Receptive Multilingualism: Linguistic analyses, language policies and didactic concepts (Hamburg Studies on Multilingualism)

Receptive Multilingualism Hamburg Studies on Multilingualism (HSM) Hamburg Studies on Multilingualism (HSM) publishes...

Author: Jan D. ten Thije | Ludger Zeevaert

258 downloads 1143 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Receptive Multilingualism

Hamburg Studies on Multilingualism (HSM) Hamburg Studies on Multilingualism (HSM) publishes research from colloquia on linguistic aspects of multilingualism organized by the Research Center on Multilingualism at the University of Hamburg.

Editors Jürgen M. Meisel Monika Rothweiler Juliane House University of Hamburg Research Center on Mulitlingualism

Volume 6 Receptive Multilingualism. Linguistic analyses, language policies and didactic concepts Edited by Jan D. ten Thije and Ludger Zeevaert

Receptive Multilingualism Linguistic analyses, language policies and didactic concepts

Edited by

Jan D. ten Thije Utrecht University

Ludger Zeevaert University of Hamburg

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data Receptive multilingualism : linguistic analyses, language policies, and didactic concepts / edited by Jan D. ten Thije & Ludger Zeevaert. p. cm. -- (Hamburg studies on multilingualism, issn 1571-4934 ; v. 6) Includes bibliographical references and index. 1. Multilingualism. 2. Communication. I. Thije, Jan D. ten. II. Zeevaert, Ludger. P115.R427 2007 306.44'6--dc22 isbn 978 90 272 1926 8 (Hb; alk. paper)

2007009072

© 2007 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

The production of this series has been made possible through financial support to the Research Center on Multilingualism (Sonderforschungsbereich 538 "Mehrsprachigkeit") by the Deutsche Forschungsgemeinschaft (DFG).

Contents About the authors Introduction Ludger Zeevaert and Jan D. ten Thije

ix 1

Part 1 Historical development of receptive multilingualism 1. Receptive multilingualism in Northern Europe in the Middle Ages: A description of a scenario Kurt Braunmüller

25

2. Linguistic diversity in Habsburg Austria as a model for modern European language policy Rosita Rindler-Schjerve and Eva Vetter

49

Part 2 Receptive multilingualism in discourse 3. Receptive multilingualism in Dutch–German intercultural team cooperation 73 Anne Ribbert and Jan D. ten Thije 4. Receptive multilingualism and inter-Scandinavian semicommunication Ludger Zeevaert

103

5. Receptive multilingualism in Switzerland and the case of Biel/Bienne Iwar Werlen

137

6. The Swiss model of plurilingual communication Georges Lüdi

159

7. Receptive multilingualism in business discourses Bettina Dresemann

179

8. Speaker stances in native and non-native English conversation: I + verb constructions Nicole Baumgarten and Juliane House

195

viii Receptive Multilingualism

Part 3 Testing mutual understanding in receptive multilingual communication 9. Understanding differences in inter-Scandinavian language understanding Gerke Doetjes

217

10. Scandinavian intercomprehension today Lars-Olof Delsing

231

Part 4 Determiningthe possibilities of reading comprehension in related languages 11. Interlingual text comprehension: Linguistic and extralinguistic determinants Renée van Bezooijen and Charlotte Gooskens 12. Processing levels in foreign-language reading Madeline Lutjeharms

249 265

13. A computer-based exploration of the lexical possibilities of intercomprehension: Finding German cognates of Dutch words Robert Möller

285

14. How can DaFnE and EuroComGerm contribute to the concept of receptive multilingualism? Theoretical and practical considerations Britta Hufeisen and Nicole Marx

307

Name index Subject index

323 326

About the authors Kurt Braunmüller Universität Hamburg Institut für Germanistik I–Skandinavistik Von Melle Park 6 D-20146 Hamburg Germany [email protected]

Anne Ribbert Dept. of English Radboud University Nijmegen P.O. Box 9103 NL-6500 HD Nijmegen The Netherlands [email protected]

Rosita Schjerve-Rindler Universitätscampus AAKH Garnisongasse 13, Hof 8 A-1090 Wien Austria [email protected]

Jan D. ten Thije Department of Dutch Language and Culture Utrecht Institute of Linguistics (UIL-OTS) Trans 10 NL 3512 JK Utrecht The Netherlands [email protected]

Eva Vetter Universitätscampus AAKH Garnisongasse 13, Hof 8 A-1090 Wien Austria [email protected] Renée van Bezooijen Dept. of Linguistics, Radboud University Nijmegen P.O. Box 9103 NL-6500 HD Nijmegen The Netherlands [email protected] Charlotte Gooskens Dept. of Scandinavian Studies University of Groningen P.O. Box 716 NL-9700 AS Groningen [email protected]. The Netherlands

Ludger Zeevaert Universität Hamburg SFB 538 Mehrsprachigkeit — Teilprojekt H3 Max-Brauer-Allee 60 DE-22765 Hamburg Germany [email protected] Iwar Werlen Universität Bern Institut für Sprachwissenschaft Länggassstrasse 49 CH-3000 Bern 9 Switzerland [email protected] Georges Lüdi Universität Basel Institut für Französische Sprach- und Literaturwissenschaft Stapfelberg 7/9 CH-4051 Basel Switzerland [email protected]

Receptive Multilingualism Bettina Dresemann Universität Erfurt Sprachenzentrum Postfach 900 221 D-99105 Erfurt Germany [email protected] Nicole Baumgarten Universität Hamburg SFB 538 Mehrsprachigkeit — Teilprojekt K4 Max-Brauer-Allee 60 D-22765 Hamburg Germany [email protected] Juliane House Universität Hamburg Institut für Allgemeine und Angewandte Sprachwissenschaft Abteilung Sprachlehrforschung Von-Melle-Park 6, II. Stock DE-20146 Hamburg Germany [email protected] Gerard Doetjes Fremmedspråksenteret Høgskolen i Østfold N-1757 Halden Norway [email protected] Lars-Olof Delsing Lunds Universitet Språk- och litteraturcentrum Nordiska språk Box 201 S-221 00 Lund Sweden [email protected] Madeline Lutjeharms Vrije Universiteit Brussel Vakgroep TALK en Instituut voor Taalonderwijs Pleinlaan 2, lokaal E210

B-1050 Brussel Belgium [email protected] Robert Möller Université de Liège Département de Langues et Littératures germaniques Place Cockerill, 3 B-4000 Liège Belgium [email protected] Britta Hufeisen TU Darmstadt Sprachenzentrum Hochschulstr. 1 D-64289 Darmstadt Germany [email protected] Nicole Marx TU Darmstadt Sprachenzentrum Hochschulstr. 1 D-64289 Darmstadt Germany [email protected]

Introduction Ludger Zeevaert and Jan D. ten Thije Receptive Multilingualism. Linguistic analyses, language policies and didactic concepts This volume reveals new perspectives from different theoretical frameworks on linguistic analyses of receptive multilingualism in Europe. Receptive multilingualism refers to the language constellation in which interlocutors use their respective mother tongue while speaking to each other. Case studies are presented from contemporary settings, along with analyses of historical examples, theoretical considerations and, finally, descriptions of didactical concepts established in order to transfer and disseminate receptive multilingual competence. Receptive multilingualism cannot (yet) be regarded as an established field within research on multilingualism, even though the economic and political developments, usually denoted as globalisation, have led to a considerable increase in international communication. In fact, it has become clear, that communicative challenges connected to these developments are hardly solvable using traditional concepts of multilingualism. Therefore, new concepts have to be developed and discussed. At the University of Hamburg and especially at the Research Centre 538: Multi lingualism pioneering work was carried out. Between 1989 and 1995 the role of semicommunication as used between speakers of Middle Low German and the Scandinavian languages was investigated.1 Receptive multilingual communication as a form of language contact that had a major impact on the development of the Mainland Scandinavian languages was discussed and in the following widely accepted in the literature on Scandinavian language history (cf. e.g. Barðdal et al. 1997: 362, Teleman 2002: 29, Josephsson 2006: 22). The language situation in contemporary Scandinavia was the subject of research in a second project,2 yielding several dissertations (Zeevaert 2004, Ház 2005, Golinski 2007 and Doetjes in prep.) and numerous further publications. In fact, only four out of fourteen articles in this volume actually come from the Hamburg research project on semicommunication. This shows how exchange and discussion on receptive multilingualism have spread over Europe and increasingly attracted attention of functionaries in all kind of institutions, various researchers and policy makers. So far, receptive multilingualism had been a typical bottom-up development, supported by official European organisations only to a certain extent compared to other EU language policies. Since the mid-nineties receptive multilingualism is promoted by the European commission on par with other possibilities of increasing the mobility of the European citizens in order to solve the structural problems within the European Union. Throughout,

Ludger Zeevaert and Jan D. ten Thije

roughly speaking, the last ten years a marked increase in the research on this topic has been observable, a fact which was not least stimulated by the challenges set by the European motto, unity in diversity, which also refers to the linguistic situation in Europe. The increasing importance of this issue has been emphasised by the appointment of a commissioner for multilingualism. Since the beginning of 2007 there are now 27 official languages in the EU. The number of languages spoken as the mother tongue by EU citizens, however, can be estimated to be between 40 and 100, depending on whether nearly extinct languages such as e.g. Karaim in Lithuania or languages that are linguistically very close to the official language of a country like Limburgish in the Netherlands are included or not. This volume challenges three tacit assumptions in the field of multilingual communication research, that are countered by the following statements: • M ultilingualism is a social phenomenon deeply embedded in European language history. • Multilingual understanding does not necessarily require near-native language competency. • English as lingua franca is not the one and only solution for interlingual communication in Europe. The first assumption refers to suggestions that multilingualism is a recent phenomenon and is mainly related to globalisation and labour mobility. In contrast to these statements it has to be said that, in actual fact, the idea of monolingualism as the standard case for individuals and societies is the result of the emergence of nation states in Europe in the eighteenth century resulting in national linguistic homogenisation. For most countries outside of Europe and North America monolingualism is a somewhat unusual phenomenon. In many countries in Asia, South America or Africa several different mother tongues are spoken. For exogamic societies such as the Vaupés in South America multilingualism is inevitable. Marriages between members of the same speech community are prohibited, meaning that every child grows up in a bilingual environment (cf. Romaine 1994: 38). In countries like India or South Africa the use of four different languages with different family members and colleagues in everyday life is quite normal (cf. the depiction of Bhatia and Ritchie 2004: 796f. , or Kamwangamalu 2004: 726f.), although a complete near native linguistic competence is not seen as a prerequisite for successful communication. The same pattern can be observed in Europe in the Late Middle Ages and in Early Modern Times when communication was multilingual by default. One purpose of this book is to reconstruct the historical developments of various multilingual constellations while focussing especially on receptive multilingualism. Scandinavia, Switzerland and the Habsburg Empire offer interesting historical material for the linguistic study of the main characteristics of this multilingual constellation under various conditions. It is not by coincidence that eight of the fourteen articles in this volume refer to multilingual constellations between Germanic languages. In several publications the

Introduction

Scandinavian language community is described as a functioning example of receptive multilingual communication and, therefore, as a model for European understanding. Already in medieval sources, e.g. the Icelandic lawbook Grágás (‘Greylag Goose’) which is passed down in manuscripts stemming from the thirteenth century, Scandinavia is described as as a common speech area in which dönsk tunga (‘the Danish language’) is spoken (cf. Melberg 1952 for a comprehensive depiction). At the same time visitors from Southern Europe (like the Spanish archbishop Rodrigo Ximenes) reported that the inhabitants of Germany, Scandinavia, the Netherlands and England all seemed to speak dialects of the same language (cf. Karker 1978: 7). Germanic languages are the official languages in nine EU countries (covering more or less the same area described by bishop Ximenes in the thirteenth century) which is more than any other language group. In sum, Germanic languages have determined multilingual history in Northern Europe considerably. In Scandinavia receptive multilingual communication was propagated as early as the first half of the nineteenth century as part of the Pan-Scandinavian movement (cf. Zeevaert 2004: 47 and Doetjes this volume). This movement gradually led to a more intensive political and cultural cooperation between the Scandinavian countries nowadays known as the Nordic Council. However, in the context of speakers of the Germanic language group outside of the Scandinavian languages this consciousness of linguistic commonalities and tradition of receptive multilingual communication is far less developed. In these cases, new didactical concepts had to be established in order to overcome any linguistic differences. It should be mentioned that already over 75 years ago Heinz Kloss (1929) designed a detailed concept of acquiring a receptive competence of the different (West) Germanic languages, aimed at establishing a mutual understanding between speakers of Afrikaans, German, Dutch, Pennsylvania Dutch, Yiddish and Frisian. At that time, his concept was mostly disregarded, seen as all attention was placed on approaches supporting the homogenisation and dissemination of national languages. Currently, a team of researchers from various European universities under the direction of Britta Hufeisen (Darmstadt) are working on the didactical implementation of the receptive multilingual approach towards the Germanic languages that we discuss more extensively below (cf. Hufeisen and Marx in prep.). The second assumption is related to the conception that only near-native language competency guarantees a successful multilingual understanding. On the contrary, the contributions presented in this volume argue in favour of native-like competence no longer being a sufficient prerequisite for adequate multilingual communication in many business and institutional settings. Consequently, this conception can or even should be replaced by a list of oral and written competencies which comprise (meta-) linguistic and intercultural understanding, action and institutional knowledge. The analyses reveal how the concept of receptive multilingualism requires more than minimal linguistic knowledge, and is neither a simple pidgin nor incomplete language learning. Instead, it represents the acquisition of receptive competencies in more than one given target language, and at the same time includes a set of specific foreign


language learning strategies on the side of the hearer in receptive multilingualism. This leads to the conclusion that passive competence is a misnomer for receptive competence. In receptive multilingual constellations as analysed in this volume the main effort has to be made by the hearer. In cases of any problems occuring the interactants have to decide whether they want to somehow solve the problem or choose a so-called let it pass-strategy (cf. Baumgarten and House, Doetjes, Dresemann, Zeevaert in this volume). In interscandinavian communicative encounters as examined by Zeevaert (2004) only few instances of accommodation strategies (slower and more accentuated pronunciation, repetitions, reformulations) could be observed. In particular cases, however, differences between (receptive) multilingual and monolingual discourses were clearly visible: in receptive multilingual discourses the terms of communication are not fixed to the socio-cultural knowledge of the members of the specific speech communities. Instead, they have to be negotiated by the participants and can be considered as being related to discursive intercultures that result from a long cooperation in cultural contact (Koole and ten Thije 1994, ten Thije 2003). Strategies aimed at ensuring mutual understanding can be negotiated during interaction, but also predefined by seeking agreement on the preferred language constellation in advance. In the case of the Dutch–German communication analysed by Ribbert and ten Thije (this volume) this was achieved by means of a team agreement between the two persons involved prior to their working relationship at the GoetheInstitute in Amsterdam. This agreement on multilingual institutional discourse fits in well with the language policy of the local institution. The institutional embedding of the interscandinavian discourses investigated by Zeevaert (this volume) enables an official formulation of written instructions for linguistic behaviour that are handed out to the participants at the interscandinavian meetings. Even in the Swiss context official rules for multilingual communicative encounters exist, but in this case they are often overridden by general customs developed by the citizens of the multilingual communities. At least in those cases in which the speakers were not able to develop communicative competences and cultural habits during earlier exolingual encounters this negotiation on multilingual understanding cannot be carried out in advance, meaning that interlingual strategies have to be tried out and executed within the discourse itself. Common for all these situations is the fact that a prerequisite necessary for the success of mutual understanding lies in the acquisition of new linguistic competences. These competences only partially overlap with those usually focussed on in foreign language teaching. Besides the competence in their mother tongue speakers have, differently elaborate, partial competences in other language varieties, but also knowledge about other languages which may be less developed or even wrong (cf. Coseriu 1988: 153ff.). A receptive multilingual competence as described by Lüdi (this volume) goes beyond pure linguistic knowledge and utilises those partial competences by means of developing multilingual communicative strategies. Such strategies can be developed by the participants within multilingual communication. The emergence of

Introduction

pidgins in situations of language contact between members of mutually incomprehensible languages shows the capability possessed by humans to establish mutual understanding if they wish to do so. A distinct improvement in communication, though, can be reached by using strategies that go beyond the utilisation of the context or of universal linguistic commonalities and by taking advantage of any given correspondences between the languages involved. As Hufeisen and Marx (this volume) point out, the term foreign languages, viewed from the perspective of receptive multilingualism, has to be considered as a misnomer seen as no language can really be regarded as foreign. However, the ability of language users to find such correspondences is limited. Hufeisen and Marx show that language learners have difficulties recognising similarities between languages automatically. Strategies of linguistic transfer have to be made obvious with the help of didactical procedures. For the Romance languages a method of inference with the help of so-called bridge languages has already been established. Horst Klein and Tilbert Stegmann developed an elaborate method (Klein and Stegmann 2000; McCann, Klein and Stegmann 2003) working with students of Romance languages at the University of Frankfurt. It was used successfully in language instruction and is based on the technique of the “seven sieves” which are used to sieve through texts in search of lexical, morphological and syntactic correspondences in the bridge language in order to make clear the similarities between the languages. The Romance languages provide especially good conditions for this procedure seen as they are much closer related to each other than e.g. the languages of the Germanic group. Intercomprehension between the Germanic languages can profit from the EuroCom method developed originally for Romance languages. For instance, ideas taken from the implementation of the EuroCom method on the Germanic languages (Hufeisen and Marx in prep.) were used successfully in an introductory Icelandic course offered to students of Scandinavian languages without previous knowledge of Icelandic at the University of Hamburg.3 The success of the method was tested. On the one hand an impressive increase in reading comprehension was reached. On the other hand special characteristics of reading in a foreign language (as described by Madeline Lutjeharms in this volume) led to less satisfactory results for certain text types. More conscious operations of linguistic transfer were required, with the accumulation of attentional processing leading to an overexertion of the working memory which in turn led to frustration and demotivation among the test persons. In sum, research on didactical methods of automatising strategies of linguistic transfer is needed. Furthermore, the practical use has to be obvious for the learners in order to keep them motivated. The disenchanting results of the neighbouring language education in Scandinavia (cf. Zeevaert 2004: 59ff.) — as a part of their education in the mother tongue, Danish, Norwegian and Swedish pupils are also instructed in the respective languages of their Scandinavian neighbours — can mainly be attributed to a lack of motivation.


A counterexample is given in the material provided by NUAS, an organisation which arranges meetings of employees from Scandinavian universities and applies receptive multilingualism at their conferences (cf. Zeevaert in this volume). This material is sent to the participants of the meetings in advance and contains most importantly information on useful strategies suggested in order to cope with the special communicative conditions. Limited complexity and a clear structure make it possible to work through the material in a short time. Zeevaert’s (2004, this volume) analyses of those meetings give the impression of a well-functioning communication. The participants use the linguistic instructions as a starting point for the development of a receptive multilingual competence. During the course of the interactions the participants gain experience and at the same time acquire metalinguistic knowledge about the languages involved, but also knowledge about successful strategies of communication which in turn can be integrated into their communicative competence. In different constellations in which the participants reveal diverse experiences those varied competences can lead to individually differing strategies, but likewise in the multilingual settings as described by Werlen (this volume) discourse traditions are established in order to cope with the peculiarities of this multilingual situation. The third — nowadays hardly tacit — assumption refers to the conception that the mastery of English as Lingua Franca is the most plausible solution for all international communication in Europe. Even though English is learned as a foreign language by the majority of the EU citizens the distribution of competence is rather erratic between and inside the different countries depending on varying traditions and levels of education, not to mention the fact that it is easier for speakers of Germanic languages to master English than for learners from other language groups. In their contribution to this volume Baumgarten and House come to the conclusion that even in discourses carried out by rather proficient speakers of English as a foreign language communicative incongruities still tend to exist mainly resulting from the speakers’ different (socio)linguistic and cultural backgrounds. Complete mutual intelligibility between the different varieties of lingua franca English cannot be taken for granted, and the results presented by Dresemann (in this volume) indicate that participants in international business encounters cannot rely solely on their knowledge of English. The contributions in this book do not dispute the importance of English as an international language. However, they exemplify how — depending on the typological distance of the languages involved, the language competencies of the participants, the given institutional preconditions and the non-linguistic purposes that have to be realised — the method of receptive multilingualism can be a far more efficient way of gaining mutual understanding than the use of English as a lingua franca. Europe is in need of a language policy that accounts for the diversified regional interests instead of issuing solutions that run counter to the actual local requirements. The learning of this lesson is one of the major challenges facing a common European language policy.

Introduction

The contents of the contributions Part 1: Historical Development of Receptive Multilingualism Kurt Braunmüller presents a survey of the linguistic situation in northern Europe in the Late Middle Ages and Early Modern Times. At that time receptive multilingualism was just one aspect of a complex diglossic/multilingual situation. It was used mainly in face-to-face trading communication. Complete bilingualism was rather an exception, and merchants more often than not only possessed partial competences or a passive competence in the second language. This communicative situation was not unusual for people in the Middle Ages seen as no standard language existed and every speaker used his own dialect and was used to having to identify sound correspondences, grammatical morphemes and divergent terms and transfer them into his own dialect. Furthermore the use of a lingua franca was influenced by the speaker’s origin, that is to say by his mother tongue, leading to syntactic and semantic differences between the speakers. In some contexts even intrasentential code mixing was not unusual. Latin was primarily the language of the church, the sciences and of higher education, in other words of written domains, whereas Low German was the lingua franca in trading situations, but also for political consultations in face-to-face communication. Low German was a prestigious language for the upper classes in northern Europe and the source of extensive lexical borrowing. Due to both the close genetic relationship and the frequent contact between Low German and the Scandinavian languages mutual understanding in oral communication was possible. Seen as they only remained in the Scandinavian towns for shorter periods the Hanseatic merchants generally did not acquire an active command of the Scandinavian languages. Common traditions of commerce and the familiarity with the terms of trade combined with processes of linguistic accommodation enabled successful face-to-face communication. Due to the fact that Middle Low German was the more prestigious language in this contact situation the accommodation was often performed by the Scandinavians and resulted in a broad lexical, but also morphological influence of Middle Low German on Danish, Norwegian and Swedish. In the course of time, however, the persistent contact led to the acquisition of deeper knowledge of the other languages improving the success rate of receptive multilingual communication even more. Finally, nationalism put an end to this way of unmediated communication between genetically closely related languages. In contrast to the European Economic Community, which developed from a trading organisation to a supranational political alliance, the Hansa was confronted with growing national efforts towards the end of the Middle Ages causing trading restrictions and ultimately the closing of all trading offices and either the migration or an integration and linguistic assimilation of the merchants. Furthermore, political power was centralised and linguistic standardisation leading to the formation of national languages initiated. It became necessary to


demonstrate one’s political loyalty by means of practising linguistic loyalty, and this development also implicated a certain loss of linguistic flexibility. The speakers were no longer confronted with dealing with deviating language varieties. A modern example for this coherence is the fact that Norwegians, who already have to deal with different spoken and written varieties of Norwegian inside their own borders, perform much better in tests on interscandinavian comprehension than their Danish or Swedish neighbours (cf. the contributions by Delsing and Doetjes in this volume). By means of three Scandinavian case studies Braunmüller’s article points out the role of receptive multilingualism as a triggering factor in the development of the modern Scandinavian languages. Receptive multilingualism was the starting point of L2language acquisition by adults. German speakers compared the definite article þann/ þan/þat with the Middle Low German article with a d- in the onset and consequently replaced the Scandinavian fricative with the familiar obstruent d. This pronunciation was taken up by Scandinavians as a kind of prestige pronunciation and ultimately led to the modern Mainland Scandinavian forms den/det/de. In a similar manner the Low German periphrastic genitive was adopted by the Scandinavian languages as the result of a reanalysis. Even the decrease of V1 patterns in main clauses is described as being influenced by adult L2-learning, viz. as a simplification of syntactic variation in the target language following the model of the source language. In this sense receptive multilingualism represents a sufficient starting point for second-language acquisition, especially for adults. A second historical example of a multilingual setting involving receptive multilingualism is contributed by Rosita Rindler-Schjerve and Eva Vetter. The authors describe the language policy of the Habsburg Empire in the nineteenth century as an exception to the general development of linguistic homogenisation that played a central role in the formation of nation states as described by Kurt Braunmüller. German as the language of the politically dominant ethnic group functioned as a lingua franca but never became an all-embracing state language. Like in many other historical multilingual situations primary data referring to the actual language use is not available. The reason for the lack of contemporary reports on the language use in multilingual settings could be seen in the fact that such situations were not considered to be unusual and thus not newsworthy. Therefore, this investigation is based on official documents of language policy that regulate the multilingual communication and can thus be seen as reactions to contemporary problems and thus as a fitting description of the situation. Even though the Habsburg language policy was based on the principles of pluralist equality and democratic participation and was aimed at meeting the linguistic requirements of the various ethnic groups throughout the different parts of the empire, it was far from being unambiguous and could differ considerably over time and geographical space. This fact is illustrated by three case studies of three specific domains — education, administration and the judiciary — in the different crown lands of Bohemia, Galicia and Trieste. The centralistic Habsburg Empire and the European Union as an economic and political alliance of equal member states can only be compared to each

Introduction

other to a limited extent. The noteworthiness of the Habsburg model of multilingualism lies in the fact that it is one of the few documented examples of linguistic pluralism and that the reasons for its success and failure can both be described and utilised in the context of the contemporary European language policy. One important reason for the failure of the implementation of principles of a pluralistic language policy in the Habsburg Empire was that on behalf of an equation between the central power and the political elites in the different territories the interests of the minorities were neglected. For a modern European language policy to be successful the lesson has to be learned that it needs to be based on a set of commonly shared values and principles constituting the ideological basis for democratic decision-making and the solving of problems. The case studies from the Habsburg Empire illustrate that it will be crucial for the EU to find a balance between the necessity for a certain homogenisation of the member states and the respect with regard to their differences. Centralistic interventions may thwart the aim of democratic equality if they disregard diversified regional interests, and it is at least doubtable whether the tendency towards ‘English only’, which at the moment can be observed within the EU and is surely looked upon especially by the smaller countries as a way to reach democratic equality in the linguistic domain, will be accepted by all members. The experiences from the Habsburg Empire are one of the few examples which can be used to build a multilingual supranational commonwealth on common political grounds in Europe.

Part 2: Receptive multilingualism in discourse Anne Ribbert and Jan D. ten Thije review three different factors that influence the occurrence of receptive multilingualism in German Dutch interlingual contact, namely factors referring to social and linguistic relations between nation states, the institutional constellations within nation states and factors related to the perspectives of the individual interactants. Compared to Scandinavian languages, German and Dutch are not as closely related (Goossens 1985). Moreover, Germany and the Netherlands have a more discordant common history resulting from the Second World War (Westheide 1997). Consequently, the willingness to exercise receptive multilingualism in German Dutch interlingual contact is — irrespective of any existing language correspondences — negatively influenced by their social history (Herrlitz 1997). Nevertheless, Beneke (1996, cit. in Loos 1997) states that in the Dutch–German border area receptive multilingualism is increasing. On the basis of a pilot study carried out at the Goethe Institute in Amsterdam the authors illustrate the occurrence and success of receptive multilingualism within institutional cooperation. Moreover, the study reveals that the existence of key words is an important prerequisite in order to successfully make use of receptive multilingualism. According to Koole and ten Thije (1994) the cognitive structure of institutional key words can be characterised as follows: they represent common institutional knowledge, they are abstract frames whose slots are filled in by concrete knowledge elements, and are connected to institutional purposes. In order

10


to investigate processes of understanding facilitated by key words the authors also address intercultural characteristics of receptive multilingualism. Research into intercultural communication has for a long time focussed on misunderstanding. Bührig and ten Thije (2006) reveal a shift of attention towards successful intercultural discourse. In actual fact, the article reveals how key words contribute to intercultural understanding by means of exemplarily reconstructing the manner in which cultural apparatuses (Rehbein 2006) are applied by the interactants. Ludger Zeevaert aims at giving a theoretical subsumption of the term receptive multilingualism and related terms such as polyglot dialogue, semi-communication and intercomprehension. It is common that these terms denote differing situations of communicative encounters between members of different speech communities. However, in contrast to interpreting, the use of a lingua franca or L2 communication, those situations are characterised by the fact that the speakers do not aim at communicating in a common discourse language but stick to their own L1 while being able to understand the respective L1 of their counterpart. Referring to Maturana’s radical constructivistic approach to communication, Zeevaert questions the opinion of receptive multilingualism being a form of ‘passive’ multilingualism. Based on examples from interscandinavian professional discourses he describes the role of the hearer in receptive multilingual communication as an active one. Thus, it is not the speaker who creates information by sending a message to a hearer. The information is far more created by the hearer in the process of integrating the speaker’s utterances into his cognitive space and thus reducing his own uncertainty. Following this model a prerequisite for successful communication is a consensual sphere common for both speaker and hearer. The suggested active role of the hearer becomes obvious particularly in the context of receptive multilingual communication seen as a common language as a part of this consensual sphere is missing. Due to either active learning (in the case of mutually unintelligible languages) or overlaps between the respective languages involved (in the case of mutually intelligible languages), parts of the linguistic systems of the participants involved in the communication are identical. In contrast to communication between speakers of the same speech community with identical codes, however, it can not be taken for granted that an utterance will always be understood. Of course the context — or the consensual sphere of speaker and hearer — can help to support successful communication. People sharing common cultural traditions, common professional knowledge or general common interests are consequently more successful in establishing mutual understanding than people without a common background. One of the few examples of institutionalised receptive multilingual communication is interscandinavian semicommunication. As shown in the contributions by Delsing and Doetjes in this volume, the level of the understanding of Danish by Swedes and Swedish by Danes does not suggest that a spontaneous, unimpeded understanding between speakers of these languages would be possible. Zeevaert’s analyses of interscandinavian discourses, however, do not reveal any severe difficulties in understanding that would justify a characterisation of those discourses as problematic. He comes

Introduction

to the conclusion that one main factor for the success of the communication is the common professional background of the participants. But also the fact that the conversations are held by larger groups, partly from the same country, helps to facilitate the communication seen as it eases the pressure on the individual speakers. In some cases even signals of second-language acquisition can be observed, mostly in discussion groups with participants more experienced in interscandinavian communication. They include strategies of dealing with trouble sources, but also metalinguistic knowledge which is at least partly acquired during the discourses. The disadvantage of L2 or lingua franca communication compared to receptive multilingualism can be seen in the discourse behaviour of speakers who are not able to use their mother tongue (mostly Finns and Icelanders who have to speak Swedish or Danish). Those speakers take part in the discourse less actively. A second European area where receptive multilingualism is practiced is Switzerland. Iwar Werlen provides an outline of Swiss multilingualism which is characterised by a demand for linguistic peace. Switzerland is a multilingual state with four official languages, even though the language borders on a whole are separated quite clearly meaning that individual bilingualism is rather an exceptional case. In contrast to Switzerland as a whole, most of the 26 cantons have only one official language. In recent years language policy has been directed at supporting receptive multilingualism by means of teaching a second national language in primary school with emphasis on the importance of receptive competences. For the communication of people from different language backgrounds the default model in Switzerland is the territoriality principle, i.e. the discourse language is the official language spoken in the respective area. In bilingual regions, however, different models are used. Werlen analyses language use in public or semi-public places in two cities in the bilingual Swiss cantons Berne and Fribourg. The situation in the German/French bilingual city of Biel/Bienne can be characterised as a double monolingualism system. Two educational systems, a French and a German one, exist, and communication with the administration can be carried out either in French or German. Since no clear separation exists between the two language groups, German-speaking children automatically acquire French and French-speaking children acquire German as an L2. In addition to this, French and German are taught as foreign languages at school from the age of 11 onwards. In Fribourg/Freiburg, another French/German bilingual city, German is a minority language. Therefore French is expected to be the default language used in informal communicative encounters between unacquainted interlocutors in public places. The results from test recordings confirm this assumption. Interestingly enough, in cases in which French speaking persons were addressed in (Swiss) German and vice versa one could observe differences between the two cities Biel/Bienne and Fribourg/Freiburg. In Biel in most cases the addressee accommodates his choice of language to the addresser, especially in service encounters in which the greeting of the customer decides on the language of discourse. Even if one of the interlocutors has limited knowledge of the respective discourse language this does not affect the usage of the Biel model. In Fribourg, however, in cases in

11

12


which a French speaking person is addressed in (Swiss) German the conversation will continue in a receptive multilingual mode unless the addresser accommodates to the addressee’s language. This article supports — like other articles in this volume — the opinion that receptive multilingualism is a significant, democratic option for multilingual societies. In multilingual settings different usages of language choice will automatically develop under different circumstances. However, the Swiss model shows that political influence on those usages is also possible. The contribution by Georges Lüdi also deals with examples from authentic receptive multilingual discourses in Switzerland. In contrast to Werlen’s observation of casual discourse he analyses the internal organisational communication in different Swiss companies. Patterns of language choice are less clear in these institutions than in the bilingual Swiss cities where traditions for language use have developed. In the examples analysed by Lüdi the interlocutors have to negotiate the language choice while making use of all communicative resources they dispose of. The status and action potentials of the participants play an important role in this kind of institutional discourse, leading to the conclusion that the Swiss model of receptive multilingualism is, in contrast to Werlen’s examples from Fribourg, by no means self-evident. Even in cases in which the official language policy of a company follows the Swiss model differing patterns of language choices can nevertheless be observed. In actual fact, bilingual bank discourses adequately illustrate those difficulties. The discourse data analysed by Lüdi originated from the fusion of a monolingual French-speaking and a monolingual German-speaking bank. The receptive competences of the interlocutors are not always sufficient in order to be able to stick to a receptive multilingual mode, and various techniques such as accommodation to the language of other interlocutors, language mixing or linguistic mediation by means of translations or short summaries by linguistically more skilled discourse participants can be frequently found in the discourses. In addition to this, a problem that appeared only marginally in the data investigated by Werlen plays an important role in Lüdi’s examples. He refers to the constellation as not being bilingual but rather trilingual, given the fact that in monolingual settings the German-speaking employees usually speak their local Swiss-German dialects, whereas the German acquired by the French-speaking employees is the written standard variety. Thus, the use of Swiss-German in certain contexts automatically excludes the French-speaking interlocutors. A different strategy was observed in the discussions at a scientific colloquium involving participants from different French, German and Swiss universities. In this constellation the language choice in discourse was not defined in advance, and the participants had to negotiate their linguistic behaviour within the discussion. The large number of instances of language crossing in which speakers do not use their L1, but instead, for reasons of politeness, switch to the L1 of their interlocutor or to the lingua franca English, show that the practical implementation of the theoretical concept of receptive multilingualism leads to different outcomes in different constellations. Nevertheless, all examples of receptive multilingual discourse investigated in the article can be described as being successful.

Introduction

Bettina Dresemann analyses the occurrence of receptive multilingualism in business communication. English is the number one language used in business communication today. However, sometimes various languages are applied within one discourse leading to various ambiguities appearing. In actual fact, interactants need to grasp utterances in a language they do not actually understand. Therefore, the author argues that participants in international business encounters cannot rely on their linguistic knowledge alone (i.e. mainly lexical and semantic knowledge), and consequently have to refer to other elements of the discourse such as tyings and cues, which enable them to relate the utterance to the situation, e.g. a business negotiation. Furthermore, discourse knowledge enables the interpretation of the utterance in combination with institutional, professional and general knowledge such as knowledge on international business constellations in general. Only taking this knowledge structure into account is it possible to explain how participants manage to understand each other and are able to act in linguistically ambiguous situations. The ability to draw conclusions from linguistic and non-linguistic cues and to combine them with other forms of knowledge, such as pragmatic and institutional knowledge, is therefore extremely essential for successful communication in multilingual (business) discourses. English as a lingua franca plays a prominent role in communication between speakers from different speech communities, not only in business discourses, but also in private encounters. Even in areas with a strong tradition of receptive multilingual understanding English has gained ground as a means of establishing communication on equal grounds (cf. the contributions by Delsing and Lüdi in this volume). Nicole Baumgarten and Juliane House compare discourses between L1–English speakers with those of speakers of English as a lingua franca while paying special respect to the linguistic construction of subjectivity in the context of I + verb constructions (e.g. I think, I don’t know, I mean). The analysis is based on Halliday’s classification of verbal process types. Baumgarten and House find differences in the expression of subjectivity between L1 English and lingua franca English discourses, but also between the different groups of lingua franca users. On the one hand the lingua franca users have a more restricted repertoire of means for the expression of subjectivity, on the other hand the distribution of the different process types also differs. The lingua franca users seem to overgeneralise and refer to the basic meanings of a structure, possibly in awareness of miscommunication, avoiding more grammaticalised structures and pragmatical usage, whereas L1 speakers of English tend to use certain constructions as verbal routines. A comparison both between different lingua franca discourses and also between different speakers reveals considerable individual differences that can partly be attributed to the speaker’s L1. Other factors, however, may also play a role in the emergence of differences. A characteristic feature shared by lingua franca English discourses and receptive multilingual communication (as described by Zeevaert in this volume) is the use of the so-called ‘let-it-pass’ strategy. Incomprehensible or inadequate utterances are often simply ignored by the discourse participants. This could be seen as a preference for a restriction of the communication to the level of a mere exchange of informa-

13

14


tional content in both settings. The study is not able to invalidate arguments that were raised against lingua franca communication and described it as less precise, monotonous, toilsome and as lacking a cultural integration (cf. e.g. Finkenstaedt and Schröder 1990). Even if in most cases of supranational communication throughout Europe it will not be possible to substitute the use of English with receptive multilingual communication due to the lack of receptive competences, the results from the study manage to raise the question as to whether differences between the usage of English as an L1 and lingua franca English actually influence the communication between L1 and L2 speakers of English. It should be seriously analysed in which contexts receptive multilingualism would actually be a profitable option.

Part 3: Testing mutual understanding in receptive multilingual communication The mainland Scandinavian languages Danish, Norwegian and Swedish are usually characterised as mutually comprehensible. In most interscandinavian encounters Scandinavians use their mother tongue expecting to be understood by their fellow Scandinavians. This impression of generally successful communication is confirmed e.g. by Zeevaert’s (in this volume) analyses. However, different studies in this field, but also understanding problems frequently reported by the participants in interscandinavian meetings suggest that receptive multilingual communication in Scandinavia is in fact not always unproblematic. In order to describe the mechanisms of receptive multilingual communication in Scandinavia objective measurements of the degree of mutual intelligibility are necessary. These measurements could also help to explain why this way of communication does not play a more prominent role in the Romance or Slavic language areas, even though the linguistic overlaps are comparable to those in Scandinavia, and to rate the importance of non-linguistic factors such as attitudes or cultural and political factors for mutual understanding in a more reliable manner. Gerard Doetjes presents an overview of the different studies of Scandinavian intercomprehension that have been performed since the 1950ies and describes the methodological problems connected to the different approaches. He comes to the conclusion that the test results are heavily influenced by the choice of method, a fact which has to be taken into account when applying the results to further research. Doetjes tries to determine how the difficulty level, but also how the type of questions used in the tests influence the results. Test types using pre-formulated answers enable the test persons to rely both on the text and on the information given implicitly in the questionnaire, whereas in the case of open questions the participants have to rely more on information gathered from the texts themselves. This leads to higher average scores in the multiple-choice test and to lower scores in tests demanding summaries. Therefore, in order to be able to determine the reliability of a test, it might be useful to recalculate the results on the basis of Doetjes’ comparison of the different methods under otherwise stable test conditions. Moreover, the results support Lutjeharms’ descrip-

Introduction

tion of the processing levels referring to the reading of related languages in such a way that longer summaries providing a wider range of possibilities to make use of formulations found in the test text led to better results compared to short summaries that require a more advanced processing of the content and a certain act of dissociating from the formal side of the test text. The INS investigation presented by Lars-Olof Delsing is the most recent analysis on the intercomprehension of Scandinavian languages. It was funded by Nordiska kulturfonden (‘the Nordic Cultural Fund’) and carried out between 2003–2005. It is aimed at describing the changes in mutual understanding in Scandinavia which have occurred since the last extensive study performed by Maurud over 30 years ago. This new investigation was felt to be necessary due to the considerable changes experienced by the Nordic countries since the 70ies. Internationalisation and globalisation have influenced the national economies, new media have amplified the choice of international TV and radio channels, and a large number of working migrants and refugees have contributed to a wide-scale change in the structure of the Scandinavian societies. Better connections such as the Öresund bridge between Denmark and Sweden or the Svinesund bridge connecting Sweden and Norway support the mobility between the Scandinavian countries and would therefore be expected to improve interscandinavian comprehension. On the other hand, changes in the school system clearly support the importance of English, and also the EU membership of Denmark, Finland and Sweden has diminished the importance of the interscandinavian cooperation. 1200 pupils from all Scandinavian countries were tested in Danish, Norwegian, Swedish and English. The investigation was enhanced by testing the parents of some of the pupils in order to be able to compare language comprehension between different generations. The test consisted of an extract from a TV show, a radio newscast and a newspaper article. The language understanding was tested by means of open questions referring to the contents and by asking for translations of certain words from the text. The results of the test more or less confirm the results from Maurud’s study. In some crucial points, however, the design of Delsing’s investigation differs considerably from that of Maurud, for example with respect to the number of participants, the geographical distribution of the test persons and also the inclusion of Greenland, the Faroe Islands, Finland and Iceland. A comparison with Maurud’s results, and also with the results from the adult control group, indicates that the level of intercomprehension in Scandinavia is decreasing, especially in Denmark and Sweden. Even though the results from the investigation may seem disenchanting with respect to the success of receptive bilingualism in Scandinavia, especially when compared to the understanding of English, the article also points out that comprehension tests cannot be seen to present a realistic picture of interscandinavian communication. In real life, context and non-linguistic behaviour positively influence understanding, and a crucial advantage of receptive multilingualism, viz. the ability to make use of the mother tongue, is not accounted for in the test.

15

16


Part 4: Determining the possibilities of reading comprehension in related languages Madeline Lutjeharms’ contribution refers to the special conditions of reading comprehension in related languages. It is focussed on the different processing levels that can be identified in the context of reading. When reading in a foreign language, lower levels of processing such as word recognition or syntactic analysis require attentional resources, in contrast to L1 reading where the processing of such form-based linguistic information generally functions automatically. Comprehension strategies comparable to those of hearers in receptive multilingual discourse can be observed (e.g. guessing, skipping parts of the text). Another problem is also reported in the context of oral communication between speakers of Danish and Swedish (cf. Teleman 1981: 105). In some cases the processing on the form level requires so much capacity that the reader is not able to notice the content. This article pays special attention to the processing of cognates. Cognates play an important role in receptive multilingual communication between related languages, and methods such as EuroCom (cf. Hufeisen and Marx in this volume) make use of cognates in order to establish a faster and more efficient access to related languages. In reading, in contrast to speaking, even non-relevant languages are activated. For the decoding of a German text by Dutch speakers not only the closely related L1, but also the genetically more remote English language has effects on word recognition.4 However, Lutjeharms was able to identify individual differences between learners in their ability to detect correspondences between related languages, and the application of the ‘seven sieves’ approach, a method of explicitly establishing conscious formal correspondences between those languages relevant for the different process levels, is a good way to improve this skill. The role of deceptive cognates (‘false friends’) is discussed in a controversial manner in the research on receptive multilingualism. On the one hand, psycholinguistic experiments suggest that cognates, as opposed to their non-cognate equivalents, show a common representation in the mental lexicon, facilitating automatic processing. On the other hand, it has been observed that deceptive cognates require conscious processing in order to prevent the activation of the (misleading) L1-meaning. Such effects must be considered when developing special methods for text-comprehension in related languages. Robert Möller presents his results from a project which simulates Dutch–German reading comprehension with the help of a computer programme. Dutch and German are closely related, but not spontaneously mutually intelligible languages. However, even though the linguistic distance is no larger than between e.g. Danish and Swedish, receptive multilingual communication is an almost unknown phenomenon between the speakers of the two languages (for a detailed explanation see Ribbert and ten Thije’s article in this volume). The analytical structures of Dutch can very often be deduced from similar German variants, whereas the German grammatical morphology often remains quite opaque for a speaker of Dutch. Due to these asymmetrical morpho-

Introduction

logical differences for a German reader of Dutch texts the threshold which has to be overcome is quite low, especially if learning aids required to deal with the lexical differences, but also with sound correspondences, are at hand. Due to this fact, one of the aims of the project is to develop such learning aids. The comparability of the recognition of words by a computer programme to the reading of a text by a human being is certainly limited, seen as human readers are able to make use of the context or even to make guesses. However, one advantage computers have over humans when objectively measuring language differences is the fact that they are less oblivious than humans. Furthermore, it is possible to reset their memory and thus perform a test under different conditions without running the risk of the results being influenced by individual differences. The programme NL-D-KOG, which contains a list of Dutch–German correspondences mainly based on the Old High German consonant shift, was used to compare the 5,000 most frequent Dutch words to their German counterparts. To determine the distance between the Dutch and German cognates, and thus the respective costs for identifying them, the Levenshtein algorithm was used. The result of the comparison was that 77% of the Dutch words were identified correctly. This suggests that, given an adequate set of correspondence rules, the majority of Dutch vocabulary should be accessible for a German reader. A comparison of the results with those of other empirical studies on Dutch–German mutual understanding leads to the conclusion that Dutch–German receptive multilingualism is actually a feasible option. Moreover, given different sets of rules, the programme can also be used to determine the distance between other languages and thus make predictions about the possibility of intercomprehension. The results from a test carried out on the reading comprehension of languages from the western branch of the Germanic language group are presented by Renée van Bezooijen and Charlotte Gooskens. The relationship between Dutch, Frisian and Afrikaans is comparable to the mainland Scandinavian situation. (West) Frisian, a language spoken in the Netherlands, was originally very closely related to English but has converged to Dutch throughout the course of time due to the strong influence of the national language. The opposite is true for Afrikaans, a language spoken by approx. 6 million people in South Africa. It originated from different Dutch dialects spoken by colonists in the Cape region in the seventeenth century, but developed into an Ausbau language in the nineteenth century. Gooskens and van Bezooijen analyse the understanding of written Frisian and Afrikaans by testing 20 native speakers of Dutch and correlate their results both with the linguistic distance between the languages involved and with the attitudes towards the speakers. The comprehension was tested by means of a cloze test based on newspaper articles. All participants were tested for both Afrikaans and Frisian. In addition to this, their attitudes towards the languages were assessed. The results show a far better understanding of Afrikaans than of Frisian, and even the attitudes towards South Africans turned out to be more positive than towards Frisians. A more detailed analysis of the results, however, did not reveal any significant correlation on an individual level. Linguistic distance was measured

17

18


on the basis of the number of cognates and non-cognates, the transparency of the lexical correspondences and the Levenshtein distance, a method used to objectively determine the similarity of words (a detailed description of this method is also found in Möller’s contribution to this volume). An important result which could be seen from those measurements is the fact that it is not the number of cognates — which is larger for Frisian and Dutch than for Afrikaans and Dutch — but rather the number of noncognates which is decisive for the degree of understanding. Even though the amount of Frisian/Dutch non-cognates is only slightly higher than the amount of Afrikaans/ Dutch non-cognates this may affect the comprehension considerably given the fact that they primarily involve content words, meaning that just one unintelligible word is enough to impede the understanding of a whole sentence. A second result, namely the higher Levenshtein distance between Frisian and Dutch compared to Afrikaans and Dutch, corresponds very well to Lutjeharms analysis of reading texts in foreign languages. It has to be assumed that the identification of cognates is less obvious and thus requires more effort for Dutch/Frisian than for Dutch/Afrikaans due to the higher Levensthein distance. Following this assumption, for a Dutch reader the processing of the content should be affected more by difficulties in identifying cognates in Frisian than in Afrikaans. On the basis of various models of L3 learning Britta Hufeisen and Nicole Marx argue that L2 acquisition differs qualitatively from the acquisition of further foreign languages. These differences are described in Hufeisen’s factor model. In addition to neurophysiological, learner external, emotional, cognitive and linguistic factors which influence L2 acquisition, the learner of a third language can also make use of previous experiences with learning techniques and strategies that can be variably successful from learner to learner. Such experiences, e.g. with different methods of learning vocabulary, should have a positive effect on the acquisition of a further language. But also the linguistic knowledge acquired when learning a language can be useful for the acquisition of further, especially related languages. The dominance of English as a first foreign language in the vast amount of European countries has the effect that other languages such as German, Spanish or French are typically acquired as L3s. This article presents different methods of L3 learning that make use of the strategic and linguistic knowledge already acquired by learners in the process of their L2 acquisition. In an experimental study carried out by Nicole Marx it was shown that this previous knowledge is not automatically activated in L3 learning, but that methods of sensitising the learners for similarities between the new language and languages already acquired clearly had a positive effect on the learning process. One such method of sensitisation is the EuroCom method which aims at optimising inference techniques in language learning. This is achieved by means of a comparison of languages on different linguistic levels, the so called ‘seven sieves’. This method was tested in an English language course at the Technical University of Darmstadt (Germany). The course was designed for students who had learned German as a foreign language but had no knowledge of English and felt the need to be able to read English texts for their studies or their

Introduction

later professions without being able to spend too much time on acquiring productive skills in a traditional language course. The positive results from the projects support the concept of receptive multilingualism. We would like to thank all authors who contributed to this volume for their efforts in promoting the research on receptive multilingualism. We are also deeply grateful to the Research Centre 538: Multilingualism for including this volume in its series Hamburg Studies on Multilingualism and to Conxita Lleó, Jürgen Meisel, Monika Rothweiler and especially Juliane House for their energetic support of this project. Very special thanks go to the numerous anonymous reviewers for their valuable advice and also to Nicholas Burke for his efforts to transform parts of the book into acceptable English. We hope that this book will encourage further research on receptive multilingualism.

