature from the late 15th and early 16th centuries means that texts in the data base from this period are in most cases slightly different in nature from the other time frames, as they contain a more varied vocabulary, and they are therefore not completely comparable to the rest. These are factors which often make comparisons difficult, but as always when dealing with historical material, we are compelled to manage with what we in fact have available. On the other hand, there is one major advantage to the use of charters as source material for studies in the field of historical linguistics. In contrast to most OSw texts, which in most cases have more or less been definitely and accurately dated by scholars, the scribes of medieval charters almost always care- fully noted down the date, year and place of issue, which renders it possible to make relatively reliable regional and chronological comparisons. This makes charters well suited for a study with regional and chronological perspectives, such as the one partly summarized and accounted for in this paper. It has not been possible to give a complete and comprehensive picture of the distribution and propagation of MLG loanwords between different scriptorial environments in the province of Närke. However, the study hopefully gives a fairly accurate representation of the occurrence, range, distribution and use of MLG loanwords in different kinds of scriptorial contexts within the province. 148 Helena Wistrand

A few possible paths of lexical diffusion between different environments and/ or social classes have been suggested in this paper, and a couple of scribes whose complete text production might be interesting to study from a language contact perspective were pointed out.

Sources and literature Sources: archives, collections and charters referred to RA = Riksarkivet (The Swedish National Archives). RAp = Pergamentsbrev (‘charters on parchment’), kronologisk huvudserie (‘chronological main series’), RA. Stu = Sturearkivet (The Sture Archive), RA. Sä = Sävstaholmssamlingen (The Sävstaholm Collection), RA. 11942 = RAp, Örebro 9 July 1381. 12079 = RAp, Segersjö 3 January 1382. 12848 = RAp, Mellösa 9 April 1385. 13031 = RAp, Riseberga 7 January 1386. 13816 = RAp, Askers kyrka 21 October 1390. 13979 = RAp, Riseberga kloster 12 November 1391. 14176 = Sä, Örebro 6 May 1393. 34517b = Stu no. 262, Örebro 28 March 1502 (loose bill enclosed with 34517 = Stu no. 262, Örebro 28 March 1502). 34680 = Stu no. 610, Örebro 31 October 1503. 34803 = Stu no. 203, Örebro 25 April 1504. 34822 = Stu no. 217, Örebro 15 May 1504. 34826 = Stu no. 625, Örebro 19 May 1504. 34863 = Stu no. 204, Örebro 24 June 1504. 34864 = Stu no. 204b, Örebro 24 June 1504. 34865 = Stu no. 204c, Örebro 24 June 1504. 35512 = Stu no. 1393, Örebro 8 September 1506. 35695 = Stu no. 875, Örebro 24 January 1507. 35808 = Stu no. 829, Örebro 9 April 1507.

References Almquist, Jan Eric, 1955: Lagmän och häradshövdingar i Sverige c:a 1350–1950. I al- fabetisk ordning uppställda. Stockholm. (Skrifter av Rättsgenetiska institutet vid Stockholms högskola.) Dahlstedt, Karl-Hampus, 1969: Vad är främmande ord? In: Karl-Hampus Dahlstedt, Gösta Bergman & Carl Ivar Ståhle (eds.), Främmande ord i nusvenskan. 2nd edn. Stockholm. (Verdandis skriftserie 17.) Pp. 1–31. Hellquist, Elof, 1948: Svensk etymologisk ordbok. 3rd edn. Lund. Jonasson, Gustaf, 1984: Medeltidens Örebro. Malmö. (Örebro studies 3.) Lasch, Agathe & Borchling, Conrad, 1928 ff.: Mittelniederdeutsches Handwörterbuch. Ed. Gerhard Cordes. Neumünster. Lovén, Christian, 1999: Borgar och befästningar i det medeltida Sverige. 2nd edn. Stockholm. (Kungl. Vitterhets Historie och Antikvitets Akademiens handlingar. An- tikvariska serien 40.) MLG loanwords in medieval charters issued in Närke 149

Moberg, Lena, 1989: Lågtyskt och svenskt i Stockholms medeltida tänkeböcker. Upp- sala. (Acta Academiae Regiae Gustavi Adolphi 58.) NEO = Nationalencyklopedins ordbok. Utarbetad vid Språkdata, Göteborgs universitet. 1–3. 1995–1996. Höganäs. Pers = Gamla papper angående Mora socken 2. Arvid Siggessons brevväxling. Ed. An- ders Pers. Commented by Lars Sjödin. 1932. Västerås. SAOB = Ordbok över svenska språket utgiven av Svenska Akademien 1–. 1893 ff. Lund. Schiller, Karl & Lübben, August, 1875–1881: Mittelniederdeutsches Wörterbuch. Bremen. Styffe, Carl Gustaf, 1911: Skandinavien under unionstiden med särskildt afseende på Sverige och dess förvaltning åren 1319 till 1521. Ett bidrag till den historiska geo- grafien. 3rd edn. Stockholm. Söderwall, K. F., 1884–1918: Ordbok öfver svenska medeltids-språket 1–2. Stockholm. (Samlingar utg. av Svenska fornskriftsällskapet. Ser 1. Svenska skrifter 27.) Söderwall, K. F., Åkerlund, W., Ljunggren K. G. & Wessén, E., 1925–73: Supplement. Stockholm. (Samlingar utg. av Svenska fornskriftsällskapet. Ser. 1. Svenska skrifter 54.) Tunberg, Sven, 1911: Studier rörande Skandinaviens äldsta politiska indelning. Uppsa- la. Utterström, Gudrun, 1968: Fem skrivare. Metta Ivarsdotters brev till Svante Nilsson. Studier i senmedeltida svenskt brevspråk. Stockholm. (Acta Universitatis Stockhol- miensis. Stockholm studies in Scandinavian philology. New series 7.) Wiktorsson, Per-Axel, 1983: Örebro historia. Personnamnen i det medeltida Örebro. Örebro. (Högskolan i Örebro. Skriftserie 30.) Wistrand, Helena, 2006: Bebrevat i Närke. Lågtyska importord i fornsvenska brev ur re- gionalt perspektiv. Uppsala. (Samlingar utgivna av Svenska fornskriftsällskapet. Se- rie 1. Svenska skrifter 88.) 150 Helena Wistrand The history of complex verbs in Scandinavian languages revisited 151 The history of complex verbs in Scandinavian languages revisited: only influence due to contact with Low German?

Kurt Braunmüller, University of Hamburg and Steffen Höder, University of Kiel

Introduction The ruling doctrine in Germanic language history states that early Germanic had verbal prefixes which, however, were lost in North Germanic in historical times (Krahe & Meid 1967, p. 42 [§ 52]): Im Nordgermanischen … schwinden sowohl beim Verbum als auch beim No- men sämtliche unbetonten Praefixe entweder völlig oder werden stark reduziert. [In North Germanic … all unstressed prefixes disappear in verbs and in nouns as well, either totally or they became greatly reduced.] Consequently, prefix verbs and other kinds of morphologically complex verbs (including particle verbs) that are found in Modern Scandinavian must reflect later innovations, particularly language contact with Middle Low German and, to a lesser extent, Medieval Latin during the (Late) Middle Ages. The introduction of verbs such as Swedish betala ‘to pay’, förstå ‘to understand’ or miss- förstå ‘misunderstand’ has indeed been considered one of the most salient replications from West Germanic that entered the (Mainland) Scandinavian languages during the era of the Hanseatic League – along with some derivational suffixes such as -he(i)t or -else (equivalent to English -ing). This general view is based on the assumption that Ancient Germanic possessed verbal prefixes, not only the prefix ga-/gi- used in the formation of the preterite participle in West Germanic, but a word-formation strategy involving verbal prefixation in general. The extant verbal prefixes are thus taken to be inherited from Ancient Germanic, and it is assumed that they were already an established part of the Germanic verbal systems in the earliest period, as is illustrated by the following quote: Grundsätzlich gilt für das Altgermanische, daß in der verbalen Komposition die Praefixe unbetont, in der nominalen betont sind. (Krahe & Meid 1967, p. 41.) [It is taken for granted as far as Old Germanic is concerned that the prefixes in verbal compositions are unstressed whereas as they are stressed in nominal ones.] 152 Kurt Braunmüller and Steffen Höder

However, some questions still remain: (a) Where did the respective derivational patterns originally come from? Were they of Germanic or non-Germanic origin? (b) How were they integrated into the verbal systems of the Scandina- vian languages? (c) Did the different prosodic patterns associated with the various types of complex verbs play a role in these processes? In this contribution, we approach these questions from a wider typological and contact linguistic perspective. Our basic claim is that all kinds of complex verb constructions in Modern Scandinavian are ultimately the result of, or have at least been heavily influenced by, language contact, via a transmission chain that involves different, and in part layered, language contact situations and both direct and indirect transfer from the source languages. Furthermore, we claim that this process is evidence for contact-induced complexification, as opposed to the simplification phenomena often observed in language contact situations such as that between Low German and Mainland Scandinavian during the Han- seatic era (cf. Braunmüller, this volume).

Prefix verbs: a contact feature? Hypotheses In contrast to the current theory, we hypothesise that the Ancient Germanic dialects did not possess any verbal prefixes or any prefixation patterns of the type that occurs in the Modern Germanic languages, and that these elements consequently were not “lost and re-invented” in the later development of Ger- manic. The main argument for this view is empirical: the fact that prefix verbs or similar structures are not attested in the oldest West and North Germanic varieties as represented by the extant sources, viz. the (nearly) 400 runic inscriptions written in the Older Futhark (of which only 117 are analysable syntactically; cf. Beuerle and Braunmüller 2004, p. 7ff.). In these inscriptions we do find morphologically simple verbs combined with directional adverbs such as ut/uti ‘out of, outside’ or i/hi ‘into, in, hither’ (Krause 1971, p. 121ff.), but no constructions that are lexicalised or show some other signs of a ‘special status’ within the language system. On the other hand, the main evidence in favour of the existence of verbal prefixes in Ancient Germanic comes (solely!) from Gothic – a language that was, however, heavily influenced by Biblical and Byzantine Greek, in particular in the syntax, and of which no non-translated texts have survived (leaving the three runic inscriptions1 aside). Indeed, Gothic has some verbal prefixes (cf. Braune and Ebbinghaus 1981, p. 134f. [§ 217a]). But the use of prefix verbs

1 Cf. the runic inscriptions of Pietroassa (on a ring, found in Romania) and the spearheads of Dahmsdorf-Müncheberg (southern Germany) and Kovel (Ukraine). None of them contains a verb. The (partly East Germanic) fibula of Charnay, found in Burgundy (France), might, however, contain a prefix verb: unþf[i]nþai ‘to find out’, but there are other divergent readings as well. The history of complex verbs in Scandinavian languages revisited 153 and the word formation pattern itself could also have resulted from the intense contact with Greek, as they were needed for an appropriate translation of Greek source texts into this (isolated) East Germanic dialect. Most obviously, prefix verbs in Gothic very often occur as translation equivalents of structurally and semantically parallel verbs in the Greek sources (such as urrinnan = ἐξέρχεσθαι ‘to go out’, anameljan = ἀπογράφειν ‘to write down’, anaqiman = ἐφιστάναι [< ἐπ-ἱστάναι] ‘to come upon’; examples from the Gospel of St. Luke 2: 1–9).2 We find this observation very instructive, as it suggests that language contact is not only responsible for the later establishment of prefix and particle verbs in Modern Scandinavian, but that even the oldest prefix verbs attested in Germanic could be the result of language contact. If the model of Gothic – a prestigious, written, dominant language with its important religious, literary, and scientific domains as well as its function as a lingua franca in the Byzan- tine Empire – led to, or at least reinforced, the emergence and use of innovative derivational patterns in a less prestigious Germanic vernacular, one should also expect similar developments in Germanic languages spoken in the area where the dominant lingua franca and literary language was Latin (a language as rich in prefix verbs as Greek is). Although direct empirical evidence for such a development is lacking, for want of comparably early sources, we can judge the hypothesis indirectly based on its predictions: if Latin is the source of prefix verb constructions in Germanic, we would expect the emergence of prefix verbs to be earliest and most thorough in those Germanic languages that were exposed to Latin earliest and in the most intense contact situations, geographically closest to the Roman Empire, Christianised first, and developed into written languages (in the sense of Kloss’s 1978 ‘language Ausbau’; cf. the model proposed by Höder 2010, p. 73ff.) earlier than others. Consequently, our hypothesis is that prefix verbs in all Germanic dialects are due to language contact, either with Ancient Greek (only for Gothic and possibly for other East Germanic dialects such as Burgundian) or with Latin (for all other Ancient Germanic dialects), and that they entered these dialects through intense contact or even via translations of (predominantly religious) texts into the Germanic vernaculars. Prefix verbs are thus replications of constructional patterns in the respective model languages, caused by the need to express more complex and differentiated verbal meanings as found in the contact languages or the source texts. This does not necessarily mean that individual lexical items were calqued or imitated by Germanic speakers – although this has happened as well – but rather that a word-formation strategy as such was adopted in the respective languages, while the lexical material was still Germanic. In addition, it is important to mention that the replication of a constructional pattern from a model language does not rule out the possibility that Germanic speakers occasionally used similar constructions independently of

2 The Greek influence is but one aspect of the contact history of Gothic. Several features of Gothic grammar seem to indicate that Gothic must have been in contact with various other languages as well (cf. van Coetsem 2000, p. 200ff.). 154 Kurt Braunmüller and Steffen Höder language contact as well. But the existence of a model construction of course reinforces (if not initiates) the stabilisation, systematic expansion, and grammatical functionalisation of such patterns. (For more details on the notion of ‘replication’, see Heine and Kuteva 2005; Johanson 2008 has coined the term ‘code copying’ for basically the same phenomenon.) Seen from a North Germanic perspective, this means that we assume a chain of repeated, direct and indirect replications of Latin constructions and their subsequent stabilisation and integration into the systems of the recipient languages, which of course involved interaction with pre-existing indigenous structures (see further below for details). The West Germanic languages, especially Middle Low German, were actually no more than a transmitting link from Latin to Scandinavian and not the original source of the emerging complex verb patterns. These are either secondary or transmitted replications from Latin via Low German or indeed, once more, primary replications of Latin models. Another consequence is that (some) prefix verbs in Faroese (and, to a lesser degree, in Icelandic as well) are actually tertiary replications, based on previous secondary replications from Danish as the dominant contact and/or colonial language over a long period of time.

Background: Ancient Germanic as a contact language Transfer mechanisms similar to the proposed replication of prefix verbs are by no means new for the Germanic dialects in general. As has been demonstrated in detail in Braunmüller (2007, 2008a,b), Ancient Germanic has to be regarded as a contact language, which emerged in a period of intense contact with people who must have spoken a very divergent (more precisely, a non-Indo-European) language, and which only became more complex during its later development. The foreigners reduced the morphology of the old (pre-)Germanic dialects considerably due to incomplete second language acquisition. The result in this case was a restructured rudimentary Indo-European contact dialect with obvious traces from a language in which ablaut played an important role in word formation and especially in verbal tense formation. (For more details see, e.g., Vennemann 2003, p. 653ff., ‘Germania Semitica’ and passim, and particularly Mailhammer 2007 on the so-called strong or irregular verbs in Ancient Germanic.) Ancient Germanic, as documented in its earliest sources, had e.g. (a) a simplified tense system (reduced to present and past tense), (b) no inherited voice distinction (only the common Germanic medio-passive verb haita ‘to be called’ has been preserved, and some residual passive forms are attested in Gothic), (c) no inherited subjunctive (only older optative) mode and reduced imperative forms, (d) fossilised and extremely simplified ablaut patterns in the verbal system (based on five e/o and two a/o classes) and (e) fixed (phonetic) word-initial stress which was not distinctive phonologically. Classical Latin and Greek were, as is well-known, far more complex and more typically Indo- European in this respect: their verbal systems were considerably more com- The history of complex verbs in Scandinavian languages revisited 155 plex, but ablaut had only a marginal status; they had an elaborated tense system, including passive voice and subjunctive forms. Moreover, Ancient Germanic shows some other features which are typical for creolised or grammatically simplified languages, such as the emergence of (f) a new past tense formation element, viz. the so-called weak or regular preterit, formed by a dental suffix (-þa/-ða; one of the two innovations specific for early Germanic; for the second one, see (h) below). This new suffix is a typical result of a far-reaching reduction and simplification process, which is not unusual for creolised languages, in which either new tense-mood-aspect (TMA) markers emerge or functional verb constructions take over tense formation, based on un- specified verbs like ‘to make’ or ‘to do (!)’ (cf. Muysken 2000, ch. 7 on ‘Bi- lingual Verbs’; for a more detailed argumentation, see Braunmüller 2008a). In addition, early runic Germanic shows (g) a tendency to develop replicated passive forms as well. As we can see from the oldest runic inscriptions in the Older Futhark, the only inherited medio-passive verb, haita ‘to be called’ (Ger. heißen, Sw. heta), in some instances gets a cliticised but sometimes redundant personal pronoun marker (-ek[a] ‘-I’). A more detailed analysis (see Braunmül- ler 2004a,b,c) suggests, however, that the writers of these inscriptions tried to imitate (or ‘replicate’) the synthetically formed passive forms of Latin that they were familiar with.3 The forms haitika (Zealand 2 / Køge) and ek … ha(i)teka (Lindholm) ‘(I) am called-I’ represent this stage of replication, which later became generalised in forms such as rAisidoka (Ellestad: ‘was erected’), tojeka (Noleby: ‘I experience/get [benefactive]’) and felAhekA or fAlAhAk (Stentoften and Björketorp, respectively: ‘is concealed’). This point shows that the Latin verbal system was in the focus of bilingual writers who tried to find and establish forms that could be taken over into their Germanic vernaculars in order to live up to the many grammatical differentiations which were a default for the model language Latin but absent in the early Germanic dialects. Therefore, these grammatical replications have to be considered parts of a more general expansion process that typically occurs when creolised languages become more widely used and have to fulfil more differentiated functions (cf. Trudgill 1992, p. 32). Last but not least, the second innovation of Germanic should be mentioned: (h) so-called weak adjective inflection. Originally occurring as a definiteness marker and a word-formation element in its own right which developed into some sort of a corollary in a definite noun phrase containing an attribute, it has to be considered a result of language contact, but this time as an instance of imperfect replication (for further details on that issue, see Braunmüller 2008b). ‘Imperfect’ because the word-formation element IE -an(t), also found in Hittite

3 We cannot go into details here. The (almost) “perfect fit” (cf. Derolez 1998, p. 112ff.) of the runic writing system with the Germanic phoneme system suggests that the few writers of the oldest Germanic runic inscriptions must have been bilingual in a way, with deep(er) insights into the grammatical system of Latin and into the wide-spread scribal practice in the Roman empire in general (all references in the text above). 156 Kurt Braunmüller and Steffen Höder with a similar specifying or singling-out function, was perceived as an inflectional ending and no longer as an Indo-European word-formation morpheme by these second language learners. This is why the weak adjective inflectional paradigms in all Germanic dialects are rather uniform but show only a few inter-paradigmatic alignments (e.g. the forms of the genitive and the dative plural are alike in all nominal paradigms). Our claim that the complex (or prefix) verbs are not of Germanic origin but replicated from a model language fits in with the context of these developments: it is not an isolated phenomenon but part of a far-reaching and deep- going expansion process of an originally simplified/creolised Indo-European variety.