Notes 1. DFG project Niederdeutsch und Skandinavien (‘Low German and Scandinavia’), 1990–1995, principal investigator Kurt Braunmüller. 2. Project Semikommunikation und rezeptive Mehrsprachigkeit im heutigen Skandinavien (‘Semicommunication and Receptive Multilingualism in contemporary Scandinavia’) at the DFG Research Centre 538: Multilingualism, 1999–2005, principal investigator Kurt Braunmüller. 3. A short survey of the results was presented by Zeevaert (2006). To test the success of the method the participants were divided up in two groups in the beginning of the course and were equipped with two different short texts taken from an Icelandic newspaper along with four open questions and four multiple choice questions. At the end of the course the test was repeated, but the texts were exchanged between the groups in order to avoid the participants already being familiar with the text. For one of the texts the percentage rate of correct answers increased from 15% to 50%, a quite impressive affirmation of the effectiveness of the EuroCom method. For the other text, however, hardly any increase in the amount of correct answers could be observed. From a discussion of the results with the respective students it became clear that in contrast to the first text, a report about scabies spread by ducks, a narrative structure facilitating the identification of the content was missing in the second text. Due to this fact, guessing strategies were not successful. 4. Similar effects were observed by Doetjes (in prep.) in tests on Danish–Swedish intercomprehension.

References Barðdal, J. , Jörgensen, N. , Larsen, G. and Martinussen, B. 1997. Nordiska. Våra språk förr och nu. Lund: Studentlitteratur. Bhatia, T. K. and Ritchie, W. C. 2004. Bilingualism in South Asia. In The Handbook of Bilingualism [Blackwell Handbooks in Linguistics], T. K. Bhatia and W. C. Ritchie (eds), 780–807. Malden: Blackwell.

19

20


Bührig, K. and Thije, J. D. ten (eds) 2006. Beyond Misunderstanding. Amsterdam: John Benja mins. Coseriu, E. 1988. Sprachkompetenz. Grundzüge der Theorie des Sprechens [UTB 1481]. Tübingen: Francke. Doetjes, G. in prep. Zur Rolle von Akkommodation bei der interskandinavischen Verständigung. Dissertation Hamburg University. Finkenstaedt, Th. and Schröder, K. 1990. Sprachenschranken statt Zollschranken? Grundlegung einer Fremdsprachenpolitik für das Europa von morgen [Materialien zur Bildungspolitik 11]. Essen: s.n. Golinski, B. 2007. Kommunikationsstrategien in interskandinavischen Diskursen [Philologia 95]. Hamburg: Dr. Kovač. Goossens, J. 1985. Was ist Deutsch - und wie verhält es sich zum Niederländischen? [Nachbarn 11]. Bonn: Kulturabteilung der Kgl. Niederländischen Botschaft. Ház, É. 2005. Deutsche und Niederländer. Untersuchungen zur Möglichkeit einer unmittelbaren Verständigung. Hamburg: Dr. Kovač. Herrlitz, W. 1997. Ist Befehl Befehl? Beobachtungen beim deutsch–niederländischen Korps in Münster. In Niederländer und Deutsche und die europäische Einigung, F. Wielenga (ed.), 66– 70. Bonn: Presse und Kulturabteilung der Kgl. Niederländischen Botschaft. Hufeisen, B. and Marx, N. in prep. EuroComGerm — Die sieben Siebe: Germanische Sprachen leichter lesen lernen [Editiones EuroCom]. Aachen: Shaker. Josephsson, O. 2006. Lågtyska och högtyska. Språkvård 2006(3). Kamwangamalu, N. 2004. Bi-/multilingualism in Southern Africa. In The Handbook of Bilingualism [Blackwell Handbooks in Linguistics], T. K. Bhatia and W. C. Ritchie (eds), 725–41. Malden: Blackwell. Karker, A. 1978. Det nordiske sprogfællesskab — historisk set. Språk i Norden 1978: 5–16. Klein, H. G. and Stegmann, T. D. 2000. EuroComRom – Die sieben Siebe: Romanische Sprachen sofort lesen können [Editiones EuroCom 1] (2nd ed., 1st ed. 1999). Aachen: Shaker. Kloss, H. 1929. Nebensprachen. Eine sprachpolitische Studie über die Beziehungen eng verwandter Sprachgemeinschaften. Wien: Wilhelm Braumüller. Koole, T. and Thije, J. D. ten 1994. The Construction of Intercultural Discourse. Team discussion of educational advisers. Amsterdam: Rodopi. Loos, E. 1997. Internationale Bedrijfscommunicatie. Reconstructief onderzoek naar het intertekstuele netwerk van Nederlandse en Duitse actoren in een bungelowpark. Utrecht. McCann, W. J. , Klein, H. G. and Stegmann, T. D. 2003. EuroComRom — The Seven Sieves. How to read all the Romance languages right away [Editiones EuroCom 5]. (2nd edn. , 1st edn. 2002). Aachen: Shaker. Melberg, H. 1952. Origin of the Scandinavian nations and languages. An introduction. In two parts. Book 1, Part 1 [Scandinavian and Celtic series I 1]. Halden: Author’s edition. Rehbein, J. 2006. The cultural apparatus: Thoughts on the relationship between language, culture, and society. In Beyond Misunderstanding, K. Bührig and J. D. ten Thije (eds), 43–96. Amsterdam: John Benjamins. Romaine, S. 1994. Language in Society. An introduction to sociolinguistics. Oxford: Oxford University Press. Teleman, U. 1981. Om förståelse i allmänhet och om förståelse via grannspråken i synnerhet. Referat och kommentar till diskussionen i sektionen Den lexikala-semantiska komponenten i den nordiska hörförståelsen. In Internordisk språkförståelse. Föredrag och diskussioner

Introduction

vid ett symposium på Rungstedgaard utanför Köpenhamn den 24–26 mars 1980, anordnat av Sekretariatet för nordiskt kulturellt samarbete vid Nordiska ministerrådet [Acta Universitatis Umensis 33]: C.-Chr. Elert (ed.), 102–20. Umeå: Umeå Universitetet. Teleman, U. 2002. Ära, rikedom och reda. Svenskt språkpolitik unter äldre nyare tid [Skrifter utgivna av Svenska språknämnden 85]. Stockholm: Norstedts Ordbok. Thije, J. D. ten 2003. Eine Pragmatik der Mehrsprachigkeit. Zur Analyse diskursiver Inter kulturen. In Die Kosten der Mehrsprachigkeit. Globalisierung und sprachliche Vielfalt; The Cost of Multilingualism. Globalisation and Linguistic Diversity, R. De Cillia, H. J. Krumm and R. Wodak (eds), 101–23. Wien: Verlag der Österreichischen Akademie der Wissen schaften. Westheide, H. 1997. Trügerische Nähe. Niederländisch-deutsche Beziehungen in Geschichte, Sprache und Kultur [Europa 2020 — Studien zur interdisziplinären Deutschland- und Europa forschung 16]. Münster: Lit. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verstän digung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač. Zeevaert, L. 2006. Isländisch sofort lesen können? Isländisch im Rahmen der germanischen Inter komprehension. Paper presented at the EuroCom-meeting in Como, 15–16 Sept. 2006.

21

Part 1

Historical development of receptive multilingualism

chapter 1

Receptive multilingualism in Northern Europe in the Middle Ages A description of a scenario* Kurt Braunmüller Universität Hamburg

This paper gives a survey of the linguistic situation in northern Europe in the late Middle Ages. It is based on three earlier research projects and summarises some of their results, especially as far as language choice and domains are concerned. The focus lies, however, on the development, role and function of receptive as opposed to productive bi-/multilingualism and language standardisation. On the basis of Peter Trudgill’s terminology of language contact, new terms for differentiating forms of linguistic convergence and divergence are proposed. Three case studies highlight the role receptive multilingualism plays for adult German learners of (eastern) Scandinavian languages, focusing on the form of the definite article, the use of the periphrastic genitive and the overlooking of V1-patterns in declarative sentences. Keywords: late Middle Ages, language contact, receptive multilingualism, Scandinavian languages

1. Issues and historical development of receptive multilingualism 1.1 A short outline of some issues of receptive multilingualism Receptive multilingualism definitely plays no salient role when discussing research on bi- or multilingualism in general. In reality it is rather the opposite: receptive multilingualism is given only marginal attention, if taken into consideration at all. The interest and focus undoubtedly lies on the various forms of ‘active’ (productive) bilingualism, their restrictions, problems and perspectives as well as on the interaction between the command of the respective languages by one and the same person and the alternate use in different contexts (known as ‘code switching’). When dealing informally with issues of bilingualism, receptive multilingualism is likely to be classified as a kind of residual form of a formerly active command of a specific language, e.g. due to the elapse of time, the lack of practice or the dominance of another language in almost all domains, often combined with the speaker changing his/her status from that of a formerly fluent native speaker to a more or less overt

26

Kurt Braunmüller

semi‑speaker (cf. e.g. Dorian 1997). Therefore, receptive multilingualism is not classified as a manifestation of bilingualism in its own right. There are, however, a few exceptions, e.g. when considering the linguistic situation in Scandinavia today. The form of interlingual communication exercised in Scandinavia, which has misleadingly been called “semicommunication” (by Haugen 1966), is considered an important political and cultural goal within inter-Nordic cooperation. In this framework, receptive multilingualism is an acknowledged form of multilingualism, based on the genetically close relationship between the three mainland Scandinavian languages Danish, Swedish and Norwegian and represents a form of interdialectal communication between these three Ausbau-languages (cf. Kloss 1978: 25). This form of direct, transnational communication is mainly applied when Scandinavians from different countries meet and although it functions quite well it is not without problems.1 Some of these problems are due to linguistic divergences emanating from language change and the dialectal parting of a formerly more coherent linguistic area. Other problems can be traced back to difficulties in trying to incorporate very divergent and therefore opaque dialects, to the lack of linguistic flexibility or to considering one’s own variety superior to those of the neighbouring languages. But, when there is a will there is a way! One of the main characteristics of (genuine2) receptive multilingualism is that it is generally applied very consciously. People who decide to make use of receptive multilingualism feel that there is, at least in some situations or for certain purposes, really no need for acquiring an active command of the respective language; an approximate or even a rudimentary understanding of this language is regarded as being sufficient. Therefore, one may intend to acquire only some reading skills or learn, at least, a few frequently occurring words and phrases of the target foreign language, in order to roughly understand what has been written in e.g. a certain publication or what a book is about. Scientists or librarians may feel such a need for at least some basic knowledge of languages they do not speak. Moreover, tourists often try to retrieve some vital information, e.g. from the menu, from traffic or other signs they are confronted with and which seem to be relevant for their visit to a country in which they do not speak or hardly understand the language. In order to grasp the approximate content of a text, written e.g. in a Romance language, one has the possibility of successfully making use of Latin word stems, if available, and/or try to understand the contents of a (shorter) text by means of internationally used words or via similarities to other more familiar Romance languages.3 All people using receptive multilingualism are fully aware of the fact that this form of grasping the essential meaning of a message encoded in an (almost) unknown language cannot be more than a temporary solution or an ad hoc compromise, but it is a method sufficient for precisely this purpose. There are always other means available if more, or more precise, information is required. In the cases mentioned above, one may look for an abstract in a lingua franca or a summary in a language you are more familiar with, or try to find a person who is

Receptive multilingualism in the Middle Ages

able to translate/interpret the respective text, maybe via a lingua franca. This process, however, does not exclude the acquisition of a deeper knowledge of the target language which, sooner or later, leads to a better intercomprehension — but not necessarily to a full and active command of the respective language. Learning another language actively is time-consuming and often unnecessary, especially when dealing with genetically closely related languages.4

1.2 Receptive multilingualism and nationalism In former times receptive multilingualism was not so unusual, especially not in faceto-face trading communication, e.g. on fairs in more distant parts of Europe. Recently, Smith (2005: 26) pointed out when analysing speaking and writing from 500 to 1000 ad, that “widespread, full bilingualism” was restricted: “Active competence in one but general passive knowledge [viz. receptive multilingualism; K.B.] of the other is one possibility; equally probable is partial ability in the second language, for example in restricted spheres of life such as mercantile activity or estate management.” A glance at modern history will help to understand why receptive multilingualism has lost its status as a form of multilingualism in its own right. The main reason for disfavouring receptive multilingualism in the last two centuries is the rise of nationalism. Since the early nineteenth century, nation, language and identity have become closely intertwined. A direct consequence of this development was linguistic standardisation, meaning that one common, written and often also spoken ‘national’ language to be used by all fellow-citizens was created in order to exhibit an overt, i.e. a visible and audible, indicator for national unity and common (ethnic) descent. The consequences of this ideology are obvious and quite simple: if you are not able to speak the indigenous and/or the corresponding national language as, or at least like, a native citizen, you cannot be regarded as a full member of this country/state. In other words, a mere understanding of vernacular varieties or even a non-‘perfect’ command of the national language is no longer considered to be sufficient even if it worked very well in practice before. The perfect command of the (dominant) vernacular language has thus taken over a shibboleth function: it is no longer communication or mutual understanding in every-day life that counts but far more the use of the appropriate linguistic code which, at the same time, serves as an identifying feature for all members of that nation. Or, to put it in another way, one can only become a citizen of a nation if one is capable of speaking its (or at least the majority’s) language as ‘perfect’ as possible, i.e. as a native (!) speaker. ‘Imperfect’ bi- or multilingualism in general, which in almost all cases can be identified by the speaker’s foreign accent, became stigmatised and receptive multilingualism was neglected because everybody now required an active command of the national i.e. the majority’s language. Nationalism thus turned the focus away from communication and managing every-day life by means of a multiple linguistic, often diglossic competence, in the direction of an idiomatically ‘perfect’ use of one politically favoured language, viz. the national language.

27

28

Kurt Braunmüller

The linguistic price of this development is very high: if one relies more or less on the standard (or national) language, one loses familiarity with local dialects and other varieties within this society, not to speak of the deteriorating ability to understand divergent but genetically closely related neighbouring languages. What is more, one expects nowadays a focused language with little or no variation, clear-cut grammatical rules and a clearly defined vocabulary and consequently disfavours all characteristics of a diffused language (in the sense of LePage and Tabouret-Keller 1985), including e.g. ad hoc word formations, incomplete or elliptic, viz. context-dependent, sentence constructions. Unified standards, launched either by chancelleries or the first book printers in the late Middle Ages, were originally created as a means to enable textual understanding and the distribution of documents and printed matters, respectively. But at the same time, they acquired a paradigmatic or leading function, especially for official documents and religious texts from the fifteenth and sixteenth century onward. Therefore, it was only inevitable consequence for (European) nationalism to proceed in this way in order to create an optimal linguistic homogeneity within a certain territory: minority languages were soon regarded as undesired languages, not only disturbing ethnic homogeneity but also the linguistic and national identity. Speakers of minority languages were therefore looked upon suspiciously seen as the majority of the population neither understood these languages nor were they interested in learning them. Moreover, the majority population felt that the members of (indigenous, not to speak of immigrant) minorities were not really interested in becoming an integral part of that nation, its goals, norms and customs. Statements such as ‘I am a Dane and my mother tongue is German’ nowadays sound, at least, contradictory, if not absurd (cf. Menke 1996). This development has, however, not yet come to an end. English as a lingua franca has bit by bit taken over the function of a supra-national standard language — leading to the same negative effects both for linguistic flexibility as well as for receptive multilingualism.5 If you find it too hard to listen and adapt to genetically related languages, you will most likely consider English as an appropriate solution of your communication problem. This strategy may certainly be effective for simpler tasks, such as ordering a meal (as a tourist) or retrieving basic information in general, but not for more complicated tasks. Furthermore, one might have the impression that the use of a lingua franca is a fair compromise seen as all interlocutors have to use the same language, i.e. no one has the privilege of speaking his/her mother tongue. But in receptive multilingualism between genetically closely related languages, the use of the native language changes with every turn in discourse. What remains is a kind of ‘principle of least effort’, often, however, accompanied by scepticism that you could have presented your message in a much better manner when making use of a language you are really familiar with. The situation in Scandinavia differs also in this respect: even in the era of nationalism, Scandinavia in toto was in the focus of patriotism (called ‘Scandinavism’) and not


only the respective countries (Denmark, Norway,6 Sweden and even Finland) themselves. In a way, it therefore seems rather natural that receptive multilingualism (or ‘semicommunication’) and nationalism did not directly interfere with each other but rather supplemented each other. But today, the impact of English as the dominant international lingua franca cannot be overlooked: many younger Scandinavians do not even try to use receptive multilingualism but prefer to converse in English from the very beginning of any interScandinavian contact. Globalisation seems to be overriding Scandinavism in the long run. One result of these normative processes, mentioned above, was that receptive multilingualism became obsolete seen as due to it favouring direct mutual understanding beyond any kind of (national or ethnic) borders sometimes even at any price, relying on the possibilities which are inherent in any linguistic diasystem. The knowledge of a diasystem implies much more than just the mastering of the grammatical rules and the vocabulary of a (standard) language: it enables an understanding of obsolete forms and sentence constructions occurring in older texts, neighbouring dialects, wide-spread sociolects and many other linguistic varieties such as jargons, languages for specific purposes and, last but not least, genetically closely related languages, which, strictly speaking, do not represent much more than dialects that have become national languages due to language policy and language cultivation.

1.3 Outline of the further discussion The main aim of this paper is to present an overview of all forms of multilingualism and language contact phenomena we have come across in Northern Europe in the late Middle Ages and in early Modern Times. I will, at first, discuss the main divergences between receptive multilingualism and (productive) multilingualism in [2.1.], followed by an overview of the functions and domains of the languages to be found in Northern Europe at that time in [2.2.], together with an (abridged) discussion on the historical development in the north European lowlands and the Baltic in general, with special emphasis being placed on language use and acquisition related to any form of multilingualism in [3.]. The paper concludes with a summary of three case studies in [4.], which reveal the impact of Low German on mainland Scandinavian languages and dialects due to language contact, mainly during the era of the Hanseatic League. These case studies relate to the morphological form of the definite article [4.1.], an ‘imported’ periphrastic genitive construction, called garpe-genitiv [4.2.], and changes in the preference in word order [4.3.]. The main triggering factor in all these instances was receptive multilingualism, accompanied by hypotheses on the structure of the target language, followed by (imperfect) L2 learning strategies and finally disseminated by a group of adults with high social prestige and influence.

29

30

Kurt Braunmüller

2. Forms of multilingualism in Northern Europe during the Middle Ages 2.1 Receptive vs. productive multilingualism Table 1 shows some of the most relevant features which distinguish receptive multilingualism from (productive) bi- or multilingualism in the Late Middle Ages and in Early Modern Times. Most of these features can also be applied to our times — if, of course, receptive multilingualism is practised at all. Issues related to the acquisition of a reading competence have not explicitly been taken into consideration due to the very restricted domains of literacy and writing in general in earlier times. The scenario for receptive multilingualism in the late Middle Ages therefore seems quite clear:7 in almost all cases the domains for this kind of asymmetric communication have been restricted to face-to-face interaction in clearly defined settings, i.e. receptive multilingualism is normally practised in purpose-oriented situations, preTable 1. Features distinguishing receptive multilingualism from (productive) bi- or multilingualism in the Late Middle Ages and in Early Modern Times Receptive Multilingualism

(Productive) Bi-/Multilingualism

Predominantly for informal communication

Both for formal and informal communication

Purpose-oriented, no (productive) acquisition of the target language is intended

Function-oriented (with reference to persons, topics or domains)

Face-to-face communication, especially in diglossic trading situations and other business contacts* may occur

No restrictions in principal but a distribution of the languages involved

Establishing communication at any price, frequent ad hoc-accommodations, no rules

Person-, topic- or domain-related language use (including code-switching, if appropriate)

Highly context- and addressee-dependent

Low mandatory context or addressee dependence

Emphasis on communication exchange and efficiency in interaction

All linguistic functions are available (if not restricted due to domains or functions)

Informal but pragmatically controlled learning by listening and speaking where no grammatical norms have to be observed

All kinds of natural-language acquisition and L2 learing, especially for the acquisition of a lingua franca (e.g. Latin or Low German)

Dominance of pragmatics and the situational context

Dominance of linguistic awareness with respect to domains, styles, norms and grammatical correctness

Includes the possibility to become a (fluent) speaker of the target language; may gradually include occasional (lexical) code switches.

Language use may be restricted to functional distribution; therefore no necessity for a full linguistic competence in all languages spoken.

*However, the origin of a pidgin differs from receptive multilingualism since the languages involved are not mutually understandable.


dominantly in trading or negotiations, in which the terms of trade and the principles of commerce have widely been acknowledged and the interaction patterns involved follow common and accepted rules. Therefore, the linguistic form, as far as grammatical and stylistic correctness or the choice of the appropriate register is concerned, only played a marginal role; with the purpose of the interaction or the facts conveyed by a message clearly prevailing. Linguistic fluency or a comprehensive understanding of the addressee’ s language were neither necessary nor expected at all. The leading norms of this kind of interaction were based on pragmatic adequacy. Even (emblematic) code switching could occur, though predominantly in long-term contacts, seen as sporadic quotes or loans from the addressee’s language not only demonstrate sufficient mutual understanding but also express a certain empathy and understanding for the interlocutor, his situation and linguistic background. Furthermore, such code switches explicitly underline the speaker’s willingness to accommodate and accept the addressee’s linguistic variety without any restriction.8 It is, however, important that the addressee (a) consents to converse in this asymmetric way following some sort of ‘let-it-pass strategy’, but (b) that he is really willing and able to practise receptive multilingualism, without (c) trying to use a lingua franca as a more appropriate means to achieve understanding on the basis of a third/neutral language, (d) nor should he prefer conversing only in his native language, due to him considering his own language to be superior to the addressee’s dialect/language or to other languages in general. Therefore, receptive multilingualism as mutually acknowledged linguistic behaviour works best when carried out between interlocutors of the same social status with at least some essential contextual conditions, as presented in table 1 being fulfilled, such as face-to-face interaction between speakers of genetically closely related languages, emphasis on informal communication exchange or predominantly purpose-oriented speech situations. In any case, you can never take for granted that receptive multilingualism will work without problems or even function at all before you have agreed on this form of communication with the addressee — unless you know in advance that receptive multilingualism is the generally accepted form of communication, as is still the case in mainland Scandinavia today. As far as the late Middle Ages and early Modern Times are concerned, one has to be aware of the fact that receptive multilingualism represented only one aspect of multilingualism to be taken into account. Functional diglossia was the default linguistic principle at that time: almost all domains were tied to certain languages, and no one really expected you to master all domains in one and the same language or you to be capable of expressing everything in any language you are familiar with (cf. e.g. Jahr 1995 or Nesse 2003). This also applied for the use of a lingua franca which was domain-related in a similar way when used in conversations with people abroad or in unfamiliar situations

31

32

Kurt Braunmüller

(see 2.2.). Therefore, nobody really expected a comprehensive or, as we nowadays prefer to say, ‘perfect’ command of such a language: if you used e.g. Latin with word order patterns descending from the syntax of one’s own vernacular, nobody would have refused such a text barely due to it not living up to the classical Latin antiquity and its stylistic norms. Since Latin was both a lingua franca and a language without any native speakers, it was open for all kinds of language contact, especially with respect to word order and semantic reinterpretation. Even dialect mixing was widely accepted during the Middle Ages and in early Modern Times. Low German (as a native language) had always been open to all kinds of (lexical) loans from the southern German speaking territories. In one of the earliest documents written in German, the famous Hildebrandslied, dialectal variants from two remote, divergent German speaking territories are found side by side in one and the same manuscript: the archaic Old High German poem was imbedded in a short, Low German, narrative frame, without any comment or justification. Moreover, intrasentential code mixing was also tolerated and could apparently be used without any notification (for more details see Braunmüller 2000a: 278ff.). Since language standardisation was not yet on the agenda in the Middle Ages, one was, basically, free to use varying dialectal forms side by side or to mix up different dialects, without anybody really blaming you for doing so, seen as no committed guidelines existed for (inter)dialectal communication. Most well-known writers tend to show (considerable) individual variation in their text production but they do not mix their dialects very often with other variants of the same linguistic diasystem, which was simply due to the fact that their own dialect was the only language they had an active command of. Receptive multilingualism also played a supplementary role: its main purpose was to learn to understand your neighbours, both those in the immediate vicinity and those living further away. The range of application for receptive multilingualism was thus always the same, namely to recognise the corresponding sounds, inflexional endings and terms in other, related dialects and to establish the appropriate equivalency rules with reference to one’s own dialect. Therefore, receptive multilingualism was very common in every-day life in the Middle Ages and beyond seen as no one spoke (or wrote) in exactly the same way nor used the same vocabulary. So, why should one strive for standardisation in a situation in which everybody was faced with linguistic variation? If an institution such as the Hanseatic chancellery in Lübeck aims to codify official texts and documents, written exclusively in Low German since the end of the fourteenth century, it is not primarily intended to create a standard language or to abolish dialectal variation far more to brand these documents as authentic (viz. written in Lübeck) and authorised by the Hansa. It would thus be more appropriate to consider this development as an early attempt to establish some kind of linguistic brand rather than a first step towards a general standardisation of Low German as a written language, in order to reduce dialectal variation. The


Hansa was not, unlike the former European Economic Community (EEC), a precursor of a political union, but remained a trading union with vacillating forms of networking.9 By the way, standardisation is actually no proper manner of minimising the burden of learning (genetically) related linguistic varieties, rather the opposite holds true: on the one hand, decoding a message will, admittedly, become much easier for the hearer/reader due to the reduction of variation; but on the other hand, the speaker/writer has to learn a large amount of new grammatical forms and equivalent words in order to make use of this trans-regional form of language. On top of that, getting used to extensive standardisation sooner or later considerably reduces linguistic flexibility and practice in understanding neighbouring dialectal variants. That is why Norwegians, accustomed to interdialectal communication in every-day life, always achieve the best marks in inter-Scandinavian comprehension tests, whereas Danes and Swedes, both acquainted to more or less standardised forms of language, perform rather badly (cf. e.g. Delsing and Åkesson 2005: 136ff.). Since the roles of production and reception change quite frequently in communication, there is actually no real winner in such a process, as far as linguistic economy and effectiveness is concerned. In any case, standardisation makes sense as far as foreign-language learning is concerned: one can rely on one, clearly defined linguistic norm, which has to be learned, but can neglect the rest of the target language’s underlying diasystem.

2.2 Functions and domains of languages in the Middle Ages and early Modern Times The linguistic situation in the (late) Middle Ages was completely different from the situation and practice we regard as normal or self-evident nowadays: As we have stated in the preceding sections, domain-, situation- and person-related multilingualism, including receptive multilingualism, was the norm, definitely not monolingualism or a dominance of national languages. Even supranational languages such as Low German (in Northern Europe and around the Baltic Sea) or global languages like Latin (but also Greek, in the Balkans and the countries surrounding the eastern part of the Mediterranean) principally showed the same functional distribution but were restricted to fewer domains. Beside their function as a lingua franca in trading situations or even in long-distance travelling (cf. Burke 1989), they were also used as languages for specific purposes (LSP). Thus, the domains of Latin were primarily the Bible, the church and Christianity in general, furthermore sciences and arts, and finally every form of advanced schooling and higher education. Referring to Smith (2005: 50), one may add “that it was in the early Middle Ages that Latin first emerged as an international mandarin language throughout most of Europe.” Furthermore, she makes clear that the culture of writing in vernaculars has its origin

33

34

Kurt Braunmüller

in Latin “as a form of cultural osmosis around the periphery of the Roman world” (p. 32) which coincides with our hypothesis on early Germanic bilingualism relating to the creation of the first runic script (see Braunmüller 2004b and Beuerle and Braunmüller 2004). Low German and, as a consequence of the Reformation, High German were able to take over some of these functions, especially as far as trading, legal texts (such as municipal laws or bylaws) or Bible texts are concerned. However, at the beginning of this development we find receptive multilingualism in face-to-face situations. This kind of trans-regional/national communication revealed a considerable amount of linguistic variation and thus required familiarity with non-standardised forms of communication. In other words, the expansion of the scope of one’s own diasystem, accompanied by receptive multilingualism as the default way of unmediated communication, made two things possible: (A) a considerable upgrading of one’s own vernacular beyond its original territory (in our case, primarily for German speakers, beyond the NorthGerman lowlands towards Scandinavia and some other parts of the Baltic area), and (B) the intense impact of (Middle) Low German on the mainland Scandinavian languages and their dialects. Therefore, language change due to language (or rather dialect) contact based on both close genetic relationship and widely practised receptive multilingualism, especially at the beginning of these contacts, not only paved the way for intensive lexical borrowings, including the adoption of new word formation patterns (cf. e.g. Diercks 1993), and minor grammatical changes but also for the incorporation of certain German word order structures into the mainland Scandinavian languages (cf. the most recent survey on this topic in Braunmüller 2004a). Table 2 illustrates the scenario of the languages found in mainland Scandinavia during the Middle Ages and in the sixteenth century. The most interesting aspect in this context is certainly the far-reaching changes that the Scandinavian vernaculars and dialects underwent as a consequence of the extensive language contact with Low, and later, with High German. However, it is not easy to decide which processes actually occurred. Suitable candidates for characterising these developments seem to be: (a) creoloids, (b) koinés or (c) some sort of interlanguage or interdialect (cf. Trudgill 2000: 79ff.). Since mutual understanding and receptive multilingualism between Germans and Scandinavians was possible one could argue to classify these developments as cases of dialect contact, but as early as the Middle Ages there was no doubt that Danish and Swedish — written forms of Norwegian gradually disappeared due to the political union with Denmark beginning in 1380 — were (national) languages and not some kind of dialect. Furthermore, there are no arguments in favour of classifying these languages as ‘jargons’ since they neither originated from double-source pidgins nor from double-source creoloids seen as they always remained what they were: Scandinavian languages in their own right — though with a vast amount of lexical loans and some grammatical extensions and restructurings.


Table 2. Scenario of the languages found in mainland Scandinavia during the Middle Ages and in the sixteenth century Languages and their use in mainland Scandinavia during the Middle Ages The pre-Hanseatic era

The early Hanseatic era and the late Middle Ages

The late Hanseatic era and the 16th century

Low German only as a vernacular (spoken in the Northern Lowlands of Europe)

Low German as a lingua franca (predominantly used as a trading language; RM was common in trans-regional communication)

Low and High German as prestigious, trans-regional languages and as LSP (RM decreasing, L2- and 2L1bilingualism prevailing)

Latin as LSP and a lingua franca



vernaculars and dialects (with some influences from Latin due to Christianity)

vernaculars and dialects (with loans from Low German)

restructured vernaculars (due to heavy impact from German)

minority languages (of various origins)

minority languages (unchanged due to diglossia)

minority languages (unchanged due to diglossia)

When analysing the impact of Low German on mainland Scandinavian languages,10 one may observe admixture, of course, and cases of simplification, but no reduction. According to Trudgill’s contact typology (2000: 82), creoloidisation occurs in language contact situations, whereas koinéisation is found in dialect contact situations. Furthermore, both categories show different degrees of admixture and simplification (with less foreign impact in koinéisation). But how can we measure the extent of that foreign impact, also in relation to the size of the respective population? Would the terms ‘interlanguage’ and ‘interdialect’, respectively, better fit the phenomena to be observed? The reason for creating an interlanguage is primarily to establish communication, whereas the emergence of an interdialect is the result of accommodation (p. 83). In any case, German–Scandinavian contact doesn’t seem to match any form of sociolinguistic classification we are aware of. However, the problem of not being able to distinguish sharply between a language (contact) on the one hand and a dialect (contact) on the other, is by no means a new one. Heinz Kloss (1978: 25ff.) therefore distinguished between Abstand- and Ausbaulanguages to terminologically dispose of the dilemma that two closely related dialectal variants within one and the same diasystem received the status of independent languages: looked at from a linguistic point of view the respective languages behave like dialects but have received a higher status due to political and/or social decisions. The latter term best fits the situation we are faced with: (Middle) Low German can be regarded as an Ausbau-language in relation to south and east German dialects — although neo-grammarians would not agree for (mainly) phonological reasons. However, it evidently has more lexical and grammatical features in common with other German dialects than e.g. with Frisian or Middle English. Danish and Swedish

35

36

Kurt Braunmüller

can also be regarded as Ausbau-languages due to their historical development; otherwise one would rather speak of ‘regiolects’ representing the two main branches of east Scandinavian. I would like to suggest the term unilateral convergeoid for a one-sided, genetically closely related convergence language, such as Danish or Swedish, and the term unilateral convergeolect for the same contact phenomenon only on a level of dialects (cf. the situation in Norway). “Converge-” is directly associated with various forms of accommodation and therefore fits all contact phenomena between mutually intelligible variants. I use the term “unilateral” because the dominant language, in this case (Low) German, has not or hardly been influenced by this development. The complementary term would be divergeoid. It could be used for characterising distance-keeping (or purist) languages such as Modern Icelandic in which the native speakers (or at least an influential group of them) for ideological reasons try to keep all foreign influences away from their mother tongue. Finnish, when considering the relationship to Estonian, could be regarded as an appropriate candidate for a divergeoid as well. The profiles of such languages are characterised by them creating or, at least, maintaining distance to all neighbouring and therefore potential contact languages. Moreover, when considering the linguistic distance between Low German and all other German dialects, you could classify Middle Low German as a divergeolect. (Modern Low German has, however, changed its status. It has changed into a convergeolect or almost a dialect of High German due to the prestigious role of this variety of German as a national/standard language.) The linguistic basis for this classification is, however, completely different from that applied by the neo-grammarians of the nineteenth century seen as it is not founded on a (predominantly) phonological split or some other bifurcating grammatical changes but has its origin in the social status of that language. The other languages located in Northern Europe do not require any comments beyond their characterisation in the table above. We will, however, come back to the functional distribution of these languages in the section 3, when comparing the Scandinavian situation with the situation in Britain during the late Middle Ages.

3. The historical situation in Northern Europe in relation to the forms of multilingualism The historical situation and the development of the various forms of multilingualism in the lowlands of northern Europe, in the Baltic and in mainland Scandinavia can be outlined as follows: A. Due to the close genetic relationship between Low German and the Scandinavian languages/dialects an unmediated, basic mutual understanding in face-to-face communication was possible. Merchants (of the Hanseatic League) speaking Low German


realised that they were not only able to understand just remote dialectal variants of Low and High German but also the Scandinavian languages and some of their dialects to a certain extent. This form of understanding was, however, heavily dependent on trading situations in which the contextual factors such as the terms of trade and the outlines of (international) business communication largely had to be the same. The historical context at that time was also optimal due to the north German merchants succeeding in gradually taking over the traditional Visby-trade across the Baltic Sea by the end of the twelfth century. Furthermore, intensive trading contacts existed between the German, Frisian and Dutch lowlands on the one hand and Jutland on the other from the very early Middle Ages onwards. On top of that, Low German and, to a minor extent, also Dutch as the most western variant of the Low German dialectal continuum, were highly respected by the upper classes all over northern Europe which led to intensive loans from Low German taking place. Even deep-rooted Scandinavian terms were replaced by loans from this prestigious ‘global’ language.11 B. Consequently, the next step was the use of receptive multilingualism as a sufficient means in specific face-to-face trading situations in which is was not necessary to actively learn the Scandinavian customer’s/addressee’s language. There was, however, not much time to become acquainted with any details of the Scandinavian languages seen as merchants from abroad were not allowed to stay over a longer period of time in the respective host countries. A stay over a period of a few weeks was normal, several months was the maximum, and that only in summer. The privilege to stay also during winter was implemented much later (cf. Brattegard 1945: 15ff. or Nesse 2002: 85f.). At the beginning, the merchants themselves travelled around, predominantly from one (coastal) trading place or important town to another, directly offering their goods and products. In later times, the vast amount of trading activities were carried out by younger representatives of the merchants. The merchants themselves preferred to stay at home directing their business via long distance communication by means of letters, messengers or trading representatives. C. The privilege of the Hanseatic merchants being allowed to settle in the Scan dinavian countries led to two diverging linguistic developments: either their mercantile agents were strictly separated from the local inhabitants of these trading places/towns like in Bergen/Norway (cf. also Nesse 2003) or Nowgorod/Russia or they settled in more or less enclosed parts of towns (cf. e.g. in Tønsberg/Norway, Køge/Denmark, or in Kalmar or Stockholm/Sweden). The first scenario evidently did not encourage (productive) multilingualism but, at most, the acquisition of a rather restricted L2-competence of the local linguistic variety (for more details on the situation in Bergen see Jahr 1995: 14ff. or Nesse 2002: 75ff., for the situation in Russia/Nowgorod see Gernentz 1988). The second scenario, however, in one way or another led to natural 2L1-bilingualism and to vivid communication and social exchange between the merchants and the indigenous population. D. In some places, especially around the Baltic Sea, High German — also referred

37

38

Kurt Braunmüller

to as Martin Luther’s and the Reformation’s language — as a kind of dialectal variant or extension of the Low German diasystem took over the function of a lingua franca in most domains. It was, however, only used in a supplementary manner. However, the quality of the German–Scandinavian language contact, based on interdialectal similarity, did not change (see Braunmüller 2000b: 10ff.). The final step in the history of the Low German–Scandinavian contacts was either the closing down of the Hanseatic offices and storehouses or the integration of the immigrant, (Low) German speaking population into the local societies. In both cases, receptive multilingualism came to an end sooner or later for different reasons (retreat vs. assimilation), or continued as sequential L2-bilingualism in the upper classes. High German, later followed by French, remained the leading foreign language and lingua franca for many Scandinavians up to the nineteenth century. It remained the easiest linguistic key for them to gain admission to European culture and technology, to trade and to contacts with the southern parts of the continent in general. If you compare the situation in the Baltic to that in late medieval London (for linguistic data see Wright 1996 and for a more detailed analysis Braunmüller 2000c), you will observe many analogies, as far as the simultaneous use of various languages and their distribution is concerned, although the historical preconditions and local circumstances differ significantly: there is, however, no basis supporting the assumption that receptive multilingualism played any role in communication in the Thames and harbour area of London at that time, which is not surprising seen as English, AngloNorman French and Latin cannot necessarily be regarded as mutually intelligible languages. The most striking observation made is related to the fact that there is the intensive and frequent macaronic use of these three languages, although the switching between the languages is far from random (cf. Wright 1997: 347). One of the most salient features of that kind of (macaronic) multilingualism is the interlinguistic use of abbreviations and suspensions. A small, p-like letter p could be used as a prefix and an apostrophe-like sign 9 as a suffix, representing the morphemes par, per, pre, pur and -re, -er, -arius respectively (see Wright 1996: 9). Furthermore, if you find a word like carpent 9 in such a text from late medieval London you first have to make sure which is the unmarked or matrix language: if it is predominantly Latin, this expression obviously means carpentarius, whereas if it is English is has to be read as carpenter. This observation reveals two general characteristic strategies of multilingualism: 1. Decoding primarily relies on the analysis of the lexical morphemes. Derivational (or inflectional) morphemes play only a marginal or supplementary role. 2. Wherever it is possible, diverging grammatical structures are avoided, in syntax,12 morphology and word formation. There is no doubt that bilinguals prefer parallel grammatical structures, especially with respect to word order (cf. e.g. Braunmüller 2001 or Nichols 1992 and Nettle 1999: 137f. for a global perspective of this observation).


Furthermore, the business records presented in Wright (1996) reveal that Latin, also as an LSP, was under permanent pressure from the two indigenous languages, the language of the upper-class (Anglo-Norman) and the colloquial language used by the rest of the population (English), the result being that Latin would be rebuilt as an analytic language. It became more and more similar to the other two vernaculars with respect to word order and (analytic) morphological structure, which can also be observed when studying Latin in its other function as a lingua franca. The motivation preceding this typological change, was — according to strategy (2) — to reduce the burden of having to manage several linguistic codes simultaneously. In Scandinavia, a similar development could be observed. Low German and later also High German word order patterns were adopted as models due to their prestigious roles in international communication and, especially High German, for its shining example of idiomatic translations of Bible texts in the sixteenth century. Yet at the same time written High German itself became more and more dependent on Latin stylistic norms, as far as OV word order in dependent clauses or the extensive use of hypotactic sentence structures were concerned. (For a recent, detailed discussion of the impact of Latin on German word order see Chirita 2003.) On top of that, Latin, Low German and a Scandinavian language (here: Swedish) may have been used in the same context and for the same purpose but with differences relating to code switching. Therefore, it may occur that Latin words or phrases are found embedded in e.g. letters with Swedish as matrix language, whereas the opposite has not (or hardly) been observed with a prestigious language such as Low German functioning as matrix language. These and other observations presented in Tiisala (2004) clearly show that multilingualism, also in writing, was default in late medieval Scandinavia but that not all languages had the same status and prestige (Tiisala 2004: 196: “The Hanse never apologises for writing in German”). Or, to put it in another way, languages like Latin or (Low) German were never dependent on loans and formulas from other languages. Moreover, they supplied the correct form for certain types of texts, e.g. letters, which contained a salutatio, benevolentiae captiatio, narratio, petitio and finally a conclusio. If you preferred to use your own language you had to obey these rules, even if the precise terminology was still missing or under elaboration.

4. Three examples for the role of receptive multilingualism in L2-language learning In the following section I will illustrate the function of receptive multilingualism in second-language acquisition, primarily by adults, with focus placed on morpho-syntax and word order. If the target language is relatively similar to one’s own variety one tries to find the easiest way to learn this language. Receptive multilingualism can be

39

40

Kurt Braunmüller

regarded as a promising starting point both for short-cuts in language use and simplified L2 language learning.

4.1 The morphological form of the definite article Both Low German and the Scandinavian languages used pre-positioned definite articles. In Middle Low German this article always had a d- in the onset (cf. de/di Man(n) ‘the man’ or dü/de/di Frū ‘the woman’), whereas in the medieval Scandinavian languages two phonologically divergent forms occur: (a) hinn/hin/hit [nom.sing.masc./ fem./neutr.] as in (h)inn gamli maþr ‘the (old) man’ and (b) sá/sú/þat or, in the later Middle Ages due to analogy, þann/þan/þat [nom.sing.mask./fem./neutr.; also occurring with the vowels æ or e, as in þæn(n) or þen(n)] as in þa/æ/enn gamli maþr ‘the (old) man’ in informal and oral communication.13 Without preceding attributes in a noun phrase the h-article always occurs post-positional. It is cliticized to the noun and the h is deleted at the juncture. The post-positioned, cliticized definite article is inflected in accordance with the noun: maþrinn ‘the man’ [nom.sing.masc.], mansins ‘the man’s [gen.sing.masc.], maninum ‘to the man’ [dat.sing.masc.] and so on. Futhermore, it can be observed that all forms of the þ-article were pronounced as [ð/__V] due to their weak stress position in syntax, as is the case in Modern English as well. Otherwise it would remain voiceless like in English, too (cf. words like thin, thermometer, thank, thorn, thumb etc.). When speakers of Low German, a language which does not have articles in postposition, realised that there was a possibility of interdialectal understanding based on receptive multilingualism between them and their Scandinavian trading partners, they most likely neglected the inflectional definite article as an inflectional ending seen as it was of minor importance for the decoding of a message as far as receptive multilingualism was concerned. However, the þ/[ð]-articles occupied the same position in noun phrases as the (German) d-articles. Therefore, German speakers would not only have analysed and understood the Scandinavian phrases correctly as forms corresponding to the definite article of their own language but they would also have interpreted these forms as being morphologically related to their d-articles. Since the allophone [ð] was not part of any medieval German phonological system, adult second language learners had (objective) difficulties in grasping this sound correctly and, furthermore, in pronouncing it in a correct manner. Due to the morphophonological and syntactic equivalence they replaced the fricatives in the Scandinavian þ/[ð]-articles with the corresponding obstruent d (as in the German definite articles). This new approach at finding a reasonable explanation for the d-onset in the mainland Scandinavian article system is based on the observation of certain principles of receptive multilingualism and L2-language learning by adults. It is claimed that the Scandinavian speakers of one of the next generations following the contact with Low German adopted the obstruent (viz. German) pronunciation of þ/[ð > d] in the onset as their own, possibly as a kind of prestige pronunciation which could be heard in the


whole of Scandinavia due to the ubiquitous presence of non-native L2-speakers of the Scandinavian languages. This approach explains adequately why the Old Scandinavian þ-article did not survive, in comparison to other articles such as the t-article as in Modern Faroese (tann/ tann/tað [nom. sing.masc./fem./neutr.] ‘the/this’) or the th-/[ð] article as in Modern English (the, but cf. also this, that etc.), but was replaced by the voiced obstruent d- (as in Mod. Scand. den/det//de [sing.utr./neutr.//pl.] ‘the/(this)’).14 An explanation based on a natural phonological change — i.e. avoiding the higher marked fricatives [ð-] in favour of obstruents [d-] — can, in principle, not be completely ruled out. But if one argues in that way, one would have to find an adequate explanation for Modern Icelandic and Modern English still having plenty of words with fricatives [θ-/ð-] in their onset. In neither of the languages can such a large impact from languages without such dental fricatives such like Low/High German be observed.15

4.2 An ‘imported’ periphrastic genitive construction The next example illustrating the role of receptive multilingualism for L2-language acquisition is the adoption of a Low German/West Germanic periphrastic possessive construction, which is informally called garpe16 genitive in Norway: Norw./Dan. far sin hat (cf. Germ. (dem [dative]) Vater sein Hut) ‘father’s hat’ [lit. ‘(the) father his hat’]. In many Indo-European languages, ‘possession’ cannot merely be expressed by genitive constructions (inflectional or periphrastic ones) or possessive pronouns (i.e. the genitive form of the personal pronouns) but also by phrases expressing a benefactive role, morphologically rendered, if available, as dative (cf. High Germ. (dem) Vater sein Hut/Buch, non-standard Germ. Wem ist dieses Buch? ‘Whom does this book belong to?’ or Lat. mihi librum est ‘this is my book’). If German merchants used such benefactive expressions in receptive multilingualism communication with Scandinavians they could rely on the interdialectal, genetic relationship connecting the two languages and on the existence of equivalent grammatical structures in both languages. Their Scandinavian counterparts would not only have understood these phrases but could have integrated them into their own language without problems or any further adoption: they clearly recognised sin ‘his/her’, occurring both in Low German and Scandinavian as a possessive pronoun, and thus treated the post-posed structural position of this part of the phrase either as a (now) obsolete syntactic position or, more likely, as a kind of apposition (in the case of Danish or Swedish) or accordingly to the default position of the possessive pronoun which was (and still is) post-positive in Norwegian (and West Scandinavian in general). Thus, they (re)analysed noun (or determiner) phrases like [[far]NP [sin hat]NP]DP, according to other, related (analytic) constructions, such as [[taket]DP [på huset]PP]DP ‘the roof of the house’, as noun/determiner phrases with modifiers to the right of the head noun. These phrases are typical for SVO languages, disregarding, however, the semantic role of far as an underlying benefactive constituent: i.e. “a hat belonging to

41

42

Kurt Braunmüller

fatherbenefactive” was re-analysed as “fatheragentive owns this [sc. his] hat”. In other words, surface structures in connection with easily identifiable, translinguistic morphemes (also called ‘diamorphs’) were re-interpreted in such a manner that they corresponded to the default structure of the other noun/determiner phrases.

4.3 Changes in the preference in word order The last example is taken from syntax. As we have argued before, receptive multilingualism can be applied if you take for granted that the target language follows the same syntactic principles you are used to from your own language/dialect. Since both Low (and High) German featured overt SVO-patterns in main clauses it was quite easy to decode most of the addressee’s utterances immediately: you could rely on the fact that the finite verb was generally placed in second position and that the unmarked order of the post-verbal noun phrases was that the indirect object was situated before the direct object. Differences in the placement of local or temporal adverbs might have occurred but that did not really prevent sufficient mutual understanding in speech situations based on receptive multilingualism. If the finite verb, however, occupied the first position in a sentence (V1), you would still be able to interpret this utterance as you might consider the sentence to be elliptic or stylistically marked in one way or another. Moreover, V1 patterns also occurred in your variety, but were restricted to questions and imperatives. In any case, you would have had the impression that the same syntactic rules you were already acquainted with applied to the target language as well. German merchants and craftsmen who settled in Denmark or Sweden felt obliged to learn to speak the local vernaculars but they obviously simplified the syntax of the respective languages, at least with respect to SVO word order, which is quite typical both for adult L2-language learning and for creoles. One of the consequences of this strategy was that the type-frequency of topicalised objects and fronted verbs (V1 sentences) decreased and that the sentence structures of main clauses became more and more uniform with respect to the sentence initial position: more than 80% of all constituents occurring in this position were now either subjects or adverbs (for further details of this investigation see Braunmüller 2006). The percentage rate of V2-structures in the Scandinavian languages before having contact with the Germans was considerably lower due to the sentence-initial rhematic position of the finite verb in prose texts (narrative as well as legal texts) being quite common. The range of these V1-structures ranged from about 8% (average number) to almost 33%.17

5. Concluding remark This tour d’horizon aimed at showing that receptive multilingualism was one aspect of a complex diglossic/multilingual situation in northern Europe in the late Middle


Ages — and in early Modern Times as well. Nationalism put an end to this way of unmediated communication between genetically closely related languages outside Scandinavia. But receptive multilingualism also represents a starting point for second-language acquisition, especially for adults. Therefore, it is important to investigate the principles and strategies of receptive multilingualism in more detail than has been the case until now. Medieval northern Europe and present-day Scandinavia represent excellent fields for further investigations.

Notes * To a large extent this survey is based on data and evidence from three research projects, all funded by the German Research Foundation DFG, (a) “Middle Low German and its role in the typological and lexical restructuring of the old Scandinavian languages” (1990–1995), (b) “Semicommunication and receptive multilingualism in contemporary Scandinavia” (1999– 2005) and (c) “Historical Scandinavian syntax in multilingual contexts” (2001–2008) — the last two projects being carried out within the Collaborative Research Centre 538 on Multilingualism at Hamburg University. Due to the very limited space only the gist and the results of these research projects can be presented here. Newer data and observations will, however, be presented and discussed in chapter 4. I would like to thank the two anonymous revisers for their suggestions and criticism which this paper has certainly profited from. 1. Cf. Braunmüller (1999: 308–47) for some basic information, Braunmüller and Zeevaert (2001) for a bibliographical survey and Delsing and Åkesson (2005) for the latest empirical investigation available. 2. This means that an active command, i.e. speaking or writing this language, has never been intended. 3. Cf. Klein and Stegmann (2000: 31ff.) for the EuroCom-project on intercomprehension of Romance languages or Munske and Kirkness (1996) as far as foreign (or loan) words in European languages are concerned. 4. Cf. the linguistic situation in mainland Scandinavia mentioned above but also between German and Dutch (see Ház 2005), Czech and Slovak, Spanish and Catalan or Portuguese, just to name a few related languages, not to speak of genuine interdialectal communication as found in Norway or in the German speaking regions of Switzerland. 5. For a comprehensive discussion of the different functions of English as a lingua franca and a native language today cf. e.g. House (2003). 6. Norway is a special case in this respect. Its nation building took place at quite a late stage (the union with Denmark came to an end in 1814, but not before 1905 did Norway become fully independent). For more details on the special conditions of Norwegian nationalism cf. Sørensen (2001) or Storsveen (2004). 7. For more details on the situation in Scandinavia see Diercks and Braunmüller (1993), Braunmüller (1995) or Nesse (2002) and for medieval London see Wright (1996). 8. More evidence on code switching in receptive multilingualism/semicommunication in

43

44

Kurt Braunmüller

Scandinavia today can be retrieved from the very comprehensive discussion in Golinski (2007: 159 ff.), who analysed various forms of code switching, predominantly between Danes and Swedes around the Öresund, and from Braunmüller (2002: 14ff.), where the phenomenon of inter-Scandinavian linguistic accommodation at any price is discussed. 9. For a more general survey of some details of this development, see Peters (2000: esp. 1500f.). 10. I.e. Danish, Swedish, Norwegian (by the end of the Middle Ages mostly represented in its dialects since the literary domains, due to the union with Denmark from 1380 onwards, were more and more taken over by Danish), but not Sami or Finnish. 11. For an outline of this scenario see Diercks and Braunmüller (1993) and esp. Engelbrecht (1993), as far as the historical background is concerned. 12. See the corpus presented in Wright (1996) and our studies on the development of Scandinavian word order in the late Middle Ages, as demonstrated e.g. in Braunmüller (1995) and (2006). 13. In Faroese only this demonstrative pronoun which also is used as definite article, tann, tann, tað ‘this; the’, shows the regular sound change from Old Norse þ to t. All other demonstrative pronouns and deictic expressions have, however, an h- in the onset (< Old Norse þ as well), due to divergent accentuation patterns, which also can be observed in Icelandic (see Petersen 2004). The reason for this one and only exception within deictic expressions has obviously been to avoid a morphological merger with the personal pronoun hann ‘he/him [acc.]’. 14. For a survey on the distribution of the definite articles in the modern Mainland Scandinavian languages and their dialects see Dahl (2004). 15. Danish, the second national language of Iceland during the union with Denmark (1380– 1944), has such a dental fricative. 16. This term is based on a pejorative Norwegian characterisation of the Low German tradesmen and means ‘boastful’ or ‘big-mouthed’ people. 17. Cf. also the discussion in Mørck (2005) or the thesis of Heusler (1921: 175–82) who considers VSO as being the normal word order in Old Norse prose texts. Data from eastern Scandinavia dating back to the pre-Low German era, viz. up to around the end of thirteenth century, are difficult to obtain due to the lack of substantial written sources. The texts available are either legal texts (landscape laws) or based on translations/adoptions from the Bible, such as the Old Swedish Pentateuch paraphrase from the mid-fourteenth century.

References Brattegard, O. 1945. Die mittelniederdeutsche Geschäftssprache des hansischen Kaufmanns zu Bergen. 2 vols. Bergen: John Griegs Boktrykkeri. Beuerle, A. and Braunmüller, K. 2004. Frühe germanische Zweisprachigkeit? Zu den frühesten Runeninschriften und den defixiones in der lateinischen Gebrauchsepigraphik. Arbeiten zur Mehrsprachigkeit, Folge B 54: 1–37. Braunmüller, K. 1995. Semikommunikation und semiotische Strategien. Bausteine zu einem Modell für die Verständigung im Norden zur Zeit der Hanse. In Niederdeutsch und die


skandinavischen Sprachen II, K. Braunmüller (ed.), 35–70. Heidelberg: Winter. Braunmüller, K. 1999. Die skandinavischen Sprachen im Überblick (2nd edn.). Tübingen: Francke. Braunmüller, K. 2000a. Højtysk som ‘naturlig’ fortsættelse af den nedertyske sprogkontakt i Norden i 1500-tallet? In Språkkontakt: Innverknaden frå nedertysk på andre nordeuropeiske språk [Nord 2000 19], E. H. Jahr (ed.), 277–88. Copenhagen: Nordic Council of Ministers. Braunmüller, K. 2000b. Voraussetzungen für die Übernahme hochdeutscher Sprachstrukturen in die skandinavischen Sprachen. In Hochdeutsch in Skandinavien, H.-P. Naumann and S. Müller (eds), 1–18. Tübingen: Francke. Braunmüller, K. 2000c. On types of multilingualism in Northern Europe in the late Middle Ages: language mixing and semicommunication. In The Nordic Languages and Modern Linguistics 10, G. Thórhallsdóttir (ed.), 61–70. Reykjavík: University of Iceland. Braunmüller, K. 2001. Verdeckte Mehrsprachigkeit. In Vulpis Adolatio. Festschrift zum 60. Geburtstag von Hubertus Menke, R. Peters, H. P. Pütz and Ulrich Weber (eds), 117–28. Heidelberg: Winter. Braunmüller, K. 2002. Semicommunication and accommodation: observations from the linguistic situation in Scandinavia. International Journal of Applied Linguistics 12: 1–23. Braunmüller, K. 2004a. Niederdeutsch und Hochdeutsch im Kontakt mit den skandinavischen Sprachen. Eine Übersicht. In Deutsch im Kontakt mit anderen Sprachen, H. H. Munske (ed.), 1–30. Tübingen: Niemeyer. Braunmüller, K. 2004b. Zum Einfluss des Lateinischen auf die ältesten Runeninschriften. Ver schränkung der Kulturen. Der Sprach- und Literaturaustausch zwischen Skandinavien und den deutschsprachigen Ländern. Zum 65. Geburtstag von Hans-Peter Naumann [Beiträge zur Nordischen Philologie 37], O. Bandle, J. Glauser and S. Würth (eds), 23–50. Tübingen: Francke. Braunmüller, K. 2006. Wortstellung und Sprachkontakt: Untersuchungen zum Vorfeld und Nebensatz im älteren Dänischen und Schwedischen. Amsterdamer Beiträge zur älteren Germanistik (Studien zum Ostnordischen) 62: 207–41. Braunmüller, K. and Zeevaert, L. 2001. Semikommunikation, rezeptive Mehrsprachigkeit und verwandte Phänomene. Eine bibliographische Bestandsaufnahme [Arbeiten zur Mehr sprachigkeit, Folge B 19]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehr sprachigkeit. Burke, P. 1989. Heu domine, adsunt Turcae! Abriß einer Sozialgeschichte des postmittelalter lichen Lateins. In Küchenlatein, Sprache und Umgangssprache in der frühen Neuzeit, P. Burke, 31–59. Berlin: Wagenbach. Chirita, D. 2003. Did Latin influence German word order? Aspects of German–Latin bilingualism in the Late Middle Ages. In Aspects of Multilingualism in European Language History, K. Braunmüller and G. Ferraresi (eds), 173–200. Amsterdam: Benjamins. Dahl, Ö. 2004. Definite articles in Scandinavian: Competing grammaticalization processes in standard and non-standard varieties. In Dialectology Meets Typology. Dialect grammar from cross-linguistic perspective, B. Kortman (ed.), 147–80. Berlin: Mouton de Gruyter. Delsing, L.-O. and Lundin Åkesson, K. 2005. Håller språket ihop Norden? En forskningsrapport om ungdomars förståelse av danska, svenska och norska. [TemaNord 2005 573]. Copenhagen: Nordic Council of Ministers. Diercks, W. 1993. Zur Verwendung prä- und postmodifizierender Morpheme im Mittelnieder deutschen und in den skandinavischen Sprachen. In Niederdeutsch und die skandinavischen Sprachen I, K. Braunmüller and W. Diercks (eds), 161–94). Heidelberg: Winter.

45

46

Kurt Braunmüller

Diercks, W. and Braunmüller, K. 1993. Entwicklung des niederdeutsch-skandinavischen Sprachkontakts. Untersuchungen zur Transferenz anhand von volkssprachlichen Texten des 15., 16. und 17. Jahrhunderts — eine Projektübersicht. In Niederdeutsch und die skandinavischen Sprachen I, K. Braunmüller and W. Diercks (eds), 9–40. Heidelberg: Winter. Dorian, N. C. 1997. Lexical Loss among the Final Speakers of an Obsolescent Language: A formerly-fluent speaker and a semi-speaker compared. [Terralingua Discussion Paper 2] (http:// www.terralingua.org/DiscPapers/DiscPaper2.html; March 8, 2006). Engelbrecht, M. 1993. Mitteleuropäisch-skandinavischer Kontakt zwischen 800 und 1600 aus historischem Blickwinkel. In Niederdeutsch und die skandinavischen Sprachen I, K. Braunmüller and W. Diercks (eds), 41–9. Heidelberg: Winter. Gernentz, H. J. 1988. Untersuchungen zum Russisch-niederdeutschen Gesprächsbuch des Tönnies Fenne, Pskov 1607. Ein Beitrag zur deutschen Sprachgeschichte. [Ost]Berlin: Akademie Verlag. Golinski, B. 2007. Kommunikationsstrategien in interskandinavischen Diskursen. Hamburg: Kovač. Haugen, E. 1966. Semicommunication: The language gap in Scandinavia. Sociological Inquiry 36: 280–97. Ház, É. 2005. Deutsche und Niederländer. Untersuchungen zur Möglichkeit einer unmittelbaren Verständigung. Hamburg: Kovač. Heusler, A. 1921. Altisländisches Elementarbuch. 2nd edn. Heidelberg: Winter. House, Juliane 2003. English as a lingua franca: a threat to multilingualism? Journal of sociolinguistics 7: 556–78. Jahr, E. H. 1995. Nedertysk og nordisk: språksamfunn og språkkontakt i Hansa-tida. In Nordisk og nedertysk. Språkkontakt og språkutvikling i seinmellomalderen, E. H. Jahr (ed.), 9–28. Oslo: Novus. Klein, H. G. and Stegmann, T. D. 2000. EuroComRom: Die sieben Siebe: Romanische Sprachen sofort lesen können. 2nd ed. Aachen: Shaker. Kloss, H. 1978. Die Entwicklung neuer germanischer Kultursprachen seit 1800. 2nd rev. ed. Düsseldorf: Schwann. LePage, R. and Tabouret-Keller, A. 1985. Acts of Identity. Cambridge: Cambridge University Press. Menke, H. 1996. ‘Ich bin eine Däne und spreche Deutsch.’ Zur Sprachgeschichte und Sprachen politik im deutsch-dänischen Grenzraum. In Sprachenpolitik in Grenzregionen, R. Marti (ed.), 137–61. Saarbrücken: SDV Saarbrücker Druckerei und Verlag. Mørck, E. 2005. Connective/narrative inversion in Middle Norwegian declarative clauses. In Contexts — Historical, Social, Linguistic. Studies in celebration of Toril Swan, K. McCafferty, T. Bull and K. Killie (eds), 263–78. Berne: Lang. Munske, H. H. and Kirkness, A. (eds). 1996. Eurolatein. Das griechische und lateinische Erbe in den europäischen Sprachen. Tübingen: Niemeyer. Nesse, A. 2002. Språkkontakt mellom norsk og tysk i hansatidens Bergen. Oslo: Novus. Nesse, A. 2003. Written and spoken languages in Bergen in the Hansa era. In Aspects of Multilingualism in European Language History, K. Braunmüller and G. Ferraresi (eds), 61– 84. Amsterdam: Benjamins. Nettle, D. 1999. Language Diversity. Oxford: Oxford University Press. Nichols, J. 1992. Linguistic Diversity in Space and Time. Chicago IL: University of Chicago Press. Peters, R. 2000. Die Rolle der Hanse und Lübecks in der mittelniederdeutschen Sprachgeschichte. In Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erfors chung. Vol. 2, W. Besch et al. (eds), 1496–1505. 2nd rev. edn. Berlin: de Gruyter.


Petersen, H. P. 2004. The change of þ into h in Faroese. RASK (Odense) 21: 51–61. Smith, J. M. H. 2005. Europe after Rome. A new cultural history 500–1000. Oxford: Oxford University Press. Sørensen, Ø. 2001. Kampen om Norges sjel 1770–1905. In Norsk Idehistorie, vol. 3. Oslo: Aschehoug. Storsveen, O. A. 2004. En bedre vår. Henrik Wergeland og norsk nasjonalitet. Dr.ar.-dissertation, University of Oslo. Tiisala, S. 2004. Power and politeness. Languages and salutation formulas in correspondance between Sweden and the German Hanse. Journal of historical pragmatics 5: 193–206. Trudgill, P. 2000. On locating the boundary between language contact and dialect contact: Low German and continental Scandinavian. In Språkkontakt: Innverknaden frå nedertysk på andre nordeuropeiske språk [Nord 2000 19], E. H. Jahr (ed.), 71–85. Copenhagen: Nordic Council of Ministers. Wright, L. 1996. Sources of London English. Medieval Thames vocabulary. Oxford: Clarendon Press. Wright, L. 1997. The records of Hanseatic merchants: ignorant, sleepy or degenerate? Multilingua 16: 339–50.

47

chapter 2

Linguistic diversity in Habsburg Austria as a model for modern European language policy Rosita Rindler-Schjerve and Eva Vetter Universität Wien

It is the purpose of this paper to show that the language policy of the nineteenth-century Habsburg Empire can be considered a promising example of multilingual management and planning because, as a model of lived multilingualism, it shows a potential that projects into present-day multilingual Europe. The present paper elaborates on Habsburg language policy, which stood in stark contrast to the dominant nineteenth-century ideology of homogeneous nation-states. As this policy was far from a unified or streamlined model, this paper investigates three specific domains — education, administration and the judiciary — in the different crown-lands of Bohemia, Galicia and Trieste, where the struggle over multilingualism and for power escalated during the nineteenth century. Keywords: language policy, Habsburg Empire, multilingualism, nation-states

1. Introduction In Part I “The Union’s Objectives” (Art.I-3) of the Treaty Establishing a Constitution for Europe it is stated that the European Union shall respect its rich cultural and linguistic diversity, and shall ensure that Europe’s cultural heritage is safeguarded and enhanced. This appears all the more important within an enlarged Union with twenty different official languages and another sixty so-called regional or minority languages. Hence, there is an immediate need for political action if the motto of “united in diversity” is not merely to be an empty wording. In this perspective, different scenarios enhancing multilingual communication have been under discussion in recent years, such as the requirement that in the future every European citizen should speak at least three community languages (as laid down in the Whitebook 1996).1 Since 2000 and the Lisbon strategy,2 the Commission has repeatedly declared multilingualism a long-term goal in order to ensure the awareness of linguistic diversity, to enhance citizens’ access to EU legislation, procedures and information and, most importantly, to promote a healthy multilingual economy. At a meeting in Barcelona3 in 2002, the Heads of State or Government of the EU called for at least two foreign languages to be taught from a very early age. In 2003 the Commission committed itself to encourage national, regional and local authorities to promote language

50

Rosita Rindler-Schjerve and Eva Vetter

learning and linguistic diversity in its “Action Plan 2004–2006 — Promoting Language Learning and Linguistic Diversity”.4 These actions complement many other forms of support such as the LINGUA and SOCRATES programmes or the European Year of Languages — 2001 and, most recently, the launching of key studies to promote debate, innovation and exchange of good practice.5 Moreover, promoting multilingualism has also become a key aim of citizen education at schools within the EU.6 Linguistic diversity has always been one of Europe’s foremost characteristic of the cultural asset and in many regions multilingual communication has a long-standing tradition. From the nineteenth century, however, the formation of the nation-states went along with cultural homogenisation that was intimately linked to literacy in the dominating national language. As a consequence, in territories where the state boundaries did not map onto the existing ethnic boundaries multilingual communities were consequently forced into national monolingualism. Language diversity as it was practiced within the multiethnic Habsburg Empire throughout the nineteenth century, however, represents quite an exception to this rule. Interethnic language use which also in the Habsburg Empire was largely regulated by the state, was generally based upon a concept of ethno-linguistic autonomy, in particular in cases where the regional societies provided for a culturally elaborated linguistic code of their own ( as e.g. the Italians). German as the language of the politically dominant group was never institutionalised as an over-all state language although there were some attempts at different periods in time to impose German in such a function. These attempts, however, were mostly inspired by the central power’s requirement to provide for an efficient communication and governance in the fields of administration, judiciary and education. Even during periods such as the pre-revolutionary Vormärz7 (pre1848) or post-revolutionary Neo-absolutism8 language policy was not explicitly conceived of as an instrument to assimilate the various ethnic groups into German although in certain regions and at certain times specific regulations by the central power were perceived as such. From 1848, and in particular from 18679 onwards language policy in the Western part of the Empire (Cisleithania)10 increasingly took into account the claims of the rising nationalities ensuring not only linguistic autonomy of the various groups but providing for a regulation which explicitly banned coercive learning of other languages that were usually spoken within the respective crown-land11 (“landesübliche Sprache”). Unfortunately, this ban on coercive language learning which was determined to ensure a policy of non-assimilation should result in escalation of the conflict among the nationalities and eventually contributed to undermine the political integration of the Empire. After the disintegration of the Habsburg state this central-European model of multicultural communication was gradually replaced by national monolingualism in most of the succession states of the Empire and it became almost completely forgotten throughout the twentieth century. It was only in the 1990s and with the fall of the iron curtain when it re-appeared again within the discussion upon future prospects and the promotion of European multilingualism, in particular, within an enlarged Union.

Linguistic diversity in Habsburg Austria

While far from being a model of a unified or streamlined language policy, the regulation of interethnic and supra-regional communication within the Habsburg Empire of the nineteenth century can still be considered a promising example of multilingual management at the state level. It is the purpose of this paper to argue why the Habsburg model carries a potential which projects into present-day European multilingualism. At this point, however, it must be stressed that comparing the multilingual policy of the Habsburg state and the EU should not ignore the fundamental structural differences that are constitutive for both contexts. First and most importantly, the EU is not a state but an alliance of states participating as equal members in a common economic and political programme that establishes the Union. The Habsburg Empire, on the other hand, was a well-defined dynastic and centralized state that nevertheless resembles the EU in certain respects. The Habsburg state reflected a multinational array that called for increased multilingual policies. Greater political integration was demanded by the central state power of the various nationalities within the state. As early as 1848 the Bohemian Charter12 formulated the principle of equal rights. This principle can be found in the Pillersdorf Constitution (1848) as ‘the inviolability of nationality’ and in the Imposed March Constitution (1849) as ‘the inviolable right of every branch of the people to the maintaining and promotion of their nationality and language’. The latter formulation is based on §1 of the Kremsier Draft Constitution (1848) which, however, was never put into practice and in which even a state guarantee of the basic principle of equal rights had been envisaged.13 In the end the principle of equal rights was taken up by the Constitutional Decree in § 19 of the State Law of 186714 as it had been formulated in the Imposed March Constitution of 1849. The integration process within the EU is based on core values such as the integrity of multiple national identities and respect for linguistic and cultural diversity with the difference, however, that from the very beginning cultural and linguistic integration has not been part of a top-down rule but instead has been evolving from the common political will of Member States. Another parallel can be seen in the lingua franca function of both German in the Habsburg Empire and English in the EU with the notable difference, however, that German was the language of a privileged part of the population, as well as the language of the state and its organs which thus gave the Habsburg state a Germanic character. English is different for several reasons since it is the official language of two Member States that joined the EC many years after its founding. Moreover, English is not a supranational state language, nor can the lingua franca version of English in the EU be said to carry an exclusively British character. Yet another important parallel can be seen in the changing premises of political decision-making in the Habsburg Empire of the nineteenth century and in the ever-enlarging EU. In the Habsburg Empire, more recently incorporated territories like Bukowina or Galicia15 differed from traditional Habsburg lands such as Trieste or Bohemia in power structure and national self-assertion. This is the reason why various

51

52


interests could not always be accommodated through universal political strategies and legal measures. Within the EU and its different processes of enlargement, most notably the one of 2004, the planning of language policies has to take into account the changing regional requirements in adapting its strategies to the reality of specific contexts. Discrepancies and double standards, as they have come into being during the last accession process when new Member States had to fulfil the Copenhagen criteria vis-àvis their ethnic minorities whereas old Member States remained exempt from this rule, are examples of this.

2. Multinational communication and the Habsburg model There are various politically and scientifically conceivable ways to manage and to approach multinational communication. Under the impact of European integration and enlargement different concepts of multilingualism have been proposed since the 1990s from which Clyne (2003) derives four major scenarios: a. Polyglot dialogue or receptive multilingualism, i.e. each speaker uses his/her first language and understands those of the interlocutors. b. Multilingual competence in related languages, i.e. speakers learn several languages which are genetically related with one language being the main code and the others being learned by contrast. c. Unspecifiedmultilingualism, i.e. speakers acquire several languages which are preferably those of the neighbouring countries. d. English as a lingua franca (Clyne 2003, 41–3). These four scenarios being an abstraction of the manifold and diversified practices of multilingual communication rather than a clear cut inventory of theoretically well defined models may be overlapping and hence can be fruitfully integrated one into another. Let us take e.g. the polyglot dialogue between speakers of genetically related languages who may communicate in their respective languages as they have acquired receptive competences of their interlocutors’ languages through interaction and maybe, also by learning. Here we refer to the inter-Scandinavian communication as it has been recently investigated by Zeevaert (2004, this volume). Taking into account the requirement of future EU citizens to speak three languages this would mean that due to the regional and supra-regional needs of EU communication speakers could draw upon these various scenarios in selecting or in combining the most appropriate among them. In the case of Austria and in view of the enlarged Union it would be advisable to intensify the learning of the neighbours’ languages which not only would enhance the communication with the neighbours but which would, at the same time, also help to overcome existing conflicts between the German speaking majority and the autochthonous minorities within the country itself. Here, it is obvious that unspecified multilingualism could be combined with multilingual competence in related


languages, and even with receptive multilingualism. Austria proves to be a good example when talking about scenarios of multilingualism, not only in terms of the newly emerged requirements of multi-national communication within the enlarged EU but also for its historical experience in politically coping with the different aspects of multilingual discourse. For lack of data this contribution will not be able to reconstruct the specific workings of receptive multilingualism within the Habsburg Empire of the nineteenth century. Taking into account the functional distribution of languages within the Habsburg state, however, we can start out from the hypothesis that receptive multilingualism must have been a reality in many domains of communication. From the present-day study of languages in contact, we know that multilingualism is intimately associated with aspects of language policy and language planning. Therefore, investigating the language policy of the nineteenth-century Habsburg Empire promises not only to show a diversified range of multilingual practices, but also to reveal the socio-political premises that led to multilingualism in past societies.

2.1 Language policy in the nineteenth-century Habsburg Austria It is important to note at the outset that, when talking about language policy in the Habsburg Empire in terms of a ‘model’ in the sense of continuous activities and strategies undertaken by the central state power to regulate multilingual communication, this does not mean that these actions were applied in equal manner and with equal measures across time and space (see also Rindler-Schjerve 2003a: 3f., Wallnig 2003: 26–9). It has already been pointed out that the interests and power structures of the various ethnic groups in the crown-lands as in Bohemia, Trieste or Lombardy were fundamentally different from those in Galicia, Bukovina or Dalmatia and could, therefore, not always be accommodated through universal legal measures and political strategies. Another point is that after 1848 the central power in Vienna followed a hard line of hegemonic self-assertion, whereas in the 1860ies and the return to constitutional principles it tried to achieve alliances with the national forces. Article 19 of the 1867 Constitution (Stourzh 1980: 1124ff.) is one of the most prominent examples for this. From this it follows that language policy in the Habsburg Empire represents the attempts of a central power to cope with a regionally diversified and conflict-laden context which arose from the differences in cultural prestige and the number of the language communities, which at different times and in different places called for specific political measures and conflict appeasement. Focusing upon this diversification in space and time we will now attempt to shortly summarise the main aspects that determine the heterogeneous character of this policy. As already mentioned, Habsburg language policy was situated in a context in which the state power was continuously challenged by the claims of the rising nationalities. It thus seems obvious that, at the level of relationship, there should have been a clear-cut

53

54


hierarchical polarisation in which the German dominated “central state” and the subordinate “nationalities” would oppose each other. A closer look at the historical data, however, shows that, rather than a steady hierarchical polarisation, there was a wide range of changing models of the exercise of power which constantly found new, fragile and temporary balances (Wallnig 2003). Particularly, in the wake of ongoing democratisation, it happened that the hegemonic power relations in the different territories of the state would themselves become unstable and would require a repeated negotiation and the establishment of a consensus in which the state power was called upon as a guarantor or an arbiter for the equal rights of different groups. From 1849 and especially after 1860 these shifting polarisations were the result of the assertion of interests within the regionally differentiated networks of the different ethnic groups, and also between the ethnic groups and the central state power itself. In Trieste and Galicia, for instance, the state sought to concede a type of pluralist multilingualism after 1849 which should have enabled the underprivileged Slovenes and Ruthenians to use their languages in the official sphere parallel to the dominant regional languages Italian and Polish. In polyglossic contexts of this kind the state power frequently attempted to negotiate between the rivalling parties taking into account the claims of the underprivileged but culturally non well-defined ethnic groups which were trying to redefine their status and to demand recognition for their discriminated language and culture. Thus, state intervention into regional power struggles was frequently read as one nationality using the state for a political move against another nationality. As interventions by the state were frequently contested by the local elites, the measures on part of the central power could remain largely inefficient. Equally, the direction of state intervention could change across time and space and, sometimes, it could also be motivated by laissez-faire strategies or even the lack of political action. Thus, it happened that language rights were not always actually implemented within the interethnic context of an individual crown-land as e.g. in Trieste where the regionally dominant Italian elites obstinately ignored the rights of the Slovenian nationality and would not admit its language in the public since they considered Slovenian not as a historically established language (Czeitschner 2003). Hence, it was only towards the end of the nineteenth century that the Slovenian language became more widely used in the public sphere. In times of crisis, such as the threat to the monarchy after 1860, this policy could also mean that the state would come to an agreement with the local elites thereby ignoring the claims of the underprivileged ethno-linguistic groups or by misjudging the actual needs of the local population. This emerges from a study on the administrative domain in Galicia where, after 1869, Vienna granted the Polish language a privileged status over Ruthenian (an underprivileged variety of Ukrainian) in order to secure its own super-ordinate hegemonic position in this land (Fellerer 2003). State intervention with the interest to impose equal language rights could also lead to the destruction of a well-functioning societal multilingualism and a well-established bilingual education


system such as in the case of Plzeň (Newerkla 2003). In Plzeň this was the result of the uncompromising implementation of the 1867 legislation regarding the principles of language equality and the prohibition of language coercion. Paradoxically around the end of the nineteenth century these gave rise to a more monolingually inspired regime, and called into question the existing multilingual practices which had been in place since the beginning of the nineteenth century in some regions. It is also worth mentioning that language policy in the Habsburg Empire of the nineteenth century was far from being unambiguous. Thus, it could happen that there was a discrepancy between an apparently pluralistic legislation and language practices which ran counter constitutional principles. This is particularly visible in the temporarily oscillating distribution of language use in polyglossic regions of the Empire, where the national paradigm frequently served as a catalyst for diverse claims on the part of socially and ethnically differentiated groups from the same region. The return to constitutional principles after 1849 did generate legal concessions to the languages “currently used in the lands” as well as to the nationalities residing within the state. However, it was not sufficiently clear for each specific crown-land which languages counted as “current” and which linguistic groups were supposed to be nationalities. Therefore legal regulations very often turned out to be a matter of interpretation in the various lands and thus provided the grounds for multiple interethnic conflicts throughout the second half of the century. Another point is that the language policy of the multi-ethnic Habsburg Empire was basically a-national. While the Habsburg Empire had evolved from a feudal-dynastic into a constitutional-democratic state in the course of the nineteenth century, this state remained fundamentally a-national.16 Since the national principle was not at focus in this policy, the state made concessions to the “ethnic groups” and “lands” only in as much as was immediately necessary for internal or external policy making. As a matter of fact and due to major ideological and political shifts, the traditional power structure of this multiethnic state had to be revised around the middle of the nineteenth century. The revolution of 1848 initially caused such radical changes in public life and public discourse that it became unthinkable for large segments of the ethnically diversified population within the state to continue being linguistically excluded from the public sphere. Consequently, the state tried to win the consensus of the established regional elites by making linguistic concessions which in cases such as Trieste further discriminated against the underprivileged nationalities which were, however, of rising regional importance. Under certain conditions such as in Galicia and during a short period even underprivileged groups such as the Ruthenians might be sponsored. The Ruthenian corpus and its status as the second regional language beside Polish were elaborated although this practice was retracted in the early 1850s. Then, with the move towards Neo-absolutism in the early 1850s the government tried — with little success — to elaborate the role of German as a state language. After 1860 the tables had turned and the state power revised its traditional hegemonic position. With the end of the neo-absolutist period, though, the conflict between

55

56


the central power and between the nationalities among each other continued to sharpen. The main reason for this was the continuing hegemonic claims of the traditional local elites so that the state was repeatedly called upon as arbiter in escalating conflicts. The conflicts escalated in the years after 1860 and 1867 and especially in those places where the central power tried to come to an arrangement with the hegemonic claims of traditional elites of the particular crown-land while simultaneously attempting to attend to the needs of underprivileged ethnic groups. Consequently, the ethno-linguistic compromises and the linguistic pluralism reached in the public domains created the impression of a degree of liberalism and tolerance that were exceptional for the time. Regarding the Habsburg language policy as a prominent model of enhancing multilingualism does not mean, however, that this policy was actually determined by pluralist views in terms of a multinational coexistence on an equal footing, and this might eventually have accounted for the inherent weakness of this model. The state forces were far from abandoning their traditional power positions, although the strategies pursued in the conservation of power were not mainly those characteristic of hierarchical domination, subordination and recrimination but rather those governed by the principles of pluralist equality, incorporation and democratic participation. In conclusion we may say that this language policy appeared to be a complex interplay of diversified practices across time and space which was perceived as a shifting balance between tolerance and pluralism on the one hand, and oppression and authoritarianism on the other. Even so, it represented a unique model of linguistic pluralism which had no comparable paradigm at a time when Europe was strongly headed towards creating linguistically unified nation states. Hence, this model of a polyglott state of nations actually stood in stark contrast to the then dominant ideology of the nation-state and of national state-formation. Yet, the interventions by the state power were doomed to failure since the apparently tolerant laws and the strategies of linguistic equality as they were laid down in the Constitution of 1867 (Staatsgrundgesetz) were neither apt to further local interethnic coexistence and tolerance nor did they stabilise the hegemony of Vienna in the long run. We do not know, how linguistic nationalism and state power were negotiated in every single case, but we know that multiple national self-assertion and statehood conceived of as a multinational entity interacted in a way which became relevant for the disintegration of the Habsburg Empire. After this short account of the main components that define the Habsburg language policy in the nineteenth century we will turn our view to three domain-specific case studies with the goal to show how this policy determined multilingual communication in specified crown-lands and across different periods in time. Here, we mainly refer to the findings in the volume of Rindler-Schjerve (2003b) which gives an outline of the interaction of language and power in the multi-ethnic Habsburg Empire. Our domainspecific case studies will draw on these findings restricting the focus to aspects of multilingual communication in the fields of education, administration and the judiciary in crown-lands such as Bohemia, Galicia and Trieste across the time from 1848 to the end of the century. The reason for the investigation of these specific domains is that


linguistic presence in these three fields should guarantee access to and participation in state power for the speakers and their communities. Thus, it goes without saying that these domains, in particular, were to become the preferred battlefields of the struggle over language and ethnic or national self-assertion.

2.2 Language policy and education Education represents a highly important domain and struggles within this domain could even give rise to the fall of a state government as in 1895.17 Concerning the promotion of multilingual education in this domain delivers a rather interesting model consisting in the teaching of different subjects in different languages under the label of “utraquismus”. The utraquist model was mostly restricted to education at the primary level where children were taught in their mother tongue and in a second language which could be a language currently spoken in the land. The mother tongue should be, however, gradually replaced by the second language in the higher grades of education. In the rare cases of utraquist secondary schools German as the so-called state language was also used in higher grades of education. This practice, though, gave rise to numerous conflicts where in lands such as Carinthia and Styria Slovenian was restricted in favour of German. A similar effect can be found in Istria and Silesia where Croatian, Polish or Czech became gradually substituted (Stourzh 1980: 1136–1141). Although this model had a high potential for enhancing interethnic communication in multilingual lands, its weak point can be seen in that it eventually would favour further assimilation into the dominant nationality of the respective land. Another point is that under the paragraphs of Article 19 in the 1867 Constitution it was laid down that all languages currently spoken in the lands were equal, and that each nationality should be entitled to schooling in its own mother tongue. Under paragraph 3 the famous prohibition of language coercion was formulated, saying that no nationality should be forced into the learning of any other language. Consequently, those utraquist schools which were resented as an instrument favouring assimilation, were gradually replaced by monolingual “nationality schools”. These monolingual schools, instead of appeasing the escalating interethnic conflicts further contributed to sharpen the emergent conflicts among the nationalities. On the other hand, it has to be mentioned that in lands such as Dalmatia and Moravia we can observe the development of regionally adapted multilingual models which intended to guarantee a certain degree of bi- and trilingual competence among the pupils. These models took into account the very requirements of local interethnic communication, ultimately ignoring the provisions regarding the ban on language coercion (cf. Burger 1995: 243f.). Comparing the multilingual policies in the Habsburg education with recent developments in multiple-language acquisition within the EU, it has to be stressed that the concept of multilingual teaching did not explicitly take into account the difference be-

57

58


ween productive and receptive competences since legal provisions concentrated first and foremost on written and spoken proficiency. With this background it was not an explicit aim of the utraquist policy to foster receptive multilingualism, although polyglot unspecified multilingualism could actually be enhanced through this model. Ultimately, as a result of the ban on language coercion, the practice of multilingual schooling was doomed to failure and, unfortunately, the utraquist model which appears to be a practical approach of both lived historical practice and modern multilingual education fell into disuse for many decades. Nowadays, and in a different form, it is being re-introduced under the label of Content and Language Integrated Learning (CLIL), especially in the context of EU education policy where it appears mainly restricted to secondary education. Another point is that CLIL targets the development of multilingual competences whereas the utraquist policy, as mentioned above, very often resulted in monolingual education. Let us now turn to Bohemia, a crown-land in which the debate on multilingual schooling was rather heated. Newerkla (2003) has thoroughly investigated the situation of bilingual education at two grammar schools in Plzeň. His findings show that the basically tolerant law of 1867 which provided for the right to have education in one’s native language and which stated that no citizen should be forced to learn the language of any other ethnic group in Plzeň, actually, produced the opposite effect and triggered off the ethnic conflicts that it had been intended to prevent. In the first half of the nineteenth century Czech was not on an equal footing with German, which had been a dominant language until 1848. After a short time of political concessions, the German dominated hegemonic forces tried again to suppress Czech self-determinism. The city parliament of Plzeň, however, managed to promote an equalizing curriculum of the utraquist type at both the German and the Czech grammar schools in order to foster bilingual competence and trust between the town’s two language communities. However, in 1866 the imperial and royal governorship in Prague ordained the discontinuation of existing utraquist teaching at the Czech grammar school and ordered the use of one single language as the means of instruction. This order was heavily opposed by the majority of Plzeň’s inhabitants who wanted both languages to continue as equal media of instruction. This struggle culminated in a bilingual protest note signed by influential individuals and presented to the city parliament. However, the city elders based on Art. 19 then determined German as the only official medium of instruction at the German grammar school. In the Czech grammar school the situation also gradually turned to the worse, although there appeared to be strong efforts to counterbalance the move towards monolingualism. In 1875/76 parallel classes were gradually dissolved and in 1877/78 there were no more parallel classes at the Czech grammar school, with the result that by that time, the school had been turned into a monolingual nationality school. At the German grammar school Czech language and literature had already become an optional subject in 1869. As a consequence, around the turn of the century German


had significantly dropped in importance. The share of Czech students at the German grammar school had shrunk to under 10% in 1905.18 The segregation was complete and the formerly dominant German language had been reduced to a subordinated means of communication (Newerkla 2003: 192). This example shows that throughout the nineteenth century there was a wide range of temporarily changing efforts in which the state power attempted to come to terms with the different nationalities in the lands and to counterbalance their diversified claims, also with the risk of sometimes misjudging the actual local needs. We see that the central power’s language regulations could run counter the actual needs of the local population and finally destroy the efforts for a peaceful coexistence between the ethnic groups. The consequent turn-over from multilingual into monolingual education triggered off ethnic conflicts which this policy had actually been intended to prevent.