Classification We define ‘complex verbs’ in the Germanic languages as conventionalised and potentially lexicalised constructions consisting of a verb and a particle in arbitrary order and with the particle either adjacent to the verb (and potentially univerbated in prosodic or graphic terms) or in a distant position. This includes various classes of complex verbs in the different Germanic languages, among others verbs with separable and inseparable prefixes in German (erwerben ‘to acquire’, anfangen ‘to start’ [but: er fängt an ‘he starts’]), prefix and particle verbs in Swedish (bebo ‘to inhabit’, angå ‘to concern’, tycka om ‘to like’), or phrasal verbs in English (e.g. get up). Complex verbs can be classified according to (a) the relative word order of their components (P[article] V[erb] vs. V[erb] P[article]) and (b) the distribution of stress (stressed verb vs. stressed particle). Stress can be indicated by a straight apostrophe as in Sw. ta 'av ‘to turn off [a road]’ vs. 'avta ‘to decrease’). These parameters yield the classification in Table 1:

Table 1: Types of complex verbs

Type Structure Examples I P'V Sw. för'klara ‘to explain’ Da. be'holde ‘to keep’ Eng. for'give Ger. zer'reißen ‘to tear [apart]’ II 'PV Sw. 'avgöra ‘to decide’ Da. 'afskaffe ‘to abolish’ *Ger. 'aufgeben ‘to give up’ [but: gibt 'auf ‘gives up’] III V'P Sw. stå 'upp ‘to stand up’ Da. se 'ud ‘to look, to seem’ Eng. get 'up *Ger. gibt 'auf ‘gives up’ [but: 'aufgeben ‘to give up’] *IV 'VP Sw. 'börja med ‘to begin with’ Da. 'tænke på ‘to think of’ Eng. 'wait for Ger. 'trauern um ‘to grieve for’ The history of complex verbs in Scandinavian languages revisited 157

As the PV order – in Germanic – coincides with adjacency, types I and II correspond to the term ‘prefix verb’, whereas type III is the ‘particle verb’ or ‘phrasal verb’ construction. Type IV contains those constructions in which the particle is normally analysed as a preposition introducing a prepositional object, i.e. a prepositional phrase as an obligatory argument of the respective verb. This type is mainly included here for the sake of completeness. In older Germanic it is, however, often impossible to distinguish between types III and IV, as there is little (if any) evidence for stress in written sources. The same holds in principle also for types I and II, although phenomena of phonetic reduction (reduced vowel inventory, loss of vowels) in pre-tonal syllables often hint at the position of stress. Additionally, two language-specific restrictions apply: in German, types II and III are coextensive because the position of the particle is not distinctive but varies depending on the clause type and the word order rules.4 In Icelandic, types I and II coincide categorically because of the non-distinctiveness of stress, which always is word-initial. If we now compare the distribution of these types among the Modern North Germanic languages, Modern German,5 Ancient Germanic, and the (hypothet- ically) relevant contact language, Latin, we get an interesting picture:

Table 2: Distribution of complex verb types

Type I (P'V) Type II ('PV) Type III (V'P) Type IV ('VP) Ancient Germanic – – ++ ++ Modern Icelandic + ++ ++ Modern Faroese + + ++ ++ Modern Norwegian ++ ++ ++ ++ Modern Swedish ++ ++ ++ ++ Modern Danish ++ ++ ++ ++ Modern German ++ ++ ++ Latin ++ –++

This distribution suggests three inferences: (a) the particle-final constructions are an inherited pattern; (b) the particle-initial constructions are innovative and may have spread geographically from the southernmost Germanic dialects to the north, reaching Insular Nordic later than the Scandinavian Mainland, result-

4 Similarly, minor language-specific rules may cause additional ambiguity. For example, types II and III are also indistinguishable in Swedish participles; cf. the following example: en politiker som alla tycker 'om är en 'omtyckt politiker ‘a politician that everyone likes is a liked [= popular] politician’. However, such rules do not obscure the classification of the respective complex verbs in general, unlike in the German case. 5 Modern German is taken to be representative for West Germanic in general, although excluding English because of its unique contact history within the Germanic family, even if many developments are apparently comparable with the continental Germanic languages (for a more comprehensive discussion, cf. Hiltunen 1983). 158 Kurt Braunmüller and Steffen Höder ing in a lesser degree of establishment of prefix verbs; (c) Latin is a potential source of this innovation, as particle-initial constructions are common, and in fact are the only possible complex verb type in this language (for the coincidence of types I and II, see below).

Contact scenarios and stages General scenario If Latin indeed is the source of Germanic prefix verb constructions, they must have spread into Germanic in general and North Germanic in particular via identifiable (i.e. sociolinguistically describable) contact situations. The most likely general scenario involves (at least) four different steps in three different contact situations and a multi-layered interplay of different replication stages. The common denominator of these four contact stages is that they all involve a politically and/or culturally dominant language already possessing prefix verbs in contact with a dominated language without prefix verbs. Step 1: contact between Latin and the oldest West Germanic dialects in Late Antiquity and the Early Middle Ages. The most important domain for this contact was the function of Latin as a lingua franca and, additionally, Christianity and the Latin educational system based on writing. The direct replication of Latin prefixation patterns resulted in the emergence of type I (P'V) in West Germanic (see below for details). Step 2: contact between Middle Low German and Old Nordic in the Middle Ages, including widespread bilingualism (of various types and degrees) in Mainland Scandinavia. The replication of Low German prefix verb constructions (i.e. the indirect replication of Latin patterns) led to the establishment of type I (P'V) in Mainland Scandinavian – this is uncontroversial. Step 3: contact between Medieval Latin and the Old Nordic languages during about the same period. Replication of Latin prefixation patterns resulted in the grammaticalisation of type II ('PV) constructions in Scandinavian, partly as a variant of corresponding constructions of the indigenous type III (V'P; see below for details on Old Swedish as an example). Step 4: contact between Danish and Insular Nordic in Iceland and on the Faroe Islands, as a consequence of the status of these countries as dependen- cies, where Danish has had (or, in the case of the Faroes, still has) official status, and where a varying proportion of the population has been (or still is) fluently bilingual over a long period of time. Replication of Danish prefix verbs reinforced the establishment of types I and II (i.e. a secondary or even tertiary replication of Low German and Latin patterns; see below for some additional remarks). Table 3 shows the timeline for the different replication stages: The history of complex verbs in Scandinavian languages revisited 159

Table 3: Replication stages

Late Antiquity/ Middle Ages Late Middle Ages/ Early Middle Ages Modern Times

Primary Latin (I~II) Latin (I~II) replication ↓ ↓ West Germanic (III) Nordic (III) + I + II

Secondary Low German (I, II~III) replication ↓ Mainland Scandinavian (III) + I

Tertiary Danish (I, II, III) replication ↓ Insular Nordic (III) + I, II

In the following two sections, we discuss some constructional, semantic, and prosodic aspects of the two stages of primary replication in more detail. This is followed by some remarks on the situation in Insular Nordic.

Latin → West Germanic Ancient Germanic, according to our hypothesis, did not have any genuine prefix verbs; rather, the only complex verb constructions were of types III and IV, i.e. particle-final constructions with the particle either adjacent to the verb or in a more distant position, as a separate phonological word. Latin, in contrast, had only prefix verbs (type I~II), i.e. particle-initial constructions such as deponere ‘to put down/away’, praedicere ‘to proclaim, to preach’, convenire ‘to come together’. The Germanic particles had, in most instances, a concrete local or directional meaning which were also kept when used in lexicalised combinations with verbs (as, say, in Modern Swedish ta bort ‘to take away’, Modern Dan- ish gå ud ‘to go out’). These constructions thus exhibited a lesser degree of semantic complexity than the prefix verbs in Latin, which were subsequently replicated in West Germanic. Latin prefix verbs could also convey concrete meanings (such as de- ‘down’ in deponere ‘to put down’ or in- ‘in[side]’ in ingredi ‘to go in, to enter’). Frequently, however, the prefix verbs had (additional) abstract meanings, which implies complex rather than compositional semantic units, since the meaning of the complex verb cannot be derived or predicted from the meaning of its components (cf. deponere in the senses of ‘to get rid of; to give up; to entrust’ or ingredi used to mean ‘to begin, to com- mence’). This often includes a difference in aspectual meaning or in argument structure (cf. [gradi + (in + accusative)] ‘to go + into sth.’ vs. [ingredi 160 Kurt Braunmüller and Steffen Höder

+ accusative] ‘to enter sth.’). The replicated derivational pattern in West Ger- manic conveys the same semantic modification, as is reflected e.g. in Modern German ver- with its aspectual (resultative) meaning in verbs like verbrennen ‘to burn [completely]’; or be-, which can turn an optional adverbial into an obligatory argument (cf. the prepositional construction in [sprechen + (über + accusative)] ‘to talk (about sth.)’ vs. the direct object in [besprechen + accusative] ‘to talk about, to discuss’). The latter use of be- is still quite productive in German (cf. bespielen ‘to play on sth. [a sports ground]’, belasern ‘illuminate sth. [with laser beams]’). The replication of Latin derivational patterns in West Germanic apparently involves not only the transfer of a word-formation strategy but also the functionalisation of different stress patterns. Similar to today’s Icelandic, Ancient Germanic had fixed stress on the word-initial syllable, i.e. the word accent was not phonologically distinctive but had a purely delimitative function. Germanic verbs, including inflected forms, were therefore typically trochees accented on the stem ('σσ); dactyls may have occurred but they were marginal (cf. OHG be- rumes ‘we bear’, 'σσσ). Phonologically speaking, the Latin stress system was quite similar. Stress was also fixed and oriented towards the end of the phonological word, and its position was clearly predictable from the segmental and prosodic context: in disyllabic words, the penultimate syllable was accented ('σσ, cf. 'ergo ‘thus’); in trisyllabic words (or words with even more syllables) the penultimate was stressed if ‘heavy’ (containing a long vowel or a vowel followed by certain consonant clusters: -σ'σσ; cf. mor'tālis ‘mortal’, mā'ternus ‘maternal’), otherwise the antepenultimate was stressed (-'σσσ; cf. 'lacrima ‘tear’). As for the prefix verbs, this had the effect that the accented syllable normally was within the verbal component, either in the stem (as in dē'pōnō ‘I put down’) or in the inflectional suffix (e.g. ingredi'untur ‘they enter’). Only in marginal cases did the stress fell on the prefix (in forms like 'ēligō ‘I elect’).6 The earliest prefix verb constructions in West Germanic are of type I (P'V) – a clear violation of the original Germanic word-initial stress pattern. If type I constructions in West Germanic, as we argue, emerged as a replication of Latin prefix verbs, then the deviant stress pattern in these constructions can be explained as a transfer of the Latin pattern that was taken over along with the constructional replication. This involves a process of functionalisation: the transfer of an accidental, phonetically determined stress pattern from Latin resulted in an innovative system in Germanic, where stress became marginally distinctive in that non-initial stress marks a structurally complex verb and also carries se-

6 The rarity of stressed verbal prefixes results from the concurrence of different factors. For a prefix to be stressed, the following conditions are required: (a) the stem syllable of the simplex has to be ‘light’ (i.e. contain a short vowel, not followed by consonant cluster); (b) the inflected simplex verb is mono- or disyllabic; (c) the inflectional suffix is monosyllabic or non-syllabic. One consequence is that while stress always falls on the verbal component of most prefix verbs, no prefix verb always gets stressed on the prefix (cf. 'ēligō ‘I elect’ vs. ē'ligimus ‘we elect’). The history of complex verbs in Scandinavian languages revisited 161 mantic information by indicating some degree of semantic complexity (aspectual modification, for example). This emergence of a more complex phonological, morphological and semantic system is another example of a contact-induced increase in complexity (cf. Dahl 2009) as a result of linguistic creativity in a contact situation.

Latin → Old Swedish The second scenario in which primary replication (directly from Latin) plays a role is contact between Latin and Old Scandinavian in the Middle Ages. This contact in general has been investigated in some detail for Old Swedish, and we will consider our discussion to be, mutatis mutandis, representative of the medieval development in mainland Scandinavia. The contact between Old Swedish and Latin is related to the beginning of the vernacular being used as a written language by a culturally elite group of Latin-Swedish bilingual speakers, mainly clerics. The main locus of language Ausbau was thus the bilingual milieu and bilingual text production (which increased rapidly after 1400) in ecclesiastical institutions, first of all in the ‘great translation workshop’ (Wollin 1991) in Vadstena Abbey (for a detailed analysis of the translational practices in Vadstena, see Wollin 1981/83). During this period, a range of (mainly syntactic) constructions were replicated from the Latin model, which resulted in the emergence of innovative structural patterns in the written variety of Old Swedish (clause-linking strategies, subjunction in- ventories, relative clause types, and more; cf. Höder 2009, 2010; Höder and Zeevaert 2008; for the role of translation in such replication processes, see Kranich, Becher and Höder 2011). Most of these innovations were restricted to (or at least specifically dominant in) the written language, a restriction that is still perceivable in today’s Swedish, where such constructions are either archaic or stylistically marked as more or less formal and thus avoided in everyday spoken language. Complex verb constructions of type I ('PV) may well fit into this category. In the (Late) Middle Ages, Swedish had only complex verbs of type III, i.e. the inherited V'P constructions, and, as a result of contact with Middle Low German, prefix verbs of the type I (P'V). Type III constructions still con- veyed relatively concrete, compositional meanings, and included for the most part motion verbs and directional particles. Concrete (but semantically modified) meanings prevailed even in type I constructions. Some Latin prefix verbs were semantically equivalent to Swedish V'P constructions, but some also represented metaphorical lexical concepts without a direct Old Swedish equivalent. Type II ('PV) constructions appeared at first as variants of corresponding type III constructions consisting of the same verbs and particles (cf. läggia nidher vs. nidherläggia, in the sense of Latin deponere ‘to put down; to give up …’). Both types, prefix verbs and particle verb constructions, formed a com- 162 Kurt Braunmüller and Steffen Höder mon class of verbs (a hybrid type II~III, similar to Modern German), without any specific lexicalisation of or semantic differentiation between the two types. However, quantitative studies (Ljunggren 1932 as well as first results from our own Old Swedish corpus7) show that by far the most occurrences of type II pattern are found in younger texts from the Late Old Swedish period (yngre fornsvenskan, c. 1375–1526). This is particularly true for texts from Vadstena Ab- bey, i.e. religious prose written or translated from Latin originals by trained bilingual scribes during the 14th and 15th centuries. Here, the majority of complex verbs of the common class (type II~III) are of type II, i.e. constructed with a prefix, while the inherited type III is in the minority. This distribution of the two types (in texts written by a highly bilingual group) in itself suggests that the replication of the Latin pattern played an important role in the stabilisation of this innovative construction. The dominance of type II in Vadstena texts suggests an early affiliation of these constructions with written, rather than spoken language. This is also reflected in the Modern contemporary Swedish tendency to perceive these constructions as being rather formal (or ‘bookish’) whenever there is no lexical differentiation between two corresponding verbs (cf. inställa vs. ställa in ‘to cancel’, utkomma vs. komma ut ‘to be published’, framkomma vs. komma fram ‘to turn out, to appear’). Furthermore, the quantitative distribution also hints at an incipient semantic differentiation between the two constructional variants – thus, the establishment of type II seems also to have been motivated by the need to express specific meanings. The innovative pattern is most prominent when a lexical gap is filled, i.e. when the lexical concept expressed by the complex verb is newly introduced. This is most evident in metaphorical loan-translations from Latin. As religious prose is the dominant genre during the Late Old Swedish period, religious metaphors are easiest to detect. Here we get a high proportion of complex verbs like (type II) upstanda meaning ‘to resurge, to rise from the dead’ as opposed to (type III) standa up ‘to rise, to get up’, um- vända ‘to convert (to Christianity)’ (vs. vända um ‘to turn round’) or uptända ‘to inflame (religiously)’ (vs. tända up ‘to set on fire’). These metaphorical complex verbs are used as (conventionalised) translation equivalents of Latin words (or, as it were, religious ‘technical terms’) that are isomorphic to their Swedish counterparts or at least similar in structure, such as resurgere ‘to resurge’, convertere ‘to convert’, and inflammare ‘to inflame’.8 This semantic distinction is strikingly parallel to the tendency in Modern Swedish to differ-

7 The Hamburg Corpus of Old Swedish with Syntactic Annotation (HaCOSSA) is a digital corpus of Old Swedish texts containing c. 150,000 words, compiled and annotated at the Collaborative Research Centre on Multilingualism at the University of Hamburg (Project H3: Scandinavian syntax in a multilingual setting). The corpus and the annotation scheme have been described in Höder (2010), and a more detailed documentation will be published online via the Zentrum für Sprachkorpora in Hamburg in 2011. 8 Note that the corresponding English translations of such terms are often Latinate – this illustrates another possible strategy of differentiating concrete and metaphorical concepts, introduced by bilingual speakers. The history of complex verbs in Scandinavian languages revisited 163 entiate structurally between rather concrete meanings, encoded as type III constructions, and abstract, metaphoric concepts, which are lexicalised as type II constructions. In some cases, this differentiation has been lexicalised (cf. avbryta ‘to interrupt’ vs. bryta av ‘to break, to tear off’; avböja ‘to reject, to refuse’ vs. böja av ‘to bend, to turn’), while in others it is optional (cf. stry- ka under ‘to underline, to underscore’ vs. understryka ‘to underline [literally]; to emphasise’).9 Finally it is worth mentioning that the emergence of a semantic distinction between type II and type III constructions may also be related to a universal cognitive mechanism, namely iconicity, viz. the principle that the form of linguistic expressions is partly motivated by their meaning. As for metaphorical complex verbs, compositionality could be a crucial point: in cases like standa up ‘to rise, to get up’, the meaning of the complex verb is entirely predictable from the meaning of its components (endocentric construction). This is mir- rored by the possibility of a discontinuous construction in Swedish, in which the particle is distant from the verbal stem or even the whole inflected verb. Metaphorical complex verbs like upstanda ‘to resurge’, on the other hand, have an unpredictable (if comprehensible) semantic relation to their components (exocentric construction). This kind of complexity is reflected in the in- separability of the verb and the particle. Thus, the emergence of type II verbs in Old Swedish illustrates that language contact phenomena are, as a rule, the result of different interacting factors, in this case: (a) a source construction in the model language Latin, (b) a related but different construction in the recipient language Old Swedish, and (c) a language-independent cognitive mechanism that affects any process of language change, whether contact-induced or not.