2.3 Language policy and administration Administration within the Habsburg Empire was a prominent field of communication since it needed to provide efficient interaction between the interests and requirements of the state power and the citizens. The state power had a basically Germanic character since German was used as the lingua franca of the public domain, at least until 1848, and as it symbolised the Empire and the dynasty. Although there was no corresponding law, as a rule German served as the universal language for all public affairs. Nevertheless, the state power had to come to terms with the multilingual reality of its citizens in order to be efficient. Already since the times of Maria Theresia,19 the state thus attempted to take into account the languages spoken in the lands, however, in varying measure and degree. The differences in admitting the respective languages in administrative procedures were determined by various factors such as time, geographical region and cultural prestige of the languages and the specific administrative context. To give an example, administration before 1848 might be more German dominated than after the revolutionary or the neo-absolutist period. Geographically speaking German could, however, be rather limited as an administrative language, if we take e.g. the Italian speaking lands where the highly prestigious Italian in many contexts even dominated German. As far as the different contexts of administration and their language use are concerned we have to consider that there was a certain graduation of interaction types depending on who communicated with whom within the institutions. Thus, e.g. interactions between the regional and the central institutions would take place in German whereas interaction between a regional institution and a citizen could be in local languages. We also note that since the seventeenth century and in view of the multilingual situation within the state there existed serious attempts on the part of the state to guarantee administrative efficiency and, hence, to regulate the language use in this domain. Let us now take a more detailed look at language use in the administration of Galicia

59

60


under Austrian rule as it was investigated in depth by Fellerer (2003, 2005). His findings show that although German had been for a long time the dominant language in this domain alongside Latin, the acts of language planning from 1812 to 1847 stipulated that the imperial authorities should employ officials with good knowledge of Polish. Yet, the use of Polish had a rather symbolic standing in this context. Polish served as second language for the communicative contexts which involved the non-German speaking population and their representatives. Things started to change after 1848 when Polish and, for a certain period, also the languages of Galicia’s under-privileged Ukrainians, called Ruthenian, entered increasingly into administrative use (Fellerer 2003: 157–64). Polish and Ruthenian were granted the right to be used as a subsidiary language in external communication, i.e. when the authority communicated with the Polish or Ruthenian speaking citizens. Traditionally speaking, though, Ruthenian had never been used before as a language of public life since it passed as an unwritten peasants’ dialect although Ruthenian speakers made up forty to fifty per cent of Galicia’s entire population. This lack of prestige should have an impact upon the further language-political development in the region. During the neo-absolutist period (1851–60) the Austrian government re-installed German as the universal language of the Empire resorting to Polish only as a second language of the land. After the return to constitutional principles in 1860, however, the situation changed and a new order developed which should last until the decline of the Empire. From the end of the 1860ies German was restricted to the interaction with the central state, i.e. the correspondence with the central institutions in Vienna and the publication of legal documents issued in Vienna. Polish and Ruthenian were additionally used for the translation of these documents. In administrative affairs at the regional and district level Polish took over and in 1866 Polish became the source language within the entire regional legislation whereas translations were provided into German and Ruthenian. By and large, Polish returned as an offical language in those contexts where the regional authorities communicated between each other or where communication took place among the officials within one single authority. As a consequence, Ruthenian remained excluded apart from the symbolic act of translating laws into this language. By the end of the decade the use of Ruthenian was back to the situation which held before 1848 (Fellerer 2003: 165 ff.). It is surprising that, after all, Polish would win the battle not only over German but, in particular, over Ruthenian, given the ethno-demographic share after which Ruthenians were considered a nationality that should not have been ignored. One may ask why Vienna did not intervene in this asymmetrical development in Galicia’s administration (on the basis of Article 19). On the other hand it is well known that the Austro-Hungarian compromise in 1867 triggered an agreement between the government and the Polish aristocracy. Hence, in times of crises when external political reasons required it, Vienna apparently renounced opposing the overall polonisation of Galicia’s administration, with the consequence that this did not leave any space for Ruthenian and Ruthenians to gain in importance.


From this language policy we can conclude that there is, at least after 1866, no clearcut polarisation between the German dominated central state and the nationalities in the lands. What we can see, is a shifting balance between pluralism and oppression which, however, appears under different guises and different motivations at different times. Thus it could happen that after 1848 and after 1866 Vienna intended to obtain different agreements with the nationalities, with the consequence that after 1848 the Ruthenians were included into this pluralist policy whereas they remained excluded after 1866. This shows that the state policy resorted to strategies which would favour the segregation of national groups in all domains of public life rather than to support individual and collective multilingualism in this area. Similar to what happened in the educational domain, also in this sphere of public life we may observe a remarkable trend towards monolingualism. Comparing language use in the administrative context of Galicia with multilingualism in EU institutions, we notice that there seems to be a parallel concerning the trend towards monolingualism in both contexts. This can be explained in that English has emerged as a preferred lingua franca in EU institutional communication , although all important EU documents must be provided in the official languages of the EU and the practice of translating is highly elaborated within the different institutions. Unlike the Habsburg language policy after 1867, linguistic nationalism is explicitly banned in this context. As far as the role of English and German as a lingua franca in this specific context is concerned, it can be said that German was more than a simple means of supranational communication since the Habsburg state had a Germanic character whereas the English-dominated EU institutions are not associated with a state power of any kind. Concerning receptive multilingualism, it must be said that there are no reliable data about how this competence would have contributed to fostering multilingual communication in the Habsburg administration. Presumably, receptive multilingualism might have been a reality in this context but what makes it different from the EU scenarios of multilingualism is that in the nineteenth century it had not yet been recognised as a resource of efficiently managing multilingual communication and of fostering supranational identities. Our data suggest, however, that there must have been a lot of unspecified multilingualism since in our polyglot context administrative affairs had to be handled in three languages in order to reach citizens. Moreover, we do also know that in accordance with the principle of equality the provincial governors were instructed that all the civil servants of the crown-lands had to acquire some knowledge of the language usually used in the land (Fischel 1910: No. 181).

2.4 Language policy in the judiciary The judiciary is an institution in which the state figures as an arbiter between conflicting parties and as a garantor for right and public order. The institutionalised resolution of conflicts in court is governed by a normative body of rules and discursive

61

62


procedures from which it follows, that language is the primary medium of interaction in this domain. In multilingual societies such as the Habsburg Empire the judiciary appears to be a very sensitive domain since it depends on the appropriate language rules in order to ensure a fair and well balanced court procedure, and particularly when it comes to bi- and multilingual testifying before the court or to the pronouncing of judgements. Therefore, the Habsburg Empire attempted to regulate the language use of the court procedures by various laws and decrees. Going back to the eighteenth century we see that there are a number of efforts by which the state tried to come to terms with the needs for multilingualism in this domain. In the following we will briefly summarise the most important regulations throughout the nineteenth century drawing on Czeitschner (1997, 2003). From the late eighteenth century onwards we observe several strategies which include the provision for German as the only official language in court (Josephinian Statute Book of 1787) and which provide, in case of need, for an interpreter to assist the parties involved in a trial (Josephinian Court Regulations 1781; Criminal Court Regulations 1788; West Galician Court Regulations 1796 and the following Penal Law 1803). Under the Pillersdorf Constitution of 1848, for the first time, fundamental rights and freedoms were granted which included the inviolability of the nationalities and their languages. In 1848 it was decreed that all the civil servants of the lands were to acquire some knowledge of the languages currently spoken in the land (Fischel 1910, No. 181), and in 1850 it was determined that a sufficient number of judges had to be proficient in the current languages of the region. The official language in the Supreme Court of Justice and in the Court of Appeal was, however, German. If trials were conducted in a language other than German, the verdict had to be read out in that language as well as in German (Ogris 1975: 550f.). In the neo-absolutist period there was no constitution nor did fundamental rights and freedoms exist. With the return to the constitutional principles in 1867 the equality of the languages currently spoken in the lands was guaranteed by the state. In the 1870s further regulations determined that documents produced in a language other than the one used in court had to be translated by an official interpreter. Witnesses who could not speak the language used in court were to be assisted by an interpreter or by the judge or minutes writer, if they were proficient in this language (Fischel 1910, No. 358). It is a major problem that the actual impact of all these measures on the multilingual practices in jurisdiction has been very little investigated. In her study on polyglott discourse and hegemony in the judicial system of Trieste Czeitschner (1997, 2003) succeeds in giving a detailed account of how the language selection was handled in this domain. From this account we may conclude that there was a discrepancy between the language use as it was legally ordained and language use as it was actually practised. The data from the State Archive in Trieste suggest that Italian remained the dominant language throughout the century and that legal regulations between the late 18th and the middle of the 19th century had no visible effect. From the 1820s onwards there are a few changes (Czeitschner 1997: 141ff. , 2003: 88–92) of the kind that


German was more frequently used in the context of internal communication within the judiciary and between different administrative authorities of Trieste. Yet, there is still a total lack of Slovene documents. From 1850 until 1918 we observe that the changes in court organisation and constitutional changes did not have an immediate impact upon the language rule in the judiciary of Trieste. Only in the 1890s the number of Slovene documents increased to a degree that in the years 1910–1918 their numbers came close to the German documents. This development shows that the constitutionally guaranteed equality of all national languages was being implemented very slowly although this practice was repeatedly monitored on part of the central authorities. An even more detailed account of the multilingual working in the judiciary can be derived from the results of the investigation of the specific role contexts that define this domain. Here, we note that Italian is dominant in all institutional interaction types regarding the handling of files and documents and court organisation as well as in communication with interested parties and with other authorities and external relations. It has to be mentioned, though, that in the field of institution- or authority-external communication Slovene was taken into account, but only towards the end of the century although the dominance of Italian remained virtually unchallenged. Within the contexts of institution- or authority-internal use, such as correspondence with authorities of higher rank we observe a dominance of Italian followed by German, and in the correspondence with authorities of equal rank a dominance of Italian and German, whereas in correspondence with authorities of lower rank German outranks Italian. As a conclusion we may say that from the 1850s onwards there is a remarkable discrepancy between the legal framework as it was defined by the state power to guarantee the equality of language use among the different nationalities on the one hand and the actual implementation of the legal provisions on the other. This means that this kind of language policy would basically open up to the nationalities within the state whereas at the same time it lacked the power to restructure multilingual communication in terms of a modern democratic society within this sensitive domain. In this context it is most significant that in contrast to what happened in domains like education and administration the judiciary shows a certain move towards multilingualism. Taking into account that this domain does not appear to be headed towards national monolingualism we may conclude that the language policy of the Empire carried some potential of shaping a supranational model of state multilingualism in which the parties should co-exist on an equal footing. Unfortunately the present data do not permit hypothesizing about the specific impact of various multilingual competences of speakers interacting in the role contexts within the judiciary. Thus, we do not know how receptive and productive competences were employed in order to communicate efficiently within this multilingual context. From various legal provisions that regulated communication in the judiciary we learn that a witness not competent in the language officially used in court could only

63

64


be examined if both the examining magistrate and the stenographer were sufficiently competent in the witness’ language. Otherwise an official interpreter had to be consulted (Czeitschner 2003: 98). The problem, however, is that we do not know what exactly was meant by “sufficiently competent” since the nineteenth-century concept of multilingualism did not yet differentiate between the various competences that are part of a modern understanding of multilingual proficiency.

3. The lesson to be learned Let us now consider what lessons can be learned from policy in the Habsburg Empire. After summarising the main points of the previous chapters we will develop an outline of the main components which defined the success and the failure of this policy. In a second step we will attempt to integrate our findings into the broader context of language policy and multilingualism in present day Europe. As to the successes and ultimate failure of the Habsburg language policy it can be said that, first and foremost, it has provided a wide range of changing models for the exercise of power in handling multilingualism and we noted that language regulations and practices were considerably diversified regionally and domain specific. It would be misleading and all too simplistic, though, to conceive of this policy in terms of a confrontation between the German dominated central power and the multi-ethnic communities conceived of as subordinate nationalities. Rather, we have to regard this policy as the continuous effort on part of the state power to cope with the evolving nationalisms within the multi-ethnic state. It goes without saying that during the nineteenth century and its multi-faceted ideological and political shifts the traditional power structure of this multiethnic state had to be re-defined and adapted to the requirements of continuous societal transformation. Obviously, most of the political efforts pursued by the state were aimed at preserving and safeguarding the traditionally existing power relations within and outside the state. From this scope it follows that the policy which was enacted in terms of compromises with and concessions to the “ethnic groups” and “lands” was not primarily determined by a conception that would account for the formation of a truly pluralistic society within a state of nationalities but, first and foremost, by strategic reasons as they emerged from everyday internal and external affairs. Here, it is important to stress that throughout the nineteenth century the ideological stance of this policy remained an a-national one. At the same time, we may say that the Habsburg policy had different outcomes since we observe a certain move, however modest, towards multilingualism within domains such as the judiciary, while quite the opposite development can be identified in the educational and administrative domains. From the discrepancies which arose from educational policies in Moravia and Dalmatia we learn that the state quite often attempted to appease the inter-ethnic conflicts in adopting inconsistent measures which further escalated the conflicts.


Another point is that the strategies defining the Habsburg model were far from being unambiguous since very often the actual implementation of language practices in the different lands ran counter the pluralistic language provisions as they issued from various decrees and, ultimately, from the constitutional law in 1867. From this it follows that, on the one hand, the state would appear as an arbiter providing for pluralist balance and equality while it ignored the rights of non-elitarian nationalities by trying to accommodate the claims of the traditional elites in the interest of maintaining central power. Turning now to the question of what we can learn from this chapter on multilingualism in central-European history we can conclude that first of all inter-ethnically diversified communities or state compounds, in general, require adequate concepts or guidelines to politically handle and to enhance multilingual communication if this communication is not to become a permanent conflict potential among the linguistically diversified communities. As linguistic access to public discourse means access to power and governance in society, the share of linguistically diversified communities in multilingual communication has to be regulated in one way or another. This is particularly true for the actual situation within EU institutions where we notice a certain tendency towards ‘English only’ that is increasingly raising concern among Member States since they fear not only the loss of status and identity, but also moderation (of having a say at the supranational level). Drawing on the Habsburg model, a second conclusion would be that conceiving of language policy in great multilingual state compounds in terms of ad hoc accommodations is doomed to failure since it runs counter crucial principles of democratic governance. Therefore — and this is still another lesson to be learned — such a language policy needs to be founded in a set of commonly shared values and principles as they form the ideological basis for democratic decision-making and problem resolution. Although the Habsburg state had evolved from a feudal-dynastic into a constitutionaldemocratic state in the course of the nineteenth century the obvious inconsistencies of this language policy originate in its insufficient foundation in such principles. If we now turn an eye on the most recent developments of the EU, we see that the newly elaborated Constitutional Treaty refers to commonly shared values, i.e. democracy, respect for human dignity, freedom, equality, the rule of law and respect for human rights, including the rights of persons belonging to minorities. These values will form the basis for constructing a multilingual Union, eventually determining prospective language policies in Europe. Since there is a great discrepancy, not only in terms of time and social setting but also in the perception of what constitutes a multilingual democratic society, it appears somewhat difficult to draw the exact lines between past and present language policies in comparing the Habsburg with the EU context. Yet, there is the principle of equality that merits some attention as it determines the quality and ultimately, the success of multilingual democratic policies at a larger supranational scale. The case studies on Trieste and Galicia are a good example of how the language rights of the different

65

66


ationalities were negotiated against the principle of linguistic equality. If we now take n into account the multilingual conflict zones in many post-modern societies we realise that little has changed in this respect since it still remains an open question of how to politically cope with equality against the background of diversity. Diversity links up with difference, i.e. difference in terms of linguistic features, speakers and their communities but also difference in the status and positions these collectives hold within the multilingual society. Hence, to cope with this conflict-laden issue will be one of the major challenges for an EU policy in which the principles of equality and diversity are high on the agenda. As mentioned above, equality has become a declared value of the Union, whereas diversity serves as its motto. If we now take regional minorities in particular, it turns out that it is extremely difficult to accommodate both principles. This is one reason why the EU has put pressure on new Member States to fulfil the Copenhagen criteria by providing for the respect and safeguard of their minorities. Another important lesson to be learned for the conceptualisation of language policies in the EU is the case study on Plzeň where it turned out that a well balanced bilingual equilibrium in the educational sphere was destroyed as the state took an all too centralistic stance implementing the principle of linguistic non-coercion (which, ironically, was meant to further democratic equality) top-down and against the actual local or regional needs. From this we learn that in large multilingual state compounds there is a need for a policy that accounts for the diversified regional needs and interests instead of enacting solutions that run counter the actual local requirements. To learn this lesson will be one of the major challenges facing a common European language policy. On the one hand, a common European language policy is actually already in the making if we consider the various steps the Commission has undertaken since the 1990s to foster multilingualism within the Union (cf. chapter 1). On the other hand, language policy remains within the domain of single Member States into which the Union does not have the legal competence to interfere. With ongoing harmonisation, particularly within the field of multilingual education, the question will be how to accommodate the diversified needs for promoting multilingualism at the supranational, national and regional levels. Here again minority languages and their integration into the Union’s multilingualism will be a touchstone for a common European language policy.20

4. Conclusion If we now turn to the four multilingual scenarios that have been proposed under the impact of European integration and enlargement (chapter 2) we can conclude that the Habsburg Empire’s multilingualism is of some interest since it shows a diversified range of practices of multilingual communication which can be related to the communicative requirements of modern multilingual societies. Concerning polyglot dialogue and receptive multilingualism, we must say that we do not know how polyglot dialogues were negotiated in detail and how receptive multi


lingualism contributed to the accommodation of the discourse between the speakers since the multilingual reality reconstructed through historical data does not provide for clear evidence referring to receptive competences. The reason might be that the concept of multilingualism characteristic of the nineteenth century did not differentiate between the various competences that define multilingual proficiency today. The results of our case studies support this hypothesis. As for multilingual competence in related languages, it may be argued that language education concentrated mainly on the languages currently spoken in the lands and on internationally prestigious languages such as French and, to a lesser degree, English. The development of multilingual competence in related languages was not a primary educational concern since language education took place, first and foremost, according to the specific communicative requirements of the particular land. Unspecified multilingualism was a most widespread phenomenon in this historical context since speakers learned several languages and preferably those of their lands and neighbours. Interestingly enough, it was the famous Article 19 of the State Law in 1867 that implicitly fostered this kind of multilingualism while, at the same time, paradoxically undermining it through the ban on language coercion (paragraph 3). Also the lingua franca scenario has to be taken into account since German, which functioned as an overarching state language, served as a supra-regional means of communication, in particular in many public domains. Yet it must be repeated once more that German as a lingua franca was more than a simple means of supranational communication since the Habsburg state had a Germanic character. Concluding, we may say that the Habsburg model can be conceived as an alternative to the still persisting trends of linguistically unified nation states and the pedagogical ideologies they have been associated with in the past. During the nineteenth century the multilingual practices within the Habsburg Empire would hardly be sustained by the current pedagogical ideologies of that time as this pedagogy was strongly headed towards the principles of language nationalism and cultural homogeneity. In the outgoing twentieth century this traditional concept has been largely revised due to the growing internationalisation and globalisation, and under the impact of democratisation and integration in Europe. Contrary to what had happened to multilingual education in the Habsburg Empire, the recent shift from mono- to multilingualism in modern European education is largely supported by theory building in this field. As European multilingual education can only draw on very little direct experience with supranational models of multilingualism, the lived multilingualism in the multi-ethnic Habsburg state may serve as a historical case study from which we learn in more detail about how great multilingual societies can organise their interaction and how they can administrate their peaceful coexistence on common political grounds. Therefore, we think that the model of the multiethnic Habsburg state — at least that of the second half of the nineteenth century — carries a potential which projects into contact and conflict scenarios in present-day European multilingualism.

67

68


Notes 1. http://www.eu.int/comm/education/doc/official/keydoc/lb-en.pdf (see: fourth general objective “proficiency in three community languages”, 47–9). 2. The member States agreed in Lisbon to establish an effective internal market to boost research and innovation and to improve education to make the EU “the most dynamic and competitive knowledge-based economy in the world” by 2010. 3. Barcelona European Council, 15 and 16 March 2002, Presidency Conclusions, Part I, 43.1. 4. COM (2003) 449 5. http://europa.eu.int/comm/education/policies/lang/key/studies_en.html 6. http://www.eurydice.org 7. This period between the Congress of Vienna in 1815 and the Revolution of (March) 1848 is usually associated with the image of a police and censorship state and embodied in the person of State Chancellor Metternich. The Vormärz is marked by a consolidation of the authoritarian state in which self-administering bodies such as the estates continued to lose ground, whereas the nobility and the middle class forces continued to be able to share in the state power apparatus. 8. The period from 1851–1860, generally referred to as ‘Neo-absolutism’, was characterised by a great deal of reformist activity concentrating on extending the powers of the centralised state. Useful innovations in the technicalities of administration were retained, such as the ministerial system and the system of district administration, but also certain liberal advances such as the abolition of land control and the freedom of teaching and learning in the universities, and because of these there was also an increase in the opportunities for the state to participate in its own development. In 1860 the implementation of a parliamentary constitution brought an end to ‘Neo-absolutism’. 9. In 1867 Hungary was incorporated into the overall structure of the Empire as a partner with equal rights: The K(aiserlich)-K(öniglich) (Royal-Imperial) monarchy became a K(aiserlich) u(nd) K(öniglich) (Royal and Imperial), which was supposed to demonstrate the upgrading of the Kingdom of Hungary. The popular names for the two parts, Cisleithania and Transleithania (deriving from the River Leitha which marked the border) were a paraphrase of the new legal situation. Hungary acquired a new government and its own parliament which only collaborated with the Viennese equivalents in matters of foreign policy, the armed forces and the joint state budget. The settlement (“Ausgleich”) was expressed in separate constitutional laws for Cis- and Transleithania. 10. The present paper focuses exclusively on the situation in Cisleithania, and in particular on multilingualism and language policies as they were practised in Czech and German speaking Bohemia, in Polish and Ruthenian speaking Galicia, in Italian, and Slovene speaking Trieste. The “Kingdoms and Territories” represented in Cisleithania were the Kingdoms of Bohemia, Dalmatia and Galicia-Lodomeria; the Archduchies of Austria above and below the Enns; the Duchies of Salzburg, Styria, Karinthia, Krain, Silesia and Bukowina; the Margravates of Moravia and Istria; the Princely Counties of Tirol and Görz-Gradisca; the Land of Vorarlberg; the City of Trieste. In 1910 the Cisleithanian population was made up as follows: 35.6% German, 23% Czechs, 17.8% Poles, 12.6% Ruthenians, 4.5% Slovenians, 2.8% Serbs and Croats, 2.7% Italians — a category which also included Ladins and Friulians — 1% Romanians, and 0.02% other groups.


11. The frequently used term “crown-land” or “land” refers in general to the fact that a country belonged to the sphere of power of a crown (e.g. the Bohemian crown of Wenceslas, the Hungarian crown of Stephen…), but especially after 1848 “crown-land” referred particularly to the countries of the Austrian crown. 12. The Charter is an imperial concession, generally seen as a reaction to the Prague revolutionary movement of 1848. 13. The Constitution was elaborated in 1848 and was the first constitution within the Habsburg Empire. Against the background of the revolutionary events (occupation of Vienna in autumn 1848) a draft constitution was elaborated by the first Reichsrat (= imperial council as the first representative body at the level of the whole state) in Kremsier some month later. A few days after the Kremsier Reichsrat had presented its constitutional project to the Emperor and to the government, the latter answered by enacting its own constitution, the so-called Imposed March Constitution (“Oktroyierte Märzverfassung”), and by dissolving the parliament. On New Year’s Eve 1851, the Emperor repealed the Imposed March Constitution. It was replaced by the socalled Silvesterpatent [New Year’s Eve Decree] which fails to mention equal national rights altogether. This is the beginning of the neo-absolutist period (1851–1860). 14. Events such as the State Law of 1867 (Staatsgrundgesetz) were considerably influenced by a complex network of internal and external political occurrences, for example the lost war against Italy and the loss of Lombardy in 1859, the state financial crisis and the constant changes in the political constellation. 15. Both lands were formerly parts of Poland and remained therefore untouched by the Habsburgs’ centralizing tendencies in the early modern period (cf. Wallnig 2003: 27f.). 16. This a-national character of the state power was also reflected in its dynasty which in most of its parts had gone through a multilingual education. Goebl (1997) gives an interesting account of the multilingual competences of the different Habsburg emperors. 17. The struggle was about the introduction of parallel classes with Slovenian and German as means of instruction in the one and only German grammar school of Celje. Celje was a German dominated town in a Slovenian dominated surrounding. The German liberal party strongly opposed this plan giving rise to an ethnically heated debate with the final result that the coalition in Vienna demissioned (Burger 1995: 159–61). 18. See Newerkla (2003: 189) for the tables of the proportional distribution of the first language at both the Czech and the German grammar school since 1850. 19. Maria Theresia (1717–1780), archduchess of Austria, queen of Hungary and queen of Bohemia (‘emperor’ only in her quality as wife of Francis Ist ), is first of all remembered for the promotion of a secular and public school system. The reforms during her reign (1740–1780) and that of her son Josef II (1780–1790) gave rise to the development of a centrally structured legal and civil administration in which civilian bodies became increasingly involved. 20. In this context it cannot be ignored that as far as linguistic minority promotion is concerned there are actually double standards since old Member States are not obliged to fulfil the Copenhagen criteria.

69

70


References Burger, H. 1995. Sprachenrecht und Sprachengerechtigkeit im österreichischen Unterrichtswesen 1867–1918. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Clyne, M. 2003. Towards inter-cultural communication in Europe without linguistic homogenization. In Die Kosten der Mehrsprachigkeit. Globalisierung und Sprachliche Vielfalt, R. de Cillia, H.-J. Krumm and R. Wodak (eds), 40–7. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Czeitschner, S. 1997. Polyglossie in der Domäne Gerichtswesen in Triest 1767–1918. Sprachpolitik und Sprachwirklichkeit in der Habsburgermonarchie. Doctoral thesis, university of Vienna. Czeitschner, S. 2003. Discourse, hegemony and polyglossia in the judicial sytem of Trieste in the nineteenth century. In Diglossia and Power. Language policies and practice in the nineteenth century Habsburg empire, R. Rindler-Schjerve (ed.), 69–105. Berlin: de Gruyter. Fellerer, J. 2003. Discourse and hegemony: The case of the Ukrainian language in Galicia under Austrian rule (1772–1914). In Diglossia and Power. Language policies and practice in the nineteenth century Habsburg empire, R. Rindler-Schjerve (ed.), 107–66. Berlin: de Gruyter. Fellerer, J. 2005. Mehrsprachigkeit im galizischen Verwaltungswesen (1772–1914). Eine historischsoziolinguistische Studie zum Polnischen und Ruthenischen (Ukrainischen). Wien: Böhlau. Fischel, A. 1910. Das österreichische Sprachenrecht, 2nd edn. Wien: Verlag Irrgang. Goebl, H. 1997. Die altösterreichische Sprachenvielfalt und -politik als Modellfall von heute und morgen. In Lingua e politica. La politica linguistica della duplice monarchia e la sua attualità. Atti del simposio Istituto Italiano di Cultura, 31 May 1996, U. Rinaldi, R. Rindler-Schjerve and M. Metzeltin (eds), 103–21. Vienna: Istituto Italiano di Cultura. Newerkla, S. M. 2003. The seamy side of the Habsburgs’ liberal language policy: Intended and factual reality of language use in Plzen’s educational system. In Diglossia and Power. Language policies and practice in the nineteenth century Habsburg empire, R. Rindler-Schjerve (ed.), 167–95. Berlin: de Gruyter. Ogris, W. 1975. Die Rechtsentwicklung in Cisleithanien 1848–1918. In Die Habsburgermonarchie 1848–1918, Verwaltung und Rechtswesen, Bd. II, A. Wandruszka and P. Urbanitsch (eds), 538–662. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Rindler-Schjerve, R. 2003a. Introduction. In Diglossia and Power. Language policies and practice in the nineteenth century Habsburg empire, R. Rindler-Schjerve (ed.), 1–11. Berlin: de Gruyter. Rindler-Schjerve, R. (ed.) 2003b. Diglossia and Power. Language policies and practice in the nineteenth century Habsburg empire. Berlin: de Gruyter. Stourzh, G. 1980. Die Gleichberechtigung der Volksstämme als Verfassungsprinzip 1848– 1918. In Die Habsburgermonarchie 1848–1918, Verwaltung und Rechtswesen, Bd. III/2, A. Wandruszka and P. Urbanitsch (eds), 975–1206. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Wallnig, T. 2003. Language and Power in the Habsburg Empire: The historical context. In Diglossia and Power. Language policies and practice in the nineteenth century Habsburg empire, R. Rindler-Schjerve (ed.), 15–32. Berlin: de Gruyter. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verständigung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.

part 2

Receptive multilingualism in discourse

chapter 3

Receptive multilingualism in Dutch–German intercultural team cooperation* Anne Ribbert and Jan D. ten Thije Radboud Universiteit Nijmegen/Universiteit Utrecht

Despite the typological proximity of Dutch and German, receptive multilingualism is seldom used in Dutch–German communication. This chapter explores the factors that influence the choice for receptive multilingualism as a mode of communication in German–Dutch encounters by reflecting on the relationship of the two languages and nation states. Furthermore, attention is paid to the institutional constellation and the relationships between the actual interactants in discourse. These theoretical preliminaries are examined in an analysis of receptive multilingualism as it occurs in the language teaching team of the Goethe-Institute in Amsterdam (cf. Roelands 2004) by analysing team discussions as a form of institutional communication (cf. Koole and ten Thije 1994). It is highlighted how institutional keywords are used in order to convey institutional knowledge that is not language-specific in order to enhance felicitous communication. The authors finally show how the discourse at the Goethe-Institute is structured interculturally, making use of Rehbein’s (2006) concept of the ‘Cultural Apparatus’. Keywords: German Dutch interaction, team discussion, interactive planning, institutional keywords, Goethe-Institute, intercultural discourse

1. Introduction This paper deals with receptive multilingualism in Dutch–German team cooperation. In order to give the reader an impression of this specific mode of communication we begin this section with a discourse fragment between the Dutchman Gerard and the German Bernd (see p. 74). The fragment is a representative extract of discourse between the Dutchman Gerard and his German colleague Bernd who both work at the GoetheInstitute Amsterdam. Gerard and Bernd have decided at the start of their working cooperation that they both want to use their respective mother tongue while speaking to each other. Both of them have adequate, that is more than minimal, knowledge of their interlocutor’s language so that they can understand each other reciprocally. The discourse stretches between Bernd and Gerard are some of the few documented cases of German–Dutch receptive multilingualism. This leads us to the question as to why the

74

Anne Ribbert and Jan D. ten Thije

Fragment 1. 130504: zaterduitsdag

use of receptive multilingualism is so rare between the Dutch and the German. Taking into account the typological closeness of the two languages and the cultural proximity of the two nation states, it is astonishing that bilingual Dutch–German discourse does not occur more often in the receptive multilingual mode. In most cases, either one of the interlocutors adapts to the language of the other, or English is used as a lingua franca (see Zeevaert, this volume, for an outline of all possible modes of multilingual communication).

Receptive multilingualism in team cooperation

In this chapter, we start by reviewing the factors that support the occurrence of receptive multilingualism. We relate these factors to the parameters introduced by House and Rehbein (2004: 3) as a means for analysing multilingual communication. Section 2 discusses the occurrence of receptive multilingualism from three different perspectives. First, we focus on the social and linguistic relations between the two nations and their standard languages. Second, we discuss the institutional constellations that enable or restrict the occurrence of receptive multilingualism. In a third step, we investigate the perspective of the individual interactants involved by describing the linguistic and cultural competences that might encourage the use of receptive multilingualism. In section 3, we discuss a case study on receptive multilingualism from the GoetheInstitute in Amsterdam in more detail. This leads to a discussion as to why receptive multilingualism is a successful mode of discourse in this specific case. We exemplify this success by analysing a stretch of planning discourse taken from team discussions (section 5). Illustrated by the use of keywords, we discuss one of the means by which understanding is established within the receptive multilingual mode. This chapter is based on a case study that was carried out by Roelands (2004). In section 6, we discuss how cultural understanding is achieved in receptive multilingual discourse. For this purpose, we will apply Rehbein’s notion of ‘Cultural Apparatus’ (Rehbein 2006). Finally, in section 7, we draw conclusions from this first case study on German–Dutch receptive multilingual discourse.

2. Receptive multilingualism as a form of multilingual communication House and Rehbein (2004) distinguish a number of parameters that constrain multi lingual language constellations. In this paper, we relate these parameters to the specific case of receptive multilingualism, which we consider as a specific mode of multilingual communication. These parameters operate on different levels, which are, however, interconnected. Besides the perspective of the individual interactants and the language contact point of view, which was discussed in House and Rehbein (2004), we consider it as inevitable to incorporate also an institutional level as an intermediate stage between the discourse of individual interactants and language contact on the societal level.

2.1 Language contact between two nations First of all, typological proximity is a parameter that acts in favour of receptive multilingualism since, as in the case of cognate languages, both interlocutors can easily gain a certain passive command of the other’s respective language (House and Rehbein 2004: 3). Möller (this volume) reveals how speakers can detect the meaning of previously unknown words in a second language on the basis of their cognate status. In actual fact, typological distance is used as a criterion for defining whether we are dealing

75

76


with receptive multilingualism in the broader sense of the word, or rather with semicommunication (see Zeevaert, this volume). Besides linguistic distance, differences in the socio-political status of the two languages and nations also play a part in the determination of receptive multilingualism as a form of multilingual communication (House and Rehbein 2004: 3). Receptive multilingualism is the most equal mode of communication, as none of the interlocutors has to adapt to the other, nor are they forced to use a lingua franca. Therefore, we would rather expect to find receptive multilingualism in situations in which the two languages have an equal socio-political status. De Swaan (2001), for instance, introduces a model that measures the status of a language on the basis of its amount of speakers and its connection to other languages (‘centrality’) via multilingual speakers. The status of the language is furthermore determined by the economic and political power of its speakers. Inequality of status will in most cases lead to the speakers of the lower status language adapting to the speakers of the language with the higher status. The status held by a foreign language in a particular country is reflected in its position in the school curriculum (Clyne 1991; Bourdieu 1992). A Dutchman would rather be expected to have some basic knowledge of German than a German of Dutch, since German is — at least traditionally — considered as an important foreign language in the Netherlands, whereas the opposite does not apply (Nelde 1988; CPB 1997). Note, however, that recently this imbalance has changed: Dutch is becoming more and more popular among German high school and university students, whereas German is losing its strong position in the Netherlands, resulting in a significant change in the language constellation between Germany and the Netherlands. This change in the traditional constellation has brought about the need for different forms of communication (Finkenstaedt and Schröder 1992). Whereas, traditionally, the Dutch would tend to adapt to the Germans, receptive multilingualism now provides a possible institutional solution for communication across the border (Herrlitz, p.c.) in those cases where the Germans have acquired receptive language skills in Dutch. The social-cultural changes in the German–Dutch relationship are a recurrent topic of various conferences (Raven and Aspeslagh 1997; Aspeslagh 1998). An additional factor to be mentioned with respect to the multilingual language constellation is the general attitude of the speakers of one language towards the speakers of the other language. In the case of the attitude towards Germans, for example, speakers of other languages occasionally still have anti-German feelings leading back to the Second World War (Zahn 1984, 1986; Lademacher 1989; Müller and Wielenga 1995). This in turn has also influenced the self-perception of Germans (Westheide 1997; Linthout 2000). The attitude and ideas of speakers about the way they are perceived might thus influence the choice of the specific multilingual discourse mode. In the case of Germans, this (expected) negative attitude can lead to an avoidance of the use of German in discourse and consequently to the use of the other person’s mother tongue (Dutch in our case), or a lingua franca (in most cases English). For instance, Prescher (2006) discusses language attrition among Germans living in the Netherlands


and relates this process to their problematic multiple identities. This can be opposed to the general positive evaluation the Dutch receive in Germany (e.g. Rudi Carrell and Linda de Mol1) (Westheide 1997).

2.2 Institutional constellation The second perspective with respect to the occurrence of receptive multilingualism relates to the institutional constellation in which the multilingual discourse takes place. Based on general research referring to language choice in Europe (Milroy and Muysken 1995; De Bot 2004) we would expect the use of receptive multilingualism to depend on the ratio of speakers of the two languages within a societal organisation (see Schjerve-Rindler and Vetter in this volume for a historical account of language choice). Furthermore, we would expect the status held by the speakers of the languages with respect to each other within the hierarchy of the institution to be an influential factor (cf. Lüdi this volume, Werlen this volume). The more equally the two groups are represented in terms of number and status, the more probable it is that receptive multilingualism will be used. Moreover, the language policy of the organisation plays a role (Herrlitz 1997; De Jongste 1997). Herrlitz and Loos (1994), for example, discuss the ‘Rudi Carrell model’. According to this model, some organisations encourage the use of multiple languages, whereas others discourage from it and prescribe the use of one single language which is considered to be the standard language of this company. It could also be the case that an organisation does not have a clear conception with regard to its language policy. In that case, it only depends on the employees which mode of communication is chosen to be used (Loos 1997, 1998). Koole and ten Thije (1994) have shown that the length of the cooperation between employees within an institution is an important factor in the establishment of a ‘discursive interculture’ (i.e. the common team knowledge that results from the multicultural long term team cooperation within the group and cannot be traced back to the sum of the individual discourse knowledge of the single participants). Due to receptive multilingualism being part of such a discursive interculture, we would expect it to potentially take place in situations in which employees have co-operated for lengthy period of time. The factors mentioned above might be enhanced or weakened depending on the country in which the institution is situated, i.e. Germany or the Netherlands (see 2.1).

2.3 The interactants’ perspective The third perspective deals with language choice from the point of view of the individual speaker and his interactions with other interactants (House and Rehbein 2004: 3). This covers both the individual’s linguistic competence in the second language (e.g. Grosjean 1982), as well as their past experiences with other cultures and languages (‘intercultural competence’). If somebody is accustomed to dealing with other cultures, he

77

78


or she will thus be better trained in communicating with people of a different origin. The knowledge of additional languages can be helpful in determining the meaning of an unknown word. Research has moreover borne out that the knowledge of several languages also enhances metalinguistic competence (Hufeisen and Marx, this volume). Besides the actual linguistic distance, the notion ‘psychotypology’ should also be considered when referring to the language constellation of multilingual communication. Psychotypology concerns the language learner’s perception of the actual linguistic distance. Work by Kellerman (1979) suggests that a language user often misestimates the actual amount of cognates existing in a closely related foreign language. Psychotypology plays a key role in language acquisition, seen as the learner subconsciously decides how much linguistic material he can transfer from his mother tongue into the second language depending on his estimation of the closeness of the two languages. Therefore it might also be a relevant concept for assessing how much the listener actually understands in the receptive multilingual mode (Lutjeharms, this volume). Finally, it should be mentioned that language users arguably have to be familiar with the phenomenon of receptive multilingualism itself in order to adequately use it. In most cases, receptive multilingualism does not arise spontaneously, but is far more the result of an agreement on the part of the speakers involved (Braunmüller and Zeevaert 2001; Zeevaert, this volume) We have discussed how various parameters referring to multilingual language constellations may influence the occurrence of receptive multilingualism in comparison to other forms of multilingual communication. In the next section, we will illustrate how these factors influenced the multilingual communication at the Goethe-Institute Amsterdam.

3. The Goethe-Institute Amsterdam The research discussed in this chapter is based on a case study at the Goethe-Institute Amsterdam (Roelands 2004; Roelands and ten Thije 2006). We will reconstruct the discourse structures observed at this international institution, paying attention to the department of language teaching, where receptive multilingualism is applied as the mode of communication. We will examine which of the factors outlined above come into consideration. The Goethe-Institute is a German organisation established to promote German language and culture abroad. It is represented in 144 cities worldwide. In our case, the German institution is situated in the Dutch capital Amsterdam. From a language contact perspective, German and Dutch are considered to be cognate languages. Linguistically, the two languages are thus closely related, a characteristic which might work in favour of receptive multilingualism. As outlined above, however, the relation between the two countries in terms of status and power is not equal. German is spoken by far more people in Germany, Switzerland and Austria. Furthermore, Germany is


economically comparatively more powerful than the Netherlands (Delsen and de Jong 1998). Consequently, German has been part of the Dutch obligatory school curriculum for a long time, whereas it was relatively uncommon for a German to speak Dutch. However, as has been stipulated above, there have recently been changes in this respect. Dutch is gaining more and more appreciation among German students, whereas on the other hand young Dutch people do not often choose to learn German any more. From an institutional perspective, the Goethe-Institute Amsterdam can be considered as a German organisation that operates internationally and, in our case, is situated in the Netherlands. In the Goethe-Institute Amsterdam German and Dutch colleagues have been cooperating closely over a longer period of time. Interestingly, due to the vast amount of highly developed language competencies of the employees in the organisation several different communicative modalities can be found when looking at the internal communication; employees either speak German, Dutch, or use receptive multilingualism in their communication. However, English as lingua franca is not used. Finally, all kinds of code switching can be detected (Auer 1998). The participants have the choice of speaking their first language or switching to other communicative modes. When looking more closely at the multilingual communication within the GoetheInstitute, we found that three departments fulfil the preconditions required for the use of receptive multilingualism, namely the department of language teaching, the department of culture, and the reception desk. However, only the department of language teaching has actually chosen to operate in the receptive multilingual mode. Functionaries in the other two teams prefer to converse in just one language, either Dutch or German. In this context, the question arises as to why the team members in the language teaching department have chosen to communicate by means of receptive multilingualism, whereas the employees working in the other departments have made different decisions with respect to their multilingual communication. Taking into consideration the biographical interviews with all employees and the audio recordings that were carried out with various analyses on distinct discourse types (Roelands 2004), we conclude that language competence, institutional position and discourse topics were the decisive factors for the decision on the respective mode of discourse used by the employees in the three departments. The following can be said about the language teaching department, which is the only department in which receptive multilingualism is used. The department is made up of two employees. The team manager Bernd is from Germany and only temporarily living in the Netherlands. He estimates his language competence in Dutch approximately at the level of B22 according to the Common European Frame Work (Council of Europe 2001). The Dutch team member Gerard is a local employee. He estimates his German language competence at the level of C2.3 Both are aware of the potential intercultural problems between the Germans and the Dutch resulting from the historical relationship of the two neighbour countries and the indirect influences of the Second World War. They unequivocally state that at the start of their working relationship three years

79

80


ago they chose to speak their first language with each other while at work. Consequently there are no discussions relating to the choice of language at the start of each interaction. Receptive multilingualism has become the institutional standard for their communication at work. In his biographical interview the Dutchman Gerard states (Roelands, 2004: 29): “It is much easier. I can express myself more easily and more precisely with respect to important details or when I am joking. It costs less energy, less effort, less reflection and I can react instantly.” (Translation by R/tT). The German Bernd mentioned an additional reason for the use of receptive multilingualism: “When everybody is allowed to speak his first language, democracy comes into being.” (ibid, 29) (Translation by R/tT). In section 2, we reviewed a number of factors that we regard to be influential with respect to the occurrence of receptive multilingualism. On the basis of the case study in the Goethe-Institute, we are given a better insight into the factors that are relevant for the constitution of this multilingual communicative mode. It turned out that the individual interactants play a major role in the choice to be made between the different multilingual modes. Within the Goethe-Institute, only one of the three departments chooses to use receptive multilingualism. The typological relationship of the two respective languages and the organisation in which the interaction takes place create various possibilities. The actual use of these possibilities, however, depends on the institutional language policy of the respective Goethe-Institute. The Goethe-Institute thus creates the possibility for its employees to choose their own mode of communication. This is referred to as the rise of an action space (Rehbein 1977). This case study raises the question as to which extent the language team in the Goethe-Institute is ‘special’. The general institutional purpose of the Goethe-Institute is to promote the German language and culture abroad. Consequently, the organisation might be more willing to develop alternative modes of multilingual communication. On the other hand, it could also be possible that such an institute has a quite strict and defined language policy with respect to the internal communication as is the case in many multinational companies. In actual fact, this case study has led us to the conclusion that the overall language policy results in an action space that the employees of the Goethe-Institute Amsterdam have filled in differently. From an interactants perspective, it might not be coincidental that, in contrast to the other employees of the Goethe-Institute, the members of the language teaching team have decided to employ the receptive multilingual mode. From their professional expertise they are aware of the impact of code switching and language choice on identity construction during communication at work. Although the members of the remaining departments have comparable language competencies in both languages at their disposal, they do not set themselves the same professional challenge to develop receptive multilingualism as an alternative mode of multilingual communication. In actual fact, the everyday receptive multilingualism used by the language teaching team exemplifies the overall general aims of the Goethe-Institute in a very original and concrete manner.


It will be necessary to conduct more research in different institutions to assess the factors responsible for language choice more precisely. Comparing Goethe-Institutes in various countries could bring about interesting findings. With the help of this procedure, we could measure the effect of the factors linguistic and cultural distance of the two language communities by means of the occurrence and development of receptive multilingualism. Furthermore, it might help to shed light on the effect of the language policy adopted by the respective Goethe-Institute for the way the different local institutes fill in the action space. It is self-evident that research in other institutions is also necessary in order to gain an all-embracing picture of the phenomenon receptive multilingualism.

4. Receptive multilingual mode of the speech action pattern “Interactive Planning” In order to reconstruct the relevance and interlacement of the factors mentioned above in the actual realisation of discourse, we analyse the discourse structures in the case study conducted in the Goethe-Institute first of all as a form of institutional discourse (Bührig and ten Thije 2005). We begin this section by introducing the functional pragmatic approach to discourse (Ehlich and Rehbein 1986; Rehbein 2001) with a focus on Koole and ten Thije’s (1994) approach to team discussions. After that we will raise the question whether this approach is applicable to discourse in a receptive multilingual mode. The functional pragmatic approach4 to discourse and text aims at the reconstruction of speech action patterns as the realisation of societal purposes. An institutional analysis includes the reconstruction of these social purposes and their transition into speech actions. Institutional discourse is analysed by means of differentiating between first order institutional knowledge (every day interactant’s knowledge) and second order institutional knowledge (professional knowledge). This difference is based on the dichotomy between clients and professionals (or ‘agents’ cf. Ehlich and Rehbein 1986) within institutions. That means that an analysis of institutional discourse begins with the reconstruction of speech action patterns while characterising the purposes of the discourse types involved. Our discourse data stem from team discussions within the language teaching team at the Goethe-Institute. According to Koole and ten Thije (1994), team discussions are instances of professional (agent-agent) interaction and can especially be characterised by two specific speech action patterns, namely the speech action pattern of interactive planning and the speech action pattern of reporting. These two patterns enable efficient knowledge transfer and action coordination within a (professional) team. The pattern of interactive planning aims at reaching a consensus on common action plans, whereas the pattern of reporting aims at the transfer and assessment of institutional knowledge in order to make decisions. By analysing the speech action pat-

81

82


tern of reporting and interactive planning in the language teaching team we focus on their main communicative tasks and thus are able to identify the receptive multilingual characteristics of German–Dutch professional discourse. As mentioned above the pattern of interactive planning has its purpose in reaching a consensus of common action plans. In practice this purpose is realised by one of the actants formulating a proposal for a plan, which is subsequently assessed by another actant. The former is referred to as the author of a plan and the latter the critic (Koole and ten Thije 1994). The critic has the opportunity to formulate an alternative plan if he does not agree with the author’s proposal. In this case the author of the original plan has to judge the alternative proposal and either persist on his original proposal or accept the alternative proposal. When one of the proposals is accepted by both author and critic a consensus is reached with the result of the pattern purpose being realised and a further speech action pattern being initiated. These subsequent crucial steps in realising the purpose of the speech action are called pattern positions (Ehlich and Rehbein 1986). The speech action pattern of interactive planning is often used as a comprehensive pattern in which other speech action patterns such as substantiating (giving reasons), explaining, rephrasing or reporting are embedded. This means that various pattern positions of the speech action pattern of interactive planning coincide with pat-

Figure 1. The speech action pattern Interactive Planning (Koole and ten Thije 1994: 122)


tern positions mentioned before. For instance, the positive or negative judgement of a proposal for a plan is realised by a report on the actual state of affairs (see figure 1). Figure 1 depicts the action structure of the speech action pattern of interactive planning (Koole and ten Thije 1994: 122). The flow-chart in figure 1 illustrates the recurrent structure of the pattern by which several judgements can be formulated about either the original or the alternative proposal before one of them is finally accepted by both author and critic. Moreover, on the basis of this flow-chart, the difference between a single pattern position and distinct speech actions must be pointed out. For instance, a proposal for a plan might be realised by one or more assertions, or even by a combination of a suggestion and an assertion whereas the acceptance of the plan might be realised either by an assent or an assertion as well. By differentiating between the analysis of the illocution of the single speech action on the one hand and the quality of the pattern position on the other, the contribution of single speech actions to the purpose of the overall pattern can be determined. The pattern of reporting has the purpose of assessing the institutional relevance of transmitted knowledge elements with the aim of making common decisions. This pattern is also characterised by the dichotomy between a reporter and an assessor (Koole and ten Thije 1994: 108). The purpose of the pattern is achieved when both accept the relevance of the transmitted knowledge for the actual speech situation. The pattern positions of the reporter are the identification, the representation and the judgement of a certain state of affairs. The pattern positions of the assessor are inquiring about and subsequently accepting the transmitted knowledge. This pattern is also recurrent in that a sequence of speech actions, i.e. a sequence of assessments and questions, can pass by before both reporter and assessor agree and accept the relevance of the transmitted knowledge. In the following, we reconstruct the discourse in our case study at the GoetheInstitute Amsterdam. We first focus on the standard traversal of these two institutional patterns in the team discussions before proceeding by analysing receptive multilingual peculiarities in this context. The question whether the realisation of the speech action pattern as introduced above (figure 1) is also applicable for the analysis of intercultural discourse has been raised before (Koole and ten Thije 1994) and was answered affirmatively. The focus of the present study is whether this also holds for intercultural discourse under the receptive multilingual mode. The following stretches of discourse were taken from the planning discussion of the language teaching team, in which the team discusses the agenda for the coming month. In this fragment the colleagues discuss an unfortunate overlapping of events within the institution. The team has the task of organising examinations on German language skills in which the teachers of the Goethe-Institute act as examiners. When going through the following month’s agenda, the team coordinator notices that the next examination overlaps the farewell party for a colleague. Therefore, the examiners could run the risk of missing the farewell party. The team coordinator’s observation of this overlapping initiates a planning discussion in order to solve this problem (see fragment 2).

83

84


Fragment 2. 210404: fare well party


85

86



If we reconstruct the traversal of the pattern of interactive planning, we notice that Bernd brings up the coincidence of the examination and the farewell party in segment 93 (s93). In the following, both colleagues assess the institutional relevance of this coincidence and agree on it having negative consequences. In s102–105, Gerard expresses a proposal for a plan by asserting: “Mijn plan . is . om . goed in de gaten te houden . dat ik die aantallen ook inderdaad VOOR Koninginnedag DOORkrijg, Dus voor eind van de komende week. Dan moeten we weten hoeVEEL . kandidaten. Want dan weten we ook hoeveel mensen we nodig hebben voor dat mondelinge examen, hoeveel teams daar moeten zijn..” (‘My intention is to keep well in mind that I receive the exact number of participants indeed before Queens’s day…That is, before the end of the coming week. Then, we have to know how many candidates.’). A four second pause (s106) follows, which can be considered as an acceptance of Gerard’s proposal. Gerard specifies his proposal by uttering (in s107): “Dan kunnen we hier meteen intern organiseren wie daar naartoe gaan” (Then we can immediately organise internally who will go there.) Thereupon Bernd reacts (s108) with the Dutch question “Meteen?” (‘at once?’). Gerard explains his proposal by saying (in s109) “Ja, RUIM van tevoren, (zodat mensen) zich daar echt op in kunnen stellen.” (Yes, long before so that everybody here can take that into account). After a pause (s110), which can be interpreted as an acceptance on the part of Bernd, Gerard notes down the deadline for subscription in the team agenda (s111–113). This marks the fulfilment of the pattern purpose, the achievement of a consensus. Fragment 2 is a typical example of the discourse which takes place in the language teaching team at the Goethe-Institute. It is striking that the patterns of interactive planning and reporting are realised so frequently. Within these patterns, the German expatriate and the Dutch local both take up the positions of the author and the critic. On the basis of this equal distribution of the positions, we can conclude that the two colleagues have a symmetric institutional relationship. It is furthermore quite striking that the code switching between Dutch and German seems to coincide with the turn taking. As we observed in fragment 1, the Dutch proposal for a plan is evaluated in German. In fact, the traversal through the pattern is not realised in one language. Therefore, the question arises as to whether one action pattern of interactive planning functions as a kind of underlying action structure for both realisations in German and in Dutch, or whether these action patterns are language-specific. Seen as both actants have a high competence in both languages and are experienced in both cultures, we cannot answer this question just on the basis of this case study. We can, however, conclude that the realisation of the purpose of these patterns is achieved in the receptive multilingual mode. This language constellation has become the standard for multilingual communication in the language teaching team. In the next paragraph, we will focus on one linguistic means used by the interactants to ensure (a more or less unproblematic) understanding within the receptive multilingual mode.

87

88


5. Institutional keywords In the following, we will elaborate on institutional keywords, one of the special means that are applied by the colleagues Bernd and Gerard at the Goethe-Institute Amsterdam in order to ensure mutual understanding. As Koole and ten Thije (1994) reveal, institutional keywords are an important prerequisite for institutional cooperation. They discuss three characteristics of the relationship between keywords and cognitive structures. First, “institutional keywords represent shared knowledge of qualified institutional actors” (ibidem; 140). Second, “the cognitive structure connected with institutional keywords is an abstract frame whose slots are filled with concrete knowledge elements. The abstract frame and the concrete knowledge elements are both interactively produced” (ibidem 140). Third, “the elements of knowledge are structured according to the institutional purpose of the keyword” (ibidem; 141). Keywords actualise common institutional knowledge and, consequently, make it easier to establish mutual understanding. Keywords help to characterise the institution in question. They can be composed of words or phrases that only exist within the particular institution, but can also contain vocabulary that is used outside the institution as well. Within the internal communication of the organisation, however, these keywords receive a special, institution-specific meaning. Table 1 contains a list of keywords used at the Goethe-Institute Amsterdam. As can be seen in the table, most keywords are German notions or abbreviations, which are used both by the German and by the Dutch team members in all the departments of the Goethe-Institute in Amsterdam. A possible explanation for this is that the GoetheInstitute is a German organisation operating on a worldwide basis. If the employees wish to refer to central aspects regarding the general institutional knowledge, it is most likely that they will choose German keywords. The frequent use of German keywords, therefore, influences the receptive multilingual communicative mode, meaning that the Dutch employees have to adapt to the foreign language more often. Some keywords such as intensiv/intensief, Saal/zaal und Sommer/zomer, however, are cognates in Dutch and German and can therefore not be precisely attributed to one of the two languages. The same holds true for abbreviations, which in a lot of cases can be derived from either of the two languages. The keywords in table 1 refer to important institutional knowledge. In fact, these words instantly activate quite specific institutional knowledge. By this procedure efficient teamwork is facilitated. In the following, we will illustrate the use of keywords by discussing an excerpt from the case study by Roelands (2004). The excerpt shows the use of a Dutch keyword by the German employee Bernd. Although some instances of code switching can be found, the discourse takes place in the receptive multilingual mode with both speakers using their first language.


Table 1. Exemplary list of keywords German keywords

Team-specific meaning

Dutch keywords

Team-specific meaning

Kurs 1–7 Niveau 1–7 N5/6

Course on seven different levels

Toetsje

Language-assessment test

Kinderkurs

German course for children

Z1, Z3, Z4

Courses on business communication

Intensivkurs

Two-week crash course

Zaterduitsdag

Saturdays German training

Sommerkurs

Summer course

Zaalagenda

Memo pad for room reservations

Wochenplan

Week planning for the whole institute

TestDaf

Test for German as foreign language

ZOP

Zentrale Oberstufen prüfung (central assessment for upper school)

PWD-Protokoll

Prüfung Wirtschaftsdeutsch International (record on the assessments for international business communication)

EDV

Elektronische Datenverarbeitung (data processing)

KDS

Kleines Deutsches Sprachdiplom (First German language certificate)

GDS

Großes Deutsches Sprachdiplom (Advanced German language certificate)

89

90


The Dutch keyword Toetsje In the excerpt ‘toetsje’ (see fragment 3), the employees Bernd and Gerard discuss the situation of a student who has requested to participate in a language course without having undertaken the required language assessment. The word ‘toetsje’ is an ordinary Dutch word that, however, activates specific institutional knowledge when mentioned by the team members of the Goethe-Institute. It refers to a short oral test in which the language competence of a student can be determined on the basis of a spontaneous conversation between an applicant and a Goethe-Institute employee. In a feedback interview, Gerard explained how the word came into use in the institute’s external communication with possible future Dutch students. The word ‘toetsje’ is used in its diminutive form in order to make it sound less threatening for a future student to undergo an examination. A word like ‘test’ or ‘examination’ would most likely frighten people pointlessly. In fact, diminutives are a typical Dutch morphological phenomenon. After having proven its function in external communication, the word ‘toetsje’ has consequently become an institutional keyword also in internal communication. Both the German and the Dutch employee are aware of its specific institutional meaning. The German team coordinator therefore uses the Dutch keyword in his German discourse. The word ‘toetsje’ which originates from external communication has thus succeeded in becoming a characteristic of the internal communication at the GoetheInstitute. Whereas the word initially only had an every day meaning, it has turned into an institutional keyword for the language teaching team (see fragment 3). The keywords we found are, apart from a few exceptions, used in one language by both interactants. In order to make sure that the same institutional knowledge is activated, the participants thus change the language code when using institutional keywords. The translation of institutional keywords is not appropriate, since a translated word may imply a change of meaning. Only when the keywords are applied in one language, all employees can be certain that the same institutional knowledge is activated. With respect to keywords, two types of code switching have been defined in the literature, namely intersentential and intra-sentential code switching (Milroy and Muysken 1995). Intersentential code switching does not occur within the contributions of the participants. However, it is found at every change in turn, since, in the case of receptive multilingualism, every switch of turn also results in a switch of code as illustrated above. Intrasentential code switching can be found more often in the discourse of the Dutchman, since he has to integrate more German keywords into his speech than his German counterpart. The analysis carried out reveals that institutional keywords enable and facilitate a certain understanding between the participants in the receptive multilingual mode. The peculiarity of keywords in receptive multilingualism is that these words do not follow the standard rules for language choice. Although the speakers choose to com-


Fragment 3. 010304: toetsje

municate in their first language, they code switch when using institutional keywords. Since these keywords inevitably activate institutional knowledge, the speakers decide not to endanger the mutual understanding by translating the keywords into their first language. Institutional keywords activate the same institutional knowledge among all interlocutors. Therefore, they help to build a common basis that is especially needed in intercultural and multilingual communication, where misunderstandings are far more likely to occur than in monolingual and monocultural discourse.

91

92


6. Receptive multilingualism and intercultural discourse Finally, we would like to discuss the relation between institutional discourse and intercultural communication (Bührig and ten Thije 2006; ten Thije 2002, 2003). This relation can best be illustrated by means of an analysis of another fragment from our corpus from the Goethe-Institute. In fragment 4, Bernd and Gerard discuss which deadline should be decided on for the enrolment for a test that is to be carried out at the Goethe-Institute. Bernd enters the planning mode by saying “So, dann PWD, wie sieht’s denn da aus mit der Terminleiste, Anmeldung ist bis circa zwei Wochen van tevoren” (“Well, coming to PWD, what about the list of dates, enrolment is possible until approximately two weeks before”). Again, an institutional keyword is used, namely PWD, meaning ‘Prüfung Wirtschaftsdeutsch’ (Test on German for economic purposes). The keyword PWD is based on the employees’ specific institutional knowledge. We also observe a code switch within Bernd’s utterance. He begins the fragment in his mother tongue German, but then switches to Dutch. By means of this intrasentential code switch to ‘van tevoren’ (‘before’), he most likely signals that he is quoting a Dutch colleague. If we read through the planning pattern that underlies this discourse, it becomes fairly obvious that the two colleagues do not agree on which deadline should be given for the exam. Bernd suggests (segment 5) that they should not be too strict on the deadline, whereas Gerard insists on handling the deadline quite strictly. According to him, enrolment shouldn’t be possible anymore after May 10 th. The planning pattern comes to an end in segment 113 when Bernd accepts Gerard’s proposal by saying: ‘Ja, gut’ (‘Yes, that’s ok.’). With respect to the intercultural apparatus at work, we would like to draw attention to the segments 108 to 114, in which Bernd states: “Die soll tien mei kommen. Eind van de maand. Einunddreißigster? April. Das sind zehn. Ach so. Stimmt. Wir sagen immer zwei Wochen. Ja gut. Hasse Recht.” (They should be there by May 10th. End of the month. April 31th. That’s 10. Oh, yes. That’s true. We always say two weeks. Okay, you’re right.) Now consider fragment 4. According to Hoffmann (1997), expressions such as ‘day’, ‘week’, and ‘year’ belong to the so-called metrical system. These expressions refer to consistent time intervals. Throughout the fragment, the actual range of these time intervals remains unclear. It finally becomes evident from the discourse that the two weeks discussed do not stand for 14, but rather ten days. A common fact, that would seem to be self-evident, is thereby put into question. This difference in the perception of everyday facts could result from cultural knowledge, as will be argued in the following. One might expect the misunderstanding regarding the length of the time interval to be solved with the use of the word ‘circa’ (approximately), but there are other hints indicating that the discourse is structured interculturally. We can observe an interesting contrast with respect to the use of the personal deictic procedure we/wir (we) in the segments 105 and 112. In segment 105: “Machen we es immer noch ganz locker”,


93



Fragment 4. 130504: vantevoren  





 



 

 





   







   

     















 



    



 



 





   

 



   











   



 

























    













 

    

 

 

















                

 





  



  



  



 











 

 



   





94


the Dutch (or colloquial German) deictic procedure ‘we’ refers to the group of speakers that includes both the speaker and the hearer and maybe even other employees of the Goethe-Institute as well. Bernd refers to a general maxim in the Goethe-Institute, namely that deadlines are generally treated fairly loosely. Bernd’s utterance could also be understood as an allusion to the proverbial Dutch informality (see e.g. Shetter 1997). In segment 112, the disagreement on the exact duration of the two weeks is brought up by Bernd. He utters: “Wir sagen immer zwei Wochen” (“We always say two weeks”). The personal pronoun ‘wir’ in this case refers to a group to which the speaker, but not the hearer belongs (‘exclusive we’). In this context, the expression ‘two weeks’ can be interpreted as a language-specific formula used by Bernd. Bernd refers to the German language community to which the hearer Gerard does not belong. This contrasting of the two groups of speakers with respect to one group-specific maxim (cf. the combination of we/wir and immer (always)) indicates that the institutional discourse is structured interculturally. Bernd’s remark suggests that he assumes that the Dutch and the German metric temporal system apparently work in a different manner. He furthermore interprets the misunderstanding relating to the time interval of two weeks as a cultural difference. Thereby, he indirectly refers to the concepts illustrated in ‘The silent Language’ by Stuart Hall (1959, 1981), which highlights the different time standards in various cultures. Gerard does not join this discussion. He acknowledges the elaborations made by his colleague by uttering the interjection ‘hm’ twice. The question now arises as to how we can account for (the discursive elaboration of) this cultural difference within a discourse-theoretical framework. For this purpose, we would like to introduce Rehbein’s (2006) concept of a Cultural Apparatus. Cultural apparatuses are discursive instruments that start to come into play when apparently self-evident cultural facts are put into question, or when problems or intercultural misunderstandings arise and remain (first and second position of the cultural apparatus, figure 2). After having reflected on a discourse problem, the interactants can either change their standard solutions and perceptions (fourth position), or stick to their original cultural standards. The outcome becomes visible in the post-history (fifth position).

Scheme of the cultural apparatus (Rehbein 2006: 74): (i)

t he underlying form of standard action (with action systems) contains a problem; (ii) negating action, which consists of two parts, suspension of discourse and identification of the problem, calls the presuppositions of (one or more of) the participants into question; (iii) there follows a (partial) restructuring of mental processes such as action practices, thought structures, forms of imagining and/or social experiences (τέχναι) (”third position”);


(iv) t his results in new forms of (speech) action through an alteration of the relationship between surface and deep structure, i.e. through the act of reflection (”fourth position”); (v) finally the post-history of the apparatus manifests with consequences etc. We would like to put forward the hypothesis that the segments 108 to 114 in fragment 4 contain the application of the cultural apparatus, leading us to the question of what happens in the fragment. Bernd reflects on Gerard’s suggestion concerning the deadline for the enrolment for the test. It is interesting to observe that he identifies the data in Dutch. He states: “tien mei” and “eind van de maand” and therefore concludes that the closing date for the deadline should be April 31th. In s109 he utters ‘das sind zehn’ (that is ten). Speaking in terms of the cultural apparatus, there is a problem being formulated. In this case, it is related to the precise interpretation of the Dutch temporal expression ‘van tevoren’ (‘before’) which was used by Bernd. The question arises as to which time interval ‘van tevoren’ exactly covers (first position). In the following, a discussion on the precise handling of the deadline for the enrolment for the test (second position) arises. After that, we can observe how knowledge is restructured with respect to the mentioned interval of time by at least one interactant, in this case Bernd. This could be interpreted as the third position of the cultural apparatus with the actual restructuring of knowledge as a result being the fourth position. However, the question arises as to whether we are really dealing with cultural knowledge that is being restructured, or whether we are merely facing a misunderstanding that could also have occurred in mono-cultural discourse. We will argue in the following, that at least Bernd interprets the misunderstanding as being due to intercultural miscommunication. He considers the German and the Dutch interpretation of the term ‘two weeks’ to be different and thus responsible for the problem. Bernd reflects on the exact period of two weeks (i.e. fourteen days) in an exothesis5 (Ehlich and Rehbein 1986) which he ends with: ‘ach so’. He then explains what the misunderstanding was based on according to him (‘Stimmt’ (s111) That’s true. ‘Wir sagen immer zwei Wochen’ (s112). We always say two weeks.). He finally accepts Gerard’s planning suggestion by uttering: Ja, gut. (Yes, okay). Hasse recht. (You’re right). If we take a look at the post-history (the fifth position of the cultural apparatus), it is interesting to find that Bernd (once again) asks what the present date is. In segment 116, he notices that the present date is April 21st, which is confirmed by Gerard in segment 117 by him repeating it ‘eenentwintig’ (21st) in Dutch. The cultural apparatus has analysed the different weighing of time as a potential misunderstanding, which results in a successful traversal of the institutional planning pattern. The receptive multilingual mode of interaction is re-established back to the institutional team standard. With respect to the general cultural system of Dutch and German presuppositions, it is striking that the two colleagues typify the exact opposite of the cultural stereotypes of the German and the Dutch (see e.g. Westheide 1997). The German proposes to handle the deadline for the enrolment quite loosely, which by no means corresponds to

95

96


the cultural stereotype of the German punctuality/precision. The Dutchman, on the other hand, wants the deadline to be handled precisely, which also does not correspond to the stereotype of Dutch informality. It is possible that the interactants adjust themselves to their perception of the other’s culture. This, however, would have to be investigated in more detail in future analyses.

7. Discussion and conclusions At the end of this paper we discuss the specific characteristics of the case study in the Goethe-Institute Amsterdam and formulate future research possibilities. We structure these considerations according to the three perspectives we introduced in section 2, namely the language contact, the institutional constellation and the interactants’ perspective. From a language contact perspective, this case concerns language contact between two closely related languages. A comparison of this case to the various situations in other Goethe-Institutes situated all over the world would make it possible to learn more about multilingualism among speakers of different languages. If other languages and nations are considered, the impact of the socio-political status of the languages and nations on the occurrence of receptive multilingualism could be analysed in more detail as well as the impact of the receptive multilingual mode on the general attitude of the speakers of the one language towards the speakers of other languages. Moreover, the changing attitudes of the neighbour countries towards Germany could be studied (Fremdling et al. , 1992; Trautmann 1991) along with the changes within small language communities in Europe (Arntz 1997). In this context, receptive multilingualism may even lead to a reduction of the use of dialects in border regions (Hinskens 1993). With respect to the second, institutional, perspective we would like to state that our observed case exemplifies a language contact type that Ehlich (1992) denotes as ‘Intektion’. With this neologism Ehlich refers to the type of language contact that comes into being when a new super-structure is constituted above two or more nation states involved such as the European Union. This language constellation is referred to as ‘sophisticated multilingualism’ (ibidem: 68). In actual fact, this multilingual communication results from the establishment of supra-national institutional structures corresponding to European unification. In this context a comparison with, for instance, the other institutional communication in German–Dutch contact would be very interesting. Since 1999, the German–Dutch border cities Dinxperlo and Suderwick have, for instance, established a joint police station that is in charge in both countries (ten Thije 2003). In this context, multilingual communication is an essential prerequisite in order to realise the main institutional purposes. Whereas the Goethe-Institute case concentrates only on internal communication, the study of this police station might expand our knowledge about the specific institutional preconditions applying to this communicative mode in external multilingual constellations (Clyne 2003).


Another restriction relating to the institutional perspective concerns the specific purpose of the Goethe-Institute. The Goethe-Institute aims at promoting German language and culture abroad. Therefore, this organisation might be generally more willing to develop alternative modes of multilingual communication. We have shown that the language teaching team realises its mean institutional purposes efficiently using the receptive multilingual mode. By expanding this case study to other Goethe-Institutes we could not only measure the effect that factors of linguistic and cultural distance between the two language communities involved have on the occurrence and development of receptive multilingualism, but we could at the same time shed light on the effect of the language policy of the Goethe-Institute on the completion of the action space in local institutes. Finally, from the interactants’ perspective, we wish to stress the fact that in this case the language competence of both interactants was comparatively high and, consequently, the receptive multilingual mode was just one of the alternatives they were able to apply within their team communication. They always had the opportunity of switching to one of the two languages involved. It would be interesting to focus on interactants with different second language skills in order to detect the minimal language competence that is needed to perform receptive multilingualism. The restrictions mentioned above prevent us from formulating general claims about receptive multilingualism that go beyond its implementation in the Goethe-Institute in Amsterdam. The perspectives for receptive multilingualism beyond the Goethe Institute have to be elaborated on in future research.

Notes * This chapter is an English elaboration of the German articles by Roelands and ten Thije (2006) and Ribbert and ten Thije (2006). We would like to thank the Goethe-Institute in Amsterdam for participating in our research. The names of the participants have been changed. We also thank Nicky Burke and two anonymous reviewers for their valuable comments on a previous version of this article. 1. The Dutch Rudy Carrell and Linda de Mol became famous as TV-presenters/actors in Germany. 2. In a recent manual the B2 level is described as follows: “Can use the language fluently, accurately and effectively on a wide range of general, academic, vocational or leisure topics, marking clearly the relationships between ideas. Can communicate spontaneously with good grammatical control without much sign of having to restrict what he/ she wants to say, adopting a level of formality appropriate to the circumstances” and “Can interact with a degree of fluency and spontaneity that makes regular interaction, and sustained relationship with native speakers quite possible without imposing strain on either party. Can highlight the personal significance of events and experiences, accounts for and sustain views clearly by providing relevant explanation and arguments.” (Council of Europe, 2003: 56)

97

98

Anne Ribbert and Jan D. ten Thije 3. In a recent manual the C2 level is described as follows: “Has a good command of idiomatic expressions and colloquialisms with awareness of connotative levels of meaning. Can convey finer shades of meaning precisely by using, with reasonable accurancy, a wide range of modification devices. Can backtrack and restructure around a difficulty so smoothly the interlocutor is hardly aware of it.” (Council of Europe 2003: 56) 4. The article ‘Discourse-Pragmatic Description’ by Bührig and ten Thije (2005) contains a survey of the different approaches to discourse and conversation analysis. 5. Ehlich and Rehbein (1986) define exothesis as the verbalization of a mental state of affairs.

References Arntz, R. 1997. Passive Mehrsprachigkeit — Eine Chance für die „kleinen Sprachen“ Europas. In Einsprachigkeit ist heilbar. Überlegungen zur neuen Mehrsprachigkeit Europas [Socio linguistica 11]. U. Ammon, K. J. Mattheier and P. H. Nelde (eds), 166–77. Tübingen: Max Niemeyer. Aspeslagh, R. 1998. Politik, Staat und Bürger: 1994–1998. In Dokumentation für die dritte Swedish–Danishändische Konferenz. Delft, den 19. und 20. März 1998, 4–11. Clingendael: Instituut voor Internationale Betrekkingen. Auer, P. (ed.) 1998. Code-Switching in Conversation. Language, interaction and identity. London: Routledge. Bot, K. de 2004. Applied linguistics in Europe. Aila Review 17(1): 57–68. Bourdieu, P. 1992. Language and Symbolic Power. Cambidge: Cambridge University Press. Braunmüller, K. and Zeevaert, L. 2001. Semikommunikation, rezeptive Mehrsprachigkeit und verwandte Phänomene. Eine bibliographische Bestandsaufnahme [Arbeiten zur Mehrspra chigkeit, Folge B 19] Hamburg: Universität Hamburg, Sonderforschungsbereich Mehr sprachigkeit. Bührig, K. and Thije, J. D. ten 2005. Diskurspragmatische Beschreibung. In Sociolinguistics– Soziolinguistik. An international handbook of the science of language and society, Vol. 2, U. Ammon, N. Dittmar, and K. Mattheier (eds), 1225- 1250. Berlin: Mouton. Bührig, K. and Thije, J. D. ten (eds). 2006. Beyond Misunderstanding. Linguistic analyses of intercultural communication [Pragmatics and Beyond New Series 144]. Amsterdam: John Benjamins. Clyne, M. 1991. Community Languages: The Australian experience. Cambridge: Cambridge University Press. Clyne, M. 2003. Towards inter-cultural communication in Europe without linguistic homogenization. In Die Kosten der Mehrsprachigkeit. Globalisierung und sprachliche Vielfalt; The Cost of Multilingualism. Globalisation and Linguistic Diversity, R. De Cillia, H. J. Krumm and R. Wodak (eds), 39–49. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Council of Europe 2001. Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge: Cambridge University Press. Council of Europe 2003. Relating Language Examinations to the Common European Framework of reference for languages: Learning, teaching, assessment (CEF). Manual Preliminary Pilot Version. Strasbourg: Langaue Policy Division. http://www.coe.int/T/DG4/Portfolio (last visited December, 12th 2006)


CPB, Netherlands Bureau for Economic Policy Analysis 1997. Challenging Neighbours. Rethinking German and Dutch economic institutions. Berlin: Springer. Delsen, L. and Jong E. de 1998. The German and Dutch Economies. Who follows whom? Heidel berg: Physica-Verlag. Ehlich, K. 1992. Kommunikationsbrüche, Von Nachteil und Nutzen des Sprachkontakts. Ziel sprache Deutsch 23 (2): 64–74. Ehlich, K. and Rehbein, J. 1986. Muster und Institution, Untersuchungen zur schulischen Kommu nikation. Tübingen: Gunter Narr. Finkenstaedt, T. and Schröder, K. 1992. Sprachen im Europa von morgen. Berlin: Langenscheidt. Fremdling, R. , Lodder, H. , Wagenaar, R. and Wielenga, F. (eds). 1992. Die überwundene Angst? Die neun Nachbarländer und die deutsche Einheit. Düsseldorf: Landeszentrale für politische Bildung Nordrhein-Westfalens. Grosjean, F. 1982. Life with Two Languages. An introduction to bilingualism. Cambridge: Harvard University Press. Hall, E. T. 1981. The Silent Language. 2nd ed. , 1st ed. 1959. New York NY: Anchor Books. Herrlitz, W. 1997. Ist Befehl Befehl? Beobachtungen zur interkulturellen Kommunikation im deutsch–niederländischen Korps in Münster. In Niederländer und Deutsche und die Europäische Einigung, F. Wielenga (ed.), 66–70. Bonn: Niederländische Botschaft. Herrlitz, W. and Loos, E. 1994. Taal en taalgebruik in de internationale bedrijfscommunicatie: Een taalwetenschappelijk perspectief. Mens en Ondernemen 2: 144–60. Hinskens, F. 1993. Dialect als lingua franca? Dialectgebruik in het algemeen en bij grensoverschrijdend contact in het Nederrijnland en Twente. In Diglossienstudien: Dialekt und Standardsprache im niederländisch-deutschen Grenzland. L. Kremer (ed.), 209–45. Vreden: Landeskundliches Institut Westmünsterland. Hoffmann, L. 1997. Zur Grammatik von Text und Diskurs. In Grammatik der deutschen Sprache, Vol 1. G. Zifonun, L. Hoffmann and B. Strecker (eds), 98–594. Berlin: de Gruyter. House, J. and Rehbein, J. 2004. What is ‘multilingual communication’? In Multilingual Communication, J. House and J. Rehbein (eds), 1–18. Amsterdam: John Benjamins. Jongste, H. M. de 1997. Aspekte der deutsch–niederländischen Wirtschaftskommunikation. In Wirtschaftswissenschaft, Anwendungsorientierte Forschung an der Schwelle des 21. Jahr hunderts. U. Gröner, H. M. de Jongste, U. Kracke and H. Senne (eds), 43–56. Heidelberg: Decker. Kellerman, E. 1979. Transfer and non-transfer. Where we are now. Studies in Second-language acquisition 2: 37–57. Koole, T. and Thije, J. D. ten 1994. The Construction of Intercultural Discourse. Team discussions of educational advisers. Amsterdam: RODOPI. Lademacher, H. 1989. Zwei ungleiche Nachbarn. Wege und Wandlungen der deutsch–nieder ländischen Beziehungen im 19. und 20. Jahrhundert. Darmstadt Wissenschaftliche Buch gesellschaft. Linthout, D. 2000. Onbekende Buren. Amsterdam: Atlas. Loos, E. 1997. Internationale Bedrijfscommunicatie. Reconstructief onderzoek naar het inter tekstuele netwerk van Nederlandse en Duitse actoren in een bungalowpark. Utrecht: Utrecht University. Loos, E. 1998. Erfolgreiche deutsch–niederländische Betriebskommunikation: Wahrheit oder Dichtung? Wirtschafts-Deutsch International. Zeitschrift für sprachliche und interkulturelle Wirtschaftskommunikation 1: 142–50.

99

100 Anne Ribbert and Jan D. ten Thije

Milroy, L. and Muysken, P. (eds) 1995. One Speaker, Two Languages. Cross-disciplinary perspectives on code-switching. Cambridge: Cambridge University Press. Müller B. and Wielenga, F. (eds) 1995. Kannitverstan? Deutschlandbilder aus den Niederlanden. Münster: agenda. Nelde, P. H. 1988. Dutch as a language in contact. International Journal of the Sociology of Language 73: 111–19. Prescher, P. 2006. Use it or lose it — First-language attrition and the “In-Between” position of Bilingual Immigrant. In Artikelen van de vijfde sociolinguistische conferentie. T. Koole, J. Nortier and B. Tahitu (eds), 377–92. Delft: Eburon. Raven, S. and Aspeslagh, R. (eds) 1997. Multikulturelles Miteinander – Eine Utopie? Die Integra tion von Ausländern in den Niederlanden und in Deutschland. Clingendael: Instituut voor Internationale Betrekkingen. Rehbein, J. 1977. Komplexes Handeln. Elemente zur Handlungstheorie der Sprache. Stuttgart: Metzler. Rehbein, J. 2001. Konzepte der Diskursanalyse. In Text- und Gesprächslinguistik/Linguistics of Text and Conversation. Ein internationales Handbuch zeitgenössischer Forschung/An Inter national Handbook of Contemporary Research, K. Brinker, G. Antos, W. Heinemann, and S. Sager (eds), 927–45. Berlin: de Gruyter. Rehbein, J. 2006. The Cultural Apparatus. Thoughts on the relationship between language, culture, and society. In Beyond Misunderstanding. Linguistic analyses of intercultural communication, K. Bührig and J.D. ten Thije (eds), 43–97. Amsterdam: John Benjamins. Ribbert, A. and Thije, J. D. ten 2006. Rezeptive Mehrsprachigkeit und interkulturelle Kommuni kation am Arbeitsplatz. In Mehrsprachigkeit am Arbeitsplatz [Forum für Angewandte Linguistik 48], B. Meyer and S. Kameyama (eds), 109–29. Frankfurt: Lang. Roelands, A. 2004. Rezeptive Mehrsprachigkeit in der institutionellen Kommunikation. Eine Fallstudie der deutsch–niederländischen Kommunikation im Goethe-Institut Amsterdam. Utrecht: Utrecht University. Roelands, A. and Thije, J. D. ten 2006. Rezeptive Mehrsprachigkeit in der institutionellen Kommunikation. Eine Fallstudie zur deutsch–niederländischen Kommunikation im Goethe-Institut Amsterdam. In Mehrsprachige Individuen – vielsprachige Gesellschaften [Forum Angewandte Linguistik 47], D. Wolff (ed.), 165–76. Frankfurt am Main: Lang. Shetter, W. Z. 1997. The Netherlands in Perspective. The Dutch way of organizing a Society and its setting. Utrecht: Nederlands Centrum Buitenlanders. Swaan, A. de 2001. Words of the World. The global language system. Cambridge: Cambridge University Press. Thije, J. D. ten 2002. Stufen des Verstehens bei der Interpretation von interkulturellen Diskursen. In Kultur(en) im Gespräch, H. Kotthoff (ed.), 61–98. Tübingen: Narr. Thije, J. D. ten 2003. Eine Pragmatik der Mehrsprachigkeit. Zur Analyse diskursiver Interkulturen. In Die Kosten der Mehrsprachigkeit. Globalisierung und sprachliche Vielfalt; The Cost of Multilingualism. Globalisation and Linguistic Diversity, R. De Cillia, H. J. Krumm and R. Wodak (eds), 101–23. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Trautmann G. (ed.) 1991. Die häßlichen Deutschen? Deutschland im Spiegel der westlichen und östlichen Nachbarn. Darmstadt: Wissenschaftliche Buchgesellschaft.

Receptive multilingualism in team cooperation 101

Westheide, H. 1997. Trügerische Nähe, Niederländisch-deutsche Beziehungen in Geschichte, Sprache und Kultur [EUROPA 2000 Studien zur interdisziplinären Deutschland- und Europaforschung 16] Münster: LIT Verlag. Zahn, E. 1984. Das unbekannte Holland. Regenten, Rebellen und Reformatoren. Berlin: Siedler Verlag. Zahn, E. 1986. Die Niederländer, die Deutschen - ihre Geschichte und ihre politische Kultur [Saarbrückener Universitätsreden 18]. Saarbrücken: Universität des Saarlandes.

chapter 4

Receptive multilingualism and inter-Scandinavian semicommunication Ludger Zeevaert Universität Hamburg

This article is aimed at providing a theoretical subsumption of the term receptive multilingualism. Receptive multilingualism is seen as a communicative practice used in situations of verbal interaction under special circumstances. It is characterised by different languages being used by the different discourse participants. This overview is illustrated with examples taken from interscandinavian semicommunication, i.e. the oral communication between speakers of Danish, Norwegian and Swedish. These examples show that the crucial difference between ‘usual’ and receptive multilingual communication lies in the fact that the discourse participants belong to different speech communities. They apply different linguistic and non-linguistic frames of reference and are thus unable to rely on an unconditional functioning of the communication. The larger the differences between the languages, the more probable is the appearance of trouble sources that can be overcome with different methods. Keywords: receptive multilingualism, interscandinavian semicommunication, speech community, trouble sources

1. Introduction The verbal exchange between speakers of one speech community is commonly seen as the default form of human linguistic communication. Putting aside cases in which the mutual understanding is impeded by extralinguistic factors (noise, hearing impairment, speaking impairment etc.), communication usually functions without problems due to processes of mutual linguistic adaptation. A speech community is here defined as a geographically and socially homogeneous group of human individuals who have direct linguistic contact with each other. In order to communicate this speech community uses a system of linguistic utterances, their language. Concerning phonology, morphology, lexicon, syntax and discourse the utterances of the individuals within this speech community are identical and comply with a conventionalised norm. The formal identity of the utterances is necessary in order to enable mutual understanding. New members of such a group (foremost children) learn this language interactively through their linguistic contact to other members of the speech community. The linguistic norm of the speech community is handed down to the next generation by

104 Ludger Zeevaert

means of steady linguistic interaction. Up to a certain amount innovations are possible as long as the speech community is able to deal with them by processes of mutual adaptation.1 The following article deals with a situation of communication that deviates from this normal form of communication, viz. the linguistic communication between members of different speech communities using different languages is analysed. In this special case the missing concordance of the languages involved has to be compensated for by extra negotiation if the communication is to be successful. In order to accomplish this task different solutions are possible dependent on the conditions implied: • A n interpreter knowing both languages mediates between the interlocutors. • Both interlocutors use an L2 (lingua franca), most often the language of a politically influential nation such as nowadays English, but also languages that are widespread for cultural or religious reasons in certain areas of the world like Latin in Europe in the middle ages or Standard Arabic in the Islamic countries may be used. Even written standard languages can play the same role, as is the case with High German used by speakers of German dialects that show a large geographical and linguistic distance. • Artificial languages such as Esperanto or naturally emerged pidgins as Tok Pisin in New Guinea are used. Another possibility is the application of a more widely known sign language as the Plains Indian Sign Language used by North American Indians or the sign language developed by Trappist monks to avoid disturbances in their contemplation caused by the noise of spoken language. • One of the interlocutors speaks the language of a different speech community as a foreign language and uses it when in contact with its members. In cases of a less proficient command of the language even a special register (foreigner talk) may be used. • The interlocutors try to reach a unilateral or bilateral linguistic approximation, e.g. by utilising a rudimentary language competence and by means of a deictic or even onomatopoetic denomination of objects. • Both interlocutors speak their own language and at the same time are able to understand the language of their counterpart (receptive multilingualism). The following considerations are based on the last-mentioned case of multilingual communication, viz. receptive multilingualism. Receptive multilingual discourses are characterised by the fact that the interlocutors involved do not aim at a common language of discourse, exactly like in cases of interpreter-mediated discourse and in contrast to all other examples of multilingual discourse constellations. However, while in the case of interpreting the communication is mediated by a third party, receptive multilingualism requires this linguistic transfer to be achieved by the interlocutors themselves. The success of such a transfer can be very different depending on the different conditions of the receptive multilingual situations. Yet a more thorough theoretical de-

Receptive multilingualism and semicommunication 105

scription of those different cases of receptive multilingualism was not performed. At a first glance, this retentiveness in relation to a theoretical penetration of the matter of investigation is not self-evident, given the large amount of literature on the subject.2 An explanation might be that a definition of receptive multilingualism would have to comprise a host of very different phenomena. In addition it has to be stated that a large amount of the existing literature on the different aspects of the subject aims at supporting the idea of receptive multilingualism and does thus not always strive for an objective description of the communication between the speakers of different languages.3 Thus, the first step that will be made in the following considerations is to differentiate between the numerous terms used in the literature in connection with situations of oral communication lacking a common language.

2. Receptive multilingualism: A theoretical overview 2.1 Receptive multilingualism and unrelated/remotely related languages Receptive multilingualism is a reasonable option of communication between languages that are unrelated or only remotely related — under the condition that all speakers involved are familiar with both languages, and provided that the speakers have only a passive competence at their disposal (either because only a passive competence was acquired, because one or more of the speakers are less advanced learners or because of a lack of language practice) or that the interlocutors prefer to use their own mother tongue in spite of an available active competence (either because they feel able to express themselves better in their first language or to mark their linguistic identity in a multilingual environment). Receptive multilingualism thus provides the opportunity to avoid linguistic discrimination in officially multilingual countries such as Switzerland or Belgium.4

2.2 Receptive multilingualism and closely related languages The mutual understanding of speakers of closely related languages, however, underlies totally different conditions. In the case related to in the following, speakers are able to understand the language of their interlocutor due to the genetic proximity of the two languages and the resulting large typological similarity. Following Haugen (1966: 153), this special case of receptive multilingualism has been termed semicommunication.5 The classical case of semicommunication is the communication between speakers of Danish, Norwegian and Swedish, but the term is also used for similar situations as in the communication between speakers of Czech and Slovak (Budovičová 1987a, 1987b) or between Middle Low German and Old Scandinavian (Braunmüller 1995).6 Semicommunication is attested for numerous language pairs, e.g. Czech–Polish (Hansen 1987), Croatian–Serbian (Haugen 1990), Hindi–Urdu (Haugen 1990),

106 Ludger Zeevaert

Icelandic–Faroese (Braunmüller and Zeevaert 2001), Portuguese–Spanish (Coseriu 1988: 140, Jensen 1989, Zeevaert 2002), Spanish–Italian (Hansen 1987), Frisian– Dutch (Feitsma 1986), Macedonian–Bulgarian (Haugen 1990) or Russian–Bulgarian (Braunmüller and Zeevaert 2001). Even the exchange between the different national dialects of Arabic may be called semicommunication (Haugen 1990). Klein (2004: 16f.) completes this list by adding Hebrew–Arabic, Turkish in Central Asia, unspecified languages in Africa and the Polynesian languages Maori, Samoan, Tahitian and Hawaiian.

2.3 Receptive multilingual communication and semicommunication Receptive multilingual communication and semicommunication are characterised by different communicative conditions and should thus be treated as different phenomena. The common characteristic of both cases is that different languages are used in one discourse. However, the underlying theoretical conception of the term language does not seem to be identical. Kloss (1967) distinguishes between abstand languages and ausbau languages. Abstand languages are languages that do not exhibit mutual intelligibility. This definition relies on the (objective) criterion of linguistic distance und thus only makes sense in comparison to other languages. In contrast to this an ausbau language may be intelligible for the members of neighbouring speech communities but at the same time can not be regarded as a dialect of the language of the neighbouring community because of its highly developed literacy. Objective criteria for the status of an ausbau language are among other things its own written standard and the existence of technical literature. Kloss’ (1967) term ausbau language is comparable to Coseriu’s concept of historical language. Coseriu founds his distinction between language and dialect on the evaluation by the speakers of the language or dialect under consideration or by the speakers of the respective neighbouring speech communities. For Coseriu (1988: 24) historical language exists as a historical cultural product and is recognised by its speakers as a language. A historical language is denoted by an adjective proper (the German, French, Italian language).7 Thus, the term receptive multilingualism can be used to describe the communication between speakers of abstand languages, but also between speakers of ausbau languages that exhibit little distance between them. The term semicommunication, however, is only appropriate in describing contact between closely related ausbau languages which has to be distinguished from cases of dialect contact such as the linguistic exchange over political borders between speakers of closely related dialects belonging to a dialect continuum, as for example in the German–Dutch borderland, or the linguistic exchange between speakers of different, less closely related varieties of minority languages in which an immediate mutual understanding is not possible (e.g. RhaetoRomance in Switzerland and Northern Italy or Saami in Norway and Sweden). At a first glance the utilisation of a purely linguistic criterion, i.e. linguistic distance or immediate mutual understanding between languages, seems to provide the best op-


portunity for a more strict definition of the different terms seen as the understanding between speakers of different speech communities is testable, for example by asking questions concerning the contents of a text.8 However, the evaluation of such tests poses several different questions: • W hich types of texts are suitable for testing language comprehension? • Which criteria should be applied in defining understanding? Did a test person understand a text only if all questions are answered correctly or is a certain percentage sufficient (and in that case, which percentage)? • Should a quantification of the degree of intelligibility rely on the numbers of test persons who pass the test or on the average results of the whole group?9 Even the two person communication game that was proposed by Ervin-Tripp (1971: 64) in order to test mutual intelligibility between different speech communities is not a suitable solution in our case. Her proposal implied that a hearer is instructed by a speaker not visible to him to perform a task, e.g. to assemble an item. However, it requires that the test persons involved did not have any kind of contact to the language of instruction seen as this would influence the receptive competence. Moreover, not the mutual understanding of languages is tested but rather the performance of an individual. Finally it has to be stated that situations, in which the success of a communicative act can be validated by a language external act as in Ervin-Tripp’s experiment, only account for a rather small part of human communication. In most cases an immediate control of success is not possible.

2.4 A consensual sphere as a condition for mutual understanding Communication models that are based on information theory such as those developed by Shannon and Weaver (1971: 34) or Eco (1988: 50)10 are suitable for describing the exchange of information between machines. However, they are only of limited use for the investigation of communication between humans, seen as human communication exceeds the pure exchange of information. Discussions about politics, the weather or a film seen recently by both interlocutors do not serve the purpose of transferring information from a speaker to a hearer. Following Ungeheuer, (1987b: 322) such cases of communication can be labelled as ‘crucial communication’ (kruziale Kommunikation). Its purpose is not the exchange of information, but far more the mediation of individually differentiated world theories11 or the transfer of mental contents from a speaker to a hearer. One model of describing ‘crucial’ communication is provided by the Chilean biologist Umberto Maturana. From his radical constructivistic point of view, he states that communication between nervous systems (e.g. humans) cannot be described as an exchange of information, because the systems involved are closed and therefore can only construct a view of the world outside the system including their interlocutor by using the neuronal activities produced by interfaces (senses of hearing, seeing, smelling etc.).