Insular Nordic The situation in the Insular Nordic languages is characterised to a high degree by contact with the colonial language Danish. On the one hand, the replication process that started in early Germanic continued after the Viking settlement in the North Atlantic in various ways. Primary replications from Latin probably played a minor role in establishing innovative complex verb types, as did secondary replications via Low German, but the most important factor in the later development of Icelandic (and Faroese) was Danish influence from the Late Middle Ages onwards: the replication chain thus goes (A) from Latin to (Low) German and then (B) to Danish and finally (C) to Insular Nordic. On the other hand, the rise of purism, particularly in Iceland, led to a ban on typical foreign elements in the language, which of course in practice meant elements that were clearly identifiable

9 For the semantic relation between type II and type III constructions in Modern Swedish, cf. Teleman, Hellberg and Andersson (1999, p. 431ff.); for the (similar) situation in Modern Norwe- gian, cf. Faarlund, Lie and Vannebo (1997, p. 83ff.). 164 Kurt Braunmüller and Steffen Höder as Danish in origin. At the same time, the same elements became much more acceptable in Faroese due to the lack of a similar purist movement, while the entire population became more and more fluently bilingual. The innovative complex verb type (PV) is attested in Modern Icelandic prefix verbs such as afturkalla ‘to revoke’, fordæma ‘to reprobate’, fyrirgefa ‘for- give’, endurbæta ‘to repair’ and undirbúa ‘to prepare’ (examples from Jörg 1989; cf. Guðrún Kvaran 2005, p. 153). However, no examples for verbal patterns with be- (or a phonetically corresponding prefix) are found. In contrast, a Low German or Danish origin did not impede the establishment and use of prefix verbs in Modern Faroese, not even in conservative written varieties. Prefix verbs of type I (P'V) with be- are fully acceptable (cf. Wittkugel 2009, p. 92ff.), and type II constructions ('PV) exist as a constructional alternative to particle verb constructions (type III, V'P; cf. Petersen 2010, p. 218f.; Höskuldur Thráinsson et al. 2004, p. 219). The difference between today’s two Insular Nordic languages can be traced back to a heterogeneous distribution in the Old West Norse varieties as early as the High and Late Middle Ages. Type I constructions are only scarcely attested in the Old Icelandic literary language as preserved in the classic prose texts. Baetke (1976), for instance, lists complex verbs beginning with for- ‘for-’ (11 entries), fyrir- (27 entries) ‘for-’ and mis- ‘mis-’ (21 entries), but none formed with salient foreign lexical material. In non-literary texts, however, such prefixes are attested, including prefixes that are clearly of Low German origin, such as be-/bí-, as has been shown by Ve- turliði Óskarsson’s (2003, p. 182ff.) investigation of the language of administrative and juridical documents before 1500. The difference between Icelandic and Faroese can thus be explained as a result of a different tradition regarding linguistic purism: in the case of Icelandic, certain verbal prefixes, especially be- and bí-, were stigmatised and became totally banned, as early as in the (Late) Middle Ages, although they had been used in non-literary text types. Other prefixes such as af- ‘from’, aftur- ‘after-’, fyrir- ‘for-’, in contrast, were accepted, especially if they were homophonous with indigenous adverbs or preposition, even if their use as prefixes was based on (or reinforced by) foreign-language models. For example, the use of fyrir- as a verbal prefix was related to the parallel use of the Latin prefix prae- in pairs of translation equivalents in text types influenced by Latin textual norms (Veturliði Óskarsson 2003, p. 187). This – seemingly paradox – situation indicates that the structural patterns of prefix verbs were not necessarily perceived as contact-induced innovations and, hence, not necessarily rejected by purists, except in those cases where the lexical material was (exclusively) Danish or Low German.10

10 The same process has also taken place during the elaboration and codification of the second written variety of Norwegian, Landsmaal, today called Nynorsk (cf. Gerdener 1986, p. 190ff.). The label for these banned derivational prefixes and suffixes is “anbeheitelse words”. These ‘forbidden’ particles are flagged as colonial Danish and thus non-Norwegian, though most of them are originally of Low German origin. The history of complex verbs in Scandinavian languages revisited 165

In the Faroe Islands, the situation is entirely different. Linguistic purism is much less important, and pragmatic factors – including bilingual communicative strategies, transfer mechanisms, and the preference of structural equivalents in the two national languages Faroese and Danish – prevail. It is therefore not surprising that borrowings from Danish into Faroese are very common. The prefix be-, for example, is almost as frequent as in Danish, which is nothing special for a totally bilingual society: Danish prefix verbs facilitate the parallel use of both languages and frequent code switching. Only a few purists try to counteract this development, inventing (or copying Icelandic) verb forms which actually do not exist in Faroese and are therefore not used by native speakers of this language.

Conclusion We have shown that the current explanation of how prefix verbs entered the Mainland Scandinavian languages does not give the full and true picture of what actually happened in Germanic language history due to various episodes of intense language contact. Rather, the transfer of prefix verbs has to be seen as part of a more complex chain of grammatical replications that took place when the creolised early Germanic dialects later underwent an intense expansion process. This process was necessary in order to match the written standards provided by the ancient model languages, Byzantine Greek (for the eastern Germanic dialects, especially for Gothic) and predominantly Latin (for all other Germanic vernaculars that came into contact with the Mediterranean civilisation). Making translations from Greek and Latin spurred the ancient Germanic elite to expand their, in part, grammatically reduced native languages. Otherwise it would not have been possible to render the full contents of the original texts into their vernaculars. As is well known for Gothic (Ulfi- las’ Bible, the Skeireins) and e.g. Old High German (Tatian) as well, these earliest translations are heavily dependent on their model texts, mostly in the syntax, but also with respect to grammatical and word-formation morphology. Various forms of transfer and grammatical replication that can be observed in these texts are, however, not restricted to translations: the increase in literacy, formal education based on script and, above all, the rapid dissemination of Christianity were the driving forces behind a systematic expansion of the grammatical systems of the early Germanic dialects. One of these innovations was the replication of prefix verbs which, were originally not part of the earliest (creolised) Germanic language. The only means of modifying verbs was to use adverbial or prepositional particles which were placed after the inflected verbs. The transfer of prefixes, which were the default way to differentiate verbal stems in Greek and Latin, into the Germanic verbal system filled a typologically empty position, the pre-verbal position. The result was a more complex verbal system, more complex even 166 Kurt Braunmüller and Steffen Höder in relation to their model languages. Thus language contact in this case did not lead to a reduction of the target language but rather to its complexification. A very similar (grammatical) process took place when Low German speakers came into intense contact with Scandinavians. Until the beginning of the Middle Ages their vernaculars had only little contact with written Latin. The main impact of a fully expanded written language on a broader scale came in the High and Late Middle Ages when the Hanseatic merchants extended their activities to the Baltic Sea area and to the Scandinavian countries. In the course of this contact, Low German prefix verbs came into the North Germanic/Scan- dinavian varieties. The effect of this contact was very similar: the Mainland Scandinavian vernaculars took over Low German prefix verbs and thus acquired approximately the same complex verbal system as to be found in any other Western Germanic language. Furthermore, contact with Latin (predominantly in the highly bilingual ecclesiastical institutions) led to the establishment of additional verb prefixation patterns in at least the written (and more formal) varieties of Scandinavian languages. Finally, due to widespread bilingualism with the colonial language Dan- ish, the innovative verbal derivation system entered the Insular Nordic languages as well and again, in principle, in the same manner. The only differences to be observed are due to sociolinguistic factors, viz. language policy (purism in Iceland) and balanced bilingualism (on the Faroe Islands), respectively. In short, the development of a complex verbal system in the Scandinavian languages is part of a much longer and more complex process, namely the expansion of a creolised proto-language and its varieties in order to match the needs of literacy and linguistic elaboration. We are not only faced with a singular replication or code-copying process but rather with a chain of grammatical replications, from Latin to West Germanic (Low German), from (Low) German to Mainland Scandinavian, and from Danish to Insular Scan- dinavian. The history of complex verbs in Scandinavian languages revisited 167

Epilogue On the plane back from the conference in Kristiansand, we found evidence that the distribution of different types of complex verbs still causes problems in daily language usage – the integration of the replicated patterns has not been completed yet (Aftenposten, 25 November 2010, Kultur, p. 6):

References Baetke, Walter, 1976: Wörterbuch zur altnordischen Prosaliteratur. 2nd edn. Darmstadt. Beuerle, Angela & Braunmüller, Kurt, 2004: Frühe germanische Zweisprachigkeit? Zu den frühesten Runeninschriften und den defixiones in der lateinischen Gebrauchs- epigraphik. In: Arbeiten zur Mehrsprachigkeit. Folge B 54. Hamburg. Pp. 1–37. Braune, Wilhelm & Ebbinghaus, Ernst A., 1981: Gotische Grammatik. Mit Lesestücken und Wörterverzeichnis. 19th edn. Tübingen. (Sammlung kurzer Grammatiken germanischer Dialekte. A: Hauptreihe 1.) Braunmüller, Kurt, 2004a: Grammatical indicators for bilingualism in the oldest runic inscriptions? In: Arbeiten zur Mehrsprachigkeit. Folge B 54. Hamburg. Pp. 39–51. Braunmüller, Kurt, 2004b. Romersk indflydelse på de ældre nordiske runeindskrifter? 168 Kurt Braunmüller and Steffen Höder

In: Lennart Elmevik (ed.), Språkhistoria och flerspråkighet. Föredragen vid ett internationellt symposium i Uppsala 17–19 januari 2003. Uppsala. (Acta Academiae Re- giae Gustavi Adolphi 87.) Pp. 63–78. Braunmüller, Kurt, 2004c: Zum Einfluss des Lateinischen auf die ältesten Runenin- schriften. In: Oskar Bandle, Jürg Glauser & Stefanie Würth (eds.), Verschränkung der Kulturen. Der Sprach- und Literaturaustausch zwischen Skandinavien und den deutschsprachigen Ländern. Zum Geburtstag von Hans-Peter Naumann. Tübingen/ Basel. (Beiträge zur Nordischen Philologie 37.) Pp. 23–50. Braunmüller, Kurt, 2007: Det ældste germansk: åbne spørgsmål og mulige svar. In: Lennart Elmevik (ed.), Nya perspektiv inom nordisk språkhistoria. Föredrag hållna i Uppsala 20–22 januari 2006. Uppsala. (Acta Academiae Regiae Gustavi Adolphi 97.) P. 31–44. Braunmüller, Kurt, 2008a: Das älteste Germanische: Offene Fragen und mögliche Ant- worten. In: Sprachwissenschaft 33. Pp. 373–403. Braunmüller, Kurt, 2008b: Observations on the origins of definiteness in ancient Ger- manic. In: Sprachwissenschaft 33. Pp. 351–371. Dahl, Östen, 2009: Increases in complexity as a result of language contact. In: Kurt Braunmüller & Juliane House (eds.), Convergence and divergence in language contact situations. Amsterdam/Philadelphia. (Hamburg studies on multilingualism 8.) Pp. 41–52. Derolez, René, 1998: On the ‘otherness‘ of the Anglo-Saxon runes and the ‘perfect fit’ of the Fuþark. In: Klaus Düwel (ed.), Runeninschriften als Quellen interdisziplinärer Forschung. Berlin/New York. (Ergänzungsbände zum Reallexikon der Germani- schen Altertumskunde 15.) Pp. 103–116. Faarlund, Jan Terje, Lie, Svein & Vannebo, Kjell Ivar, 1997: Norsk referansegramma- tikk. Oslo. Gerdener, Wilhelm, 1986: Der Purismus im Nynorsk. Historische Entwicklung und heutiger Sprachgebrauch. Münster/W. (Münsterische Beiträge zur deutschen und nordischen Philologie 1.) Guðrún Kvaran, 2005: Íslensk tunga. Vol. 2: Orð. Handbók um beygingar- og orðmyn- dunarfræði. Reykjavík. Heine, Bernd & Kuteva, Tania, 2005: Language contact and grammatical change. Cam- bridge. Hiltunen, Risto, 1983: The decline of the prefixes and the beginnings of the English phrasal verb. The evidence from some Old and Early Middle English texts. Turku. (Turun yliopiston julkaisuja. Sarja B 160.) Höder, Steffen, 2009: Converging languages, diverging varieties. Innovative relativisa- tion patterns in written Old Swedish. In: Kurt Braunmüller & Juliane House (eds.), Convergence and divergence in language contact situations. Amsterdam/Philadel- phia. (Hamburg studies on multilingualism 8.) Pp. 73–100. Höder, Steffen, 2010: Sprachausbau im Sprachkontakt. Syntaktischer Wandel im Alt- schwedischen. Heidelberg. (Germanistische Bibliothek 35.) Höder, Steffen & Zeevaert, Ludger, 2008: Verb-late word order in Old Swedish subordinate clauses. Loan, Ausbau phenomenon, or both? In: Peter Siemund & Noemi Kintana (eds.), Language contact and contact languages Amsterdam/Philadelphia. (Hamburg studies on multilingualism 7.) Pp. 163–184. Höskuldur Thráinsson et al., 2004: Faroese. An overview and reference grammar. Tórs- havn. Johanson, Lars, 2008: Remodeling grammar. Copying, conventionalization, grammati- calization. In: Peter Siemund & Noemi Kintana (eds.), Language contact and contact languages. Amsterdam/Philadelphia. (Hamburg studies on multilingualism 7.) Pp. 61–79. The history of complex verbs in Scandinavian languages revisited 169

Jörg, Christine, 1989: Isländische Konjugationstabellen [Icelandic conjugation tables]. Hamburg. Kloss, Heinz, [1952] 1978: Die Entwicklung neuer germanischer Kultursprachen seit 1800. 2nd edn. Düsseldorf. (Sprache der Gegenwart 37.) Krahe, Hans & Meid, Wolfgang, 1967: Germanische Sprachwissenschaft. Vol. 3: Wort- bildungslehre. Berlin. (Sammlung Göschen 1218b.) Kranich, Svenja, Becher, Viktor & Höder, Steffen, 2011: A tentative typology of translation-induced language change. In: Svenja Kranich et al. (eds.), Multilingual discourse production. Amsterdam/Philadelphia. (Hamburg studies on multilingualism 12.) Pp. 11–43. Krause, Wolfgang, 1971: Die Sprache der urnordischen Runeninschriften. Heidelberg. Ljunggren, Karl G., 1932: Studier över förhållandet mellan verbalpartikel och verb i fornsvenskan. Lund. Mailhammer, Robert, 2007: The Germanic strong verbs. Foundations and development of a new system. Berlin/New York. (Trends in linguistics 183.) Muysken, Pieter, 2000: Bilingual speech. A typology of code-mixing. Cambridge. Petersen, Hjalmar P., 2010: The dynamics of Faroese-Danish language contact. Heidel- berg. (Germanistische Bibliothek 37.) Teleman, Ulf, Hellberg, Staffan & Andersson, Erik, 1999: Svenska Akademiens grammatik. Vol. 3: Fraser. Stockholm. Trudgill, Peter, 1992: Introducing language and society. London. van Coetsem, Frans, 2000: A general and unified theory of the transmission process in language contact. Heidelberg. (Monographien zur Sprachwissenschaft 19.) Vennemann, Theo, gen. Nierfeld, 2003: Europa Vasconica – Europa Semitica. (Edited by Patrizia Noel Aziz Hanna.) Berlin/New York. (Trends in linguistics 138.) Veturliði Óskarsson, 2003: Middelnedertyske låneord i islandsk diplomsprog frem til år 1500. København. (Bibliotheca Arnamagnæana 43.) Wittkugel, Jana, 2009: Der färöische Sprachpurismus und seine Auswirkungen. Eine Darstellung anhand einiger dänischer Affixe im Färöischen [unpublished master thesis]. Hamburg. Wollin, Lars, 1991: The monastery of Vadstena. Investigating the great translation workshop in medieval Scandinavia. In: Roger Ellis (ed.), The medieval translator II. London. Pp. 65–88. Wollin, Lars, 1981/83: Svensk latinöversättning. Lund. (Samlingar utgivna av Svenska fornskriftsällskapet. Serie 1: Svenska skrifter 74.) 170 Kurt Braunmüller and Steffen Höder Low German influence and typological change in Swedish 171 Low German influence and typological change in Swedish: some results from a research project

Ludger Zeevaert, Háskóli Íslands

Background For approximately the last 20 years, the department of Scandinavian studies at Hamburg University has been among the central research facilities on Middle Low German–Scandinavian language contact (for a detailed overview, see Kurt Braunmüller’s article in this volume), although the focus of research has steadily shifted towards different contact interactions (Danish–Faroese, Swe- dish–Latin). A dominant feature of the research carried out in Hamburg is its focus on Middle Low German influence on the typology of the Scandinavian languages, whereas traditional research in this area concentrated primarily on the lexicon – Middle Low German loanwords, and occasionally also word formation and morphological loans. More contemporary theories of language contact, however, stress the importance of including syntactic loans, because in widely accepted typologies of contact such as Thomason and Kaufman’s (1988), syntactic loans have been evaluated as evidence for intensive language contact. This article is intended as a summary of the results of research concerning Middle Low German influence on the syntax of Swedish that was carried out as a part of the project “Scandinavian syntax in a multilingual context” between 2004 and 2008 at the Collaborative Research Centre at Hamburg University.

Typological change from Ancient Nordic to Modern Scandinavian The fundamental question for the project was to examine the role of bilingualism in the typological restructuring of the Scandinavian languages from the first written sources in the 2nd century to the modern Scandinavian languages Danish, Norwegian and Swedish (Sonderforschungsbereich 538 Mehrspra- chigkeit. Finanzierungsantrag für den Förderungszeitraum Juli 2008 – Juni 2011, p. 278). One of the aims of the research was to determine the influence of Low German on the order of constituents (subject, finite verb, object) of the 172 Ludger Zeevaert

Scandinavian languages, using Greenberg’s word-order typology (Sonder- forschungsbereich 538 Mehrsprachigkeit. Finanzierungsantrag für den För- derungszeitraum Juli 2002 – Juni 2005, pp. 329, 339, 341). Differences between Ancient Nordic and the modern Scandinavian languages exist on all linguistic levels (phonology, morphology, lexicon, syntax; Pettersson 1996, p. 70f.), but not all of these differences can be described as typological changes, and not all of them derive from the influence of Middle Low German. The influence of the Hanseatic League, which is usually associated with Low German–Scandinavian language contact, is limited to the period between 1250 and 1550. Linguistic changes that can be observed earlier in the written sources, i.e. changes from the language of the Ancient Nordic runic inscriptions to the language of the younger runes and the texts from the classical Old Swedish period written with the Latin alphabet, cannot be attributed to Middle Low German. But even for changes from the time of Hanseatic influence, Middle Low German is rather implausible as the source if similar changes can be found in Icelandic, given the fact that the Faroe Islands and Iceland were not part of the Hanseatic sphere and thus were not influenced by Low German (Kjartan Ottósson 1990, p. 14). Therefore, the examination of typological influence from Low German has to focus on changes which can be observed after approximately 1200 in the continental Scandinavian languages, but not in Ice- landic, and which correspond to structures which are found in Middle Low German.

What is typological change? The identification of Middle Low German influence on the Scandinavian languages appears to be a feasible task, but it is less easy to define what should be counted as typological change. Language typology deals with the examination of systematic linguistic variation. Usually, the aim of typological research is the description of linguistic universals and of principal differences between different language types. Thus, in order to characterise linguistic change as typological change, there has to be a change in the language system. However, the criteria used to determine whether a change is regarded as systemic or not do not seem to be very well defined (Thomason 2001, p. 1641ff.). Generally, lexical loans should not be evaluated as an indication of contact- induced typological change, nor should single grammatical differences. Unfor- tunately, holistic approaches in typology that aim at establishing a small number of language types characterised by a bundle of common linguistic traits which, on the surface, seem to be independent of each other but which exhibit a deeper relationship and thus appear to be connected have not yet provided convincing results (Comrie 2001, p. 26). Contemporary typological research seems to focus on contrasting single grammatical differences in selected languages. Low German influence and typological change in Swedish 173

Morphology One characteristic difference between Icelandic and the Scandinavian languages on the mainland is to be found in the morphological domain. Tradition- ally, languages are described as belonging to different morphological types (isolating, agglutinating or inflecting; Comrie 2001, p. 26). It is not disputed that all the different geographical and historical varieties of the Scandinavian languages and German belong to the inflecting type. The differences in the inflectional systems of e.g. Swedish and Icelandic lie only in the different complexity of the systems, and the reduction of complexity of Swedish inflectional morphology, which began long before the Middle Low German period (Pet- tersson 1996, p. 65), cannot be attributed to the influence of a language from a different morphological type. It is commonly assumed that contact with Low German played an important part in this development (Jahr 1995, p. 12); nevertheless, this has to be seen as a universal language contact phenomenon and not as typical of contact between Low German and the Scandinavian languages. There is, however, one morphological change that definitely can be attributed to Low German influence, viz. the import of a group of affixes for word formation in Danish, Swedish and Norwegian, the so called anbeheitelse affixes (Akselberg 2005, p. 1829). These affixes are not found in the earliest written sources but appear first in texts from the Hanseatic period, and they are not used in Icelandic (at least not as a productive means of word formation, but merely as elements of lexical loans from Danish). In the mainland Scandina- vian languages, these affixes were first borrowed as parts of purely lexical loans and became productive later on (Wessén 1954, p. 19). Generally, however, the use of prefixes and suffixes in word formation is definitely not atypical for the Scandinavian languages as a whole, so that one would subsequently hesitate to speak of a typological change in morphology.