108 Ludger Zeevaert

Understanding would thus be the process of reconstructing the reality constructed by the nervous system in using such externally and internally generated neuronal activities.12 In this view the hearer’s role is as active as that of the speaker. The sense of an utterance is derived by a hearer by placing it in his own cognitive space. Knowledge has to be created by the hearer (Maturana 1998: 22). It is his own choice how to adjust his cognitive space by internally developing his own mental state. His choice is triggered by the “message”, but the orientation created is independent from the sense the “message” represents for the speaker. Strictly speaking there is no transfer of thoughts from a speaker to a hearer. Information is created by the hearer in reducing his own uncertainty by his interactions in his own cognitive space.13 The basis for the orientation of the hearer is his own reference system, viz. the state of activity caused by his interactions. This state forms the context for linguistic interaction (Maturana 1998: 60). This context is neither an objective nor a static variable. It changes in the course of a linguistic interaction. Ambiguity in utterances or misunderstandings that can be detected by an observer not involved in the communication do not exist for the interlocutors. Due to the current context, the meaning of an utterance is always unambiguous, even if this context may change as a result of the linguistic interaction or even by means of self reflection.14 This in turn implies the necessity of a consensual sphere common for both speaker and hearer, at least if communication is to be regarded as successful under the condition that the hearer creates the meaning actually intended by the speaker. On the one hand this consensual sphere comprises the language, and on the other hand the context, i.e. the cognitive frame of reference (cf. Maturana 1998: 121ff.). The regular interaction of individuals which includes nurture and education creates a structural linkage leading to the formation of a common language, but also to the construction of a common reality.15 Those commonalities are enlarged by means of linguistic interaction.

2.5 Receptive multilingualism and multilingual discourses Speech communities consist of individual speakers, therefore the extent of the individuals’ contact influences the distance between languages. A political separation of related languages usually leads to a larger linguistic distance,16 whereas their unification in one state normally results in linguistic approximation.17 In addition to this, members of a communication community18 do not only dispose of a common language but also of a common system of experiences derived from their upbringing in the same environment. By this, communication is facilitated.19 From this follows that a fundamental distinction has to be made between the communication of speakers using different languages and the conditions of this communication. In contrast to the usual form of monolingual discourses they can be described as multilingual discourses (the special case of interpreted discourses is, though mentioned in table 1, not discussed in the remainder of the paper).


Table 1. Monolingual and multilingual discourses Monolingual discourse

Multilingual discourse Interpreted discourse

Receptive multilingual discourse

A and B share a common L1 Interpreter is able to speak And understand l1 of a and b

A is able to understand L1 of B, B is able to understand L1 of A

A und B share a common L2 Interpreter is able to speak And understand l2 of a and (Divergent) l2 of b


A speaks L1 of B as L2


Interpreter speaks and Understands l1 of a and (Divergent) l2 of b

Not all cases presented in table 2 were treated separately in the literature because some of them might be regarded as special cases of phenomena already described. One matter this overview is meant to show is that some authors stress the actual discourse situation (e.g. Haugen 1966), whereas others focus on more general linguistic conditions (as Dahlstedt 1971). It is also true that a precise delimitation of different forms of multilingual discourses is not feasible. Especially a precise discrimination between mutually intelligible and unintelligible closely related languages is not possible. Moreover, the acquisition or improvement of receptive skills is much easier in a related language than in an unre-

Table 2. Different types of multilingual discourses A is able to understand L1 of B, B is able to understand L1 of A

A is able to understand A is able to L1 of B, B is able to understand L2 understand L2 of A of B, B is able to understand L2 of A

Because the languages are closely related (mutually intelligible)

semicommunication secondary speech (Haugen 1966) community (Dahlstedt 1971)

peripheral speech community (Börestam 2001)

Because the languages are related (mutually unintelligible) and A and B have acquired receptive skills

adjoining languages (Kloss 1929)

special case of intercomprehension

intercomprehension (cf. Meissner 2004)

Because the languages are not or not closely related (mutually unintelligible) and A and B have acquired receptive skills

polyglot dialogue (Augustin 1997)

special case of polyglot dialogue

receptive multilingualism (cf. Hansen 1987)

110 Ludger Zeevaert

lated one. In other words, ‘total unconditional intelligibility’ and ‘no possibility of understanding’ are the end points of a continuum (see figure 1). German Dutch

Icelandic

Russian

Japanese

Figure 1. Mutual intelligibility between German and other languages

The distances between the languages do not represent the results from language distance measurements, but are meant to display an impressionistic representation of the fact that the possibility of mutual understanding between languages is dependent on the proximity of their relationship. However, even between unrelated languages such as German and Japanese mutual understanding is not completely excluded.20 The different degrees of understanding may thus be described as a continuum where cases of communication in which speaker and hearer attribute exactly the same meaning to an utterance mark the one end, whereas the other end is marked by cases of communication in which a hearer is not capable of attributing any meaning to an utterance. It has to be pointed out that the meaning attributed to an utterance by a hearer does not solely rely on the decoding of speech signals. To derive a meaning from an utterance a hearer has a large context to his disposal — comprising common knowledge, culture specific knowledge, common communicative competence, situational context, prosody, body language etc. On the other hand the understanding of utterances in the mother tongue may be slightly obscured by the use of special technical terminology, unknown abbreviations or a lack of contextual knowledge. Finally, the capability of decoding linguistic utterances in closely related languages can be improved by different methods, a fact that is instrumentalised by the method of intercomprehension. In this context commonalities, partly identical elements and similarities between languages are used systematically in order to acquire a receptive competence in a neighbouring language. The EuroComRom-method21 developed by Klein and Stegmann (2000) is a successful practical application of this concept. Corresponding projects were soon to follow, cf. Zybatow (2002) and Duke, Hufeisen and Lutjeharms (2004).22

3. Interscandinavian semicommunication: An authentic example 3.1 Preliminary remarks The idea of analysing communication between closely related languages based on receptive competence is not completely new. 75 years ago Heinz Kloss (1929) designed a detailed, but by and large disregarded, concept of acquiring a receptive competence of the different (West) Germanic languages, aimed at establishing a mutual understanding between speakers of Afrikaans, German, Dutch, Pennsylvania Dutch, Yiddish and

Receptive multilingualism and semicommunication

Frisian. In the last ten years or so the political and social changes in Europe have led to a massive increase of research in this field. The main emphasis lies on the evaluation of the linguistic, political and social conditions of receptive multilingualism and on the development of didactic concepts.

3.2 Interscandinavian communication In this context it is very often referred to the Interscandinavian communication as a motivation and model.23 Traditionally, even in official contexts such as political consultations of the Scandinavian countries or in conferences with participants from different Scandinavian countries, the Scandinavians do not use a lingua franca as English or agree to use one of the languages involved, but rather every participant uses his own mother tongue.24 This form of receptive multilingualism is patronised by the so-called neighbour language education in Scandinavian schools and is based on the widespread feeling of historical, cultural and political commonalities between the Scandinavian people. It has to be pointed out that there is a qualitative difference between Interscandinavian communication and the mutual understanding of languages in the Romanic, the Slavic or the complete German language group. On the one hand the speakers of those languages concerned by the intercomprehension research are far less conscious of the possibility of this way of communication than people in Scandinavia and have thus not developed a tradition of semicommunication. On the other hand the distances between languages such as French and Rumanian, Polish and Russian or English and Icelandic are much larger than the differences between the Mainland Scandinavian languages Danish, Norwegian and Swedish. Nevertheless, scholars in the field of Interscandinavian communication agree on the fact that, due to the phonologic and lexical differences between Danish and Swedish, a spontaneous, unimpeded understanding does not always work between speakers of the two languages. Tests conducted by e.g. Maurud (1976) and Bø (1975) show that spoken Danish is understood by Swedes to less than 50% of all cases.25 In other words, there is a contradiction between the degree of understanding predicted on the basis of the linguistic conditions of receptive multilingual communication and the fact that obviously successful receptive multilingual communication can be observed between the speakers of the different Scandinavian languages. One can think of various possibities to explain the fact that Interscandinavian communication apparently works in practice despite problems arising from theoretical considerations: • Th e demands of the aim of the communication are reduced in comparison to monolingual communication, i.e. perhaps a successful communication is only simulated. • Communicative problems are disambiguated with the help of extralinguistic context.

111

112

Ludger Zeevaert

• Th e speakers use special strategies in dealing with trouble sources that differ from those used in monolingual communication. • The success of the communication is a result of (controlled or uncontrolled) language acquisition. Knowledge about the linguistic divergences is acquired by the speakers who are thereby enabled to overcome those divergences. In the following lines I would like to try to test those different possibilities on genuine examples from Interscandinavian communication in order to obtain conclusions about the feasibility of receptive multilingualism even in different contexts.

3.3 Interscandinavian work groups and panel discussions Zeevaert (2004) investigated group work sessions within meetings of The Nordic Association of University Administrators26 which besides lectures are an integral part of those conferences. In those sessions up to eight participants from different Scandinavian countries discuss questions concerning the work of the university administrations from the perspective of their countries of origin. The aim of the discussion is first and foremost the exchange of information — discourse languages are Danish, Norwegian and Swedish. The aim of Zeevaert’s (2004) investigation was to find out whether the participants were able to establish a functioning communication despite the problems of understanding that were attested in theoretical, experimental and demoscopic approaches, and if certain strategies, different from those used whithin a speech community, were developed. The methodological approach was based on conversation research and communication theory; transcriptions of six discourses were analysed, the transcriptions being supplemented with computer based quantitative evaluations of the discourse behaviour in digitalised versions of the original recordings. An examination of both the length and the frequency of pauses, back channel behaviour, treatment of trouble sources and turn taking gave the impression of a functioning communication. The results from earlier investigations that revealed severe understanding problems for special language combinations (Maurud 1976) or described Interscandinavian communication as fragmentary or problematic (Haugen 1966, Börestam 1994) could thus not be confirmed. Seen as the discourses generally appear to be semantically coherent, the possibility of a simulation of successful communication, which in reality is based on a lack of mutual understanding, can be excluded. The work groups and panel discussions investigated by Zeevaert (2004) mainly represent ‘crucial’ communication.27 Their aim is placed on the mediation of mental contents and not the instigation of non-linguistic actions. This implies on the one hand that a direct control of success is usually not given and on the other hand that misunderstandings not detected by the participants do not have severe consequences. It is quite possible, that trouble sources discovered by a discourse analysis do not constitute communicative problems for the discourse participant as excerpt 1 suggests.


Table 3. Discource participants in excerpt 1 Token

Discourse language

Mother tongue

m/f

IM KT GR

Swedish Danish Norwegian

Swedish Danish Norwegian

f f f

Excerpt 1.

113

114 Ludger Zeevaert

This discourse was recorded at a NUAS-conference in Sweden and is mainly concerned with the prospective economical development of universities and the development of planning and cooperation between the university administrations. In the excerpt from the discourse presented here the Swedish woman IM (for a list of the participants see table 3) gives examples for the problems that arised when computers were introduced in the administration. The description of the grotesque coincidence that the monitor of a computer-hostile employee starts burning does not mainly serve the function of informing but far more fulfils an expressive function; from IM’s utterance it can be seen that she knows that her narration is a digression. The recurrent laughing documents the social function of the communicaton, and finally even the aim of improving the receptive competence is fulfilled: the typical Swedish term bildskärm (‘monitor’) is framed by pauses to provide an opportunity for questions if the term should cause any problems in understanding. Obviously the Danish woman KT is able to linguistically decode the utterances of the Swedish woman IM. This may be derived from the fact that neither pauses, nor measures to ensure the understanding can be observed, and KT’s proposal to solve the problem has to be regarded as being coherent, at least if only the direct context is considered. However, if one looks at the whole discourse context, the contribution of KT seems to be less adequate. The actual subject of the discourse section, namely how to motivate computer-hostile employees, is not picked up by KT, and her concrete proposition seems to be influenced by the fact that she does not grasp the ironic dimension of the anthropomorphic depiction given by IM (the computer seems to take revenge on the employee who does not like him).28 At this point the communication has to be regarded as textually successful but pragmatically unsuccessful. Of course the interlocutors are not only simulating successful communication. However, this part of discourse shows that the demands on receptive multilingualism have to be reduced in comparison to monolingual communication, even if no problems in linguistic decoding are observed at the surface.


3.4 The role of a common background The work group sessions that were recorded in Scandinavian conferences did not give the possibility of solving problems of understanding by means of extralinguistic context. In contrast to for example lectures, in which by means of the use of visual media it is possible to provide the listener with information over different channels, the participants of the group sessions had to solely rely on the spoken word to support the linguistic decoding. A use of gestures exceeding the amount of non-verbal communication in monolingual communication was not observed. However, the discourses give hints that the common professional background, viz. the occupation in the university administration, has a positive effect on the mutual understanding. In excerpt 2 the participants (see table 4) inform each other of how the allotment of money is handled at their universities. Excerpt 2.

115

116 Ludger Zeevaert

Even if in the different countries different terms are used for the different methods of accountancy the speakers succeed in finding the counterpart of the terms in their own language due to their common background (see excerpt 3). The Swedish speaker RA asks if the other universities use the kassaprincip. The Norwegian speaker JM makes sure that he has understood RA correctly by reformulating (with question intonation) the Swedish term kassaprincip to the Norwegian term kontantprinsipp. RA affirms this by repeating kontantprincip which is a form also possible in Swedish. Even the term nettobudsjettering which is introduced into the discourse by the Norwegian speaker JM and is unknown to the Swedish speaker RA is not clarified by a direct request (e.g. “what does nettobudsjettering mean in Swedish?”). RA rather approaches the problem with a slightly long winded exothesis (“vad menar de nu med nettobudgetering?”) that is continued and completed by JM when RA breaks off his considerations. Finally the Danish speaker OJ paraphrases the Norwegian term nettobudsjettering with the Danish term nettobevilling. The consensual part, i.e. the common professional context including knowledge of procedures and processes in the university administrations that are common in all countries, makes it possible for the speakers to lead a successful communication at potentially problematic points. Table 4. Discourse participants in excerpt 2 Token

Discourse language

Mother tongue

m/f

RA JM TE OJ

Swedish Norwegian Norwegian Danish

Swedish Norwegian Norwegian Danish

m m m m


Excerpt 3.

3.5 Dealing with trouble sources As mentioned above, the discourses treated by Zeevaert (2004) do not exhibit a particularly high amount of trouble sources and repair compared to monolingual communication. However, the organisers of the conferences do not predict the discourse situation to be unproblematic, which becomes clear from the fact that NUAS established a task force concerned with language questions that developed several manuals aimed at improving mutual understanding. The manuals suggest to choose suitable rooms with good acoustics and to provide the speakers with technical and other resources (overhead projector, microphones, facilities to copy handouts and word lists) to support the understanding of the spoken words with the help of visual and written information. Even the choice of adequate

117

118

Ludger Zeevaert

referents and a generous time schedule (to prevent people from speaking too fast due to time pressure) are proposed. Much more interesting, though, are strategies chosen to deal with trouble sources that appear in the course of the discourse and show differences to monolingual discourse situations, seen as such strategies might also be useful for other situations of receptive multilingualism. Is was mentioned above that the discourses analysed by Zeevaert (2004) have to be classified as crucial communication,29 meaning that it is most often not possible to control whether or not an utterance was actually understood. Not every problem of understanding has to be verbalised instantaneously, in fact very often the use of a let it pass-strategy (waiting for a clarification of the problem later in the discourse) can be a succesful choice. Of course this strategy cannot be illustrated by examples because it is characterised by not dealing with a trouble source. Even if one speaker doesn’t frequently take part in a discourse, a fact that can be derived from a quantitative analysis of the number and length of the utterances of the different speakers, it does not automatically have to be assumed that he has problems in understanding the other speakers. As up to eight persons take part in the discourses the responsibility of keeping the discussion alive can be divided between the speakers, and not every participant has to actively take part throughout the whole discussion. It is significant, though, that speakers who are not able to use their mother tongue (Finns, Icelanders, Faroe Islanders, Greenlanders) have a lower stake in the discourses compared to the speakers using their native language (Danes, Norwegians, and Swedes). Even this cannot be regarded as significant evidence for problems of understanding, but it shows that in receptive multilingual communication, in situations in which one speaker is able to use his mother tongue, whereas the other speaker has to switch to an L2, the non mother tongue speakers appear to be impeded. The absence of back-channel signals, however, has to be looked at as a more obvious sign for problems of understanding. Back-channel behaviour30 demands far less mental activity compared to complete utterances and can therefore be performed more or less automatically. It is thus not an indicator for problems of language production but rather for problems with respect to reception. This does not automatically imply that the hearers producing little back-channel behaviour have problems in understanding their interlocutors. Exactly as for the active participation in a discourse, one has to act on the assumption of individual, gender specific and culturally dependent differences in the production of back-channel signals. A lack of received back-channel signals, on the contrary, is a rather strong hint towards the assumption that the utterances of the respective speaker tend to cause problems of understanding for the other hearers. In order to assess if his own utterances are understood, a speaker in receptive multilingual communication should not only regard requests for repair by the hearers but also take into consideration whether or not the hearers perform a normal back-chan-


nel behaviour. The reason for this is that one must assume that the participants apply the same communicative rules they are used to from monolingual communication correspondingly in receptive multilingual communication, implying that too many requests for clarification could be conceived as a face-threatening act,31 afflicting the communication too much. Typical strategies of linguistic accommodation including a slow and enunciated speech, lexical accommodations to the target language, the avoidance of colloquial expressions and an increase in redundancy by means of repetition and reformulation can only be identified in the discourses in question in a few instances. This means that even if linguistic accommodation can not be regarded as being a typical trait of Interscandinavian communication it is able to give important information about the functioning of the communication in those cases in which it can actually be observed. This will be illustrated by the following analysis.

3.6 Neighbouring-language acquisition The discourse investigated in the following was recorded at a conference of university administrators in Sweden. The work group discusses questions regarding the future economical development of the universities and the development and planning of cooperation in the university administrations. Within the discussions more general questions are also brought up — different structures in the direction of the departments, different approaches in covering the demand for buildings, the increasing international competition between universities and experiences with electronic data processing are discussed, partly in an anecdotal manner. Three of the participants speak Swedish as a mother tongue,32 two Norwegian, and one Danish (see table 5). The overall percentage rate of Norwegian is largest, but no language is clearly over- or underrepresented. The Danish stake is lowest, but one has to consider that the Danish speaker KT does not have any support from compatriots and is solely responsible for the Danish contribution. A more thorough inspection of the utterances of the different speakers (cf. table 5) shows that the three female speakers GR (Norwegian), KT (Danish) and IM (Swedish) are the three participants mainly engaged in the discourse. Table 5. Discourse participants and their share in dicourse Speaker

Discourse language

M/f

Stake in dis course/min

Stake in dis course/turns

Back-channel signals

LF ÅS IM KT KÅ GR

Swedish Swedish Swedish Danish Norwegian Norwegian

m f f f m f

1.0 0.5 10.5 11.0 3.5 14.0

38 7 98 165 50 127

52 4 47 148 28 85

120 Ludger Zeevaert

The impression of a rather balanced discourse is supported by the fact that the turns change mainly between the different languages (81%). A preference for turn taking between speakers of the same language that was observed in comparable discourses33 cannot be found. The transitions between the turns are fluent. No conspicuous pauses are to be seen. Very often there is an overlap between the utterances of the different speakers. Such overlaps presuppose that the hearer is able to follow the speaker in the planning of his utterance seen as the initiative for taking over the turn has to be provoked before the speaker has ended his utterance. They can be regarded as signals for a good working mutual understanding. The same holds true for the back-channel behaviour observed. The speakers signal understanding and attention by regularly back-channelling (every eight seconds on average). Especially the Danish speaker KT shows by her large amount of back-channel signals (cf. table 5) that she is able to understand the utterances of the Norwegian and Swedish participants. But there are further indications suggesting that KT is more concerned with safeguarding the understanding than her interlocutors. A comparison with Danish speakers in other discourses shows that KT speaks extraordinary slowly and enunciated,34 her utterances are also shorter than those of the Norwegian and Swedish discourse participants (four seconds on average compared to six seconds for the Norwegians and five seconds for the Swedes). KT is the only participant in the discourse who makes efforts to lexically adapt to the other discourse languages in order to avoid trouble sources. One example of this effort can be found in excerpt 4, where KT realizes a trouble source during her utterance, namely the Danish word sovepude (‘pillow’). She obviously assumes that the word will not be understood by the Swedes, and therefore adds the Swedish synonym kudde. This cannot be seen as a typical example of a repair as described in conversation or discourse analysis,35 but rather as the anticipatory avoidance of a trouble source. Excerpt 4.


Excerpt 5.

The strategy used consists of a double utterance of the trouble source, first in Danish and then once more in Swedish. In excerpt 5 KT applies the same strategy and actually uses a Swedish idiom, ta hand om något (‘to take care of something’). In this discourse KT is the only participant adapting linguistically. The other speakers’ lack of accommodation can of course not be illustrated by examples, but possibly it can be demonstrated in the context of other initiated repairs. Even if they only occur occasionally (five instances in the whole discourse) it can be stated that GR has a preference for repairing by means of repetition (see excerpts 6 and 7). In this case the effort of repairing the trouble source with a repetition is successful seen as the Danish term for ‘currant’, ribs, is nearly identical to the corresponding Norwegian term, so that a mere perception problem may be causing the problem. Swedish uses a different word, vinbär, but this does not lead to a problem in this case

Excerpt 6.

121

122 Ludger Zeevaert

Excerpt 7.

because GRs utterance in excerpt 6 is not part of the group work itself but part of the small talk in a short coffee break. The trouble source in excerpt 7 is related to a Norwegian place name. In Inter scandinavian communication place names very often cause problems because in most cases no linguistic context can be used for the required decoding. The Norwegian town Kongsvinger is obviously not known to the Danish speaker KT. The Norwegian speaker uses a repetition to repair the trouble source specifying that Kongsvinger is a town (Danish/Norwegian by), KT then confirms her understanding by means of a repetition. GR’s short hesitation may originate from the fact that she activates her knowledge that by means ‘village’ in Swedish (‘town’ would be stad). However, it is not necessary to deal with this trouble source because the initiative comes from the Danish speaker KT so that a reformulation in Swedish would be the wrong strategy in this context. In excerpt 8, however, GR applies a reformulation because KT is signalling a problem of understanding by asking a question. In this case the subject is an administration handbook that had been discussed with its advantages and disadvantages earlier in the discourse. After a request for clarification by KT, GR starts with a repetition that she breaks off again in order to choose a reformulation instead. KT, though, confirms the success of the repair before GR comes to the actual trouble source in her reformulation. In the only repair that is requested to be made by KT she uses a reformulation (see


Excerpt 8.

excerpt 9). The Swedish speaker IM intervenes because she does not understand the Danish term uddannede (‘educated’). Consequently KT reformulates with the Swedish term utbildning. The success of the repair is then confirmed by IM. Excerpt 10 shows how such a relatively high level of competence (definitely exceeding a receptive competence in the case of KT) can be acquired in practice. The Swedish speaker IM proposes to fetch some coffee and avoids using the typical Swedish term fika36 (‘have a cup of coffee’). The Danish speaker KT remembers from an earlier occasion that Swedes have a special word for drinking coffee and asks for it. Both Swedes in the group react to her question by naming the word and explaining it. LF actually uses the Danish term wienerbrød (‘Danishes’).

124 Ludger Zeevaert

Excerpt 9.

Excerpt 10.


Excerpt 11 may indicate that KT improves her knowledge of Swedish by reading novels in the neighbouring language. IM cites a proverb often used by one of her colleagues (‘no spying no idea’),37 GR signalises her understanding by her laughter, but a confirmation of understanding is not given by the other hearers, so GR initiates a repair of the trouble source. After that KT and LF confirm their understanding, and after a short break KT provides the source of her language competence, the novels of the Swedish author Jan Guillou who is first of all famous for his spy novels. Of course such discourse behaviour by far exceeds a purely receptive multilingualism, but it shows that the acquisition of active knowledge is by all means a good way of improving the communication.

126 Ludger Zeevaert

Excerpt 11.

4. Conclusion In this article an approach was made to describe theoretically, and with the help of examples, a form of human linguistic communication that is commonly not only assumed to be deviant from the normal situation of linguistic exchange between members of a speech community, but also from normal foreign language communication because in this context different languages are used simultaneously within one discourse. This form of communication denominated receptive multilingualism, however, is far more than just a linguistic curiosity. It is successfully exercised in different regions of the world and provides, for example in the context of the European Union, a possibility of mastering the linguistic challenges arising from the advancing globalisation without relinquishing cultural and linguistic diversity.38 As is shown in the different articles in this volume the conditions for such a receptive multilingualism can be very different depending on the languages involved, but also depending on the communicative prerequisites. For this reason, I aimed at carrying out a differentiation of this term in the first part of this article. Taking a theoretical model of linguistic understanding as a starting point it seemed to be important to first of all distinguish between the (linguistic) conditions of such a form of communication and its practical use in multilingual discourse situations. Furthermore, based on the criterion of linguistic distance, three prototypical scenarios of receptive multilingual communication were sketched, which in reality however, are not clearly separable in practice or, in other words, form a continuum.


Following those theoretical considerations a documented case of receptive multilingual communication was investigated, namely the communication at conferences with participants from the different Scandinavian countries. It was shown that this form of multilingual communication that has a long tradition in Scandinavia and is institutionally embedded can be regarded as an equivalent alternative to lingua franca communication or interpreting even if it, like other forms of communication between members of different speech communities, does not correspond to monolingual L1 communication. This communication does not only include the understanding of speakers of mutually intelligible dialects of a (hypothetic) roofing language (‘Scandinavian’), which is the case for the linguistic relation between Norwegian and Swedish. It also includes the communication between closely related, but not unconditionally intelligible languages such as Danish and Swedish, between speakers of remotely related, mutually fairly intelligible languages such as Icelandic and Swedish and even between speakers of unrelated and totally unintelligible languages such as Swedish and Finnish or Danish and Greenlandic. Even if Scandinavia, due to its special historical conditions, is very suitable for this particular form of multilingual communication this case of functioning receptive multilingualism can serve as an example even in a larger context of multilingual European understanding. Like the Scandinavian countries, the countries in the European Union share a common cultural and religious tradition, and in both cases the relationship between the countries was not free from conflicts including wars. Nowadays the concordances are so comprehensive that political, economical and cultural tasks formerly seen as national responsibilities are conducted in common institutions, which of course in many cases requires a certain willingness to compromise. Given this willingness such a ‘unity in diversity’ might also be feasible in the linguistic domain.

Transcription conventions: “...” “/” “ . “ “((3s))” “((laughs, 2s))” “[laughing” “( )” “(2s)” “(ingenting)” “<S [...] >” [[00: 07: 30]]

interrruption repair pause, less than 1 second pause, 3 seconds verbal, nonphonological activity nonphonological activity during utterance not understood not understood, lenght of utterance in brackets good guess switch of language with abbreviation for the language switched to time elapsed in discourse

128 Ludger Zeevaert

Notes 1. Cf. Karam (2000: 121). The study of such processes of linguistic adaptation form the starting point for the Communication Accommodation Theory developed by Howard Giles. (cf. e.g. Giles, Coupland and Coupland 1991). This theory is based on the Speech Accommodation Theory (SAT) which was established at the beginning of the 1970s as a reaction to Labov’s work. It is inspired by the discovery that interlocutors adapt to each other with regard to speed of speech, pauses, length of utterances, choice of stylistically salient phonological variables, body language etc. in order to signalise mutual approval. In the beginning the SAT was primarily interested in social conditions of linguistic variables. From the 1980s on it was modified to the more sociolinguistically targeted Communication Accommodation Theory (CAT). 2. Cf. the bibliography by Braunmüller and Zeevaert (2001). 3. A typical example is Ohlsson’s (1977) refusal to use the term semicommunication as a description for the communication between Scandinavians with different mother tongues — not because it is wrong, but because the negative connotation of the prefix semi- might have a negative effect on the efforts to achieve a better mutual understanding between the Scandinavian countries. 4. Cf. the articles by Georges Lüdi and Ivar Werlen in this volume. Finkenstaedt and Schröder (1990) describe receptive multilingualism even for Alsatian and French in Alsace and for French and English in Canada. Following Hansen (1987), receptive multilingualism between Danish and Japanese is documented at least in one case. 5. The problematic nature of this term is discussed by Braunmüller (2001: 5f.) and Zeevaert (2004: 66ff.). 6. C.f. also the article by Braunmüller in this volume. 7. “Wir verstehen unter einer “historischen Sprache” eine Sprache, die als historisches Kulturprodukt vorhanden ist und von ihren eigenen Sprechern und denen anderer Sprachen als Sprache anerkannt wird. Wir erkennen eine historische Sprache vor allem daran, daß sie einen eigenen Namen hat, durch ein adjectivum proprium bezeichnet wird, z. B. die deutsche, französische, italienische Sprache. ” (Coseriu 1988: 24) 8. Methods to measure the mutual intelligibility between language pairs are described in the articles of Delsing, Doetjes, van Bezooijen and Gooskens and Möller in this volume. Cf. for further tests of mutual understanding between dialects Wolff (1966), Kirk (1970), Biggs (1957), Voegelin and Harris (1951), Brodkey (1972), Casad (1974) or Yamagiwa (1967). Even concerning the understanding between Scandinavian languages different tests were made, cf. Bø (1975), Maurud (1976), Börestam (1986), (1987a), (1987b), Ridell (2000), Jörgensen and Kärrlander (2001), Doetjes (2003), Doetjes (2004) — a summary of the different tests is given by Golinski and Doetjes (2005): 14ff. 9. The results from existing tests show that this problem is not trivial. A study carried out by Jörgensen and Kärrlander (2001), testing the understanding of spoken Danish by Swedish and of spoken Swedish by Danish adolescents, reveals a rather high individual variability in the number of correct answers. Even for the different questions large differences in the number of correct answers occur, due to the fact that some questions imply a much higher level of knowledge of the specific cultural background of the two countries. In Maurud’s (1976: 117) study Norwegians gave 94% correct answers to questions regarding the content of a Swedish


text, whereas mother-tongue speakers reached only 91,2% in answering questions on the same Swedish text. In other words, the Norwegians had a better understanding of the neighbouring language Swedish than the Swedes of their own mother tongue. Comparable problems are reported by Delsing in this volume. 10. Shannon and Weaver’s model was originally designed on behalf of the American telephone company Bell. Like Eco’s model it describes the transmission of a signal from a sender to a receiver through a data channel. 11. “[...] die Vermittlung individuell differenzierter Welttheorien” (Ungeheuer 1987a: 58). 12. Thus understanding also includes mental processes not directly triggered by a speaker. By this, cases of human communication can be explained in which a hearer succeeds in understanding an utterance with temporal delay but without any additional information being given by the speaker. In these cases understanding has to be led back to the interaction of the hearer with his own mental states. 13. “Es ist dem Orientierten überlassen, wohin er durch selbständige interne Entwicklung auf seinen eigenen Zustand seinen kognitiven Bereich orientiert. Seine Wahl wird zwar durch die »Botschaft« verursacht, die so erzeugte Orientierung ist jedoch unabhängig von dem, was diese »Botschaft« für den Orientierenden repräsentiert. Im strengen Sinne gibt es daher keine Übertragung von Gedanken vom Sprecher zum Gesprächspartner. Der Hörer erzeugt Information dadurch, daß er seine Ungewißheit durch seine Interaktionen in seinem kognitiven Bereich reduziert.” (Maturana 1998: 58f.) 14. As an illustration, Maturana (1998: 60) uses the example they are flying planes, a modification of an example from Chomsky (1965: 21). Thus, it is a common trait of ambiguous utterances and optical illusions such as Necker’s cube (cf. Nørretranders 1994: 267f.) that they allow for different interpretations, although only one interpretation can be activated at a time. 15. Cf. Ackermann (1997: 55). 16. Typical examples of such a development are Danish and Swedish, cf. Zeevaert (2004: 43f.), Serbian and Croatian, cf. Radovanović (2000) and Škiljan (2000) or even Luxemburgian and German, cf. Moulin (2006). 17. Following Wandruszka (1998: 156), this holds true for the Intergerman understanding that is facilitated by the fact that the genuine dialects were by and large replaced by a system of regional dialects that allow different levels of approximation compared to spoken and written standard German. (“daß heute kaum noch irgendwo jemand außerhalb der engsten Dorfgemeinschaft die angestammte unverfälschte Ortsmundart spricht, weil man eine ganze Skala landschaftlicher Annäherungsformen an die allgemeine deutsche Umgangssprache und an das Schuldeutsch zur Verfügung hat.”) 18. It seems to be necessary to distinguish between a speech community (germ. Sprachgemeinschaft) consisting of all speakers of a language on the one hand (in the sense of a historical language, cf. Coseriu 1988: 24), and in contrast, a communication community (germ. Kommunikationsgemein schaft) consisting of the speakers of a community who, over a longer period of time, have direct linguistic contact with each other and, as a result, have a common linguistic norm at their disposal usually representing the first language of the speakers. Karam (2000: 121) describes a village community as a typical example. This corresponds to Grimshaw’s (1971: 116) description of a speech sub-community. Lenke, Lutz and Sprenger (1995: 69) use the term Kulturgemeinschaft (‘cultural community’) in order to explicitly include even common moral concepts, behaviour-

130 Ludger Zeevaert al and politeness systems, technologies and common knowledge other than the common language. 19. Börestam’s (1986: 26) investigation of the understanding of Danish by older and younger Swedes shows that such a common system of experiences may influence the level of understanding in receptive multilingual communication. Börestam is able to show that the mutual understanding of speakers of Swedish and Danish works best between members of the same generation. Werlen (1998) investigates problems in the communication between speakers of German from Germany and from Switzerland that are not based on linguistic differences, but rather on different systems of experience that she describes as different mentalities, cf. e.g. Werlen (1998: 75). 20. A test described by Andersen (1978: 25) shows that Danish listeners are able to reach a partial understanding of spoken texts in East Asian and African languages due to the use of international loan words within those texts. 21. For a more detailed description c.f. the article by Hufeisen and Marx in this volume. 22. In the literature, the term intercomprehension is used for different phenomena. Klein (2004: 15) uses it as an expression for the (linguistic) conditions for a receptive competence in a related language, whereas Meißner (2004: 42) on the other hand descibes it as the mental process leading to understanding while reading texts from a related language (“das ‚Verstehen bzw. die Dekodationsfähigkeit von nicht erlernten, aber gleichwohl den Individuen mehr oder weniger transparenten Sprachen’. [...] das interkomprehensive Lesen [bleibt] eng an den Vorgang der verbalen Zeichenerkennung gebunden. [...] Allein beschränkt sich die Interkomprehensionsfor schung auf die verbale Zeichenidentifikation und –deutung im Sinne der etymologischen und interlingualen Transparenz.”). For Van de Poel (2000: 102) intercomprehension is a method of learning foreign languages: “Since communication relies on comprehension, intercomprehension as a learning principle is a way to save time and energy when learning a language.” 23. Cf. e.g. Duke, Hufeisen and Lutjeharms (2004: 111). 24. In reality the matter is a bit more complex. Only Danes, Norwegians and Swedes are able to use their mother tonuges, whereas Faroe Islanders, Finns, Greenlanders, Icelanders and Sami, who also take part in the cooperation between the Scandinavian countries, have to use one of the three Mainland Scandinavian languages in order to be understood. Therfore in the research on the topic a distinction is made between a primary speech community consisting of Danes, Norwegians and Swedes and a secondary speech community including even Finns and Icelanders, cf. Dahlstedt (1971) and Doetjes in this volume. 25. Cf. for those and other tests of language understanding Zeevaert (2004: 77ff.), Golinski and Doetjes (2005: 14ff.) and the articles by Delsing and Doetjes in this volume. 26. Det nordiska universitetsadministratörssamarbetet (NUAS). 27. In the sense of Ungeheuer (1987b), cf. 2.4. 28. This seems to be a typical problem for Interscandinavian communication. An investigation by Teleman (1977) shows that mother tongue speakers of Swedish, who usually had no problems in linguistically interacting with Danes, had difficulties when encountering Danes in phatic communication, due to the fact that in the given context on the one hand less textual and extratextual context was at their disposal, and on the other hand that irony, plays on words, jokes and indirectness are used very often in phatic communication.


29. Cf. 2.4. 30. Cf. Duncan (1973: 38f.). 31. Vgl. Scollon and Wong Scollon (1995: 34ff.). 32. LF and IM are Swedish citizens, ÅS belongs to the Swedish spreaking minority in Finland. 33. Cf. Zeevaert (2004: 237). 34. KT produces 139 words per minute on the average, Danish speakers in other discourses in the corpus produce 220 words per minute. 35. Cf. Schegloff, Jefferson and Sacks (1977) or Rehbein (1984). 36. Other Swedish speakers in Interscandinavian discourses use this term, cf. Zeevaert (2004): 273ff. 37. The Swedish spaning is a euphemism for spying, comparable to the English term intelligence. 38. The European commission tries to follow the European motto ‘unity in diversity’ and does not strive for the enforcement of a lingua franca in Europe, but favours multilingualism, cf. A new framework strategy for multilingualism, http://europa.eu.int/comm/education/policies/ lang/doc/com596_en.pdf [seen 7.12.2005]. See also Rindler-Schjerve and Vetter in this volume.

References Ackermann, S. 1997. Organisches Denken. Humberto Maturana und Franz von Baader [Spektrum Philosophie 5]. Würzburg: Ergon. Andersen, E. 1978. Dansk, norsk og svensk i mundtlig form. Nogle synspunkter på attituder, identitet og strategier. In Mødet mellem sprogene i det dansk-tyske grænseområde. Foredrag og indlæg fra en konference i Aabenraa 13–15 september 1978, 14–37. Aabenraa: Institut for grænseregionsforskning. Augustin, V. 1997. The polyglot dialogue: A new approach in cross language communication and its putting into practice. In La compréhension multilingue en Europe. Actes du Colloque des 10 et 11 mars 1997, tenu à Bruxelles, sous l’egide de la Commission européenne (DG-XXII), M. Slodzian and J. Souillot (eds), 93–6. Paris: s.n. Biggs, B. 1957. Testing intelligibility among Yuman languages. International Journal of American linguistics 23(2): 57–62. Börestam, U. 1986. Dansk-svensk språkförståelse vid vägskäl? In Svenska i tid och otid. Vänskrift till Gun Widmark från doktoranderna i Uppsala, C. Anderson Östman and A. Malmberg (eds), 18–27. Uppsala: s.n. Börestam, U. 1987a. Dansk-svensk språkgemenskap på undantag. Nordisk språkförståelse i nutidshistoriskt och regionalt perspektiv belyst av svenska gymnasieungdomars förståelse av äldre och nutida talad danska [FUMS rapport 137]. Uppsala: s.n. Börestam, U. 1987b. Språkbron över Öresund — ideal eller verklighet? Rapport från en experimentell undersökning av svenska ungdomars förståelse av äldre och nutida talad danska. In Att förstå varandra i Sydnorden. Nordiskt seminarium 7–9 november 1986 Schæffergården, Gentofte, Danmark [Nordiska språk- och informationscentret. Seminarierapport 21], 23– 38. S.l.: s. n.

131

132 Ludger Zeevaert Börestam Uhlmann, U. 1994. Skandinaver samtalar. Språkliga och interaktionella strategier i samtal mellan danskar, norrmän och svenskar [Skrifter utgivna av Institutionen för Nordiska Språk vid Uppsala Universitet 38]. Uppsala: Institutionen för Nordiska Språk. Börestam Uhlmann, U. 2001. Nordisk språkgemenskap — centrum och periferi. In Moderne lingvistiske teorier og færøsk, K. Braunmüller and J. í L. Jacobsen (eds), 199–225. Oslo: Novus. Bø, I. 1975. Elevenes forståelse av nabospråkene, deres behov for og innstilling til undervisning i nabospråkene. In «Nordisk Projekt» — en undersøkning av grannespråksundervisningen i byene Esbjerg, Eskilstuna og Stavanger. Rapport, 63–123. S.l.: s.n. Braunmüller, K. 1995. Semikommunikation und semiotische Strategien. Bausteine zu einem Modell für die Verständigung im Norden zur Zeit der Hanse. In Niederdeutsch und die skandinavischen Sprachen II [Sprachgeschichte 4], K. Braunmüller (ed.), 35–70. Heidelberg: Winter. Braunmüller, K. 2001. Semicommunication and Acommodation. Observations from the linguistic situation in Scandinavia [Arbeiten zur Mehrsprachigkeit, Folge B 17]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehrsprachigkeit. Braunmüller, K. and Zeevaert, L. 2001. Semikommunikation, rezeptive Mehrsprachigkeit und verwandte Phänomene. Eine bibliographische Bestandsaufnahme [Arbeiten zur Mehr sprachigkeit, Folge B 19]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehr sprachigkeit. Brodkey, D. 1972. Dictation as a measure of mutual intelligibility: a pilot study. Language Learning — A Journal of Applied Linguistics 22(1): 203–20. Budovičová, V. 1987a. Literary languages in contact (A sociolinguistic approach to the relation between Slovak and Czech today). In Reader in Czech sociolinguistics [Linguistic & literary studies in Eastern Europe (LLSEE) 23], 156–75. Amsterdam: John Benjamins. Budovičová, V. 1987b. Semikomunikácia ako lingvistický problém. Studia Academica Slovaca 16: 49–66. Casad, E. H. 1974. Dialect Intelligibility Testing [Summer institute of linguistics of the university of Oklahoma, publications in linguistics and related fields 38]. Mexico City: s. n. Chomsky, N. 1965. Aspects of the Theory of Syntax. Cambridge MA: MIT Press. Coseriu, E. 1988. Sprachkompetenz. Grundzüge der Theorie des Sprechens [UTB 1481]. Tübingen: Francke. Dahlstedt, K.-H. 1971. Den nordiska språkgemenskapen. Språkvård 1971(4): 10–16. Doetjes, G. 2003. De rol van taalvariatie en taalafstand in de communicatie tussen Zweden, Noren en Denen. Toegepast Taalwetenschap in Artikelen 69(1): 131–42. Doetjes, G. 2004. Auf falscher Fährte in der interskandinavischen Kommunikation [Arbeiten zur Mehrsprachigkeit, Folge B 53]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehrsprachigkeit. Duke, J. , Hufeisen, B. and Lutjeharms, M. 2004. Die sieben Siebe des Eurocom für den multi lingualen Einstieg in die Welt der germanischen Sprachen. In Neuere Forschungen zur Europäischen Interkomprehension [Editiones EuroCom 21], H. G. Klein and D. Rutke (eds), 109–34. Aachen: Shaker. Duncan, S. J. R. 1973. Toward a grammar for dyadic conversation. In Semiotica 9: 29–46. Eco, U. 1988. Einführung in die Semiotik [UTB 105]. München: Wilhelm Fink. Ervin-Tripp, S. M. 1971. Sociolinguistics. In: Advances in the Sociology of Language. Vol. I: Basic concepts, theories and problems: Alternative approaches [Contributions to the sociology of


language 1], J. A. Fishman (ed.), 15–91. Den Haag: Mouton. Feitsma, A. 1986. Interlingual communication Dutch Frisian, a model for Scotland? In Scottish Language and Literature, Medieval and Renaissance. Fourth International Conference 1984 — Proceedings — [Scottish Studies 4], D. Strauss and H. W. Drescher (eds), 55–62. Frankfurt: Peter Lang. Finkenstaedt, T. and Schröder, K. 1990. Sprachenschranken statt Zollschranken? Grundlegung einer Fremdsprachenpolitik für das Europa von morgen [Materialien zur Bildungspolitik 11]. Essen: s.n. Giles, H. , Coupland, J. and Coupland, N. (eds). 1991. Contexts of Accommodation. Developments in applied sociolinguistics [Studies in Emotion and Social Interaction]. Cambridge/Paris: Cambridge University Press/Éditions de la Maison des Sciences de l’Homme. Golinski, B. and Doetjes, G. 2005. Sprachverstehensuntersuchungen im semikommunikativen Kontext [Arbeiten zur Mehrsprachigkeit, Folge B 64]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehrsprachigkeit. Grimshaw, A. D. 1971. Sociolinguistics. In Advances in the Sociology of Language. Vol.I: Basic concepts, theories and problems: Alternative approaches [Contributions to the sociology of language 1], J. A. Fishman (ed.), 92–151. Den Haag: Mouton. Hansen, E. 1987. Det nordiske sprogfællesskab. In The Nordic languages and modern linguistics 6. Proceedings of the Sixth International Conference of Nordic and General Linguistics in Helsinki, August 18–22, 1986, P. Lilius and M. Saari (eds), 7–20. Helsinki: Helsinki University Press. Haugen, E. 1966. Semicommunication: The language gap in Scandinavia. In Explorations in Sociolinguistics, S. Lieberson (ed.), 152–69. Den Haag: Mouton. Haugen, E. 1990. Babels forbrødring. Om tospråklighet og språkplanlegging [Det Blå Bibliotek]. Oslo: Universitetsforlaget. Jensen, J. B. 1989. On the mutual intelligibility of Spanish and Portuguese. Hispania 72: 849–52. Jörgensen, N. and Kärrlander, E. 2001. Grannspråksförståelse i Öresundsregionen år 2000. Gymnasisters hörförståelse. Lund: Institutionen för nordiska språk. Karam, F. X. 2000. Investigating mutual intelligibility and language coalescence. International Journal of the Sociology of Language 146: 119–36. Kirk, P. L. 1970. Dialect intelligibility testing: The Mazatec study. International journal of American linguistics 36: 205–11. Klein, H. G. 2004. Frequently asked questions zur romanischen Interkomprehension. In: Neuere Forschungen zur Europäischen Interkomprehension [Editiones EuroCom 21], H. G. Klein and D. Rutke (eds), 15–37. Aachen: Shaker. Klein, H. G. and Stegmann, T. D. 2000. EuroComRom — Die sieben Siebe: Romanische Sprachen sofort lesen können. 2. edn. [Editiones EuroCom 1]. Aachen: Shaker. Kloss, H. 1929. Nebensprachen. Eine sprachpolitische Studie über die Beziehungen eng verwandter Sprachgemeinschaften. Wien: Wilhelm Braumüller. Kloss, H. 1967. ‘Abstand Languages’ and ‘Ausbau Languages’. Anthropological Linguistics 9(7): 29– 41. Lenke, N. , Lutz, H.-D. and Sprenger, M. 1995: Grundlagen sprachlicher Kommunikation. Mensch– Welt — Handeln — Sprache — Computer [UTB 1877]. München: Wilhelm Fink Verlag. Maturana, H. R. 1998. Biologie der Realität. Frankfurt am Main: Suhrkamp. Maurud, Ø. 1976. Nabospråksforståelse i Skandinavia. En undersøkelse om gjensidig forståelse av tale- og skriftspråk i Danmark, Norge og Sverige [Nordisk utredningsserie 13]. Stockholm: s.n.