Typology and syntactic change Since Greenberg’s 1963 article “Some universals of grammar with particular reference to the order of meaningful elements”, the description of different language types based on syntactic criteria has played a central role in research on language typology, probably due to the fact that a description of word order constitutes a rather straightforward way of characterising the world’s languages without a deeper knowledge of the languages involved (“Japanese is a typical SOV language, Irish is a typical VSO language”). Moreover, syntactic typology allows for a combination of different syntactic criteria, viz. the order of elements in phrases and in clauses, and establishing a logical correlation between them (see Braunmüller’s 1982 distinction between left- and right-modifying languages). In this way it is possible to organise the diversity of human languages and to reduce the verbalisation of human thoughts to a few simple 174 Ludger Zeevaert and logical principles. This approach also provides the possibility of explaining the differences that diverging word-order patterns display when teaching foreign languages (e.g. the word order in German subordinate clauses for speakers of French, or in Swedish subordinate clauses for speakers of German), and generally the fact that a change of word order is seen as a typical result of language contact and especially of substratum interference. Very often, however, languages cannot be described as being consistent when it comes to syntactic typology. Word order can vary between main and subordinate clauses, and frequently differences can be found between the order of modified and modifying elements in phrases and clauses (Vennemann 1974, p. 347). Such inconsistencies can be explained as stages in the development of one language type to another; this was the background for the afore-mentioned research project.

Results from the project on Low German and Scan- dinavia The starting point for my work was results from an earlier project on Low Ger- man–Scandinavian language contact that was carried out at Hamburg Univer- sity between 1989 and 1995 (for a detailed description, see Braunmüller in this volume). The central hypotheses for this project were the following (Braun- müller 2000, p. 14): – Contact between Middle Low German and the Scandinavian languages depended mainly on semicommunication. – The syntactic structures of the languages involved were very similar. – The Scandinavian languages borrowed syntactic structures from Low Ger- man which led to further reduction regarding the differences.

As a result of a syntactic analysis, Braunmüller (1993) established that syntactic loans from Low German could not be found in his corpus of Early Modern Danish and Swedish chapbooks, but that oral contact between Northern Ger- man merchants and the inhabitants of Scandinavian villages favoured word- order patterns which were common to all the contact languages, but which were unmarked in Low German and marked in Swedish and Danish. One exception, however, was word order in subordinate clauses. Braunmül- ler (1993, p. 254f.) gives several examples of verb-final subordinate clauses from the Swedish chapbook Grisilla which cannot be attributed to translation interference but which are obviously derived from Middle Low German structures (“Strukturen mit offensichtlich mittelniederdeutscher Herkunft”, Braun- müller 1993, p. 255). Low German influence and typological change in Swedish 175

Syntax in contact Part of the project “Scandinavian syntax in a multilingual perspective” examined whether word order in the Scandinavian languages was influenced by Low German–Swedish bilingual speakers favouring syntactic constructions common to both linguistic systems and thus bringing about a syntactic paral- lelisation in the languages. One key question was the extent to which this influence can be seen as a dominant factor in the typological change from the oldest written sources of Northern Germanic (the Ancient Nordic runic inscriptions) to the Modern Scandinavian languages.

From SOV to SVO? When studying typological syntactic change from Ancient Nordic1 to Modern Scandinavian, the question of a transition from one language type to another is of special interest. Such a change can be described using Green- berg’s typological model for syntax. Building on earlier research, Greenberg argues that word order is characterised by an orderliness that is universal and which can be described using universally valid rules (Greenberg 1990, p. 40). According to Greenberg, the dominant word-order pattern in declarative sentences characterises a language type. Elements of this basic order (Greenberg 1990, p. 42) that are relevant to the Germanic languages are the order of subject, predicate2 and object in the main clause, and the order of noun and adjective, noun and demonstrative, noun and numerals, and noun and genitive in the noun phrase. The description of this basic order seems to be unproblematic for the languages involved. According to Greenberg (1990, p. 67), Modern Swedish is a language of the SVO/AN/GN type and Modern German is of the SVO/AN/NG type.3 The word-order type of Ancient Nordic, the oldest predecessor of Mod- ern Swedish known from written sources, is less easy to determine using Greenberg’s system. Based on a quantitative analysis of all runic inscriptions from the Ancient Nordic period that can be analysed syntactically, Beuerle and Braunmüller (2004) describe Ancient Nordic as SVO/AN/GN.4 However, an analysis of the same corpus performed by Antonsen (2002) leads to the opposite result, namely the order SOV/NA/NG. This is not, as might be expected from the different theoretical approaches used by Antonsen and Beuerle/ Braunmüller, a result of differences in the definition of the notion of verb (which in any case would not explain the differences in the analysis of the order of adjective and noun and genitive attribute and noun). In both analyses, V de-

1 Cf. Braunmüller (2002, p. 649) for a discussion of this term. 2 The English/American term is verb. 3 Subject before predicate before object; attributive adjective before noun; noun before genitive attribute. 4 In the sense of Greenberg’s dominant basic word order in declarative main clauses. 176 Ludger Zeevaert notes the finite verb and not, as is common in syntactic analyses in the generative framework, the non-finite part of the predicate. The different results stem partly from variations in the composition of the corpus. The corpus for Beuerle & Braunmüller’s (2004) analysis consists of all inscriptions in the older fuþark that can be syntactically analysed, while Anton- sen rejects some of the inscriptions as not belonging to the Ancient Nordic period. Given the rather small size of the corpus, this leads to substantial differences in the percentage of word-order types. However, what is even more important for the fundamental differences between the two analyses (Beuerle and Braunmüller: 68% SVO, Antonsen: 71% SOV) is the fact that difficulties in deciphering older runic inscriptions in particular can result in diverse interpretations of an inscription. Even in cases where the transcription of the individual runes is uncontroversial, differing analyses of the parts of speech of an inscription can lead to rather different results concerning word order. For example, the word fahi in the Noleby inscription is analysed as an adjective (ʻsuitableʼ) by Antonsen (2002, p. 288), but as a finite verb (ʻI paintʼ) by Beuerle & Braunmüller (2004, p. 12). Vg 63 Fyrunga Stora Noleby: runo fahi raginakudo tojeka5

The possibility cannot be completely excluded that, in such cases, Antonsen’s interpretations are biased partly by his view on the typological classification of Ancient Nordic and Proto Germanic.6 Such difficulties in determining Ancient Nordic word order may be one reason for the fact that other researchers and even reference books refrain from carrying out their own quantitative analysis and define the basic word order of Ancient Nordic based on only one inscription, viz. the inscription on one of the golden horns of Gallehus (e.g. Penzl 1989; Lehmann 1993), which is generally interpreted as SOV (e.g. Faarlund 2001, p. 1707). Gallehus: ek 'hlewagastiz 'holtijaz 'horna 'tawido7

5 Translation according to Beuerle and Braunmüller (2004, p. 12): “[Eine] Rune male ich, eine von den Ratern (sc. den Göttern) stammende” ((A) rune, I paint, one descended from the coun- celors (i.e. the gods)). Translation according to Antonsen (2002, p. 288): “I prepare [a/the] suitable, divinely-descended ʻruneʼ [message]”. 6 “According to the theory proposed by Greenberg (1966) [cited as Greenberg 1990 in this article] and refined by Vennemann (1974, p. 345), XV languages should display noun-modifiers before the noun. The older inscriptions, however, have them uniformly after the noun, with the exception that genitive attributes can appear either before or after the head noun, depending on the nature of the latter. It is also conceivable, indeed highly probable, that adjectives could have appeared before nouns when they were nonrestrictive in function, but we have no direct evidence for this position in the inscriptions. On the basis of this evidence, I conclude that Northwest Germanic [= Ancient Nordic in this article] was still basically an XV language, with indications (particularly in the substantival groups, but also occasionally in the position of the finite verb) of an incipient shift to VX order (cf. also Lehmann 1978)” (Antonsen 2002, p. 295). 7 ‘I, Hlewagastir [son] of Holt, made the horn’. Low German influence and typological change in Swedish 177

However, it is questionable whether just this single inscription should be relied on for this purpose. The original inscription was destroyed some 200 years ago, although fairly good copies of the text exist. A major difficulty for determining the word order lies in the fact that the inscription is not written in prose but consists of a Germanic long-line with a fixed pattern of stressed syllables (') and alliteration (_) which largely determine the word order. The two stressed words in the first half-line and the first stressed word in the second half-line alliterate, which means that the only place left for the second stressed word in the second half-line, the finite verb tawido, is in final position. Thus, OV order is the only word order permitted by the metre. Summing up, it has to be stated that the oldest Germanic sources, i.e. the inscriptions in the older fuþark, cannot be used as a reliable source for determining the word order of Ancient Nordic. This means that an analysis of typological change (in the sense of Greenberg’s word-order typology) based on a written corpus has to start with the more reliable sources from c. 1100 onwards.

The basic word order of Old West Norse and Old East Norse In contrast to the situation in Ancient Nordic, the basic word order of Old Swedish is uncontroversial. In historical grammars and handbooks it is uniformly described as a V2 language, and Larsson’s (1931) study shows that this already held true for the Swedish runic inscriptions in the younger fuþark (c. 950–1100). Findings from the project “Scandinavian syntax” confirm the earlier research results. In main clauses in different Swedish texts from the 14th to the 16th century, the finite verb takes either the first or the second position; main clauses with finite verbs in a later position were not found. The percentage of V1 clauses varied between 5% and 26%, and the percentage of V2 clauses between 95% and 74%.8 This corresponds to the situation in Modern Swedish and Modern German which, according to Greenberg (1990, p. 67), are classified as SVO languages. In contrast, the word order of the noun phrase in main clauses is less clear. According to Christoffersen (2002, p. 184f.), “the order of elements in the noun phrases seems to vary” both in the inscriptions in the younger fuþark and in Old Norse. A systematic investigation of the word order in Old Swedish noun phrases has not yet been carried out as part of the project, but an analysis of the position of genitive attributes and possessive pronouns is able to confirm the assumption of variation in the order of attribute (A) and head noun (H) in this domain, although there is a rather clear preference for the order attribute–noun (table from Zeevaert 2006):

8 For a more detailed overview, see Zeevaert (2006). 178 Ludger Zeevaert

genitive possessive attribute pronoun AH HA AH HA Rök Runestone Ög 136 (c. 800) 6 5 0 1 Law code of Uppland (Upplandslagen) (1296) 86 7 35 45 Old Swedish Genesis (c. 1300–1350) 144 2 158 2 Old Swedish Legendarium (c. 1300–1350) 88 10 205 0 Autograph of St Bridget of Sweden (c. 1360) 11 0 22 0 Swedish Charters (1375) 146 0 481 0 Old Swedish Charlemagne’s saga (Karl Magnus) 13 5 124 4 (c. 1400–1450) Old Swedish Didrik’s chronicle (Didrik av Bern) 23 0 102 1 (c. 1450–1500) Gospel of St. Mark (1526) 147 1 280 0 Gospel of St. Luke (1541) 247 1 496 1

Middle Low German influence on the basic word order of Old Swedish The basic word order of Middle Low German, viz. the word order in declarative main clauses, is described as SVO/AN/NG in the research literature (e.g. Härd 2000; Nissen 1884), even though some variation can be observed in the noun phrase. Thus, the assumption of a typological change in Swedish caused by Low German influence is rather implausible, at least if the basic word order is seen as the crucial typological parameter. The position of the finite verb does not change from Old Swedish to Modern Swedish, and it conforms to the position of the finite verb in (High and Low) German. Even a change from SOV in Proto-Germanic to SVO in the younger Germanic dialects, which is widely assumed in the literature, would have been completed before the era of Middle Low German influence on the Scandinavian languages, if it took place at all (cf. the arguments against this assumption presented above). In addition, the word order in noun phrases did not change fundamentally from Old Swedish to contemporary Swedish. What can be observed is a reduction in variation, which in one case (the order of adjective and noun) goes in the direction of Low German, whereas in the other case (genitive attribute and noun) the development goes in the opposite direction.

Inconsistency in word order and typological change Inconsistencies in the basic word order of a language such as Swedish (modifying to the right in the clause but modifying to the left in the noun phrase) is Low German influence and typological change in Swedish 179 generally seen as a sign of the language changing from one type to the other (Lehmann 1978, p. 37ff.; Vennemann 1974, p. 347ff.). The background for this assumption is Greenberg’s (1990) statistical examination that led to the formu- lation of his 45 universals. Only a small number of those universals are relevant to the question of typological change in the Scandinavian languages due to contact with Low German, and some of them are inconsistent with the typological characteristics of German and the Scandinavian languages. Therefore, it is mainly only the general conclusions about the basic word order of languages drawn from Greenberg’s examination of over 30 languages that are of interest here. With respect to word order, Greenberg (1990, p. 60) distinguishes between harmonic and disharmonic language types; the criterion for doing so seems to be the order of modifying and modified structures. According to Greenberg, linguistic harmony is a characteristic of languages which have prepositions and the orders noun–genitive, noun–adjective and finite verb–subject. This implies that languages with prepositions, such as the Germanic languages, are harmonic only if they follow the word order VSO or VOS. Admittedly, Green- berg’s reasoning about the consistency of languages remains rather vague, obviously because his material does not contain a clear statistical tendency towards harmony in languages. However, Greenberg’s description was subsequently taken as a linguistic law, especially in the research on word order in ancient Germanic dialects (e.g. Lehmann, Vennemann, Braunmüller). It does not seem unreasonable to compare Greenberg’s approach to the neogrammarian tradition of research on the older stages of the Germanic languages. In his article on the regularity of word-order change, Vennemann (1974, p. 339) departs from Otto Behagel’s third law on the order of specifying and specified elements as stated in his Deutsche Syntax (“Ein drittes Gesetz fordert, daß das unterscheidende Glied dem unterschiedenen vorausgeht” [The third law demands that the differentiating element precedes the differentiated one]). Vennemann provides examples which show that Behagel’s law does not even hold for his mother tongue, German. However, Vennemann does not question the idea of an iconic relationship between logical relations and syntactic structure which forms the basis of Behagel’s approach, but rather revises the definition of ‘modifier’ and ‘modified’ (‘operator’ and ‘operand’ in Venne- mann’s terminology). The fact that only a few languages show consistency in the order of operator and operand is seen by Vennemann (1974, p. 349) as evidence for the correctness of the principle rather than for its incorrectness. The reason for this is that syntactic changes which can be triggered by e.g. language contact (1974, p. 351) can lead to a disturbance of the natural serialisation in a language. This leads to a levelling out of this linguistic inconsistency over the course of the next 3000 to 4000 years. Also Lehmann’s (1972) reconstruction of the word order of Proto-Germanic as SOV, which was accepted by Antonsen for Ancient Nordic, is based on the assumption that inconsistency is a sign of language 180 Ludger Zeevaert change. From the fact that the Modern Germanic languages are inconsistently SVO languages, it follows that they previously were consistently SOV languages.

Why should a language be type-consistent? The reason languages should be type-consistent, or in Braunmüller’s (1982, p. 30) terminology pre- or postmodifying, is not entirely clear from the research literature. Greenberg seems to aim for a natural scientific foundation of psy- cholinguistic research by formulating scientific laws9 that can be used to explain the acquisition and the processing of language. Greenberg’s underlying assumption seems to be that acquisition and use of language lead to a language following universal laws because it is easier to learn, speak and understand a language with a consistent word order. According to Hawkins (1994, cited in Comrie 2001, p. 28), type consistency (head-initial: SVO, AdNP, NG, NA) is a result of parsing preferences which favour combinations that are easier to analyse. It seems to be the case that this assumption is not entirely based on empirical facts but merely on intuitive plausibility (Braunmüller 1982, p. 42). Greenberg’s universals, for example, do not permit any conclusions about SVO or SOV languages at all. The only empirical element in the discussion (e.g. in Vennemann 1974 and Lehmann 1978) is the fact that type-consistent languages such as the SOV language Japanese, the VSO language Arabic and the SVO language English, do exist. The fact that there are not only “such marvellously consistent XV and VX languages as Japanese and Arabic” (Vennemann 1974, p. 349), but also a much larger number of inconsistent languages is, according to the research on typological change, an argument that supports the principle of consistency. Changes towards both typological consistency and typological inconsistency are explicable. An example of language change leading to a larger degree of consistency is Vennemann’s (1974) description of the development of negation in French. According to Vennemann, in the process of language acquisition the originally emphatic pas (ʽstepʼ) was interpreted as negation because in a language like French, which is developing in the direction of the SVO type, the operator of a verb is expected to follow the verb. Developments leading away from typological consistency are usually explained as examples of contact- induced change. A typical case is the development of the position of the finite verb in German subordinate clauses, which can be looked upon as a rather clear example of type inconsistency, and which according to Lehmann (1972) is a loan from Latin. This assumption does not lack a certain degree of irony, given

9 “In addition to its importance for the interdisciplinary field of psycholinguistics and psychology proper, this study of language universals is intimately connected with the establishment of scientific laws in the linguistic aspects of human behavior” (Greenberg, Osgood & Jenkins 1973, p. xxiv f.). Low German influence and typological change in Swedish 181 the fact that German subordinate clauses with the finite verb in final position are frequently taken as an argument that Proto-Germanic also had SOV order in main clauses. In other words, the literature on historical typology is not completely free of contradictions and circular reasoning. A manifest example is Lehmann’s reconstruction of Proto-Germanic as a consistent SOV language. On the one hand, his approach requires the acceptance of the principle of consistency in order to establish SOV as the word order of Proto-Germanic through internal reconstruction. On the other hand, the development from this reconstructed word order to the SVO order of modern Germanic languages is used as proof that this principle has been at work in the development of the Germanic languages. Most typologists seem to be aware of flaws in the theoretical and empirical foundations of the typological method. Braunmüller (1982, pp. 41–44) gives a quite clear description of these deficiencies; nevertheless, he decides to apply a typological approach in his research.10 According to Vennemann (1984), ideal linguistic typologies are of no theoretical value but should be exclusively used as a benchmark for a typological description of a specific language.11 A holistic typological approach which tries to establish a limited number of language types characterised by a number of common structures connected to each other by a non-obvious, deeper interrelationship appears to have failed (Strömsdörfer & Vennemann 1995; Comrie 2001, p. 26), resulting in contemporary typological research which concentrates on a contrastive description of single grammatical differences between selected languages.

Preliminary conclusion With reference to the project’s central question, i.e. the role that Middle Low German played in the typological restructuring of the Scandinavian languages, the following conclusions can be drawn. For the Swedish language, there is no basis for stating that a fundamental typological change occurred. Both Old Swedish and Modern Swedish are languages of type II (SVO) in Greenberg’s syntactic typology. The earliest written

10 Parts of Vennemann’s (1974) groundbreaking work on syntactic typology and (Germanic) word order are capable of giving the impression of an ironic dissociation: “As a matter of fact, these very inconsistencies may turn out to be the strongest argument for the universals and the principle to those who are not already convinced by Greenberg’s statistics, by the existence of such marvellously consistent XV and VX languages as Japanese and Arabic, and by the power of a principle which reduces the entire basic word order structure of a language to a single rule of overwhelming transparency and simplicity” (Vennemann 1974, p. 349). 11 “[W]hat is the purpose of an ideal typology? My answer is: a purely practical one” (Vennemann 1984, p. 600). “What about locutions, abounding in the recent historical linguistic literature, such as that certain languages are, or have been, developing toward a certain type, e.g., that English has become, during its recorded history, an ever more consistent SVO language? In my opinion such statements have a purely orientative value …” (Vennemann 1984, p. 606). 182 Ludger Zeevaert

Scandinavian sources, the Ancient Nordic inscriptions in the older fuþark, are only of limited value for an evaluation of typological development, given the facts that the text corpus is rather limited (117 inscriptions) and a syntactic analysis is very often ambiguous or impossible. Interestingly, an examination of all syntactically analysable inscriptions carried out for the project (Beuerle & Braunmüller 2004) gave the result that Ancient Nordic was already a language of Greenberg’s type II. Concerning the noun phrase in Scandinavian, a type-consistent development, i.e. a tendency to right modification that would be expected for an OV-language according to typological theories, cannot be detected. Even though the amount of variation in word order that can be observed in the oldest Swedish texts was reduced in later stages of Swedish, the result was not consistency, but type-inconsistent left modification. Evidence for influence from Middle Low German on the basic word order of Swedish could not be found. Moreover, with regard to the decrease in noun phrase word order variation from Old Swedish to Modern Swedish, contact with Low German influence is not a convincing explanation, since adjective–noun order is in accordance with Low German whereas genitive–noun order is not.

Word order in subordinate clauses Though the assumption, based on earlier research on the word order of Ancient Germanic and typological change, that the basic word order of Swedish was influenced by contact with Middle Low German could not be verified, Braun- müller (1993, p. 254f.) was able to show such an influence in Swedish subordinate clauses: ... , att du aldrig därut slitas skallt, ... that you never from there fatigue-INF-PASS shall-2SG.PRES (Braunmüller 1993, p. 255, from the Swedish chapbook Grisilla)

In the context of contact-induced language change, the word order in subordinate clauses is of special interest because a structural difference can be found between the two clause types in Modern Swedish, but not in Modern Icelandic, a language that was basically not influenced by Low German. Both in Modern and Old Icelandic the finite verb is in second position in main and subordinate clauses. In contrast, Modern Swedish exhibits a difference between clause types which, according to Vennemann (1984), can be denoted as ‘verb-early’ and ‘verb-late’. Therefore, contact between Swedish and Low German was discussed in the research literature as a source for this change. De Boor (1922, p. 171) assumes that the tendency to place finite verbs in subordinate clauses not in second position but in a position later in the clause, which can be observed in Later Old Swedish, developed under the influence of German. Larsson (1931, p. 145ff.) showed that the new structure can be found particularly in translations from Low German and in texts written by Swedish–Low German Low German influence and typological change in Swedish 183 bilinguals. However, he does not see Low German as the source of the new structure but rather as a factor that reinforced an already existing tendency.12 The same opinion is expressed by Wessén (1992).13

The new word order in subordinate clauses: a result of oral contact? In the research literature, contact between Middle Low German and Scandina- vian is usually described as oral contact between speakers of Low German and Scandinavian languages in Scandinavian towns (e.g. Törnqvist 1955; Marold 1980; Simensen 1989; Braunmüller 2007). Interestingly, an examination of word order in contemporary Swedish subordinate clauses carried out by Jör- gensen (1978) demonstrates convincingly that the typical word order adverb– finite verb in subordinate clauses first developed in the written language and has not yet been completed in the spoken language. Jörgensen’s corpus of (spoken) interviews contains about four times more conjunctional clauses with the order finite verb–adverb than with the order adverb–finite verb, i.e. the word order of the written standard language (see the table in Jörgensen 1978, p. 63). In his corpus of written language, however, the proportion of conjunctional clauses conforming to the standardised word order is 92%. According to Jör- gensen (1978, p. 189), this is due to the fact that spoken language, at least in this case, is more conservative than written language. This would imply that the word order of modern Swedish subordinate clauses is an innovation, rooted in the written language. A comparison of the position of the sentence adverb in Swedish and Icelan- dic conjunctional clauses in internet texts (Zeevaert 2009, p. 288) showed that the verb–adverb order is especially frequent in blogs, i.e. texts that are characterised by oral language use, whereas subordinate clauses with the finite verb in final position (the extreme case of verb-late) can be found mainly in formal written contexts.14 Icelandic texts with the non-standard word order adverb– verb were found exclusively in rather formal texts (parliamentary speeches and theological texts). These few exceptions can be explained as remnants of the chancellery language, with its roots in the 17th and 18th century, which was influenced by Danish.

12 “I denna lågtyskans likhet med vårt språk torde man böra se en av anledningarna till att det tyska inflytandet kunnat bli så kraftigt, som faktiskt synes vara fallet.” [This likeness of Low German and Swedish can be seen as one reason for the fact that the German influence became as strong as it in fact seems to be the case.] (Larsson 1931, p. 173.) 13 “Under y. fsv. tid har tyskt inflytande kraftigt främjat bruket av denna bisatstyp med predikatet sist, även i svenskt skriftspråk.” [During the time of Younger Old Swedish German influence has strongly strengthened the use of this type of subordinate clause with the predicate at the end of the sentence also in written Swedish.] (Wessén 1992, p. 333.) 14 Platzack (1983, p. 47): “On the whole, this word order [i.e. final position of the finite verb in subordinate clauses] is lost in YNSw, although it may be found now and then in officialese ...” 184 Ludger Zeevaert

Middle Low German as a source for SOV word order? It seems to be the case that in Middle Low German, subordinate clause word order was also affected by the written language. According to the research literature on Middle Low German syntax (Härd 2000; Nissen 1884, p. 5; Marelli 2004), all three word-order options (finite verb in the second position, after the second position, and in final position) were possible. Chirita (2003, p. 189ff.), however, was able to show that the word order in chapbooks, i.e. texts representing an authentic oral style quite close to the spoken language, differ from chancellery language by a less consistent use of the verb-final order in subordinate clauses. Rösler’s (1997) diachronic examination of the position of the finite verb in Middle Low German subordinate clauses, which was carried out on a larger corpus, was able to show an increase in the use of verb-late word order in her material. According to Rösler, this increase can be explained by a growing influence of the chancellery language on written language use. Even if it is assumed that a prestige variety of Swedish that developed from the spoken variety used between German merchants and craftsmen and native inhabitants of Swedish towns, which was characterised by borrowings from German (Diercks & Braunmüller 1993, p. 10), later influenced the written standard, it is less plausible that oral contact was the source of the new word order in Swedish subordinate clauses. Complex hypotactic constructions are rather atypical in spoken language, and verb-late word order was not yet obligatory in German during the earlier period of contact between Low German and Swedish. Thus, the development of modern Swedish word order in subordinate clauses seems to be mainly a phenomenon of the written language.

Diachronic variation Another argument against the Low German provenance of verb-late word order is the fact that verb-final subordinate clauses were already being used in the oldest Swedish sources written in the Latin alphabet, the provincial laws. The percentage of such verb-final clauses varies rather dramatically between different texts (see de Boor 1922, pp. 122ff., 171ff.; Jörgensen 1987, p. 122), and the widespread opinion that verb-final clauses are the remnant of an earlier SOV stage of Scandinavian seems to be built on circular reasoning as the result of an inappropriate mixing of different approaches to the dating of texts (paleogra- phy, history, linguistics). According to Wenning (1930, p. 31), more archaic (“ålderdomligare”) provincial laws like Äldre Västgötalagen (Older law code of Västergötland, oldest manuscript from c. 1280) and Dalalagen (Law code of Dalecarlia, oldest manuscripts from around 1350), exhibit a high percentage of verb-final clauses, which can be taken as a strong indication of this word order being especially conservative. The explanation for the fact that (according to Wenning) more modern texts like Upplandslagen (Law code of Upland, oldest manuscripts from the first Low German influence and typological change in Swedish 185 half of the 14th century) and Magnus Erikssons Landslag (oldest manuscripts from the second half of the 14th century) show even more verb-final clauses is, according to Wenning, that they are examples of a standardised official juridical language which exhibits an even more conservative character.15 Thus, verb-final word order is used as a criterion for dating the texts, but at the same time the fact that verb-final word order is found mainly in texts described as linguistically conservative is taken as a proof of the archaic nature of this structure. A comparison of Swedish law texts and the oldest law text preserved in Ice- landic, Grágás – the oldest Icelandic manuscripts are from the middle of the 13th century, but according to historical sources, the Icelanders began to write down their laws in the 12th century (Kristján Árnason 2003, p. 254) – is not able to confirm the age of verb-final subordinate clauses. According to Jörgensen (1987, p. 122), Grágás has only 0.4% verb-final subordinate clauses and would thus represent a much younger type of language than the Swedish law texts, which were written down later than the Icelandic texts. It is safe to say that the Icelandic laws are strongly based on an oral tradition (Kristján Árnason 2003, p. 254). According to Landnámabók and Íslendinga- bók, before Iceland became a literate society, the law was recited orally at the Alþingi. Thus, early Icelandic writing is the continuation of an oral tradition based on contemporary language use, but also on a standardised style for oral tales, and this already existing Icelandic oral literary standard could be used to develop a written standard language. So the linguistic development from the language of the younger runic inscriptions to the classical Icelandic sagas can be looked upon as comparably unbroken. Such an oral tradition is also assumed for the Swedish provincial laws,16 although the evidence is less clear than in the case of the Icelandic laws. Therefore, the assumption that the word order in Ice- landic subordinate clauses is a linguistic innovation, while the Early Old Swe- dish word order represents a more conservative linguistic state, is less probable.17 Another argument against this is the fact that verb-final subordinate clauses are not found in Swedish texts from the 14th century, and they appear again in the written sources much later, increasing in frequency until the 17th century. The later examples, however, are usually looked upon as not belonging to the core of the Swedish language system, but as a typical phenomenon of the written language (Delsing 1999, p. 214). The development of word order in subordinate clauses can definitely not be described as a steady change which hap-

15 Åkerlund (1944, p. 3); concerning the number of subordinate clauses with verb-late order in the Old Swedish provincial laws, see Jörgensen (1987, p. 122). 16 For Delbrück (1918, p. 3) there is no doubt that the Older Law code of Västergötland, like the other provincial laws, was transmitted orally and recited publicly at the Thing assembly. (Es kann “nicht zweifelhaft sein, daß dieses Landrecht [Äldre västgötalagen] wie alle andern seit alter Zeit mündlich festgehalten und in Thingversammlungen vorgetragen worden ist.”) 17 Magnus Lagabøters landslov, which was in use in Norway since 1274, also has, according to Christoffersen (1997, p. 44) verb-second order in main and subordinate clauses. 186 Ludger Zeevaert pened in all types of texts at the same time. The provincial laws have up to 18% verb-final subordinate clauses; religious texts from the early 14th century have 0%; Swedish diplomas from the second part of the 14th century have 13%; nar- rative texts from the 15th century have only 3%; and Swedish translations of the Bible from the 16th century have up to 27% verb-final subordinate clauses. Therefore it is reasonable to presume that the word order in Swedish subordinate clauses was not so much influenced by (oral) contact with Low German but rather by literacy in Latin. Latin as the language of the Church played a crucial role in developing a tradition of writing in the vernacular in Sweden after the adoption of Christianity. The high percentage of verb-final subordinate clauses in Upplandslagen (the provincial law of Uppland) (32%, against only 5% in the oldest Swedish text, the provincial law of Västergötland from c. 1225; de Boor 1922, p. 162ff.) can be explained very easily in this way. According to the preface (lat. confir- matio), Upplandslagen is by no means a precise written record of an oral tradition, but rather a text that was revised by a committee of lawyers and theolog- ians appointed by the Swedish king, including the lagman Birger Persson and the Provost of Uppsala cathedral Andreas And (Lönnroth & Delblanc 1987, p. 52ff.). On account of their international education, these men were skilled in Latin. It is reasonable to assume that the specific word order for subordinate clauses in Early Old Swedish texts is not a remnant of the Proto-Germanic past, but rather a typical feature of Swedish chancellery language that was formed through literacy in Latin and Latin text traditions; it is evident especially in Swedish official letters and somewhat in juridical texts. This tradition was continued in the language of Bible translations in the 16th century, which were heavily influenced by Martin Luther’s High German Bible. Luther’s translation in turn was influenced by German chancellery style, being itself strongly based on Latin models.18

Conclusion One finding of the “Scandinavian Syntax” project is that the influence of Low German on the typological development of Mainland Scandinavian word order is rather marginal in comparison with the fundamental morphological and lexical changes that were induced by this language contact.19 Low German definitely did not influence the syntactic typology of Swedish, and this research shows that the changes which can be observed in the word order of subordinate

18 See Zeevaert (2002, with further bibliographical information) and Besch (2000, especially p. 1722) concerning the origin of Martin Luther’s German in the chancellery language of Upper Saxony. 19 Cf. however Stefan Mähl (in this volume) on Low German influence on Swedish syntax. Low German influence and typological change in Swedish 187 clauses were not caused by oral contact between speakers of Low German and Swedish. However, the emergence of written standard languages in Scandina- via at the beginning of the 12th and 13th centuries has not been explored exhaus- tively, and work remains to be done in order to reach a better understanding of the processes that caused the differences in the word order of Swedish and Ice- landic subordinate clauses.

References Åkerlund, Walter, 1944: Om det finita verbets plats i den fornsvenska bisatsen. In: Arkiv för nordisk filologi 57. Pp. 1–67. Akselberg, Gunnstein, 2005: Trends in the linguistic development since 1945 III: Nor- wegian. In: Oskar Bandle et al. (eds.), The Nordic languages. An international handbook of the history of the North Germanic languages. Volume 2. Berlin, New York. (Handbücher zur Sprach- und Kommunikationswissenschaft 22.2.) Pp. 1823–1831. Antonsen, Elmer H., 2002: Runes and Germanic linguistics. Berlin, New York. (Trends in linguistics. Studies and monographs 140.) Besch, Werner, 2000: Die Rolle Luthers für die deutsche Sprachgeschichte. In: Werner Besch et al. (eds.), Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung. 2. Teilband (2nd ed.). Berlin, New York. Pp. 1713– 1745. (Handbücher zur Sprach- und Kommunikationswissenschaft 2.2.) Beuerle, Angela & Braunmüller, Kurt, 2004: Frühe germanische Zweisprachigkeit? Zu den frühesten Runeninschriften und den defixiones in der lateinischen Gebrauchs- epigraphik. In: Arbeiten zur Mehrsprachigkeit. Folge B 54. Pp. 1–37. Braunmüller, Kurt, 1982: Syntaxtypologische Studien zum Germanischen. Tübingen. (Tübinger Beiträge zur Linguistik, 197.) Braunmüller, Kurt, 1993: Syntaktische Divergenzen und Transferenzen. Ein strukturel- ler Vergleich anhand mittelniederdeutscher und altskandinavischer Volksbücher. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 231–270. Braunmüller, Kurt, 2000: Historisk syntaks og flersprogethed. In: Nordlyd 28. Pp. 1–23. Braunmüller, Kurt, 2002: The Ancient-Nordic linguistic system from a typological point of view: Phonology, graphemics, morphology, syntax and word order. In: Os- kar Bandle et al. (eds.), The Nordic languages. An international handbook of the history of the North Germanic languages. Berlin, New York. (Handbücher zur Sprach- und Kommunikationswissenschaft 22.1.) Pp. 649–656. Braunmüller, Kurt, 2007: Receptive multilingualism in Northern Europe in the Middle Ages. A description of a scenario. In: Jan D. ten Thije & Ludger Zeevaert (eds.), Re- ceptive multilingualism. Linguistic analyses, language policies and didactic concepts. Amsterdam, Philadelphia. (Hamburg studies on multilingualism 6.) Pp. 25– 47. Chirita, Diana, 2003: Did Latin influence German word order? Aspects of German- Latin bilingualism in the Late Middle Ages. In: Kurt Braunmüller & Gisella Ferra- resi (eds.), Aspects of multilingualism in European language history. Amsterdam, Philadelphia. (Hamburg studies on multilingualism 2.) Pp. 173–200. Christoffersen, Marit, 1997: Saken er BIFF! Negasjonens plass i leddsetninga i eldre og nyere norsk. In: Jarle Bondevik (ed.), Språket er målet. Festskrift til Egil Pettersen på 75-årsdagen 4. mars 1997. Bergen. Pp. 44–52. Christoffersen, Marit, 2002: Nordic language history and research on word. In: Oskar 188 Ludger Zeevaert

Bandle et al. (eds.), The Nordic languages. An international handbook of the history of the North Germanic languages. Berlin, New York. (Handbücher zur Sprach- und Kommunikationswissenschaft 22.1.) Pp. 182–191. Comrie, Bernard, 2001: Different views of language typology. In: Martin Haspelmath et al. (eds.), Language typology and language universals. Volume 1. Berlin, New York. (Handbücher zur Sprach- und Kommunikationswissenschaft 20.1.) Pp. 25–39. de Boor, Helmut, 1922: Studien zur altschwedischen Syntax in den ältesten Gesetzes- texten und Urkunden. Breslau. (Germanistische Abhandlungen 55.) Delbrück, B[erthold], 1918: Germanische Syntax IV. Die Wortstellung in dem älteren westgötischen Landrecht. Leipzig. (Abhandlungen der philologisch-historischen Klasse der Königl. Sächsischen Gesellschaft der Wissenschaften 36, 1.) Delsing, Lars-Olof, 1999: Från OV-ordföljd till VO-ordföljd. En språkförändring med förhinder. In: Arkiv för nordisk filologi 114. Pp. 151–232. Diercks, Willy & Braunmüller, Kurt, 1993. Entwicklung des niederdeutsch-skandinavischen Sprachkontakts. Untersuchungen zur Transferenz anhand von volkssprachli- chen Texten des 15., 16. und 17. Jahrhunderts – Eine Projektübersicht. In: Kurt Braunmüller & Willy Diercks (eds.), Niederdeutsch und die skandinavischen Sprachen I. Heidelberg. (Sprachgeschichte 3.) Pp. 9–40. Faarlund, Jan Terje, 2001: From Ancient Germanic to modern Germanic languages. In: Martin Haspelmath et al. (eds.), Language typology and language universals. An international handbook. Volume 2. Berlin, New York. (Handbücher zur Sprach- und Kommunikationswissenschaft 20.2.) Pp. 1706–1719. Greenberg, Joseph H., 1990: Some universals of grammar with particular reference to the order of meaningful elements. In: Joseph H. Greenberg, On Language. Selected Writings (edited by Keith Denning and Suzanne Kemmer; before in: Joseph H. Greenberg (ed.), 1963: Universals of language. Cambridge, Mass. Pp. 58–90). Stan- ford. Greenberg, Joseph H., Osgood, Charles E. & Jenkins, James J., 1973: Memorandum concerning language universals. In: Joseph H. Greenberg (ed.), Universals of language. 2nd ed. 1st ed. 1963. Cambridge (Massachusetts), London. Pp. xv–xxvii. Härd, John Evert, 2000: Syntax des Mittelniederdeutschen. In: Werner Besch et al. (eds.), Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung. 2. Teilband. 2. ed. Berlin, New York. (Handbücher zur Sprach- und Kommunikationswissenschaft 2.2.) Pp. 1456–1463. Jahr, Ernst Håkon, 1995: Nedertysk og nordisk: språksamfunn og språkkontakt i Hansa- tida. In: Ernst Håkon Jahr (ed.), Nordisk og nedertysk. Språkkontakt og språkut- vikling i Norden i seinmellomalderen. Oslo. Pp. 9–28. Jörgensen, Nils, 1978: Underordnade satser och fraser i talad svenska. Funktion och byggnad. [Lund.] (Lundastudier i nordisk språkvetenskap. Serie C 10.) Jörgensen, Nils, 1987: Studier över syntax och textstruktur i nordiska medeltidslagar. Lund. (Samlingar utgivna av Svenska fornskrift-sällskapet 76.) Kjartan G. Ottósson, 1990: Íslensk málhreinsun. Sögulegt yfirlit. Reykjavík. (Rit Íslenskrar málnefndar 6.) Kristján Árnason, 2003: Icelandic. In: Ana Deumert & Wim Vandenbussche (eds.), Germanic Standardizations. Past to Present. Amsterdam, Philadelphia. (Impact: Stu- dies in language and society 18.) Pp. 245–279. Larsson, Carl, 1931: Ordföljdsstudier över det finita verbet i de nordiska fornspråken. I. Diss. phil. Uppsala. (Uppsala universitets årsskrift 1931. Filosofi, språkvetenskap och historiska vetenskaper 1.) Lehmann, Winfred P., 1972: Proto-Germanic syntax. In: Frans van Coetsem (ed.), To- ward a grammar of Proto-Germanic. Tübingen. Pp. 239–285. Lehmann, Winfred P., 1978: The great underlying ground-plans. In: Winfred P. Low German influence and typological change in Swedish 189