134 Ludger Zeevaert Meißner, F.-J. 2004. Transfer und Transferieren. Anleitungen zum Interkomprehensionsunter richt. In Neuere Forschungen zur Europäischen Interkomprehension [Editiones EuroCom 21], H. G. Klein and D. Rutke (eds), 39–66. Aachen: Shaker. Moulin, C. 2006. Grammatisierung und Standardisierung des Luxemburgischen. Eine grammatikographische-sprachhistorische Annäherung. In Perspektiven einer linguistischen Luxemburgistik. Studien zu Diachronie und Synchronie [Germanistische Bibliothek 25], C. Moulin and D. Nübling (eds), 305–39. Heidelberg: Winter. Nørretranders, T. 1994. Spüre die Welt. Die Wissenschaft des Bewußtseins. Reinbek: Rowohlt. Ohlsson, S. Ö. 1977. Nabospråksförståelse. Nabospråk eller grannsprog. Mål & Mæle 4(1): 19–24. Radovanović, M. 2000. From Serbo-Croatian to Serbian. Multilingua 19(1/2): 21–35. Rehbein, J. 1984. Reparative Handlungsmuster und ihre Verwendung im Fremdsprachenunter richt [Rolig Papir 30]. Roskilde: Universitetscenter. Ridell, K. 2000. Språkbroar. En studie av dansk-svenska radiosamtal i Öresundsregionen (D-uppsats). Uppsala: Uppsala Universitet, Institutionen för nordiska språk. Schegloff, E. A. , Jefferson, G. and Sacks, H. 1977. The preference for self-correction in the organization of repair in conversation. Language 53: 361–82. Scollon, R. and Wong Scollon, S. 1995. Intercultural communication. A discourse approach [Language in society 21]. Oxford: Blackwell. Shannon, C. E. and Weaver, W. 1971. The Mathematical Theory of Communication. 12th edn. , 1st edn. 1949. Urbana IL: University of Illinois Press. Škiljan, D. 2000. From Croato-Serbian to Croatian: Croatian linguistic identity. Multilingua 19(1/2): 3–20. Teleman, U. 1977. Samspråk på grannspråk. In Dialectology and Sociolinguistics. Essays in Honor of Karl-Hampus Dahlstedt. 19 April 1977 [Acta Universitatis Umensis 12], C.-C. Elert, S. Eliasson, S. Fries and S. Ureland (eds), 234–43. Umeå: s.n. Ungeheuer, G. 1987a. Was heißt ‚Verständigung durch Sprechen?‘ In Gerold Ungeheuer: Kommu nikationstheoretische Schriften I: Sprechen, Mitteilen, Verstehen [Aachener Studien zur Semi otik und Kommunikationsforschung 14], J. G. Juchem (ed.), 34–69. Aachen: Rader. (Before in Gesprochene Sprache. Jahrbuch 1972 [Sprache der Gegenwart. Schriften des Instituts für deutsche Sprache 26], 7–38. Düsseldorf: Schwann 1974.) Ungeheuer, G. 1987b. Vor-Urteile über Sprechen, Mitteilen, Verstehen. In Kommunikations theoretische Schriften I: Sprechen, Mitteilen, Verstehen [Aachener Studien zur Semiotik und Kommunikationsforschung 14], J. G. Juchem (ed.), 290–338. Aachen: Rader. Van de Poel, K. 2002. Motivating as a key factor in promoting multilingualism. In Eurocom. Mehrsprachiges Europa durch Interkomprehension in Sprachfamilien. Tagungsband des Inter nationalen Fachkongresses im Europäischen Jahr der Sprachen 2001. Hagen, 9–10 November 2001, G. Kischel (ed.), 102–11. Hagen: Fernuniversität. Voegelin, C. F. and Harris, Z. S. 1951. Methods for determining intelligibility among dialects of natural languages. Proceedings of the American philosophical society 95: 322–9. Wandruszka, M. 1998. Die europäische Sprachengemeinschaft. Deutsch–Französisch–Englisch– Italienisch–Spanisch im Vergleich [UTB 1588] (1st ed. 1990). Tübingen: Francke. Werlen, E. 1998. Sprache, Kommunikationskultur und Mentalität. Zur sozio- und kontaktlinguistischen Theoriebildung und Methodologie [Germanistische Linguistik 194]. Tübingen: Max Niemeyer. Wolff, H. 1966. Intelligibility and inter-ethnic attitudes. In Language and Culture in Society. A reader in linguistics and anthropology, 2nd edn. , D. Hymes (ed.), 440–5. New York: Harper


and Row/John Weatherhill. (Before in Anthropological Linguistics 1(3), 1959: 34–41.) Yamagiwa, J. K. 1967. On dialect intelligibility in Japan. Anthropological Linguistics 9(1): 1–17. Zeevaert, L. 2002. Insertionaler Sprachwechsel in semikommunikativen Diskursen. In: B. Meyer and L. Zeevaert: Sprachwechselphänomene in gedolmetschten und semikommunikativen Diskursen [Arbeiten zur Mehrsprachigkeit, Folge B 36], 3–17. Hamburg: Universität Ham burg, Sonderforschungsbereich Mehrsprachigkeit. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verständi gung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač. Zybatow, L. 2002. Die slawistische Interkomprehensionsforschung und Eurocomslav. In Euro com. Mehrsprachiges Europa durch Interkomprehension in Sprachfamilien. Tagungsband des Internationalen Fachkongresses im Europäischen Jahr der Sprachen 2001. Hagen, 9–10 November 2001, G. Kischel (ed.): 357–71. Hagen: Fernuniversität.

135

chapter 5

Receptive multilingualism in Switzerland and the case of Biel/Bienne Iwar Werlen Universität Bern

Switzerland is a multilingual country with four different national languages. A more detailed look, however, shows that on the level of the individual cantons multilingualism is rather an exeption. Only four of the 26 cantons have more than one official language, and even here the areas in which different languages are spoken are separated sharply. This article analyses examples taken from such exceptional cases, Biel/Bienne and Fribourg/Freiburg. The results from the two studies indicate that the ‘Swiss model’ of receptive multlingualism fostered by the Swiss language policy is utilised by the population, even if the default language model preferring the majority language is more widespread. In most cases, however, speakers do not stick to such static models and make use of their linguistic repertoires according to the communicative needs in a given setting. Keywords: Switzerland, receptive multilingualism, French, German, Swiss model

0. Introduction What has been termed “receptive multilingualism” by the editors of this volume is at the heart of Switzerland’s multilingualism. Since at least 1975, it is a central objective of the Swiss Cantonal Ministers of Education to foster national cohesion and communication by introducing a second national language at the primary school level. In the 1975 recommendations, the Ministers decided to introduce French as the first foreign language in the German speaking area and German as the first foreign language in the French speaking area. For the cantons of Grisons and Ticino, special solutions were proposed. The measures recommended, however, were only slowly implemented, especially in the German speaking area. Some cantons did not introduce the new system before 1999, and one canton resisted completely. In the late 1990s, a new discussion concerning the role of English arose (see Acklin Muji 2003 and Stotz and Meuter 2003 for details). The main divergence between advocates of English and defenders of the second national language as the first foreign language was the prioritization of economic versus national reasons for learning foreign languages. English was seen as the language of globalization, international communication and economic success; French and German were seen as languages needed for national cohesion and communication

138 Iwar Werlen

and for communication with neighboring countries such as France and Germany. In recent years, the debate has lost much of its importance. The ministers decided that two foreign languages should be introduced during primary school, namely a second national language and English. But which of the two should be taught first? While the French speaking cantons decided to start with German and introduce English at a later stage, the German speaking cantons in the east of the country chose to start with English, while those in the western part decided to start with French. This, of course, will create a barrier for inter-cantonal mobility of families with school-age children. In order to understand the background of these discussions, we will first provide a short sketch of Swiss multilingualism, followed by a discussion of four models of interlingual communication in Switzerland (see also Lüdi, this volume). We will then turn our attention to an atypical Swiss city, Biel/Bienne, which is officially bilingual and practices a bilingual way of communicating that differs from the model observed in other bilingual areas. Biel/Bienne will be contrasted to another bilingual city, Fribourg/ Freiburg. Receptive multilingualism in the Swiss context shows various forms which are differentiated due to varying context factors.

1. The context: Switzerland as a multilingual nation Switzerland is a democratic federal state with four national languages. Article 4 of the country’s 1999 Constitution asserts that German, French, Italian and Romansh are the national languages. This prominent place was deliberately chosen by parliament to show that multilingualism is an essential part of Switzerland’s self-conception. The first three languages are also declared “official languages” (in paragraph 4 of article 70), while Romansh is an official language only for people whose main language1 is Romansh — it is a so called “restricted” official language. The distinction of national and official languages may be seen in terms of status and usage: all four languages have the same status as national languages, but only three are regularly used by the federal administration. The four national languages are valid only for the confederation as a whole, not for its 26 member states, the cantons. According to article 70 paragraph 2 of the Constitution, each canton has the right to define its own official language or languages. Traditionally, 22 of the 26 cantons have only one official language (1 Italian, 4 French, and 17 German). Only three cantons (Berne, Fribourg, Valais) accept two official languages (French and German). However, in these bilingual cantons, the language areas are sharply separated. Only the bilingual district of Biel/Bienne and some communities in the canton of Fribourg are exceptions to this pattern, as will be shown in the ensuing discussion. The Canton of Grisons is the only trilingual canton of Switzerland with German, Italian and Romansh as official cantonal languages. With the exception of the Romansh speaking area, the linguistic borders are relatively clear and the languages form geographically distinct areas across cantonal bor-

Receptive multilingualism in Switzerland 139

Table 1. Main languages per language area in percentages for the whole resident population (source: Federal Office of Statistics, Federal Census, 2000) Main language German French Italian Romansh Non-national lang’s Resident population

Language area German

French

Italian

Romansh

86.6 1.4 3.0 0.3 8.7 5,221,135

5.1 81.6 2.9 0.0 10.4 1,720,365

8.3 1.6 83.3 0.1 6.7 320,247

25.0 0.3 1.8 68.9 4.0 26,263

ders. In each of the four language areas, the local language is absolutely dominant in terms of numbers of speakers. This can be shown by the figures in Table 1. The figures show very clearly that the local language is spoken by more than 80 percent of the resident population (Swiss citizens and foreigners) in each area. Again, the only exception is the Romansh-speaking area where the percentage of the local language is smaller due to the high number of German speakers. Non-local languages are in a minoritarian situation in all four language areas. The figures also show that there is a clear disproportion between the absolute number of speakers, with German speakers forming the vast majority, and Romansh speakers representing the other end of the spectrum with only a very small minority. This disproportion is mirrored in the distribution of the languages on the national level as shown in Table 2. Table 2. Main languages of the resident population of Switzerland in percent and absolute numbers (source: Federal Office of Statistics, Federal Census, 2000) Main language

Percentage

Absolute

German French Italian Romansh Non-national languages

63.7 20.4 6.5 0.5 9.0

4,640,359 1,485,056 470,961 35,095 656,539

The figures show that 91.0% of the resident population speak one of the four national languages as their main language. But the proportions of these languages are very unevenly distributed. The speakers of German are by far the largest group, followed by French speakers. Italian speakers constitute only 6.5% of the whole population and the speakers of Romansh are by far the smallest group. 9% of the total population speak a non-national language. In Table 3, the six numerically most important non-national languages are shown. The non-national languages are essentially the languages of migrating workers and of people seeking political asylum. The speakers of Serbian/ Croatian and Albanian — languages of former Yugoslavia — were immigrant workers until the beginning of the Balkan wars that marked the end of Yugoslavia; later on they

140 Iwar Werlen Table 3. Non-national main languages of the resident population of Switzerland in percent and absolute numbers (source: Federal Office of Statistics, Federal Census 2000) Non-national language

Percentage

Absolute

Serbian, Croatian Albanian Portuguese Spanish English Turkish Other languages

1.4 1.3 1.2 1.1 1.0 0.6 2.4

103,350 94,937 89,527 77,506 73,425 44,523 173,271

immigrated as asylum seekers. Speakers of Portuguese, Spanish and Turkish were and are immigrant workers. The relatively small group of English speakers consists essentially of two different groups: highly qualified professionals and politically motivated migrants from African or Asian countries. It is a commonplace today to say that Switzerland is no longer a quadrilingual but a multilingual country (see e.g. Dürmüller 1997), given the figures of the 2000 census cited in Tables 2 and 3. Even though this is correct in demographic terms, it is still incorrect in legal terms. Furthermore, there is a strong tendency to neglect non-national languages — except English — in Swiss language policy. This policy is shaped by two conflicting principles: the principle of territoriality and the “freedom of language” principle — the latter being what is internationally referred to as the “personality principle”. The principle of territoriality is the consequence of the right of the cantons to define their official languages. The personality principle says that every person has the right to use the languages of their choice, and this right may only be restricted by the territoriality principle when there is a legal base, an overwhelming public interest and if the restriction is appropriate. Since the cantons are responsible for the educational system, their official language is also the official language of the educational system. Given the territoriality principle, a Canton may not accept public schools in other languages than the official one. This policy clearly stabilizes the status quo of the national languages and their borders (with the exception of the Romansh area which is in an ongoing process of Germanisation; see e.g. Furer 2005). As stated in the introduction to this paper, the Swiss Conference of the Ministers of Education2 decided to install a policy of receptive multilingualism in the sense that a second national language is taught at primary school level. In their latest program of activities, dated June 6, 2006, they declare that they want “to foster forcefully the language competence of the pupils in the local language (first national language) from the beginning, to give solid knowledge of a second national language and of English to all pupils and to give them the possibility to learn a third national language” (EDK 2006, 3, our translation). In a project called HarmoS (EDK 2004) the Ministers will define the minimal competences pupils should achieve after grades 2, 6 and 9 in L1, L2 and L3 defined in terms of the Common European Framework


of Reference for Languages. The receptive competences should receive more attention than the productive ones. At the end of grade 9 (i.e. at age 16) pupils should be able to communicate in a second national language and in English. In an ongoing research project,3 we found that French was in fact the L2 most often indicated by Swiss German speakers while in the French speaking area German and English were equally frequent L2’s. This means that the linguistic policy of the Ministers of Education impacts the linguistic competences of the adult population. But other factors, for example economic ones, play a role, too. Various configurations of majorities and minorities are one of the main problems of Swiss linguistic policy making. On the national level, German is the majority language, the three other languages being in a minority situation. Obviously, in the monolingual cantons, the majority language is always the official language. But in the bilingual cantons the situation is more complex. Two thirds of the population in the two cantons of Fribourg and Valais speak French and one third speaks German. In these Cantons, French is the majority language and German the minority one. This creates a double constellation of majority and minority, German being majority language on the federal level but minority language in the canton, French being minority language on the federal level but majority language in the canton. The third bilingual canton, Berne, has a large German speaking majority (about 84%) and a small French speaking minority (about 8%). Most French speakers here live in three northern districts and in the bilingual district of Biel/Bienne. The trilingual canton of Grisons has a German speaking majority (about 65%), followed by the Romansh speaking minority (about 15%) and the even smaller Italian speaking minority (about 11%). Minorities tend to be more sensitive to unequal treatment, majorities tend to “forget” the minorities. This may lead to conflict situations. Since the key metaphor for Switzerland’s linguistic policy is “linguistic peace”, conflict situations are sought to be avoided. Thus, Swiss multilingualism is essentially based on a policy of consensus where the rights of the minorities are sought to be respected. Before turning our attention to Biel/Bienne, the linguistic situation of the German speaking part of Switzerland has to be specified. What Kolde (1981: 65) described as medial diglossia is the situation in which the choice of either the H- or the L-variety depends on the medium of communication, i.e. speaking versus writing. Roughly speaking, it can be said that Swiss German dialects are spoken and Standard German is written. The Swiss German dialects belong to the Alemannic dialects spoken also in Alsace, Baden-Württemberg (Germany), Vorarlberg (Austria) and Liechtenstein. The Swiss German dialects are mutually intelligible even if some mountain and more peripheral dialects such as those of the Upper Valais are said to be harder to understand than others. As already mentioned, Standard German is primarily a written language. Its relation to the Swiss German dialects is a controversial topic (see Hägi and Scharloth 2005). The defenders of the diglossic approach see Standard German and the dialects as two different forms of the same language system. The opposite position holds that Standard

142 Iwar Werlen

German is a closely related foreign language. In the aforementioned ongoing research project on the linguistic competences of Swiss adults,4 a surprising 52% native Swiss German dialect speakers mentioned Standard German as their first foreign language. It comes therefore as no surprise that many Swiss adults do not like to speak Standard German at all and prefer switching to French or English.

2. Four models of interlingual communication in Switzerland Given the linguistic situation of Switzerland as a whole, a key question arises as to how Swiss people with different language backgrounds communicate together? In his comparative study on language contact in the two bilingual cities of Biel/Bienne and Fribourg/Freiburg, Kolde (1981) described two models of language choice in a multilingual setting, given a contact between two unacquainted persons with two distinct main languages in a public, neutral situation. He first distinguished the choice of the first speaker (the initial choice). Normally, this will be his or her first language.5 The second speaker reacts in one of two ways. In the first model — called Swiss model by Kolde (1981: 237) — the second speaker will answer in her or his own first language. In the second model — Kolde’s Biel/Bienne model (1981: 239) — the second speaker will answer in the language of the first speaker. We can add a third model, tentatively called the default model, normally valid for monolingual settings in the context of the territoriality principle: Everybody will choose the default language of the territory. A fourth model is the choice of a lingua franca that is neither participant’s first language — e.g. English — by both the first and the second speaker. The Swiss model is a clear instance of receptive multilingualism: Everyone speaks their own language and expects the other speaker to understand them. In Switzerland, this model is especially valid for the two major linguistic groups — German and French. Romansh speakers are expected to speak and understand at least German, Italian speakers at least French. The Swiss model is applied on the federal level, e.g. in parliamentary commissions, or in official meetings of the federal administration. The medial diglossia of the German speaking part of Switzerland, however, is one of the disturbing factors in this model: most Swiss German speakers do not like to speak Standard German and most French speakers do not understand Swiss German dialects. Thus, Swiss German speakers prefer to speak French — even on German speaking territory. The Biel/Bienne model is a bilingual model: Every person is expected to speak and to understand both languages (German and French). In the Biel/Bienne context, the spoken form of German is the Swiss German dialect and, restricted to special cases, Standard German. This is a somewhat puzzling situation since French speakers learn Standard German at school but are expected to speak and understand Swiss German dialect. Bilingual speakers in Biel/Bienne often speak both the Biel/Bienne German dialect and French.


The default model is a monolingual model: Everyone should speak and understand the local default language. In monolingual settings6 this will be the local language. For the Romansh speaking area, the default language in more touristically developed communities seems to be German. For Biel/Bienne, there is no default language, for Fribourg/Freiburg, the default language is French. The lingua franca model (see Watts and Murray 2001) is also a bilingual model implicating the acquisition of English as a foreign language by everybody. There are some domains where spoken and written English already is a lingua franca in Switzerland, e.g. in scientific communication or in major international companies (see Murray 2003). But overall, the lingua franca model is not very widespread. A second lingua franca model with Italian as lingua franca was current under labor migrants with different first languages (e.g. Spanish, Portuguese, Serbian/Croatian, Albanian) in the German speaking area in the second half of the last century; it is not clear if this lingua franca still exists or is about to disappear given the fact that there are practically no new Italian labor migrants today. It seems that Spanish is in the process of replacing Italian in the function of a migrant worker’s lingua franca.

3. The case of Biel/Bienne — an officially bilingual city Biel/Bienne and the small community of Evilard-Leubringen together form the bilingual district of Biel/Bienne which is part of the officially bilingual canton of Berne.7 The other districts of this canton are monolingual: either German or French speaking (for more information on this canton and Biel/Bienne see Werlen 2000). Biel/Bienne became a bilingual city in the second half of the nineteenth century when French speaking workers from the surrounding countryside came into the city to work in the newly established factories. This French “colony” obtained the right to have a French school. Since 1900 the percentage of French speakers has been relatively constant at a level of 28% to 31%. Table 4 shows the situation in Biel/Bienne at the end of 2000 in terms of the main languages. 55.4% of the population are German speaking, 28.1% French speaking. The largest group with a main language other than the two official ones are the speakers of Italian.This is a result of the earliest group of work migrants Table 4. Main languages of the resident population of Biel/Bienne in percent and absolute numbers (source: Federal Office of Statistics, Federal Census, 2000) Main language

Percent

Absolute

German French Italian Others Total

55.4 28.1 6.0 10.4 100.0

26,957 13,695 2925 5078 48,655

144 Iwar Werlen Table 5. Non-national main languages of the resident population of Biel/Bienne in absolute numbers (source: Federal Office of Statistics, Federal Census 2000) Main language

Absolute number of speakers

Spanish Albanian Portuguese Serbian, Croatian East-asian lang’s Turkish English Arabic Various

1066 681 555 483 430 420 263 219 961

in the second half of the last century. Other languages comprise some 60 languages. Table 5 shows the languages which are mentioned by more than two hundred people. The city’s administration does not take into account languages other than the two official ones for the simple fact that the administration wants to know in which language people should be addressed. In 2003, the city’s statistics service published the figures of Table 6. Table 6 shows that the foreign resident population is divided into two equally large groups while the Swiss resident population has a proportion of roughly 2: 1. Spanish, Italian and Portuguese foreigners often find French easier than German; they will therefore choose French. The consequence of this choice is that their children will attend the French schools. Albanians, Serbians, Croatians and Turks on the other side prefer German. So the Latin-based languages feed the French minority. The institutional bilingualism of Biel/Bienne may be seen as a sort of double monolingualism (see Elmiger 2005): Each person has the right to communicate with the administration in one of the two official languages and get answers in the same language. In the case of Biel/Bienne that means that every officer should have a good command of both languages, but their clients are allowed to speak either French or German. This concept of double monolingualism leads to a doubling of the school system: There are German and French schools from kindergarten to high school.8 These two systems were totally separated until recent years. The other official language of the city was Table 6. German and French as the official languages of the Swiss and foreign resident population of Biel/Bienne in percent and absolute numbers (Source: Statistics service of Biel/Bienne 2003) German

French

absolute

percent

absolute

percent

Swiss Foreigners

23,975 6,808

47.4 13.6

12,692 6,735

25.3 13.4

Total

30,783

61.3

19,427

38.7


taught as a foreign language in both systems. Only recently, bilingual classes were introduced, but most pupils attend monolingual classes. This concept of double monolingualism (with bilingual officers) contrasts sharply with the common idea of individual bilingualism. People born in Biel/Bienne are expected to be bilingual at least to a certain extent (see Werlen 2006). Since there is no areal separation between the linguistic groups, even children with monolingual parents normally will have contact on playgrounds or other places with children speaking the other language of the city. Therefore, they will learn the second language “dans la rue” (on the street) (for a more detailed account of this subject see Py 2006). This leads to the effect that bilingual children with French speaking parents will learn the local German dialect as their second first language. If such children attend French schools they will additionally learn Standard German (starting in grade 4, i.e. at age 11) as a foreign language. The consequence is that they will speak French and the Biel dialect as native speakers, Standard German, however, as their first foreign language, usually with a French accent. Children with German speaking parents, on the other hand, will learn Standard German in its Swiss variety (starting at grade 1 of the primary school) and French as a foreign language at grade 5.

4. The research project bil.bienne: bilinguisme à bienne — Kommunikation in einer zweisprachigen Stadt In order to observe the patterns of communication and bilingualism in Biel/Bienne, a team of researchers carried out a research project (see Conrad, Matthey and Matthey 2002, Conrad 2005, Elmiger 2005, Werlen 2005, Elmiger and Conrad 2006) which consisted of three distinct modules. In the first module, a series of 40 interviews with people living in Biel/Bienne were conducted. The questions asked concerned the city’s bilingualism, the language biographies of the interviewees and their attitudes towards bilingualism. Their social networks were also recorded in order to see what languages they mentioned as used with friends, neighbors and co-workers. There were always two interviewers, one speaking German (normally in the dialect), the other speaking French. Both interviewers were bilingual but they kept to the use of only one of their languages (with a few exceptions). The interviewees were free to choose their language. In most interviews, there is a certain amount of code switching on the side of the interviewees. Much of the above information on Biel/Bienne and its bilingualism are extracted from these interviews. In a second module, conversations in public and semipublic places were recorded. Students and researchers asked people the way, bought something in a shop, etc. After the recording, people were informed that the conversation had been recorded for a research project. Most people accepted the recording; if this was not the case, the recording was deleted immediately. People were then asked some questions about their first language, their place of residence and other topics. The aim of these recordings was to

146 Iwar Werlen

see in which language the person answered. The most interesting cases were conversations between people with different languages. Module three was similar to module two, but the recordings were done in the bilingual city of Fribourg/Freiburg in order to compare the two cities. In the following analysis we will concentrate on the recordings of module two.

5. Language choice in a Swiss multilingual setting In this section we will present the essential findings of modules two and three; for a more detailed analysis see Conrad (2005). The choice of language in the conversations recorded in Biel/Bienne and Fribourg/Freiburg are of special interest here. Based on the models discussed in section 2 of this paper there were two expectations, i.e. that in Biel/Bienne it should be the first speaker who would determine the language choice and that in Fribourg/Freiburg the default language would be French. We therefore present in Table 7 the number of recordings made in both cities and the language choice made by students and researchers. Table 7. Languages used by students and researchers at the beginning of the recordings in Biel/ Bienne and Fribourg/Freiburg Language used by researcher

Biel/Bienne

Fribourg/Freiburg

Swiss-German dialect French Standard German

22 44 4

68 2 0

The distribution of the languages used at the beginning of the recordings is not really comparable. In Fribourg/Freiburg, the recordings in a Swiss German dialect far outnumber those in French, and no recordings were done in Standard German. In test recordings in French we always got answers in French. We therefore concentrated on recordings beginning in Swiss German since the German speaking group is in a clear minoritarian situation as Table 8 shows. Only a fifth of the resident population of Fribourg/Freiburg are German speakers; the vast majority is French speaking. This Table 8. Main languages of the resident population of Fribourg/Freiburg in percent and absolute numbers (source: Federal Office of Statistics, Federal Census 2000) Main language

Percent

Absolute

German French Italian Others Total

21.2 63.3 3.8 11.4 100

7520 22,603 1359 4065 35,547


Table 9. Language choice of addressees in Biel/Bienne and Fribourg/Freiburg Language choice of addressee

Biel/Bienne

Fribourg/Freiburg

Accommodation to addresser’s Swiss German dialect Accommodation to addresser’s Standard German Accommodation to addresser’s French No accommodation to addresser’s Swiss German dialect No accommodation to addresser’s Standard German No accommodation to addresser’s French Addresser’s Swiss German dialect is addressee’s L 1 Addressee’s L 1 is Swiss German dialect, addressed in Standard German Addresser’s French is addressee’s L 1

11 3 32 1 1 2 9 2

21 0 2 25 0 0 21 0

9

1

minoritarian situation is reflected by the fact that the city of Fribourg/Freiburg shows no clear manifestations of bilinguality. While street names in Biel/Bienne are bilingual, the street names in Fribourg/Freiburg are written in both languages only if there are historical sources showing a German name. For a person arriving in Fribourg/Freiburg without knowing its bilingual status, the city looks like other French speaking towns in Western Switzerland. So it was decided to start more conversations in the minoritarian language in order to see how people reacted. In Table 9, the language choice of the people recorded are shown. In Fribourg, the 70 recordings are evenly divided in three major categories: in 21 cases the addressee accommodated his or her French to the Swiss German dialect. In 25 cases the addressee did not accommodate his or her French. In these cases the addresser may choose to switch to French (in 9 cases), to Standard German or to continue in Swiss German dialects (in 16 cases), applying the Swiss model in the latter cases. In 21 cases, the addressee reacted to Swiss German dialect in Swiss German dialect the addressee being a native speaker of a Swiss German dialect. Only in one case was a French speaking addressee addressed in French. In the only two cases with French as the language of the addresser, both addressees accommodated to French. In Biel/Bienne on the other hand, accommodation to the addresser’s language took place in 46 of the 70 recordings. There were only four cases with no accommodation and these cases were recordings on the street, not in shops or offices, i.e. in non-service encounters; in three of the four cases people did not know the other language well enough (and they apologized for their lack of knowledge). These three people were visitors to Biel/Bienne normally living in other places. In 18 cases, both speakers had the same L 1 and in 2 cases a person addressed in Standard German answered in Swiss German dialect. The overall picture that emerges is that, in Biel/Bienne, the normal situation is that the addresser’s language leads to an accommodation of the addressee. In Fribourg/ Freiburg, however, French is the default language; if the addresser does not switch to French, the conversation will be in both languages (Conrad 2005) as predicted by the

148 Iwar Werlen

Swiss model. The only exception to this model is given in cases where addresser and addressee are both native speakers of Swiss German dialects. The four models mentioned above may roughly be described as norms in the sense of mutually assumed rules of behavior. But it is clear that such norms can be cancelled, in which case there will be explicit negotiations of the language(s) to be used in the contact situation. In special situations, there will also be special norms. In Biel/Bienne, one of these special norms is that the customer is king, meaning that in shops or restaurants as well as in the offices of the municipal administrations it is the customer who decides on the language spoken (and not the person who speaks first, perhaps greeting the customer). There is ample evidence of this in the recordings we made in offices and shops.

6. How multilingual communication works in Biel/Bienne and in Fribourg The following examples stemming from module 2 in Biel/Bienne (examples 1–3) and from Fribourg (example 4) illustrate the points made so far. The first example shows how the choice of French by the student researcher leads to an accommodation of the clerk (Im). The transcription begins (see example (1a)) when Anouk (An), the student researcher, enters the office of a job agency. (For readers not familiar with Swiss German dialects and French, Swiss German dialects will be printed in Courier New Example 1a.


and French in Lucida Console. The rules of transcription are given in an appendix at the end of this paper.) The clerk (Im) is greeting her in Swiss German dialect (line 1). She answers in French (line 2). Promptly, the clerk repeates his greeting in French. Thus he accepts the language choice of the customer. This small dialogue is typical of the Biel/Bienne situation in that the choice of the greeting in French or German by a customer is pivotal for the language to be used by the clerk. The language throughout the ensuing conversation is French even if the clerk has a marked accent in it. Our interpretation is confirmed by the clerk later on in the conversation when Anouk asks him why he switched from Swiss German to French (see example (1b). Anouk mentions that she realized his greeting in the Biel German dialect but that she answered in French (line 91). She asks him why he started in German and then continued in French (line 96). He answers that he tried to speak German because his French is really bad (line 98). Later on he even calls his French “le petit français qu’on a” (the little amount of French that one has) alluding perhaps to the idiomatic “le petit nègre” (gibberish). He then tells Anouk that his first language is Italian, the Biel dialect Example 1b.

100

150 Iwar Werlen

Example 1c.

being his second language. He prefers to speak the German dialect but then he clearly states the rule: The general rule — one will always try to speak the language of the “candidate” (line 102) — overrides his own preference of the German dialect. His deliberate attempt to determine the language of the conversation by greeting the customer in Swiss German failed in the case of Anouk and will fail in any other case in which a customer chooses French. In the next example, we will find a more complex situation created by our research design. Anouk, the student researcher, enters a drugstore. Anouk’s L1 is French; German is her first L2. She is greeted by the saleswoman in French (line 1 in example (2a)). She reciprocates the greeting in French (line 2) but then switches to Standard German (printed in Courier New). There are two observations to be made. The first one concerns the greeting bonjour. It can be used in both languages, the difference being the pronounciation. French has Example 2a.

Receptive multilingualism in Switzerland

a voiced fricative, a nasalized first vowel and an uvular /R/ at the end. German speakers normally use a voiceless fricative, many of them do not nasalize the first vowel and have an apical /r/ at the end. bonjour may therefore be used to test the linguistic affiliation of an unknown person (see Conrad 2005 and Werlen 2006 for examples). The second observation concerns Anouk’s switch to German after having heard the French pronounciation of bonjour and replying to it with French pronounciation. For a native German speaker, Anouk has an obvious French accent, but it seems that the saleswoman does not realize this accent. Not only Anouk’s accent is clearly nonnative, she also has some problems in defining her request. There is even a mistake in lines 2 and 3 (the noun unterschied has masculine gender, not neuter). Nevertheless, the saleswoman switches to German in line 5. After a long and painstaking conversation in German, Anouk explains that she recorded the conversation and why she did it. Anouk tries to find the German translation of analyse conversationnelle (line 52). She does not succeed and switches to French (line 53) without switching back to German. This is the point (line 54) where the saleswoman realizes that Anouk is a French speaker and she identifies herself as a romande, stemming from the Suisse romande. Later on in the conversation, she tells Anouk that she always tries to speak French at the beginning of Example 2b.

151

152 Iwar Werlen

a conversation saying bonjour. If the customer answers in German, she has to switch to German; otherwise, she will speak French. This is exactly the same strategy as the strategy of the clerk in the first example with the only exception of the preferred language. In this case, the norm of the customer as king leads to an exolingual communication situation9 in which two native speakers of French communicated in two differently fossilized varieties of German.

Example 3. Biel/Bienne: dr ober quai The next example shows that the choice of the language of conversation is readopted after an interruption caused by a question in the other language to a passer-by. The beginning of the recording is practically incomprehensible but the students noted that they started the conversation in French. The transcription begins when FI is greeting a female passer-by obviously known to her. That the passer-by is an acquaintance of FI’s Example 3.


follows fromFl’s greeting which implicates a tu-relation.10 The word sali that she uses is originally a French greeting (salut) but is used here in the local German dialect form. The passer-by answers with tschou, the Biel and Bernese form of Italian ciao which, in contrast to the Italian farewell, may also be used as a greeting. Both greetings are Swiss German dialect forms. In line 3, FI begins in French (le quai) but then immediately switches to the Biel dialect and asks the other woman where she thinks that the upper quai is. The answer is given in line 5 in a Bernese dialect form. In line 6, FI repeats part of the answer in German (a securing move) and then, addressing Cornelia (Co), immediately switches back to French. In the following lines she explains the way to the upper quai in French using such colloquial forms as et pis. In line 12 she does not immediately find the word ruisseau but it is not clear whether or not this is due to her restricted knowledge of French. The Biel/Bienne model explains the switching from French to German back to French: Given the start in French, the addressee was forced to switch back to French when speaking anew to the addresser. We already saw in the examples cited that the Biel/Bienne model does not require near-native competences in the L2. As long as the goal of the communication can be reached with the participants’ restricted linguistic means there is no need to change the Biel/Bienne model. Only in four cases did the conversations in public places lead to problems (see above) due to restricted language knowledge. However, there may be more problematic situations such as medical consultations where restricted language knowledge may produce serious problems. We did not make recordings in such settings, but, in the interviews of module 1, experiences of this type were referred to by the interviewees.

Example 4. The Swiss model in Fribourg: poireau As already said, we also found instances of the Biel/Bienne model in Fribourg, but not as often as in Biel/Bienne. Overall, Fribourg prefers the default-language model. This

153

154 Iwar Werlen

Example 4.

may lead to a bilingual communication of the receptive multilingualism sort. This can be illustrated by an example from Fribourg (see example 4).

7. Conclusion Receptive multilingualism is one of several possible solutions to the problem of communicating in a multilingual setting in Switzerland. It is officially fostered by the Ministers of Education. The figures of an ongoing research project show that the official language policy of the Ministers had a successful impact on the population. But receptive multilingualism is only one of four possible models. The lingua franca model is another possibility which, however, is not very frequently used. More widespread is the default language model: Even in multilingual settings, the language of the majority is used. In bilingual places such as Biel/Bienne, however, rules other than the one defining the Swiss models are adapted, leading to a specific Biel/Bienne model. The idea of such models, however, may be criticized as being too static and not taking into account the dynamics of the speakers in a communication situation. As was shown in some of the examples, speakers may use their linguistic repertoires according to their communicative needs resulting in code-switches, bilingual communications (called parler bilingue by Lüdi and Py 2002: 131ff.) and — sometimes, even though not exem-


plified in our corpus — to the use of a lingua franca as e.g. English with tourists from a Scandinavian country. This flexibility of the speakers in multilingual settings remains a challenge for research on multilingualism, but also for language policy.

Notes on the transcriptions Transcriptions are made according to the Gesprächsanalytisches Transkriptionssystem (GAT, see Selting et al. 1998) with the exception of the notation of Swiss German dialects and French. Speakers are indicated by short forms; addressees are indicated by the letter I (interviewee) followed by m (male) or preceded by F (female). Researchers and students are indicated by initials or the first two letters of their first name. Swiss German is transcribed according to the proposals of Dieth (1986); it is a broad phonemic transcription using only the characters of the German (Swiss) typewriter keyboard. Standard German and French are transcribed more or less orthographically with the exception of hesitation signals (such as French eu) and assimilations (such as French ya for il y a). An interlinear literal translation is added in italics; this translation is as close as possible to the original and therefore often ungrammatical in English. As already mentioned in the text, for readers not familiar with Swiss German dialects and French, Swiss German dialects are printed in Courier New and French in Lucida Console.

Notes 1. The term “main language” was chosen in the 1990 and 2000 censuses instead of the formerly used term “mother tongue”. It was defined as “the language you think in and you have the best command of ”. In the censuses, people could indicate only one language as their main language. People who considered themselves bilingual had to choose one of their languages. For more information on the distribution of languages see Lüdi and Werlen (2005). 2. The Swiss Conference of the Ministers of Education does not publish in English; the publications cited are written in German and translated by us. The sources are cited with the German abbreviation of the Conference (EDK). 3. The project “Language competences of the adult population in Switzerland” (directed by I. Werlen) is part of the National Research Program 56 “Linguistic diversity and language competences in Switzerland”. It started in January 2006 and will be finished in December 2007. 4. Cf. note 3. 5. It is important to distinguish between contact situations in monolingual and in multilingual settings. The term “monolingual setting” does not imply that there are no speakers of other languages, but it means that the local official language is strongly preferred even if it is not the first language of the speaker. In multilingual settings, the situation is less clear, as we will show later on in the context of Biel/Bienne. 6. As explained in note 5.

155

156 Iwar Werlen

7. This is the legal situation according to the cantonal Constitution. The regulations of the Constitution are currently under revision. If this revision is accepted by the people, Biel/Bienne will be the capital of a bilingual region comprising a lot more communities than now. 8. Since school systems are very different in English speaking countries, the term high school may be misleading. The German term is Gymnasium, the French one gymnase; both lead to the Matur or maturité (known in Germany as Abitur). 9. For a more refined analysis of exolingual vs. endolingual and bilingual vs. unilingual communication see Lüdi and Py (2002): 161. 10. As is well known, both French and German distinguish between formal and informal forms of address, often called vous-form and tu-form.

References Acklin Muji, D. 2003. Le débat suisse sur l’enseignement des langues étrangères. Vers une réflexion sur l’auto-constitution du collectif helvétique. In anglais, Englisch, inglese, Englais … English! [Bulletin VALS-ASLA 77], H. Murray (ed.), 67–81. Neuchâtel: Institut de linguistique. Conrad, S.-J. , Matthey, A. and Matthey, M. 2002. Bilinguisme institutionnel et contrat social: le cas de Biel-Bienne (Suisse). Marges linguistiques 3, http://www.marges-linguistiques.com. Conrad, S.-J. 2005. Zweisprachige Kommunikation: Biel/Bienne und Freiburg im Vergleich. In Villes bilingues – Zweisprachige Städte – Bilingual Cities. Akten des Kolloquiums in Biel/ Bienne, 19.-20. März 2004 [Bulletin VALS-ASLA 82], S.-J. Conrad and D. Elmiger (eds), 43–62. Neuchâtel: Institut de linguistique. Conrad, S.-J. 2006. Verständigung im öffentlichen Raum von Biel/Bienne. In Le projet bil.bienne –bilinguisme à bienne * kommunikation in biel, D. Elmiger and S.-J. Conrad (eds). TRANEL (Travaux neuchâtelois de linguistique 43), juin 2006: 5–22. Neuchâtel: Institut de linguistique. Dieth, E. 1986 Schwyzertütschi Dialäktschrift: Dieth-Schreibung. Aarau: Sauerländer. (2nd edn., edited by Ch. Schmid-Cadalbert). Dürmüller, U. 1997. Changing Patterns of Multilingualism: from Quadrilingual to Multilingual Switzerland. Zürich: Pro Helvetia. EDK [Sch weizerische Konferenz der kantonalen Erziehungsdirektoren] 2004. HARMOS. Zielsetzungen und Konzeption. Juni 2004. http://www.edk.ch/PDF_Down loads Harmos /Harmos_Weissbuch_d.pdf. EDK [Schweizerische Konferenz der kantonalen Erziehungsdirektoren] 2006. Tätitgkeits programm der EDK 22.6.2006. http://www.edk.ch/PDF_Downloads/LLTG/tgpro_d.pdf. Elmiger, D. and Matthey, M. 2006. La diglossie vu du “dedans” et du “dehors”: l’exemple de Bienne et d’Evolène. In Le projet bil.bienne – bilinguisme à bienne * kommunikation in biel, D. Elmiger and S.-J. Conrad (eds). TRANEL (Travaux neuchâtelois de linguistique) 43, juin 2006: 23–47. Neuchâtel: Institut de linguistique. Elmiger, D. and Conrad, S. J. 2006. Introduction. In Le projet bil.bienne – bilinguisme à bienne* kommunikation in biel, D. Elmiger and S.-J. Conrad (eds). TRANEL (Travaux neuchâtelois de linguistique) 43, juin 2006: 1–4. Neuchâtel: Institut de linguistique. Elmiger, D. 2005. L’orientation de Bienne comme ville bilingue: entre protection du monolin-


guisme et promotion du bilinguisme. In Villes bilingues – Zweisprachige Städte – Bilingual Cities. Akten des Kolloquiums in Biel/Bienne, 19.-20. März 2004. S.-J. Conrad and D. Elmiger (eds). Bulletin VALS-ASLA 82: 17–29. Neuchâtel: Institut de linguistique. Furer, Jean-Jacques 2005. Die aktuelle Lage des Romanischen. [Volkszählung 2000, Struktur erhebung der Schweiz]. Neuchâtel: Office fédéral de la statistique. Hägi, S. and Scharloth, J. 2005. Ist Standarddeutsch für Deutschschweizer eine Fremdsprache? Untersuchungen zu einem Topos des sprachreflexiven Diskurses. Linguistik online 24(3), http://www.linguistik-online.com/24_05/haegiScharloth.html. Kolde, G. 1981. Sprachkontakte in gemischtsprachigen Städten. Vergleichende Untersuchungen über Voraussetzungen und Formen sprachlicher Interaktion verschiedensprachiger Jugendlicher in den Schweizer Städten Biel/Bienne und Fribourg/Freiburg i.Ue. [Zeitschrift für Dialektologie und Linguistik, Beihefte H. 37]. Wiesbaden: Steiner. Lüdi, G. and Werlen, I. 2005. Sprachenlandschaft Schweiz. Unter Mitarbeit von S. Colombo et al. [Volkszählung 2000, Strukturerhebung der Schweiz]. Neuchâtel: Office fédéral de la statistique. Lüdi, G. and Py, B. 2002. Etre bilingue (2nd ed.). Bern: Lang. Murray, H. 2003. The status of English in Switzerland. In Living with Languages. The contemporary Swiss model, J. Charnley and M. Pender (eds), 87–106. Bern: Lang. Py, B. 2006. Devenir bilingue à Bienne. In Le projet bil.bienne - bilinguisme à bienne * kommunikation in biel, D. Elmiger and S.-J. Conrad (eds). TRANEL (Travaux neuchâtelois de linguistique) 43, juin 2006: 71–85. Neuchâtel: Institut de linguistique. Selting, M. et al. 1998. Gesprächsanalytisches Transkriptionssystem (GAT). Linguistische Berichte 173: 91–122. Stotz, D. and Meuter, T. 2003. Embedded English: Integrating content and language learning in a Swiss primary school project. In anglais, Englisch, inglese, Englais . . . Englis, H. Murray (ed.), Bulletin VALS-Asla No 77: 83–101. Neuchâtel: Institut de linguistique. Watts, R. J. and Murray, H. (eds) 2001. Die fünfte Landessprache? Englisch in der Schweiz. Zürich: vdf Hochschulverlag AG an der ETH. Werlen, I. (ed.) 2000. Der zweisprachige Kanton Bern. Bern: Haupt. Werlen, I. 2005. Biel/Bienne – Leben in einer zweisprachigen Stadt. In Villes bilingues – Zwei sprachige Städte – Bilingual Cities. Akten des Kolloquiums in Biel/Bienne, 19.–20. März 2004. S.-J. Conrad and D. Elmiger (eds). Bulletin VALS-ASLA 82: 5–16. Neuchâtel: Institut de linguistique. Werlen, I. 2006. Formulierungen von Normen kommunikativen Handelns in Biel/Bienne. In Le projet bil.bienne – bilinguisme à bienne * kommunikation in biel, D. Elmiger and S.‑J. Conrad (eds). TRANEL (Travaux neuchâtelois de linguistique) 43, juin 2006: 87–109. Neuchâtel: Institut de linguistique.

chapter 6

The Swiss model of plurilingual communication Georges Lüdi Universität Basel

The Swiss have decided to teach three languages, the local language, a second national language and English, from Primary school on. This should enable communicative encounters between persons from different linguistic backgrounds. An asymmetry between productive and receptive competences should foster the receptive bilingualism in a form sometimes called “Swiss model”: everyone speaks his or her own language and understands the other. But real communicative practices do not match this stereotype fully. This paper analyses the complex dynamics of face to face interaction in some key examples of authentic cross-linguistic communication at work in Switzerland. It shows how native and non-native speakers take mutual profit from all the languages they know. Neither monolingual models nor receptive bilingualism strictu sensu dominate; “mixed” forms of exploiting the respective repertoires emerge. In order to understand this language use, the paper questions the traditional representations of language competence. Keywords: receptive competence, Swiss model, cross-linguistic communication, Switzer land

1. Introduction We all know that Switzerland is quadrilingual and basks in peaceful linguistic harmony. But this image is misleading and needs to be corrected in many respects. First, the Federal Constitution mentions four national languages (German, French, Italian and Romantsch), but, in addition, there are increasing unrecognized linguistic minorities of migrants who live in Switzerland (see table 1).1 Table 1. Linguistic minorities of migrants in Switzerland German French Italian Rhaeto-Romance Other languages

1950

1960

1970

1980

1990

2000

72.1 20.3 5.9 1.0 0.7

69.4 18.9 9.5 0.9 1.4

64.9 18.1 11.9 0.8 4.3

65.0 18.4 9.8 0.8 6.0

63.6 19.2 7.6 0.6 8.9

63.7 20.4 6.5 0.5 9.0

160 Georges Lüdi Table 2. Most important languages of immigration in 2000 Serbian/Croatian Albanian Portuguese Spanish English Turkish Tamil Arabic Dutch Russian Chinese Thai Kurdish Macedonian

in %

in absolute figures

1.4 1.3 1.2 1.1 1.0 0.6 0.3 0.2 0.2 0.1 0.1 0.1 0.1 0.1

103,350 94,937 89,527 77,506 73,425 44,523 21,816 14,345 11,840 9‘003 8,279 7,569 7,531 6,415

Table 2 gives an overview of the most important languages of immigration in 2000 (in absolute figures and in percentage of the whole resident population). Second, the territoriality principle allows or rather constrains the use of only one of the official languages in each of the language regions of the country with the exception of a few overlap areas such as Bienne, Fribourg or the federal capital Berne. Juridically, Switzerland is thus a mosaic made up of largely monolingual regions in which the other national languages enjoy more or less the same status as, say, Spanish or English. The school system, with obligatory language teaching in a second national language from the 3rd/5th grade of primary school onwards, does make an attempt to correct this picture. But even if every adolescent in Switzerland has acquired a basic knowledge of one of her/his neighbouring languages, we can hardly speak of bilingualism, let alone multilingualism. We will come back to this point. Third, German speaking Switzerland provides the stage for a diglossic situation between standard German (often called High German or written German) and Swiss German, the functional distribution of which can be roughly described as ‘medial’ (Kolde 1981). In almost all social situations throughout German-speaking Switzerland dialects are spoken in all social classes without exception (in 2000, 86.1% of all inhabitants of the German speaking region, including the speakers of other languages, spoke Swiss German [exclusively or together with another variety] at home, but only 9.0%, mostly foreigners, spoke Standard German [exclusively or together with another variety]). The diglossic situation and the fact that the dialect can be shown to have adopted an increasing number of functions in the electronic media and in written communication (chat rooms, short messages)2 in the last decades3 have led to considerable unrest among the Romance-speaking minorities. Fourth, as everywhere else in Europe, English plays an increasing role as a foreign language taught at school, as a language at work and as a language present in the “lin-

The Swiss model of plurilingual communication 161

guistic landscape”. In 2000 (Federal census of the population with questions on language use at work), 23.4% of the active population in the German speaking part of the country indicated that they used English at work on a regular basis. In 1990 they were 17.4%. The corresponding figures are 17.7% in 2000 and 13.0% in 1990 for the French and 11.0% in 2000 and 7.7% in 1990 for the Italian speaking parts of the country For obvious reasons (shared institutions, existence of a national economy, frequent mobility of parts of the population, etc.), communicative encounters between persons from different linguistic backgrounds, may they live in the same linguistic region or communicate across the language borders, are frequent and necessary. Another frequently reproduced myth says that communication is based on the principle “everyone speaks his or her own language” because all Swiss citizens are competent in several national languages, at least in the three major ones. This paper pursues the question whether real communicative practices match this stereotype or not.