Lehmann (ed.), Syntactic typology. Studies in the phenomenology of language. Sus- sex. Pp. 3–55. Lehmann, Winfred P., 1993: Theoretical bases of Indo-European linguistics. London, New York. Lönnroth, Lars & Delblanc, Sven, 1987: Den svenska litteraturen. Från forntid till fri- hetstid. 800–1718. [Stockholm]. Marelli, Paolo, 2004: Sintassi e semantica dei casi nella Sächsische Weltchronik. Pisa. (Medievo tedesco. Studi e testi 10.) Marold, Edith, 1980: ‘Mischsprache’ oder Kontinuum? Die skandinavisch-deutschen Sprachbeziehungen im Mittelalter im Licht der neueren Forschungen zum Sprach- kontakt. In: Heinz Rupp & Hans-Gert Roloff (eds.), Akten des VI. Internationalen Germanisten-Kongresses Basel 1980. Bern, Frankfurt am Main, Las Vegas. (Jahr- buch für Internationale Germanistik. Reihe Kongreßbericht 8.) Pp. 142–148. Nissen, C[arl] A[dolf], 1884: Forsøg til en middelnedertysk syntax. Diss. phil. Copen- hagen. Penzl, Herbert, 1989: Die Gallehusinschrift: Trümmer der nordisch-westgermanischen Ursprache. In: Heinrich Beck (ed.), Germanische Rest- und Trümmersprachen. Ber- lin, New York. (Reallexikon der Germanischen Altertumskunde – Ergänzungsbände 3.) Pp. 87–96. Pettersson, Gertrud, 1996: Svenska språket under sjuhundra år. En historia om svenskan och dess utforskande. Lund. Platzack, Christer, 1983: Three syntactic changes in the grammar of written Swedish around 1700. In: Struktur och variation. Festskrift till Bengt Loman 7.8.1983. Åbo. Pp. 43–63. Rösler, Irmtraud, 1997: Satz – Text – Sprachhandeln. Syntaktische Normen der mittelniederdeutschen Sprache und ihre soziofunktionalen Determinanten. Heidelberg. (Sprachgeschichte 5.) Simensen, Erik, 1989: Zum Kontakt zwischen Niederdeutsch und Nordisch im Mittel- alter unter besonderer Berücksichtigung der Frage nach einer niederdeutsch-nordischen Mischsprache. In: Karl Hyldgaard-Jensen, Vibeke Winge & Birgit Christen- sen (eds.), Niederdeutsch in Skandinavien II. Akten des 2. nordischen Symposions ‚Niederdeutsch in Skandinavien’ in Kopenhagen, 18.–20. Mai 1987. Berlin. (Bei- hefte zur Zeitschrift für Deutsche Philologie 5.) Pp. 65–77. Sonderforschungsbereich 538 Mehrsprachigkeit. Finanzierungsantrag für den För- derungszeitraum Juli 2002 – Juni 2005. Hamburg. Sonderforschungsbereich 538 Mehrsprachigkeit. Finanzierungsantrag für den För- derungszeitraum Juli 2008 – Juni 2011. Hamburg. Strömsdörfer, Christian & Vennemann, Theo, 1995: Ziele der syntaktischen Typologie. In: Joachim Jacobs et al. (eds.), Syntax. An international handbook of contemporary research. Volume 2. Berlin. (Handbücher zur Sprach- und Kommunikationswissen- schaft 9.2.) Pp. 1031–1043. Thomason, Sarah Grey, 2001: Contact-induced typological change. In: Martin Haspelmath et al. (eds.), Language typology and language universals. An international handbook. Volume 1. Berlin, New York. (Handbücher zur Sprach- und Kom- munikationswissenschaft 20.1.) Pp. 1640–1648. Thomason, Sarah Grey & Kaufman, Terrence, 1988: Language contact, creolization, and genetic linguistics. Berkeley, Los Angeles, London. Törnqvist, Nils, 1955: Till frågan om den tysk-svenska språkblandningen i Sverige under medeltiden. In: Nysvenska studier 34. Pp. 98–123. Vennemann, Theo, 1974: Topics, subjects, and word order: From SXV to SVX via TVX. In: J[ohn] M. Anderson & C[harles] Jones (eds.), Historical linguistics I. Syn- tax, morphology, internal and comparative reconstruction. Proceedings of the First 190 Ludger Zeevaert

International Conference on Historical Linguistics, Edinburgh 2nd–7th September 1973. Amsterdam, Oxford. (North-Holland linguistic series 12a.) Pp. 339–376. Vennemann, Theo, 1984: Verb-second, verb late, and the brace construction. Com- ments on some papers. In: Jacek Fisiak (ed.), Historical syntax. Berlin, New York, Amsterdam. (Trends in linguistics. Studies and monographs 23.) Pp. 627–636. Wenning, Alf E., 1930: Studier över ordföljden i fornsvenskan. Predikatets bestämning- ar i äldre och yngre fornsvenskan. Diss. phil. Lund. Wessén, Elias, 1954: Om det tyska inflytandet på svenskt språk under medeltiden. [Stockholm]. (Skrifter utgivna av Nämnden för svensk språkvård 12.) Wessén, Elias, 1992: Svensk språkhistoria. III. Grundlinjer till en historisk syntax. (1. ed. 1956.) (Nytryck i Nordiska språk och svenska 6.) Zeevaert, Ludger, 2002: Multilingualismus und Sprachwandel. Zum möglichen Einfluß deutsch-schwedischer Mehrsprachigkeit auf die Gustav-Vasa-Bibel. In: John Ole Askedal & Hans-Peter Naumann (eds.), Hochdeutsch in Skandinavien. II. Interna- tionales Symposium, Oslo, 19.–20. Mai 2000. Frankfurt am Main. Pp. 101–125. Zeevaert, Ludger, 2005: Zur typologischen Entwicklung der skandinavischen Sprachen vom Anfang bis zur Neuzeit. Talk given at the Research Center on Multilingualism, Hamburg University, Zeevaert, Ludger, 2006: Variation und kontaktinduzierter Wandel im Altschwedischen. Hamburg. (Arbeiten zur Mehrsprachigkeit. Folge B 74.) Zeevaert, Ludger, 2009: Deutscher Einfluss und syntaktischer Wandel im Schwe- dischen. In: Lars Wollin, Dagmar Neuendorff & Michael Szurawitzki (eds.), Deutsch im Norden. Akten der nordisch-germanistischen Tagung zu Åbo/Turku, Finnland, 18.–19. Mai 2007. Frankfurt. Pp. 279–306. Low German and Nynorsk – a strained relationship? 191 Low German and Nynorsk – a strained relationship? A glimpse into Norway’s most recent language history

Erik Simensen, University of Oslo

Introduction A good deal is known about contact between Low German and the Nordic languages during the Middle Ages. Much of this knowledge is documented through research published in the conference series called “Niederdeutsch in Skandinavien” (1985–2001). This linguistic contact was extensive, deep and intense, and has left many traces in most of the Nordic languages (Icelandic being an exception in this respect), in the lexicon as well as in word formation. An ensuing question may be asked: what happened next? What was the fate of the Low German loans after the Middle Ages? In 1989 I examined a limited section of Norwegian vocabulary in a study of loanwords from German containing the prefix be-. These belong to a group of Low German, and partly High German, loanwords – the “anbeheitelse words”, as they are generally labelled by Norwegians, referring to the most widely used prefixes (an-, be-) and suffixes (-heit, -else) imported from German. I have tried to find out more about the status of these words in the vocabulary of Nynorsk (for short: NN, lit. ‘New Norwegian’, one of the two official written standards of Norwegian) in recent years, more precisely the period from 1993 to 2006 inclusive. In this paper I have not considered the situation in Bokmål (for short: BM, lit. ‘book language’, the other official written standard of Nor- wegian). My aim has been to see to what extent, and how, these words have been included in – or excluded from – standard NN. The handling of these words has been, and to some extent still is, the subject of heated discussion among Norwegians. This subject has also interested linguists outside Norway. Kurt Braunmüller has dealt with it in his survey of the Nordic languages (Braunmüller 1991, pp. 166–169), and it has been treated with profound insight and thorough discussion by Wilhelm Gerdener in his 1986 doctoral dissertation. My own impression is that these loanwords have become gradually more acceptable in standard NN, particularly in recent years. The question is whether this impression has an empirical basis. In order to discover possible phases and lines of development in greater detail, I have studied different editions of two widely used works embodying the official standards of NN: 192 Erik Simensen

Nynorskordboka (Dictionary of New Norwegian, henceforth NB) 2nd edition (1993) 3rd edition (2001) 4th edition (2006, electronic version 2010). Alf Hellevik’s Nynorsk ordliste (New Norwegian Word List, henceforth AH) 7th edition (1993) 9th edition (2000), with Kåre Skadberg (KS) and Aud Søyland (AS) as co-editors 10th edition (2005), edited by KS and AS alone.

AH has been revised and reissued many times since its first appearance 1938. It has been – and still is – the standard NN speller for Norwegian pupils and students. It evokes strong emotions, being loved or hated with great intensity, but negative feelings tend to prevail, as a headline in Norway’s biggest news- paper, Aftenposten, revealed on 25 October 2010: “Spynorsk mordliste” (lit. ‘Spew Norwegian murder list’). When the different editions of these works are compared, it becomes appar- ent that significant changes have taken place. The course of changes are most easily traced in the editions of AH (with KS and AS), and especially by comparing the changes in the Supplement (containing NN equivalents to BM), which was first introduced in the 3rd edition (1962) and has been an inseparable component in AH ever since. In the following surveys, compounds have not been counted.

Nynorskordboka affix 2nd ed. (1993) 3rd ed. (2001) 4th ed. (2006) an- 30 41 41 be- 99 130 130 -else 81919 -heit 24 103 103

While there is a small increase between the 1993 edition and the later two editions for the an- and be- prefixes, the -else entries are more numerous, and the biggest increase was found for the -heit suffix.

Alf Hellevik’s Nynorsk ordliste The findings from comparing the AH editions are more interesting. In the 1962 edition, there is a comment in the preface on the use of BM forms (and ipso fac- to the anbeheitelse words): Neither can it be said that all these BM words are unacceptable in NN. (Sometimes a tilde is added in order to indicate that the BM word may be allowed, at least in cer- Low German and Nynorsk – a strained relationship? 193

tain meanings. … It seems appropriate to stress that guidance [italics in original] is given here, not instructions and prohibitions (p. 10) [my translation]. A corresponding comment appears in the 1993 edition (p. 29), where the entire last sentence is italicized. In the same edition, notes on each of the an-, be-, -else and -heit affixes have been included as footnotes in the main section of the text. In the 1993 edition we read that “More sparsely used in NN are words like anbefale …, anfall …, angi or angje …” (p. 39). In the 2000 edition this had been changed to: “Formerly not much used in NN, but many words beginning in an- are well established in Norwegian dialects and are now common in NN, e.g. words like anbefale [etc. as earlier]” (p. 32). In 1993, AH stated that be- was: Common in spoken language and in BM. Even in written NN several be- words are used (see the word list), but we use many of them preferably in certain specialized meanings, see, e.g. belysning, bestemme. Other be- words are sparsely or not at all used in NN. We write, for instance, nytte instead of benytte, tyde instead of bety, sty- re for bestyre, svare på for besvare; see the Supplement (p. 49). In the 2000 edition (p. 47), the second sentence has been changed to: “Even in written NN several be- words are used (see the word list), but we use some [italicized by the editors] of them preferably in certain specialized meanings, see, e.g. belysning.” This is repeated in the 2005 edition (p. 50). In the 1993, 2000 and 2005 editions it was said that -else is “used in a few words, mainly referring to concrete objects: …Especially -ing has replaced -else, e.g. ending … .” Where -heit is mentioned the warning is clearer: Words ending in -heit (or -het) are common in the dialects, but have been sparsely used in ordinary NN prose. Some derivatives with -heit are now entered in the word list … but students should be permitted to use even other -heit derivatives which are common in spoken language. But it is important that they also know and are able to use the other derivational suffixes that are more frequently used in NN rather than -heit, especially -dom, -leik and -skap … . We often recommend using the adjective and not the derived noun (1993, p. 116; 2000, p. 141). In the 2005 edition “Some” is replaced by “Several” (p. 143). As we can see, the door has gradually been opened for more words in these categories, which means that they have become more acceptable in standard NN. In the following detailed survey of AH, I have distinguished between three types of entry: those in the main text of the word list, those in the Sup- plement, and those entered in both. Some words are entered only in the Sup- plement but are marked as acceptable in NN (after some adjustments having been made). 194 Erik Simensen

Words with an- 1993 (AH) MAIN TEXT, 27: anbefale, anbefaling, anbod, andakt, andektig, anfall, anfekting, angi, angivar, angrep, angripar, angripe, anheng, anlegg, anløpe, anmarsj, annamme, anordning, anretning, ansikt, anslag, anstalt, anstand, anstendig, antrekk, anvise, anvisar. SUPPLEMENT, 68: anbringe, anbud, anbyder, andel, andragende, anerkjenne, anerkjennelse, anføre, anfører, anførsel, angi, angivelig, angjeldende, angå, an- gående, anhang, anholde, anklage, ankomme, ankomst, anledning, anlegge, an- liggende, anløp, anløpe, anmelde, anmeldelse, anmelder, anmerke, anmerkning, anmode, anmodning, annamme, anrette, anrik(n)ing, anrop, anrope, ansatt, anse, anseelse, anselig, ansette, ansettelse, anskaffe, anskaffelse, anskuelig, anskuelse, anslag, anslå, anspenne, anstendighet, anstrenge, anstrengelse, anstøt, ansøke, ansøkning, anta, antakelig, antakelse, antall, antenne, antrukken, antyde, antydning, anvende, anvendelig, anvise, anvisning. MAIN TEXT AND SUPPLEMENT, 6: anbod (a NN form which is only orthographically adjusted from the BM form is counted together with the latter as one lemma), angi, anløpe (semantic difference BM/NN), annamme (different shades of meaning BM/NN), anslag (semantic difference BM/NN), anvise (semantic difference BM/NN).

2000 (AH, KS & AS) MAIN TEXT, 46: the same as in 1993, plus: anbydar, anerkjenne, angå, anledning, anløp, anretning, ansats, anslå, anstendigheit, anstrengjande, anstrengje, an- støyt, anta, antakeleg, antyde, antydning, anvend*, anvende, anvendeleg. Many of these entries are still in the SUPPLEMENT. The word marked with * was not included in the 1993 SUPPLEMENT. SUPPLEMENT, 66: the same as in 1993, plus: anvendt; minus: ansøke, ansøkning, antegne. MAIN TEXT AND SUPPLEMENT, 23: the same as in 1993, plus: anbydar, anerkjenne, angå, anledning, anløp, anslå, anstendigheit, anstrengjande, anstrengje, anstøyt, anta, antakeleg, antyde, antydning, anvend, anvende, anvendeleg.

2005 (KS & AS) MAIN TEXT, 47: the same as in 2000, plus: antaste. SUPPLEMENT, 66: the same as in 2000. MAIN TEXT AND SUPPLEMENT, 24: the same as in 2000, plus: anløpe.

Words with be- 1993 (AH) MAIN TEXT, 101: bebuar, bedageleg, bedekkje, bedekning, bedra, bedrag, bedra- gar, bedrageri, bedragersk, bedrift, bedømme, bedømming, bedøve, bedøving, befal, befale, befaling, befolkning, befrukte, befrukt(n)ing, begeistra, begeistre, begeistring, begripe, begripeleg, begynnar, begynne, begynning, behag, be- hageleg, behald, behaldar, behaldning, behandlar, behandle, behandling, be- herska, beherske, behov, belaste, belast(n)ing, belegg, beleggje, belysning, beløp, bemanne, bemanning, beordre, beredskap, berekne, berekning, berykta, Low German and Nynorsk – a strained relationship? 195

berømt, berøring, besetning, besett, besifre, besifring, besjeling, beskjed, beskrive, beskriving, beskøyt, beslag, beslagleggje, beslå, bestand, bestemd, bestemme, bestemming, bestikk (n.), bestikk (m.), bestille, bestråle, bestråling, bestå, besynderleg, besøk, besøkje, betale, betaling, betene, betening, betenkt, betennelse, betent, betjening, betjent, beundrar, beundrarinne, beundre, beund- ring, bevare, bevaring, beverte, bevertning, bevilling, bevis, bevise, bevis(e)leg, bevisst. SUPPLEMENT, 245: bearbeide, bebodd, beboelig, beboer, bebreide, bebreidelse, bebude, bebygge, bebyggelse, bebyrde, bedekke, bedekning, bedervet, bedrift, bedrøve, bedrøvelig, bedrøvelse, bedugget, bedyre, bedømmelse, bedøvelse, bedåre, beedige, befare, befatning, beferdet, befeste, befestning, befinne, befippelse, beflitte, befolke, befolkning, befordre, befordring, befrakte, befraktning, befrakter, befri, befrielse, begavelse, begavet, begi, begivenhet, begjær, be- gjære, begjæring, begjærlig, begrave, begravelse, begredelig, begrense, be- grensning, begrep, begripelse, begrunne, begrunnelse, begunstige, begunstigelse, begynnelse, begå, behendig, behjelpelig, behold, beholde, beholdning, be- høve, bekjempe, bekjenne, bekjennelse, bekjent, bekjentgjøre, bekjentskap, beklage, beklagelig, bekle, bekledning, bekomme, bekostning, bekranse, bekrefte, bekreftelse, bekvem, bekvemme, bekymre, bekymring, belage, beleilig, beleire, belesse, belest, beleven, beliggende, beliggenhet, belivet, belyse, belysning, belære, belønne, belønning, beløp, beløpe, bemektige, bemerke, bemerk- ning, bemyndige, bemyndigelse, benekte, benevne, benevning, benytte, benåde, benådning, beordre, beplante, beplantning, beramme, berammmelse, berede, beredskap, beredt, beredvillig, beregne, beregning, bereist, beretning, berette, berettige, berettigelse, berike, beriktige, bero, berolige, beruse, berømme, berømmelig, berømmelse, berømt, berøre, berøve, beråd, besatt, bese, besegle, beseire, besetning, besette, besettelse, besiktige, besiktigelse, besindig, besindighet, besinne, besitte, besittelse, beskadige, beskaffen, beskatning, beskatte, beskikke, beskjeden, beskjeftige, beskjeftigelse, beskjære, beskrive, beskrivelse, beskylde, beskyldning, beskytte, beskyttelse, beslaglegge, beslektet, beslutning, beslutte, besluttsom, besnære, besparelse, bestalling, bestanddel, bestandig, bestemmelse, bestige, bestikke, bestikkelse, bestille, bestilling, bestjele, be- strebe, bestride, bestyre, bestyrer, bestyrerinne, bestøvning, besvare, besvarelse, besvime, besvimelse, besvær, besværlig, besynderlig, besørge, beta, beta- gende, betakke, betatt, betegne, betegnelse, betenke, betenkelig, betenkelighet, betenkning, betids, betinge, betingelse, betjene, betone, betrakte, betraktning, betrekk, betro, betrodd, betroelse, betryggelse, betuttet, bety, betyde, betydelig, betydning, bevege, bevegelig, bevegelse, bevendt, beverte, bevertning, bevilge, bevilling, bevirke, bevisst, bevissthet, bevitne, bevitnelse, bevokte, bevoktning, bevæpne, beære. MAIN TEXT AND SUPPLEMENT, 20: bebuar, bedekkje, bedekning, bedrift, befolkning, behald, behaldning, belysning, beløp, beordre, beredskap, berømt, besett, beskrive, beslagleggje, bestille, besynderleg, beverte, bevertning, bevilling. SUPPLEMENT ONLY, BUT MARKED AS ACCEPTABLE IN NN, 2: befeste (not included in the 2000 ed.), bestilling.

2000 (AH, KS & AS) MAIN TEXT, 132: the same as in 1993, plus: bebreide, bebreiding, bederva, bedrøva, bedrøveleg, begjær, begynning, behøve, beklage, beklageleg, bekrefte, bekymre, bekymring, belage, belyse, benåde, beordring, beskjeden, beskjedenheit, bestikke, bestikking, bestilling, bestøve, betakke, betone, betutta, bety, betydeleg, betydning, bevege, bevegelse (i.e. mainly the forms from the 1993 Sup- 196 Erik Simensen

plement, but in NN orthography). Still not included: bedyre, berøre, bestand (‘strip of woodland’). SUPPLEMENT, 239: the same as in 1993, minus: befrakte, befrakter, befraktning, beløp, beråd, bestalling, bestyrerinne; plus: belemre. MAIN TEXT AND SUPPLEMENT, 52: bebreide, bebuar, bedekkje, bedekning, bederva, bedrift, bedrøva, bedrøveleg, befolkning, begjær, behald, behalde, behaldning, behøve, beklage, beklageleg, bekrefte, bekymre, bekymring, belage, belyse, belysning, benåde, beordre, beredskap, berekne, berekning, berømt, besetning, besett, beskjeden, beskrive, beslagleggje, bestikke, bestikking, bestille, bestilling, besynderleg, betakke, betene, betone, betutta, bety, betydeleg, betydning, bevege, bevegelse, beverte, bevertning, bevilling, bevisst, bevisstheit. SUPPLEMENT ONLY, BUT MARKED AS ACCEPTABLE IN NN (after the adjust- ment -lig > -leg), 10: bearbeide, begavelse, behjelpelig, beskytte, beskyttelse, bestandig, betrakte, betraktelig, betraktning, bevendt.

2005 (KS & AS) MAIN TEXT, 139: the same as in 2000, plus: begavelse, behjelpeleg, beskytte, beskyttelse, betrakte, betrakteleg, betraktning. SUPPLEMENT, 235: the same as in 2000, minus: beedige, belesse, belivet, betids. MAIN TEXT AND SUPPLEMENT, 53: the same as in 2000, plus: bedyre. IN THE SUPPLEMENT ONLY, BUT MARKED AS ACCEPTABLE IN NN, 1: bestandig.

Words with -else 1993 (AH) MAIN TEXT, 8: bakkels(e), betennelse, fordøyelse, hakkels(e), røykjelse, spøkjelse, stiftelse, stivelse. SUPPLEMENT, 185: aktelse, anelse, anerkjennelse, anmeldelse, anseelse, ansettelse, anskaffelse, anskuelse, anstrengelse, antakelse, atspredelse, avbrytelse, avgjørelse, avvikelse, bebreidelse, bebyggelse, bedrøvelse, bedømmelse, bedø- velse, befippelse, begavelse, begravelse, begripelse, begrunnelse, begunstigelse, begynnelse, bekjennelse, bekreftelse, bemyndigelse, berettigelse, berømmelse, besettelse, besiktigelse, besittelse, beskjeftigelse, beskrivelse, beskyttelse, besparelse, bestemmelse, bestikkelse, besvarelse, besvimelse, betegnelse, betingelse, betroelse, betryggelse, bevegelse, bevitnelse, deltakelse, drøftelse, en- delse, ervervelse, feiltakelse, forargelse, forbannelse, forbauselse, forbindelse, forbitrelse, forbrytelse, fordøyelse, foreteelse, forjettelse, forkjølelse, forkyn- nelse, forlatelse, forlengelse, forlovelse, fornemmelse, fornøyelse, forpliktelse, forsakelse, forseelse, forsendelse, forsinkelse, forskrekkelse, forsnakkelse, for- ståelse, forsørgelse, fortapelse, fortegnelse, fortsettelse, fortvilelse, foryngelse, fristelse, følelse, gjengivelse, gjengjeldelse, gjentakelse, gudsdyrkelse, heftelse, helbredelse, henrettelse, henvendelse, hevelse, hukommelse, iakttakelse, inn- bydelse, innflytelse, innrømmelse, innsigelse, innskytelse, irettesettelse, kjen- nelse, kvestelse, ledelse, ledsagelse, lettelse, lidelse, lignelse, meddelelse, min- nelse, misunnelse, mottakelse, nytelse, omgivelse, omvendelse, oppbyggelse, oppdragelse, opphisselse, opprinnelse, oppsettelse, oppstandelse, opptuktelse, overensstemmelse, oversettelse, overskridelse, oversvømmelse, overtalelse, overtredelse, overveielse, overvinnelse, påtaleunnlatelse, redegjørelse, refselse, rettelse, rystelse, salvelse, sammensvergelse, selvangivelse, selvantennelse, selvbeherskelse, selvoppholdelse, siktelse, skapelse, skikkelse, skrivelse, skuf- Low German and Nynorsk – a strained relationship? 197

felse, splittelse, stavelse, størrelse, svekkelse, takksigelse, tildragelse, tilføyelse, tillatelse, tilsigelse, tilskyndelse, tilsnikelse, tilstedeværelse, tiltredelse, til- værelse, tykkelse, understøttelse, undersøkelse, undertrykkelse, unnseelse, ut- førelse, utsettelse, utskeielse, utsvevelse, uttalelse, varetakelse, vedtakelse, vek- kelse, velsignelse, vemmelse, verdiforringelse, vielse, villelse, villfarelse, værelse, ydmykelse, ytelse, ærekrenkelse, øvelse. MAIN TEXT AND SUPPLEMENT, 1: fordøyelse.

2000 (AH, KS & AS) MAIN TEXT, 17: the same as in 1993, plus: bevegelse, fornøyelse, følelse, hevelse, prøvelse, salvelse, skikkelse, skuffelse, størrelse. SUPPLEMENT, 184: the same as in 1993, plus: forferdelse, misligholdelse, prø- velse; minus: ledsagelse, oppsettelse, tilsigelse, vedtakelse. MAIN TEXT AND SUPPLEMENT, 10: the same as in 1993, plus: bevegelse, fornøyelse, følelse, hevelse, prøvelse, salvelse, skikkelse, skuffelse, størrelse. SUPPLEMENT ONLY, BUT MARKED AS ACCEPTABLE IN NN, 4: begavelse, beskyttelse, innflytelse, oppstandelse.

2005 (KS & AS) MAIN TEXT, 21: the same as in 2000, plus: begavelse, beskyttelse, innflytelse, oppstandelse. SUPPLEMENT, 183: the same as in 2000, minus: skrivelse. MAIN TEXT AND SUPPLEMENT, 14: the same as in 2000, plus: begavelse, beskyttelse, innflytelse, oppstandelse.

Words with -heit 1993 (AH) MAIN TEXT, 25: drektigheit, dumheit, fleirspråklegheit, frekkheit, fromheit, frukt- barheit, godheit, griskheit, grådigheit, gyldigheit, hyppigheit, leilegheit, lystigheit, mog(e)legheit, myndigheit, nyheit, nøyaktigheit, offentlegheit, tom- heit, tospråklegheit, travelheit, trøysamheit, verdigheit, ømheit, øverheit. SUPPLEMENT, 241: aktsomhet, allmennhet, alminnelighet, anstendighet, barm- hjertighet, begivenhet, bekvemmelighetsflagg, beliggenhet, besindighet, betenkelighet, bevissthet, billighet(serstatning, -hensyn), bitterhet, bluferdighet, blyghet, dumhet, dyktighet, dødelighet, elendighet, enfoldighet, enhet, enighet, enkelhet, ensomhet, evighet, falskhet, fasthet, feighet, ferdighet, festligheter, fin- het, forbindtlighet, fordragelighet, forfengelighet, forgjengelighet, fornødenhet, forsiktighet, forstemthet, frekkhet, frihet, fruktbarhet, fuktighet, fullkommenhet, gavmildhet, gjenvordighet, gjerrighet, gjestfrihet, godhet, grusomhet, grådighet, gudfryktighet, gyldighet, hastighet, heftighet, helhet, hemmelighet, herlighet, hevngjerrighet, holdbarhet, hyppighet, høyhet, håpløshet, kjedeligheter, kjed- som(melig)het, kjærlighet, klarhet, knapphet, kostbarhet, kristenhet, kyndighet, kåthet, ledighet, leilighet, letthet, lettsindighet, likegyldighet, likhet, lovkyn- dighet, lydighet, mangfoldighet, medlidenhet, menighet, mildhet, minnelighet, mistenksomhet, modenhet, morsomhet, munterhet, myndighet, narraktighet, nid- kjærhet, norskhet, nyfikenhet, nysgjerrighet, nærhet, nødvendighet, nøyaktighet, offentlighet, omstendighet, omtenksomhet, oppmerksomhet, oppriktighet, oppsetsighet, overbærenhet, overhøyhet, personlighet, platthet, redelighet, renhet, renslighet, rettferdighet, rettighet, riktighet, romslighet, rådighet, rådløshet, 198 Erik Simensen

rådvillhet, sakkyndighet, saktmodighet, samhørighet, samvittighet, sannhet, sannsynlighet, sedelighet, seighet, selskapelighet, selvbevissthet, selvfølgelighet, selvstendighet, sendrektighet, severdighet, sikkerhet, siktbarhet, sindighet, sjeldenhet, sjenerthet, skjendighet, skjødesløshet, skjønnhet, skrøpelighet, skyl- dighet, skånsomhet, slapphet, slibrighet, sløvhet, smakløshet, sorgløshet, spar- som(melig)het, spedalskhet, spenstighet, sprekhet, spydighet, stadighet, stahet, stedighet, stivhet, stolthet, storhet, stridighet, sunnhet, svakhet, svimmelhet, særdeleshet, særegenhet, sømmelighet, søvnløshet, takknemlighet, taktløshet, tankeløshet, taushet, tenksomhet, tetthet, tilbakeholdenhet, tilbøyelighet, tilfel- dighet, tilfredshet, tilgjorthet, tilhørighet, travelhet, treenighet, trefoldighet, treghet, treskhet, tretthet, trofasthet, troløshet, troskyldighet, troverdighet, trygghet, tverrhet, tydelighet, tørrhet, tålmodighet, tålsomhet, tåpelighet, uavhengighet, uendelighet, uhumskhet, uleilighet, underdanighet, unnfallenhet, uroligheter, ustøhet, utholdenhet, utidighet, utlendighet, uvitenhet, vaktsomhet, valgbarhet, vanførhet, vanskelighet, varighet, varsomhet, veldedighet, velgjø- renhet, veltalenhet, vennlighet, vennskapelighet, verdighet, viderverdighet, vik- tighet, vilkårlighet, villighet, virkelighet, virksomhet, visshet, vitebegjærlighet, vitterlighet, vittighet, voldsomhet, ydmykhet, ytterlighet, ærbødighet, ærgjerrig- het, ærlighet, ødselhet, ømfintlighet, ømhet, øvrighet, åpenhet, årvåkenhet. MAIN TEXT AND SUPPLEMENT, 15: dumheit, fleirspråklegheit, frekkheit, frukt- barheit, godheit, grådigheit, gyldigheit, hyppigheit, leilegheit, mog(e)legheit, myndigheit, nøyaktigheit, offentlegheit, travelheit, ømheit.

2000 (AH, KS & AS) MAIN TEXT, 70: the same as in 1993, plus: anstendigheit, beskjedenheit, bevisstheit, bitterheit, dyktigheit, dødelegheit, einigheit, ferdigheit, fordragelegheit, forlegenheit, forsiktigheit, fortrulegheit, haldbarheit, hemmelegheit, høflegheit, landflyktigheit, lesbarheit, likegyldigheit, nysgjerrigheit, openheit, pålitelegheit, reinslegheit, rådigheit, saklegheit, sikkerheit, skjønnheit, slappheit, sluheit, spydigheit, stadigheit, staheit, stoltheit, sunnheit, svakheit, svimmelheit, sårbarheit, sårheit, takknemlegheit, tausheit, tilgjengelegheit, vennlegheit, vennskapelegheit, vittigheit, ytterlegheit, øvrigheit. Several of these (beskjedenheit, forlegenheit, fortrulegheit, høflegheit, landflyktigheit, lesbarheit, pålitelegheit, saklegheit, sluheit, sårbarheit, sårheit, tilgjengelegheit) were not even mentioned in the 1993 Supplement. SUPPLEMENT, 243: the same as in 1993, plus: høflighet, landflyktighet, lesbarhet, saklighet, sårbarhet, sårhet, tilgjengelighet; minus: siktbarhet, spedalskhet, treskhet, vanførhet, vitterlighet. MAIN TEXT AND SUPPLEMENT, 48: the same as in 1993, plus: anstendigheit, bitterheit, dyktigheit, dødelegheit, einigheit, ferdigheit, fordragelegheit, forsiktigheit, haldbarheit, hemmelegheit, høflegheit, likegyldigheit, nysgjerrigheit, saklegheit, sikkerheit, skjønnheit, slappheit, spydigheit, stadigheit, staheit, stoltheit, sunnheit, svakheit, svimmelheit, sårbarheit, sårheit, tausheit, tilgjengelegheit, vennlegheit, vennskapelegheit, verdigheit, verkelegheit, vittigheit, ømfintlegheit, øvrigheit; minus: fleirspråklegheit, moglegheit. SUPPLEMENT ONLY, BUT MARKED AS ACCEPTABLE IN NN (after adjustments: -het > -heit, -lig > -leg, menig- > meinig-, se- > sjå-, sjelden- > sjeldan-), 16: enkelhet, forfengelighet, hastighet, kostbarhet, menighet, munterhet, narraktighet, nødvendighet, platthet, riktighet, romslighet, rådighet, severdighet, sindighet, sjeldenhet, underdanighet. Low German and Nynorsk – a strained relationship? 199

2005 (KS & AS) MAIN TEXT, 104: the same as in 2000, plus: alminnelegheit, elskverdigheit, enkel- heit, forfengelegheit, hastigheit, hurtigheit, innbilskheit, kostbarheit, lovmes- sigheit, munterheit, nødvendigheit, oppriktigheit, plattheit, romslegheit, sin- digheit, sjeldanheit, smidigheit, standhaftigheit, særeigenheit, tapperheit, til- bøyelegheit, tilfredsheit, tydelegheit, tåpelegheit, uhyrlegheit, underdanigheit, varigheit, viktigheit, vilkårlegheit, villigheit, ærbarheit, ærbødigheit, ærgjerrig- heit, ærverdigheit. SUPPLEMENT, 242: the same as in 2000, minus: utlendighet. MAIN TEXT AND SUPPLEMENT, 48: the same as in 2000. SUPPLEMENT ONLY, BUT MARKED AS ACCEPTABLE IN NN (after adjustments: -het > -heit, menig- > meinig-, se- > sjå-, valg- > val-), 9: heftigheit, meinigheit, narraktigheit, nidkjærheit, riktigheit, sjåverdigheit, særdelesheit, ustadigheit, valbarheit.

Discussion If we disregard the words not allowed in NN, i.e. those entered only in the Sup- plement, the words with these affixes fall into two large classes: (1) an “upper class” of words included only in the main text and thus considered fully acceptable; and (2) a “lower class” of words included both in the main text and the Supplement, but in the latter section supplied with NN equivalents (more or less, depending i.a. on context), suggested alternative phrases and a tilde, which means that the word in question is acceptable in NN under certain conditions. Even lower in status are those words in the Supplement which are not entered in the main text. Examples: befeste, bestilling in 1993; bearbeide etc. in 2000; heftigheit etc. in 2005. This situation is probably best explained as a reflection of the linguistic purism which has been part of the NN movement ever since the time of Ivar Aasen. The rejection of loans from German had mainly two motives. One was national: these words and word elements were imported, thus marking Norwe- gian as a partly foreign language, carrying the hallmark of a dependent language (i.e. dependent on Danish, which is in turn dependent on German in this respect). International loanwords of Greek and Latin origin were more readily accepted because they represented a common European heritage and were not especially marked as Danish (see Bakken 2008, p. 35). The other motive was based on considerations internal to Norwegian linguistic structure: these elements obstructed or ousted the use and development of native derivational mechanisms, thus curbing the independent growth of the Norwegian language itself. This linguistic movement has been a powerful factor in the history of the Norwegian language all the way from Ivar Aasen to Marius Hægstad, Nikolaus Gjelsvik, Alexander Seippel and Gustav Indrebø and down to the present. At the same time it cannot be denied that many of these words were, and still are, frequently used in spoken language in many dialects. Since NN claims to be an 200 Erik Simensen adequate written representation of the language of the Norwegian people, it should accept many of these words in written NN. Puristic efforts may explain why these words have been treated so differently in NN dictionaries and word lists. The situation has become even more difficult for language users because the grouping of these words has led to differences with respect to which related forms are allowed in written NN. Hegge- lund discussed this problem in an article in Maal og Minne (1985, p. 231). In the 1993 edition of AH, certain verbs were entered in the main text (e.g. anløpe, anvise), whereas the corresponding nouns (anløp, anvisning) appeared only in the Supplement; or vice versa: the noun appeared in the main text (anretning, anlegg), but the verb only in the Supplement (anrette, anlegge). In the 2000 edition as well, some verbs were entered in the main text (anerkjenne, anstrengje, anta) while the corresponding nouns appeared only in the Supplement (anerkjennelse, anstrengelse, antakelse). In an attempt to remedy this situation, over the years some of the most frequent pairs of words related in this way have been kept together and accepted: e.g. beskytte and beskyttelse, alminneleg and alminnelegheit, anstendig and anstendigheit, etc. In general, the course of the changes between 1993 and 2005 has been such that many words which first appeared only in the Supplement gradually “advanced” to be accepted in the main text. As is evident from this survey, most of the changes in acceptance levels seem to have taken place at the end of the 1990s. There is a smaller difference in this respect between NB 2001 and 2006 than between NB 1993 and 2001. Similarly, the differences between the various editions of AH are more numerous and largest between 1993 and 2000, less so between 2000 and 2005. This is probably a consequence of an important resolution passed by the Language Council of Norway on February 3, 1999 about general guidelines regulating the adoption of words imported from German and Danish into official standard NN.1 This resolution stated that imported words which are widely used in Nor- wegian dialects should not be excluded from NN word lists and dictionaries – with some limitations, the most important being the exclusion of nouns ending in -else and -heit where synonyms in -ing, -nad and -skap, or short forms, are available (thus: beskriving but not beskrivelse; vørdnad but not anseelse; lik- skap but not likheit; samvit but not samvittigheit). The distribution of these words across dialects is an important consideration, as expressly stated in the resolution. But in addition, I think it is fair to say that some classical NN words are becoming obsolete and sound rather bookish, e.g. brune for betennelse, tame for ferdigheit and togn for tausheit. In an intermediate position, so far as I can judge, are e.g. semje and medvit, for einigheit and bevisstheit, respectively. The latter group does not have such a literary ring. At the same time, NN is constantly exposed to influence from BM, especially in its lexicon. The concept of ‘the language of the Norwegian people’ (Norw. folkemål) covers a

1 I am indebted to Helge Sandøy for this information. Low German and Nynorsk – a strained relationship? 201 wider range of usage than before. And the opposition against the anbeheitelse words sometimes looks like a delayed revenge on language history, as Lars Vikør once wrote (1983, p. 54). It has been said that NN has an identity problem: it must not become too similar to BM, and it cannot alienate itself from its literary tradition. For these reasons purism is still alive, in various forms and degrees. (In an article from the year 2000, Kjell Venås gave a historical overview and a balanced evaluation of the different positions.) The adherents of høgnorsk (‘High Norwegian’), based in the circles of Vestmannalaget (‘The Westerners’ Association’), Ivar Aasen-sambandet (‘The Ivar Aasen Society’) and Norsk måldyrkingslag (‘The Norwegian Society for Language Cultivation’), still maintain their views. On the other hand, the anbeheitelse words are not very productive, although words with be- and -heit have become somewhat more numerous over the years. As regards the -heit words, I think their increase is due to the fact that they have an important function as abstract adjectives. For the -else words, -ing is an old, well-incorporated and much used alternative. Some affixes are very frequent in compounds, e.g. betennelse and leiligheit. But on the whole the German loanwords are so old and deeply rooted in the spoken Norwegian language that they are now not felt to be foreign (see also Askedal 2005, p. 1601). We have to accept, I think, that the linguistic instinct, i.e. the feeling for what should be allowed as acceptable NN, has changed in this respect amongst the majority of Norwegians. I hope that the commission which has recently been appointed to revise the official NN standard will address this problem and come forward with more satisfactory solutions.

Conclusion In the title of this article, I raised a question. What is my answer, then? It seems to me that the relationship between Low German and Nynorsk is not so strained as it used to be. The debate has calmed down. At first, I thought of using a different title: from enmity to reconciliation. But the situation is not as simple as that. Another problem, which is complex and definitely more serious, is the massive influx of (especially technical) terms and phrases from English which threatens to dominate entire sectors of Norwegian vocabulary (e.g. in the fields of the oil industry and finance). But that is a different story. I do not pretend to have given a full and objective presentation of this special part of Norwegian language history – if objectivity is at all possible in a matter of this nature, where I am a participant and a spectator at the same time. Some readers will find my description subjective and biased, and they may be right, to some extent. Let me answer by paraphrasing the famous words of Émile Zola: it is a piece of language history, seen through a temper.2

2 I thank Oddrun Grønvik and James E. Knirk for valuable comments on an earlier version of this article. 202 Erik Simensen

Sources and literature Almenningen, Olaf, 1989: Ord med førestavinga be- i nynorsk skriftmål: Språkleg frigjering eller ugras i målåkeren? In: Ord og Mål. Festskrift til Magne Rommetveit 4. oktober 1988. Oslo. Pp. 39–57. Askedal, John Ole, 2005: The standard languages and their systems in the 20th century, III: Norwegian. In: The Nordic Languages. An International Handbook of the His- tory of the North Germanic Languages. Handbücher zur Sprach- und Kommunika- tionswissenschaft 22.2. Berlin & New York. Pp. 1584–1602. Bakken, Kristin, 2008: Det nynorske ordforrådet – sett frå K. In: Nog ordat? Festskrift till Sven-Göran Malmgren den 25 april 2008. Gothenburg. (Meijerbergs arkiv för svensk ordforskning 34.) Pp. 33–41. Braunmüller, Kurt, 1991: Die skandinavischen Sprachen im Überblick. Tübingen. (UTB für Wissenschaft, Uni-Taschenbücher 1635.) Gerdener, Wilhelm, 1986: Der Purismus im Nynorsk: historische Entwicklung und heutiger Sprachgebrauch. Münster. (Münstersche Beiträge zur deutschen und nordischen Philologie 1.) Heggelund, Kjell T., 1985: Bruken av ord på an-, be-, -else og -he(i)t i nynorsk. In: Maal og Minne. Pp. 227–248. Heit strid om nynorsk: Dokument og meiningar 1980–83. Ed. by Arne Lauvhjell 1983. Oslo. Hellevik, Alf: Nynorsk ordliste: Større utgåve. 3rd ed. 1962; 7th ed. 1993; 9th ed. 2000, ed. by Alf Hellevik, Kåre Skadberg and Aud Søyland; 10th ed. 2005, ed. by Kåre Skadberg and Aud Søyland. Oslo. Norsk språkråd (‘The Language Council of Norway’). The resolution of February 3 1999 is accessible on Internet: http://sprakradet.no/nb-no/Sprakhjelp/Rettskrivning- _Ordboeker/Ordtilfanget/Retningslinjer/. Nynorskordboka: Definisjons- og rettskrivingsordbok. Ed. by Marit Hovdenak et al. 2nd ed. 1993, 3rd. ed. 2001, 4th ed. 2006 (electronic version 2010). Oslo. Simensen, Erik, 1992: Einige Bemerkungen zur Geschichte der mit be- präfigierten Wörter in Norwegen. In: Niederdeutsch in Skandinavien III: Akten des 3. nordischen Symposions ‘Niederdeutsch in Skandinavien’ in Sigtuna 17.–20. August 1989. Berlin. (Beihefte zur Zeitschrift für deutsche Philologie 6.) Pp. 139–150. Venås, Kjell, 2000: Nynorsk skriftmål og dei lågtyske lånorda i norsk. In: Språkkontakt – Innverknaden frå nedertysk på andre nordeuropeiske språk, ed. by Ernst Håkon Jahr. Copenhagen. (Skrift no. 2 from the project Språkhistoriske prinsipp for lånord i nordiske språk. Nord 2000:19.) Pp. 289–302. Vikør, Lars: Nynorsk for folket? In: Heit strid om nynorsk: Dokument og meiningar 1980–83. Oslo. Pp. 52–68. Low German and Nynorsk – a strained relationship? 203

ACTA ACADEMIAE REGIAE GUSTAVI ADOLPHI 1. Ivar Modéer: Småländska skärgårdsnamn. 1933. 2. Inger M. Boberg: Sagnet om den store Pans død. 1934. 3. G. Granberg: Skogsrået i yngre nordisk folktradition. 1935. 4. Gertrud Areskog: Östra Smålands folkmål. 1936. 5. B. Ohlsson: Blekingskusten mellan Mörrums- och Ronnebyån. 1939. 6. Johan Götlind (och Samuel Landtmanson): Västergötlands folkmål. 1–4. 1940–1950. 7. Lennart Björkquist: Jämtlands folkliga kvinnodräkter. 1941. 8. Sven Rothman: Östgötska folkminnen. 1941. 9. Erik Brevner: Sydöstra Närkes sjönamn. 1942. 10. Verner Ekenvall: De svenska ortnamnen på hester. 1942. 11. Lars Levander och Ella Odstedt: Övre Dalarnes bondekultur. 1–4. 1943–1953. 12. Helmer Olsson: Folkliv och folkdikt i Vättle härad under 1800-talet. 1945. 13. Folke Hedblom: De svenska ortnamnen på säter. 1945. 14. Lars Forner: De svenska spannmålsmåtten. 1945. 15. Johan J. Törners »Samling af Widskeppelser». Med inledning och anmärkningar utg. av K. Rob. V. Wikman. 1946. 16. Carl Ivar Ståhle: Studier över de svenska ortnamnen på -inge. 1946. 17. Bengt Holmberg: Tomt och toft som appellativ och ortnamnselement. 1946. 18. Oskar Loorits: Grundzüge des estnischen Volksglaubens. 1–3. 1949–1960. 19. Daniel Harbe: Folkminnen från Edsbergs härad. 1–2. 1950–1956. 20. Harry Ståhl: Kvill och tyll. En studie över några i svenska ortnamn ingående ord med betydelsen ’åmöte’, ’ågren’ o. dyl. 1950. 21. Lars Hellberg: Inbyggarnamn på -karlar i svenska ortnamn. 1. 1950. 22. Per Wieselgren: Ortnamn och bebyggelse i Estlands forna och hittillsvarande svenskbyg- der. Ostharrien med Nargö. 1951. 23. Einar Törnqvist: Substantivböjningen i Östergötlands folkmål. 1–2. 1953. 24. Valter Jansson: Nordiska vin-namn. En ortnamnstyp och dess historia. 1951. 25. Nils Tiberg: Ståndssamhället [i Svensk-Estland]. 1951. 26. Gunnar Linde: Studier över de svenska sta-namnen. 1951. 27. Gideon Danell: Ordbok över Nuckömålet. 1951. 28. Andrus Saareste: Petit atlas des parlers estoniens. 1955. 29. Sigurd Fries: Studier över nordiska trädnamn. 1957. 30. Nils von Hofsten: Eddadikternas djur och växter. 1957. 31. Aleksander Loit och Nils Tiberg: Gammalsvenskbydokument. 1958. 32. Lars Alfvegren: r-genitiv och are-komposition. 1958. 33. Gösta Franzén: Runö ortnamn. 1959. 34. Nils von Hofsten: Segerlöken, Allium victoralis, i folktro och folkmedicin. 1958. 35. Nils Tiberg: Runöbondens ägor. 1959. 36. Nils von Hofsten: Pors och andra humleersättningar och ölkryddor i äldre tider. 1960. 37. Sockenbeskrivningar från Hälsingland 1790–1791. Med efterskrift och register utg. av Nils-Arvid Bringéus. 1961. 38. Nils Tiberg: Estlandssvenska språkdrag. 1962. 39. Elis Åström: Folktro och folkliv i Östergötland. 1962. 40. Olof Gjerdman and Erik Ljungberg: The Language of the Swedish Coppersmith Gipsy Johan Dimitri Taikon. Grammar, Texts, Vocabulary and English Word-Index. 1963. 41. Per Wieselgren: Ormsö ortnamn och bebyggelsehistoria. 1962. 42. Gösta Franzén: Laxdælabygdens ortnamn. 1964. 43. Jöran Sahlgren: Valda ortnamnsstudier. 1964. 44. Allan Rostvik: Har och harg. 1967. 45. Karl Axel Holmberg: De svenska tuna-namnen. 1969. 46. Kustaa Vilkuna: Kainuu – Kvänland, ett finsk-norsk-svenskt problem. 1969. 47. Fridolf Isberg: Supplement till G. Danells Ordbok över Nuckömålet. 1–2. 1970–1971. 48. Dag Strömbäck: Folklore och filologi. Valda uppsatser utgivna av Kungl. Gustav Adolfs Akademien 13.8 1970. 204 Erik Simensen

49. Claes Åneman: Om utvecklingen av gammalt kort i i ord av typen vidja i nordiska språk. 1–2. 1970. 50. Gustav Ränk: Die älteren baltischen Herrenhöfe in Estland. 1971. 51. Nils Tiberg: Estlandssvenska husdjursnamn. 1972. 52. Sven Söderström: Om kvantitetsutvecklingen i norrländska folkmål. Gammal kort stavelse i Kalix- och Pitemålen och målen i Nordmalings och Ragunda socknar. 1972. 53. Maj Reinhammar: Om dativ i svenska och norska dialekter. 1. Dativ vid verb. 1973. 54. Vidar Reinhammar: Pronomenstudier. 1975. 55. Roger Johansson: Svensk rommani. 1977. 56. Evert Melefors: Byngen, Smissen och Listar. Inbyggarbeteckningar och husbondenamn på Gotland. 1. Typologi och ordbildning. 1983. 57. Stefan Brink: Sockenbildning och sockennamn. Studier i äldre territoriell indelning i Nor- den. 1990. 58. Lena Moberg: Lågtyskt och svenskt i Stockholms medeltida tänkeböcker. 1989. 59. Bengt Odenstedt: On the Origin and Early History of the Runic Script. Typology and Graphic Variation in the Older Futhark. 1990. 60. Staffan Fridell: Ortnamn på -ryd i Småland. 1992. 61. Isländsk folkdiktning. Med inledning och kommentar av Hallfreður Örn Eiríksson. 1992. 62. John Insley: Scandinavian Personal Names in Norfolk. A Survey Based on Medieval Records and Place-Names. 1994. 63. Estlands svenskar och svenskbygd. Bibliografi sammanställd av Stig Appelgren. 1997. 64. Halvar Nilsson: De värmländska medeltidsbreven. Regester med kommentarer. 1997. 65. Gun Widmark: Stora vokaldansen. Om kvantitativa och kvalitativa förändringar i forn- svenskans vokalsystem. 1998. 66. Sven B. Ek: Polletik eller det föråldrade samhällets slingerbukter. 1998. 67. Eberhard Löfvendahl: Post- och järnvägsstationers namn i Götaland 1860–1940. Namn- givning i spänningsfältet mellan allmänna och enskilda intressen. 1998. 68. Annika Karlholm: Folkligt bildspråk. En studie över djurbenämningar i metaforisk användning i svenska dialekter. 2000. 69. Katharina Leibring: Sommargås och Stjärnberg. Studier i svenska nötkreatursnamn. 2000. 70. Eva Nyman: Nordiska ortnamn på -und. 2000. 71. Food from Nature. Attitudes, Strategies and Culinary Practices. Proceedings of the 12th Conference of the International Commission for Ethnological Food Research, Umeå and Frostviken, 8–14 June, 1998. Ed. by Patricia Lysaght. 2000. 72. Dag Strömbäck: Sejd och andra studier i nordisk själsuppfattning. Med bidrag av Bo Almqvist, Hans Mebius och Gertrud Gidlund. 2000. 73. Prosten Carl Nyréns Afhandling om Östgöthiska dialecten. Utgiven med inledning och kommentarer av Maj Reinhammar. 2000. 74. Etnologin inför 2000-talet. Föredrag och artiklar med anledning av uppsalaetnologins 50-årsjubileum den 8 juni 1998. Red.: Gösta Arvastson, Birgitta Meurling & Per Peter- son. 2000. 75. Gösta Holm: Uppväxtmiljö och språk. Agneta Horns språk i dialektgeografisk belysning. Ett bidrag till textlokaliseringens metodik. 2000. 76. Gun Widmark: Det språk som blev vårt. Ursprung och utveckling i svenskan. Urtid – Runtid – Riddartid. 2001. (Nytryck 2004.) 77. Per Vikstrand: Gudarnas platser. Förkristna sakrala ortnamn i Mälarlandskapen. 2001. 78. Sigfrid Svensson som folklivsforskare. En minnesskrift i anledning av hundraårsdagen av hans födelse den 1 juni 1901 under redaktion av Nils-Arvid Bringéus. 2001. 79. Ulla-Britt Kotsinas: Lever Ekenssnacket? Om äldre och nyare Stockholmsslang. 2001. 80. En 1700-talsordlista från Piteå. Utgiven med inledning och kommentarer av Maj Rein- hammar. 2002. 81. Lennart Larsson: Varifrån kom svenskan? Om den svenska vokabulären i en fyrspråkig ordbok utgiven i Riga 1705. 2003. 82. Mats Rydén: Botaniska strövtåg. Svenska och engelska. 2003. 83. Carl Göran Andræ: Sverige och den stora flykten från Estland 1943–1944. 2004. Low German and Nynorsk – a strained relationship? 205

84. Ella Odstedt: Norrländsk folktradition. Uppteckningar i urval och med kommentarer av Bengt af Klintberg. 2004. 85. Sockenbeskrivningar från Gästrikland 1790–1791. Utg. av Nils-Arvid Bringéus. 2004. 86. Anna Westerberg: Norsjömålet under 150 år. 2004. 87. Språkhistoria och flerspråkighet. Föredragen vid ett internationellt symposium i Upp- sala 17–19 januari 2003. Utg. av Lennart Elmevik. 2004. 88. Suffixbildungen in alten Ortsnamen. Akten eines internationalen Symposiums in Upp- sala 14.–16. Mai 2004. Hrsg. von Thorsten Andersson und Eva Nyman. 2004. 89. Maj Reinhammar: Ord för begreppet ’hos’ i äldre svenska och svenska dialekter. 2005. 90. ”Slipp tradisjonene fri – de er våre!” Red. av Bente Alver och Ann Helene Bolstad Skjel- bred. 2005. 91. Andreas Nordberg: Jul, disting och förkyrklig tideräkning. Kalendrar och kalendariska riter i det förkristna Norden. 2006. 92. François-Xavier Dillmann: Les magiciens dans l’Islande ancienne. Études sur la repré- sentation de la magie islandaise et de ses agents dans les sources littéraires norroises. 2006. 93. Festmåltid och vardagsmat. Föredrag vid ett symposium till Anders Salomonssons minne 3–4 juni 2005. Redigerad av Mats Hellspong. 2006. 94. Nils-Arvid Bringéus: Carl Wilhelm von Sydow som folklorist. 2006. 95. Staffan Fridell: Ortnamn i stilistisk variation. 2006. 96. Kristina Neumüller: Vattensjön och Vattenån. Samband mellan sjönamn och ånamn i Medelpad. 2007. 97. Nya perspektiv inom nordisk språkhistoria. Föredrag hållna vid ett symposium i Upp- sala 20–22 januari 2006. Utgivna av Lennart Elmevik. 2007. 98. Masks and Mumming in the Nordic Area. Ed. by Terry Gunnell. 2007 99. Stefan Mähl: geven vnde screven tho deme holme. Variablenlinguistische Unter- suchungen zur mittelniederdeutschen Schreibsprache in Stockholm. 2008. 100. Carl Renmarck: Plurima Lingvæ Gothicæ Rudera. Utgiven med inledning och kommentarer av Maj Reinhammar. 2008. 101. Karin Wilson: Markusevangeliet i Lars Rangius samiska översättning från 1713. 2008. 102. Tingsprotokoll från Svärdsjö socken åren 1545–1619. Utgivna av Allan Rostvik. 2008. 103. Nils-Arvid Bringéus: Åke Campbell som etnolog. 2008. 104. Kartan i forskningens tjänst. Föredragen vid ett symposium i Uppsala 23–25 november 2006. Utgivna av Lars-Erik Edlund, Anne-Sofie Gräslund och Birgitta Svensson. 2008. 105. Samlade visor. Perspektiv på handskrivna visböcker. Föredrag vid ett symposium på Svenskt visarkiv 6–7 februari 2008. Utgivna av Gunnar Ternhag. 2008. 106. Folkkultur i fokus. Tretton jubileumsföreläsningar. Redaktör: Maj Reinhammar. 2009. 107. Döden speglad i aktuell kulturforskning. Symposieföredrag utgivna av Anders Gustavs- son. 2009. 108. Thet Gothlendska Tungomålet. Språkkapitlet i Lars Neogards Gautauminning (1732). Utgivet med inledning och kommentar av Lars Wollin. 2009. 109. Svenska etnologer och folklorister. Redaktörer: Mats Hellspong och Fredrik Skott. 2010. 110. Thorsten Andersson: Vad och vade. Svensk slåtter-, rågångs- och arealterminologi. 2010. 111. Klas-Göran Selinge: Språket i landskapet. Om runstenar, rågångar och byar. 2010. 112. Probleme der Rekonstruktion untergegangener Wörter aus alten Eigennamen. Akten eines internationalen Symposiums in Uppsala 7.–9. April 2010. Herausgegeben von Len- nart Elmevik und Svante Strandberg. 2010. 113. Studier i svenska språkets historia 11. Förhandlingar vid Elfte sammankomsten för svenska språkets historia i Uppsala 23–24 april 2010. Utgivna av Maj Reinhammar under medverkan av Lennart Elmevik, Staffan Fridell, Mats Thelander och Henrik Williams. 2010. 114. En medeltida ordspråkssamling på fornsvenska. Utgiven av Inger Lindell. 2011. 115. Rut Boström: Anders Sparrmans brev till Carl von Linné. En kulturhistorisk och språklig undersökning med naturvetenskapliga inslag. 2011. 116. Studier i dialektologi och sociolingvistik. Föredrag vid Nionde nordiska dialektologkon- ferensen i Uppsala 18–20 augusti 2010. Utgivna av Lars-Erik Edlund, Lennart Elmevik och Maj Reinhammar. 2011. 206 Erik Simensen

117. Anders Wepsäläinen: Stalotomterna. En kritisk granskning av forskningsläget rörande en omdiskuterad fornlämningstyp. 2011. 118. Matti Mörtbergs värmlandsfinska uppteckningar. Sammanställda och kommenterade av Torbjörn Söder. 2011. 119. Rolf Kjellström: Nybyggarliv i Vilhelmina. 1. Träd och växter som resurs. 2012. 120. Thorsten Andersson: Gamla strand- och önamn i Bråviksbygden. 2012. 121. Contact between Low German and Scandinavian in the Late Middle Ages. 25 Years of Research. Lennart Elmevik and Ernst Håkon Jahr (editors). 2012.

The editor of the various series is the Secretary to the Academy, Associate Professor Maj Reinhammar, [email protected].

Distribution: Swedish Science Press Box 118 SE-751 04 Uppsala ISSN 0065-0897 UPPSALA 2012 E-post: [email protected] ISBN 978-91-85352-97-5