2. The educational context As already mentioned, most Swiss citizens are monolingual during their childhood (if we exclude the migrant population with bi-/plurilingual repertoires including the local language and one or various non-national languages). Therefore, the basis for the communication between the language groups must be laid by the educational system. A first major reform of foreign language teaching started in 1975 when the Conference of Cantonal Ministers of Education (short EDK or CDIP) decided that (a) all Swiss children (and not only those “gifted” for language learning and attending secondary education with enlarged requirements) should have the opportunity to learn a second national language, (b) this learning should start in Primary School (4th or 5th grade) and (c) it should be based on a communicative approach and not on the grammar-translation-method, which dominated the Swiss educational system at that time (CDIP 1975). Eleven years later, having realized that the next school level did not account for the skills acquired at Primary School in an appropriate way, the same Conference introduced so called “Meeting points” (CDIP 1986) in order to coordinate the teaching of the languages at different levels. Even more explicitely than in 1975, the accent was laid on communicative skills not only for the Primary School, but for the whole period of compulsory education. Receptive skills should be significantly higher than productive ones at the end of the 9th grade, and orality more important than reading and especially writing skills. Even if it is easier to control knowledge than aptitudes, the accent should be laid on the latter. The vision explicitly aimed at by this asymmetry between productive and receptive competences was that — because the whole population would have acquired a second national language — crosslinguistic encounters would be possible for everybody at least between the German and the Swiss speaking French (the others would have to acquire active knowledge in at least one of those two) along the lines of the “Swiss model” (Kolde 1981, CDIP

162 Georges Lüdi

1986). With this “plan”, we are manifestly at the heart of the debate about receptive multilingualism. Even if by the mid nineties most of the cantons had adopted the 1975 and 1986 recommendations, a general dissatisfaction with the quality of the L2 competences acquired by the young generation was felt (only the elites were really able to apply the “Swiss model”, e.g. at national expert meetings in the Swiss Academy of Human and Social Sciences or in the Council of Research of the Swiss National Foundation). This is due to different reasons, which we cannot discuss in detail here, e.g. the Romands’ dissatisfaction with the fact that the language spoken by their German compatriots does not really resemble the German they learn at school, too few hours dedicated to L2-teaching, inappropriate methods and ways of testing, inappropriate teacher training etc. But a new major challenge had appeared: learning a second national language and communication along the lines of the “Swiss model” were becoming more and more overruled by the presence of English as an international lingua franca. Why should children in Zurich learn French — and those in Geneva German — if they could easily communicate in English? And what about the thousands of migrant children whose L1 is not the local official school language? With all these problems in mind, the CDIP commissioned a Holistic language-policy plan (“Gesamtsprachenkonzept”) that should make proposals for a deeply rooted reform of language education in Switzerland comprising the local national language, a second national language, English, languages of immigration and other ancient and modern languages. The plan was presented in Summer 1998;4 another six years of political discussions were necessary until the corresponding political decisions were taken in March 2004. The plan is based on a sociolinguistic, pedagogical and political analysis of the situation, and on the following specific premises: • Th e ability to express trains of thoughts precisely, to argue in a differentiated way, and to communicate ideas by using the medium of language is essential for the cultural and political life as well as for the generation and dissemination of knowledge. • Linguistic and cultural diversity is deeply rooted in the history of Switzerland and Europe. It is part of the continent’s identity and of the idea that Swiss people have of their country. Its cultivation and maintenance is an explicit goal of the Swiss (cf. article 70 in the Constitution) and European policies. • Due to numerous movements of migration, but also to the expansion of private tourism, Switzerland, historically quadrilingual, became multilingual. • The knowledge of neighbouring languages or partner languages not only permits cross-border communication, but also and especially contributes to the development of mutual comprehension and of an attitude of tolerance for other cultures and thus to peace. • As an additional professional qualification, proficiency in various languages becomes more and more important in a world characterised by mobility and globalisation.


• I n terms of educational policy, the study of foreign languages has a central meaning as an element in the education for citizenship. • During the obligatory school period, the foundations have to be laid so that each individual and the society as a whole can take on those challenges and meet them with success. • Building on their original monolingual or plurilingual competences, the repertoires of the pupils should be broadened in the direction of functional plurilingualism, within the realms of an integrated language pedagogy and without augmenting of the total strain. In summary, great importance is attributed to the development of the local language (in its standard variety in diglossic contexts), the learning of additional national languages, the languages of the neighbours (among them also the languages of immigration) and languages of larger diffusion because this contributes to building up a stable society, functioning plurilingually and open to a multicultural world. The recommendations contain clear references regarding the number of foreign languages (at least one second national language and English) and the general objectives that should be achieved (the standards being formulated on the basis of the Common European Framework scale). This meets the requirements of a management oriented to efficiency. The goals formulated are demanding and exceed in many ways the results previously obtained by the educational system. Therefore — and knowing that the educational system was not fully successful in reaching the more limited goals — ways and means are described for enhancing the efficiency of language teaching in order to achieve these objectives, e.g. starting with second language learning at grade 3, two languages at primary school, meeting points guaranteeing vertical (between the levels) and horizontal (between schools and cantons) coherence, integrated language pedagogy, teaching subjects through the media of foreign languages (CLIL), exchange pedagogy, general diversification of teaching/learning methods in order to meet the needs of different student profiles and of different combinations of languages (teaching Italian as L2 being for instance different from teaching Italian as L3), “intercomprehension” methods teaching several related languages simultanously, etc. These recommendations are starting to be implemented, some of them meet political resistances: Learning two foreign languages at primary school is questioned by some teacher organisations, offering of languages of immigration is rejected by local nationalist politicians). Further, there is a heavy debate going on about the order (English > Second national language or Second national language > English). There are voices advocating in favour of intercommunity communication in English in Switzerland (cf. Watts and Murray 2001). Nevertheless, we might say that — as in most political declarations throughout Europe — there is consensus among Swiss that the solutions for the communication problems in an increasing multilingual country may not be monolingual (one dominant national language or English), but must be plurilingual (several national languages and English and other languages).

164 Georges Lüdi

3. French–German intercommunity communication: from myth to reality We said already that the “Swiss model” might be an idealisation. Let us become more precise now. When talking about communication in situations where the partners rely on asymmetric competences, applied linguists frequently speak about “exolingual communication”, characterized not only by the difference in itself, but also by the fact that all partners are conscious of this situation and use different strategies to compensate for these differences and to avoid misunderstanding or non-understanding (see already Noyau and Porquier 1984, Py and Alber 1985, Lüdi 1989). In most cases, the underlying representation of how communication works is monolingual: One linguistic variety is chosen that is at least partly shared by all the partners and both, i.e. the partners with a weaker and those with a better competence of the language respectively, make an exclusive use of this variety. A Morroccan and a French partner in Paris will exclusively speak French, an Albanian and a German in Munich German, etc. In cases where no one has even little knowledge of the other’s language, a third code may be used as lingua franca, e.g. English between a Chinese and a German. However, many researchers have argued since several decades that this conception of cross-linguistic or intercultural communication is biased by what we may call “homoglossic” views of the society and “monolingual” representations of the individuum (Lüdi 2005c). For many years, a bulk of research has shown that the interacting partners often have the choice between several more or less shared language competences because more than half of mankind is plurilingual by birth and/or has formally or informally learned additional languages in multilingual settings. In fact, pluri- and multilingualism is the normal case in most parts of the world, unilingualism can be defined as an exception. The partners might still want to insist on monolingual ways of communicating e.g. for social reasons, because one and only one variety is perceived as the “legitimate” one in a given situation (Bourdieu 1982). But while it is their choice to define the situation as “monolingual” despite having alternatives, they can also decide to mutually define the situation as “bilingual” or “plurilingual” even if their competences are not equal. This does not only involve reframing of the social situation, but also another way of activating the languages in the brain in the so-called “bilingual mode” (Grosjean 1985, 2001). This is to say that the factual symmetry or asymmetry between the speakers’ competences does not automatically entail specific ways of behaviour, but that the interacting partners have considerable freedom in resolving the communicative task and in defining the situation. This applies particularly if we swap a perspective that is centred on modelling the behaviour of speakers or hearers for an interactive perspective. Interactivity means a reciprocal definition of the communicative situation by all interacting partners (Schutz 1967). Based upon Schegloff ’s definition of “discourse as an interactional achievement” (Schegloff 1982), we claim that the choice of the language mode is locally negotiated and defined by the partners — in the compass of systems of social values, but sometimes with the explicit intention of questioning and reshaping these systems.


Many people still associate “mixed” forms of speech with insufficient competences for choosing a monolingual mode. This, definitely, is not always the case, as a huge amount of research has shown in more than 30 years (see for example Milroy and Muysken 1995). Today, code-switching as a recognised form of use of bi-/plurilingual repertoires is largely documented by research (cf. Lüdi 2005a for an overview) and has led to highly valued forms of multilingual literature (cf. Lüdi 2005b). On the other hand, it is evident that “interlinguistic compensation strategies” (Faerch and Kasper 1983a and b) are frequently used by beginning learners (even in socially monolingual situations) not because they actively choose a bilingual mode, but because they just cannot do better. There are good reasons for the hypothesis that code-switching (bilingual technique) and translinguistic wording (exolingual technique) can be distinguished (Lüdi 2003). But is this distinction relevant when it comes to explaining how people really interact in daily life? A research project on Language Dynamics and Management of Diversity, financed by the European Union and starting on October 1st 2006, is intended to generate a broad and in-depth evaluation of competing monolingual and plurilingual scenarios of language use in different professional and institutional settings. It will compare the advantages and drawbacks of these scenarios with due attention to the associated effects in terms of efficiency and fairness. Expected research results will constitute essential groundwork towards a deeper and more integrative understanding of the complex processes linking key dimensions of language learning and language use as basis for formulating language policy proposals. The project is founded on new conceptions of ‘multilingual competences’. They are viewed less in terms of formal (syntactic) knowledge and more as a set of linguistic resources available to members of a community for socially significant interactions. The totality of these resources constitute the linguistic repertoire of a person or a community (Gal 1986), not in an “additive” sense (the multilingual person as a kind of combination of various monolingual components), but in a holistic view (Grosjean 1985; Lüdi and Py 2003) which values the integration of partial competences (Council of Europe 2001). A positive assessment of partial competences also sheds new light on the “exolingual” techniques for assuring intercomprehension in the case of asymmetric competences mentioned before. But the notion of competence itself has undergone deep changes in the course of the twentieth century. The “Hymesian revolution” made it possible to go beyond the reductionist conception of linguistic competence introduced by Chomsky. It recognised the plural nature of communicative capacities and allowed for accounts of the discursive and pragmatic dimensions of language use. For Hymes and his successors (for example Canale and Swain 1980), the focus remains however on the observable competence of the learner and on the know-how deposited in his or her brain. There are several intents to move beyond these approaches that tend to misjudge the situational sensitivity of competences (regardless of whether this competence is linguistic or pragmatic) and the complexity of the processes in which these are implemented, as part of the practical activities of learners, users and their interlocutors.

166 Georges Lüdi

An interactionist interpretation of competence, by contrast, is based upon a contextualised and collective conception of activities and of human cognition, and focuses on the central role that practical communication (and, therefore, social action) play in their formatting. This perspective deconstructs, in a radical manner, any individualising, decontextualised and isolating definition of linguistic and communicative competences. Research on language acquisition confirms the high degree of unpredictability and the complex dynamics of face to face interaction in authentic interaction (Pekarek 1999). At the same time, it sheds new light on the ways native and non-native speakers of a language take mutual profit from their entire repertoires using their respective first languages and other support languages (e .g. English when speaking French as L2, French, English, Spanish and Latin when speaking Italian etc.). We thus need a language theory that gives special prominence to the ways the interlocutors exploit all the resources that are at their disposal. We fully agree with Franceschini (1999) who summarises her considerations as follows: I conceive the linguistic system from the perspective of its use in interaction. In human evolution, the System Language emerges from its use, in a phylogenetic as well as in an ontogenetic perspective. Language developed on a biological base, but always in the context of social interaction. (1999, 272 my translation; see also Tomasello 1999, 2003 and Bates 2003).

According to this concept, the elaboration and the mobilisation of competences are configured in the course of practical activities that are linked to specific sociocultural contexts and to particular forms of action, interaction and intersubjectivity (Berthoud, Grin and Lüdi 2005, Pekarek Doehler 2005, Lüdi 2004, 2006).

4. Examples of intercommunity communication at work It is relative to this background and with a view to elaborating such considerations that we recorded many hours of oral interaction between actors from very different linguistic and social backgrounds, in and outside Switzerland, in order to get a better understanding of the alternatives present in potentially plurilingual situations (Lüdi, Py et al. 1995, 2003, Franceschini 1999, Mondada 2003, 2004 a and b, Miecznikowski and Mondada 2001). In the following, we will concentrate on examples of internal communication at work in Switzerland. The research question is whether monolingual models on one hand (the use of only one national language or, progressively, the exclusive use of English) and on the other the “Swiss model” of receptive bilingualism — ‘every one speaks his or her language and understands the other, i.e. mobilises exclusively passive competences in the other’s language’5 — are realised in real life or, in other words, how these ideal types are reflected in daily interactions. We formulate the hypothesis that in a situation of shared, even asymmetric, bilingualism, other forms of communication will emerge. We conceive the analysis of key examples as part of a search for empirical evidence without any claim for statistical relevance.


The first example was recorded in Montreux by Bettina Wetzel-Kranz (Wetzel-Kranz 2000). ‘Walter’ (all names changed) is a young German Swiss informatician mandated to install a new computer program in the office of a small publishing house where ‘Marinette’ and ‘Jeanne’, two young French Swiss women, work as professionals. He is training them in front of a computer where the program is running. The default language choice seems to be in contradiction to the Swiss model: the German speaker chooses French, French speaking Marinette often German and sometimes French. Only Jeanne mainly sticks to her dominant language French. The three of them make sporadic use of English words or expressions, always in relation to what they observe on the screen. In the second part of the example (unrecorded speech of several minutes between lines 34 and 36)‚ Horst, the German boss of Marinette and Jeanne, joins the scene (line 44). The personal constellation changes, and so does the linguistic behaviour. In the unrecorded part and from line 36 onwards, all participants were in an exolingual-monolingual French mode. When Walter addresses himself to Horst, he uses (Standard) German. After a short period of exploratory language choice, i.e. alternation between French, German and English, the base language changes to German in line 55 (even if Jeanne continues intervening in French): Example (1)

5

10

15

20

25

Marinette: Walter: Marinette: Walter: Marinette: Jeanne: Marinette Walter: Marinette Walter: Marinette: Walter: Marinette: Walter: Marinette: Walter: Marinette: Walter: Jeanne:

exploring? oui [on va] sur exploring eh et couper data c’est où? . là? äh non . oben au-dessus ( ) G . als äh ah là . ici . ouais ja genau . und äh . products livreur livreur . texte [was?] nein . Text . unten ouais partie A hier hast du A ah okay und alle Sektionen comment fonctionne l’Europe ((3 sec.)) comment ça- comment fonc-tionne l’Europe’ ça c’est . comment fonctionne l’Europe- c’est le seul qui ne soit pas de la partie B Marinette: mais ça c’est la partie A Walter: il est de partie A ça? Jeanne: ah vous êtes en partie A là . ah non je croyais qu‘on était/ qu‘on faisait la partie A . B excuse-moi Walter: comment fonctionne Europe . c’est dedans maintenant? . comment fonctionne . c’est tout en anglais ici Marinette: äh

168 Georges Lüdi

30

35

40

45

50

55

60

65

Walter:: Europe place . European Union ((6 sec.)) Marinette: how Europe works ((3 sec.)) ah . Walter: mais pas ici Marinette: non mais . comment fonctionne l’Europe . c’est . es ist diese Titel Walter:: ah okay Marinette: äh . schau mal wieder bei exploring Walter:: oui ((the conversation goes on for several minutes)) Marinette: les images en fait peuvent nous parvenir en principe de chaque personne concernée par l’article . si l’article c’est par exemple l’article c’est le comité européen . Walter: oui Marinette: ben l’article c’est Walter: attends attends . peut-être je . je . Horst’ könntest du mal kommen’ . weil jetzt gehts um die Ressourcen Horst: ja Walter: maintenant il est en train de faire ça ((Horst joins the group)) you have made the ressources here’ Horst: yeah yes . now we have here the images from editor Walter: Horst: mhmh Walter: now it will be received from Taylor . the first . aso Jeanne: sent to Taylor Walter: sent to Taylor or received’ Horst: received Marinette: no . sent .. the images ((3 sec.)) Horst: kommen von Taylor . gehen an die Grammatek Jeanne: ah ah ah Marinette: ja . aber zuerst Jeanne: (le troisième) Marinette: zuerst wir bekommen die Fotos von eh die Leuten nein . das wo der Taylor verantwortlich is kriegt er sie . das ist mir so gesagt Horst: worden . Walter: et ça maintenant Marinette: ja weil . aso Horst: und es macht ja auch Sinn . weil der Taylor muß sie ja erst mal sehen ob es gut is Marinette: klar . aber . zum Beispiel die Adressen ist immer uns . und dann . wir schicken . weil zum zum Beispiel . ich hab das Problem gehabt . weil . wir haben ein Fotos bekommen äh äh und jetzt müssen wir das zu Taylor schicken (recording and transcription: Bettina Wetzel-Krantz)

Generally speaking, the overall language choice is inconsistent as one can expect for multilingual speech (Lüdi and Py 2003). The four persons make a very pragmatic use of all the communicative resources they share. Walter and Marinette display a fairly good active interactive competence in the second national language; Jeanne’s German


is only receptive as is Horst’s knowledge of French. “English only” could also be an alternative as shown in the intermediary part of the interaction. But normative representations of either monolingual language use or of what the “Swiss model” could be seem totally absent. At the same time, we get a first glimpse of existing power relations between Horst and his collaborators. Communication does not, of course, occur in a “neutral space”; multilingual speakers exploit their repertoire in order to press the maximum gain out of their language choice. Even if Horst’s preferential language choice is normally English (he often imposes English to the others), here he switches to German in line 55, entailing a general change of the base language of the communication. Nevertheless, it is evident that multilingual forms of speech can be very rewarding, as illustrated for example in the linguistic behaviour of Walter. Furthermore, this interaction, which illustrates various ways of mobilising multilingual repertoires, is obviously successful. In the list of activities presented in the Common European Framework (Council of Europe 2001), it would enter into the category “practical goal-oriented co-operation”. Thus, in order to measure the success of this communicative event, we do not have to assess the quality of the language(s) spoken, but we have to answer the question whether the goal of the interaction has been achieved. An analysis of the mutual reactions of the interacting partners to the ongoing conversation shows that this is the case in the perception of the members themselves. We get further confirmation in the activity report written by Walter at the end of his stay in Montreux. Let us now turn to a second example. The setting is in a regional bilingual bank that we have analysed in detail (Heiniger 2005, Lüdi and Heiniger 2005). This institution resulted from a merger between a former monolingual French-speaking bank in the canton of Jura and a monolingual German-speaking bank operating in the Laufental. From the beginning, the new company was supposed to work in a geographical area covering the French-German language border, with local agencies working in the languages of the respective local clients while the head offices merged and thus came to work in both languages. For those responsible it was always clear that the new bank would be bilingual, the explicit language strategy formulated by the head of the bank being that every employee should continue speaking his or her language, even in joint meetings. The reality as observed in an ethnolinguistic study turned out to be slightly different. Let us first say that the new bank is very successful commercially. Quite evidently, communication in the merged back-offices between persons from different language backgrounds works quite well. But it is not the “Swiss model” in its ideal sense. “Ich habe einfach festgestellt, dass das [sc. Jeder spricht seine Sprache] nicht ausreicht, weil, vor allem die welsche Seite noch ein bisschen…, noch nicht so weit ist, oder nicht, dass sie uns verstehen. Darum muss man halt einfach jeweils meistens auch auf Französisch probieren [I have realised that it (i.e. each one speaks her or his language) is not sufficient, because, mainly the French speaking side still has to ... , is not ready, I mean, to understand us]”, says one informant. For different reasons (lack of status and knowledge on both sides of the border in dominantly rural areas,6 lack of competences with

170 Georges Lüdi

some actors in the bank, etc.) English plays no role at all in the internal communicative networks. The three varieties (French, German and Swiss German) do not appear to be chosen randomly, but neither is there an evident preferential choice for each person. Werner, the highest ranking participant in the discussion, switches between the varieties, frequently chooses the “other” language, translates or asks for translation. Jean always speaks French; Heiri uses the emotionally loaded Swiss dialect knowing about the risk that Jean will not understand; Werner uses different bridging techniques. Here one short example: Example (2) Werner Jean Werner Jean Werner Heiri Werner

Aso, dann, dann werden wir das, das Kombi-Flex Produkt werden wir auch in diesem Prospekt abbilden, he? On veut, Michel, on Oui. On laisse ce produit combi-flex Hypothek dans le prospectus. Ah, il va venir justement ce que l’on fait ici on aura, ce prospectus et puis on va offrir ce produit, eh Ok … Jäjä, wenn eine frogt, scho. parce qu’on dit ici falls gewünscht, si on souhaite. (Recording and transcription: Monica S. Heiniger)

In short: even if ‘every one speaks her/his language’ is the generally accepted communicative maxim, it is however sometimes combined with or replaced by other techniques: • S ome actors (mainly speakers of German) accommodate to the French-speaking interlocutor and choose French from the beginning or switch to French when a communicative problem appears. • German speaking persons normally do not use the default oral variety but make an effort to speak standard German; however, Swiss German is chosen in emotionally important moments even if the Romands do not understand. • (More) bilingual colleagues are asked to translate or at least to make a summary in the other language (see De Stefani, Miecznikowski and Mondada 2000). • Very often, the utterances in either language are not pure, but mixed: concepts, lexical items or short phrases of an embedded language are splashed as “translinguistic markers” (Lüdi 2003) into utterances in the base language. This happens with particular frequency in cases where German-dominant persons speak French. We could add many similar examples of talk at work to this list. They do not really contradict the model of receptive plurilingualism. However, they show how an abstract concept is put into practice, how, from the level of a model, it is translated into real life. Several reasons may cause deviations from the ideal model: (a) someone’s receptive competence in the other language is not sufficient to ensure communication; (b) some

The Swiss model of plurilingual communication

participants willingly use a variety some of their interlocutors do not understand. It is a particularity of polyadic conversations, that one can make a language choice that excludes some of the participants; (c) accommodation patterns lead someone to the use of the interlocutor’s language as preferential choice, etc. The main lesson to draw from these examples is, however, that there is a lot of freedom in the individual linguistic behaviour, but also that responsibility for ensuring communication is shared by all participants in the interaction. The awareness of specific constraints in plurilingual interactions is also visible in our third example, documented in a slightly different context. Participants from several European countries and language backgrounds meet at a colloquium on history of the antiquity organised by the Upper Rhine Universities. The whole colloquium was tape recorded in the framework of a research project on plurilingualism in scientific encounters (Mondada 2003). Example (3) ostman

5

10

15

steiner ((2sec)) steiner ((8 sec)) ostman

i may therefore conclude . with a questionmark\ .. might . the byzanz/ . xxx section/ about which . xxxx extensively adds where . and which . was so much at odds with nnnn/s principal interest . itself . be another example of how much . the present volume bears the imprint . of (yussouf . acüla)\ .. °thank you°\ nous vous remercions/ beaucoup/ nous avons . cinq minutes/

thank you\ ((rires)) steiner il y a certainement des ques[tions/ (liberman) [INDENT} [wow/ x (you) x °xx° liberman eh x in english . eh what interest me (what) . intrigued me\ .. is the problem of the (1) outside stimulation of turkish historiography . which came from russia from the immigrants from russia/ (e9/OT15058/am1B/CD5:12.37- 13.36)

A frequently reproduced stereotype says that international scientific encounters make a preferential or even exclusive use of English. English is very present in this meeting indeed, but the choice of the lingua franca is in no way dominant. Instead, many “plurilingual techniques” are employed. And it is manifest that the participants are aware of their mobilisation of plurilingual repertoires as can be observed from their frequent mentioning of their language choice. In our example, Ostman concludes his presentation in English by making a metadiscursive announcement of a question, the question and the usual thanks to the auditors (lines 1–5). The chairman Steiner thanks

171

172 Georges Lüdi

in French and opens the floor for discussion (6–9). Because nobody takes the floor, Ostman repeats his thanks ironically and earns laughter (lines 10f.). Steiner asks again for questions (12). The first question is by Liberman, in the language of the paper, but he announces explicitly that he will use this language (15). We retain from these examples that the receptive competence of all participants is mutually judged to be sufficient for using either French or English as a preferential choice. The next example from the same colloquium shows that this is also true for German. This time, three languages (at least) constitute the shared repertoire: Example (4) Müller

5

10

aber ansonsten . ist da leider gottes wenig rauszuholen was eh . <mich ((en riant))>= x =((petit rire)) Müller =dann enttäuscht hat aber na ja besser man merkt sowas nach drei wochen beschäftigung[wie wenn man ein . halbes jahr eh: schon drangesessen ist\ . vielen dank\ Steiner alors eh: écoutez ces problèmes/ paraît stimule:r/ des exposés tout à fai:t animé:s [xxx/ [INDENT} [((quelques rires)) Steiner ça a des effets/ positifs\ . bon euh: on a encore deux minutes Durand oui/ euh: . euh °eh° i . i think (i) don/t think you should really drop . (drop) that\ because there are there are so many things to do/ with these eh . magazines

(recording and transcription Mondada and team)

After a paper presented in German, Steiner thanks again in French and opens the floor, Durand asks the first question in English. But he does not do this without a preparation sequence (“oui/ euh: . euh °eh° i . i think”). This type of behaviour shows that the participants are aware of the plurilingual character of their meeting and of the difficulties raised by that fact, but also and mainly that they are willing to use specific techniques for the management of the linguistic diversity and, simultaneously, for guaranteeing the success of the interaction. In this case Durand, a French-speaking scholar, prefers English to French, probably due to politeness, because he assumes that Müller’s receptive French is weaker than his receptive English. Mondada and her team found many instances of language crossing, i.e. using another than the preferential language for many different functions. Again, the success of the plurilingual speech cannot be assessed in terms of linguistic norms, but only with regard to the scientific outcome of the conference, which was very good in accordance with what participants reported.

5. Perspectives As I said before, my purpose is not to blame the communicative maxime underlying the model of receptive multilingualism, but rather to show which forms this maxime


adopt in real life. Even if we find similar tokens of multilingual behaviour in many other parts of the world, it still seems legitimate to speak of a “Swiss model”. In fact, one cannot deny that external conditions exert an influence on the speakers’ behaviour. We think that long-term Swiss experiences in language use in French–German intercommunity communication contributed to the construction of a communication culture which might entail a higher acceptance of “mixed” speech than in other countries. Thus, an existing system of linguistic values determines which “linguistic capital” the knowledge and use of one or the other variety convey (Gumperz 1982, Bourdieu 1982). However, the fact that language choice is highly unstable, very dynamic and permanently renegotiated among the participants, suggests that we should question the traditional representations of what a — monolingual or multilingual — language competence is. One might want to continue the reflection in two directions: a. If we conceive the multilingual competence as an integrated whole, formed by partial competences in all the varieties (languages and dialects) that the repertoire of the multilingual person consists of, then we have to include a “strategic competence” (Council of Europe 2001), i.e. the mastery of exolingual and multilingual communication techniques as we have observed above. In other words, we continue here our plea for a new linguistic focus on the multilingual speaker/hearer (cf. Lüdi 2004). We would even dare to interpret the “perfect” knowledge of one language as a “partial competence”. b. If we reject the “essentialist” conception of language that underlies most modern linguistic theories at least since Herder, then we will have to focus, on the contrary, on the situated, locally negotiated use of variable linguistic resources that draw upon the multiple repertoires of all participants in the interaction (Lüdi 2006). Communicative competences are extremely sensitive to context. In the light of new models that stress the collective dimension of social action and cognition, the claim seems justified that linguistic resources are constructed, formatted and implemented in the framework of concrete, practical, common linguistic activities of the users. Following Franceschini (1998, 1999), Pekarek Doehler (2005) and others, we are sceptical about individualising, context free and autonomous (i.e. from other dimensions of social competence) conceptions of language competence. This is particularly important for constructing models of plurilingual communication. A broad consensus starts establishing itself that “the human language faculty has an endowment for multilingualism” (Meisel 2004). In the light of new research results, this cannot mean the predominant use of one single language at a time. The mobilisation of the participants plurilingual resources happens in situ, not consecutively but in one common movement, is frequently not determined in advance, but selforganised, negotiated among the participants. The latter put together all their possibilities, which are at the same time constantly reconfigured.

174 Georges Lüdi

In this sense, the described model of plurilingual communication — which might not be very “Swiss”, indeed, — can serve as a kind of anticipation of models of plurilingual interaction to cope with the growing diversity of the European Union, in order to find a “plurilingual answer” (Umberto Eco) for the communicative problems in a world characterised by growing globalisation. The colloquium organised by EUCOR, the association of the Upper Rhine universities Strasbourg and Mulhouse (French), Karlsruhe and Freiburg im Breisgau (German) and Basle (Swiss), clearly points in this direction. These considerations, of course, open the path to many new and intriguing questions. Let us just mention one: An increasing number of universities throughout Europe require students to show competences in foreign languages, usually in the form of language certificates. How does a new conception of plurilingual competences and plurilingual interaction become compatible with this requirement? Let us, first of all, reconsider a couple of statements of the Common European Framework for languages: a. The competence in a foreign language is never “perfect”, but can be situated on a differentiated scale going from beginners to independent users. b. The framework encourages profiling of objectives and competences, which constitute the answer to very many different communicative needs. A researcher may for example only need comprehension knowledge. He will have to answer questions like: What sort of things will he be listening to or reading? Under what conditions will he have to act? What knowledge of the world or of another culture will he need to call on? (Council of Europe 2001, 44). The response could for example be that this person needs listening and reading competence at the B2 level in his or her scientific discipline, but that the other skills might be situated at a much lower level. However, even if language courses more and more aim at objectives of the kind we just suggested, exams and certificates more scarcely do so. This is particularly true for oral interaction. Indeed, “the focus on the ability of the candidate in conventional approaches within second language assessment views the candidate in a strangely isolated light; it is he or she who held to bear the bunt of the responsibility for the performance; in this sense the inevitable gap between a test and real life appears unusually stark.” (McNamara 1997: 452) Research on the relation between interviewer style and candidate performance in oral exams (Brown and Hill 2003) sheds light on the particular context sensitivity of oral skills. If one adds to all that the plurilingual character of the workplace interactions we have observed (more than two languages), the frequent deviations from the “receptive multilingualism” model (oral production in foreign languages) and the particular awareness for the instability of the situation, it seems obvious that there is no secure procedure to test the capacity for successful plurilingual communication as it was described above, and that the best way for acquiring it is learning by doing.


Notes 1. For further details on the linguistic landscape of Switzerland in 2000 see Lüdi, Werlen et al. (2005). 2. See for example the work of Samuel Spycher (2004) on SMS communication and of Beat Siebenhaar (n.d.) on Swiss chat rooms. For early overviews already Ris (1990) and Siebenhaar (1996). 3. The results of the Swiss National Census show that the use of Swiss German continued increasing between 1990 and 2000, namely in the educational systems, but that the medial diglossia (one speaks Swiss Dialect and writes Standard German) is being weakened (Werlen 2004). 4. http://www.romsem.unibas.ch/sprachenkonzept (last visited January, 30th 2007) 5. One might call this model another manifestation of a monolingual ideology in the sense that accepting to understand another language could be the condition for refusing to speak it actively. 6. In 2000, i.e. two years before the merger, only 15.0% of the working population of Laufen ever used English at work and even only 6.7% of those working in Delémont. For the banking sector the respective figures were 19.4% and 11.5% respectively in the two townships compared with an overall rate of use of English of 51.8% and 42.2% in the banking sector of the respective language region.

References Bates, E. 2003. Natura e cultura nel linguaggio. In Frontiere della biologia. Il cervello di Homo sapiens, E. Bizzi, P. Calissano and V. Volterra (eds), 241–65. Rome: Istituto della Enciclopedia Italiana fondata da Giovanni Trecanni S.p.A. Berthoud, A.-C. , Grin, F. and Lüdi, G. 2005. La gestion de la diversité linguistique dans des contextes professionnels et institutionnels. EU project proposal. Bourdieu, P. 1982. Langage et pouvoir symbolique. Paris: Fayard. Brown, A. and Hill, K. 1998. Interviewer style and candidate performance in the IELTS oral interview. IELTS Research Reports 1: 1–19. Canale, M. and Swain, M. 1980. Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics 1: 1–47. CDIP 1975. Recommandations et décisions concernant l’introduction, la réforme et la coordination de l’enseignement de la deuxième langue nationale pour tous les élèves pendant la scolarité obligatoire, 30 octobre 1975. CDIP 1986. Points de rencontre: enseignement des langues étrangères à la charnière des scolarités obligatoire et postobligatoire, 30 octobre 1986. Council of Europe 2001. Common European Framework of Reference for Languages: Learning, teaching, assessment. Cambridge: Cambridge University Press. De Stefani, E. , Miezcnikowski, J. and Mondada, L. 2000. Können sie vielleicht kurz übersetzen: les activités de traduction dans des réunions de travail plurilingues. Revue Française de Linguistique Appliquée 5(1): 25–42.

176 Georges Lüdi Faerch, C. and Kasper, G. 1983a. Plans and strategies in foreign language communication. In Strategies in Interlanguage Communication, C. Faerch and G. Kasper (eds), 20–60. London: Longman. Faerch, C. and Kasper, G. 1983b. On identifying communication strategies in interlanguage production. In Strategies in Interlanguage Communication, C. Faerch and G. Kasper (eds), 210–38. London: Longman. Franceschini, R. 1998. Code-switching and the notion of Code in linguistics: proposals for a dual focus model. In Code-Switching in Conversation. Language, interaction and identity, P. Auer (ed.), 51–74. London: Routledge. Franceschini, R. 1999. Italiano di contatto: Parlanti occasionali e riattivazioni di conoscenze non focalizzate. Habilitation Basle: Philosophische Fakultät der Universität Basel. Gal, S. 1986. Linguistic repertoire. In Sociolinguistics. An international handbook of the science of language and society, U. Ammon et al. (eds), 286–92. Berlin: Walter de Gruyter. Grosjean, F. 1985. The bilingual as a competent but specific speaker-hearer. Journal of Multilingual and Multicultural development 6: 467–77. Grosjean, F. 2001. The bilingual’s language modes. In Language Processing in the Bilingual, J. L. Nicol (ed.), 1–25. Oxford: Blackwell. Gumperz, J. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Heiniger, M. S. 2004. La gestion du plurilinguisme dans une entreprise: la perspective des acteurs. Basle: Romanisches Seminar (final paper). Kolde, G. 1981. Sprachkontakte in gemischtsprachigen Städten. Vergleichende Untersuchungen über Voraussetzungen und Formen sprachlicher Interaktion verschiedensprachiger Jugendlicher in den Schweizer Städten Biel/Bienne und Fribourg/Freiburg i.Ue. Wiesbaden: Steiner. Lüdi, G. 1989. Aspects de la conversation exolingue. In Actes du xviiie Congrès International de Linguistique et de Philologie Romanes. Université de Trèves (Trier) 1986, T. VII. D. Kremer (ed.), 405–24. Tübingen: Niemeyer. Lüdi, G. 2003. Code-switching and unbalanced bilingualism. In Bilingualism: Beyond basic principles. Festschrift in honour of Hugo Baetens Beardsmore. J.-M. Dewaele, A. Housen and Li Wei (eds), 174–88. Clevedon: Multilingual Matters. Lüdi, G. 2004. Pour une linguistique de la compétence du locuteur plurilingue. Revue française de linguistique appliquée IX (2): 125–35. Lüdi, G. 2005a. Code-Switching/Sprachwechsel. In Sociolinguistics — Soziolinguistik. An International Handbook of the Science of Language and Society. 2nd completely revised and extended edn. , Vol. I. , U. Ammon et al. (eds), 341–50. Berlin: Walter de Gruyter. Lüdi, G. 2005b. «Parler bilingue» et discours littéraires métissés. Les marques transcodiques comme traces d’expériences interculturelles. In: Des cultures en contact. Visions de l’Amérique du Nord francophone, J. Morency, H. Destrempes, D. Merkle and M. Pâquet (eds), 173–200. Québec: Editions Nota bene. Lüdi, G. 2005c. L’unilinguisme est-il compatible avec la démocratie? Romanistische Zeitschrift für Literaturgeschichte 29: 89–107. Lüdi, G. 2006 Multilingual repertoires and the consequences for linguistic theory. In: Beyond Misunderstanding. Linguistic analyses of intercultural communication. K. Bührig and J. D. ten Thije (eds), 11–42. Amsterdam: John Benjamins. Lüdi, G. and Heiniger, M. S. 2005. L’organisation de la communication au sein d’une banque régionale bilingue. Sociolinguistica 19: 82–96.


Lüdi, G. , Py, B. et al. 1995. Changement de langage et langage du changement. Aspects linguistiques de la migration interne en Suisse. Lausanne: L’Age d’Homme. Lüdi, G. and Py, B. 2003. Etre bilingue. 3rd edn. Berne: Lang. Lüdi, G. , Werlen, I. et al. 2005. Le paysage linguistique en Suisse. [Statistique de la Suisse. Recen sement fédéral de la population 2000]. Neuchâtel: Office Fédéral de Statistique. McNamara, T. 1997. ‘Interaction’ in second language performance assessment : Whose performance ? Applied Linguistics 18(4): 446–66. Meisel, J. 2004. The bilingual child. In The Handbook of Bilingualism [Blackwell Handbooks in Linguistics 15], T. K. Bhatia and W. C. Ritchie (eds), 91–113. Oxford: Blackwell. Miecznikowski, J. and Mondada, L. 2001. Comment construit-on des objets de savoir dans des réunions de recherche plurilingues? In Comunicare in ambiente professionale plurilingue / Communicating in professional multilingual environments. Atti del Convengo VALS-ASLA, Lugano, 14–16.9.2000, S. Cigada, M. Matthey and S. Gilardoni (eds), 217–40. Lugano: Edi zioni dell‘USI. Milroy, L. and Muysken, P. (eds) 1995. One Speaker, Two Languages: Cross-disciplinary perspectives on code-switching. Cambridge: Cambridge University Press. Mondada, L. 2003. La science polyglotte: Conditions et possibilités des interactions scientifiques plurilingues. In Langues et production du savoir. Actes du Colloque de l’Académie Suisse des Sciences Humaines, Lugano, 14.6.2002, 33–42. Bern: SAGW. Mondada, L. 2004a. La construction du savoir scientifique dans l’interaction plurilingue. Pratiques et représentations langagières dans la construction et la transmission des connaissances, Cahiers du Français Contemporain 9: 15–27. Mondada, L. 2004b. Ways of ‘doing being plurilingual’ in international work meetings. In Second Language Conversations, R. Gardner and J. Wagner (eds), 27–60. London: Continuum. Noyau, C. and Porquier, R. (eds) 1984. Communiquer dans la langue de l’autre. Paris: Université de Paris VIII. Pekarek, S. 1999. Leçons de conversation: Dynamiques de l’interaction et acquisition de compéten ces discursives. Neuchâtel: Delachaux & Niestlé. Pekarek Doehler, S. 2005. De la nature située des compétences en langue. In: Repenser l’enseignement des langues: Comment identifier et exploiter les compétences? J.-P. Bronckart, E. Bulea and M. Puoliot (eds), 41–68. Villeneuve d’Ascq: Presses universitaires du Septen trion. Py, B. and Alber, J.-L. 1985. Interlangue et conversation exolingue. Université de Lausanne. Cahiers du Département des langues et des sciences du langage 1: 30–47. Ris, R. 1990. Dialektologie zwischen Linguistik und Sozialpsychologie. In La Suisse face à ses langues — Die Schweiz im Spiegel ihrer Sprachen, J.-P. Vouga (ed.), 40–9. Aarau : Sauerlän der. Schegloff, E. A. 1982. Discourse as an interactional achievement: some uses of and other things that come between sentences. In Analysing Discourse: Text and talk, D. Tannen (ed.), 71–93. Washington D.C.: Georgetown University Press. Schutz, A. 1967[1962]. Collected Papers I: The problem of social reality. The Hague: Nijhoff. Siebenhaar, B. 1996. Das Verhältnis von Mundarten und Standardsprache in der deutschsprachigen Schweiz. www.germanistik.unibe.ch/siebenhaar/pdf/Siebenhaar_Mda_Std_unv1996. pdf. Siebenhaar, B. (n.d.): Das sprachliche Normenverständnis in mundartlichen Chaträumen in der Schweiz. www.germanistik.unibe.ch/siebenhaar/pdf/SiebenhaarChatGL.pdf.

178 Georges Lüdi Spycher, S. 2004. ‘I schribdr de no . . .’ – Schweizerdeutsche Umgangsformen in der SMS-Kommu nikation. Networx 36: 3–33. http://www.mediensprache.net/networx/networx-36.pdf. (last visited January, 30th 2007) Tomasello, M. 1999. The Cultural Origins of Human Cognition, Cambridge MA: Harvard University Press. Tomasello, M. 2003. Constructing a Language: A usage-based theory of language acquisition. Cambridge MA: Harvard University Press. Watts, R. J. and Murray, H. (eds) 2001. Die fünfte Landessprache? Englisch in der Schweiz. Zürich: vdf Hochschulverlag AG an der ETH. Werlen, I. 2004. Zur Sprachensituation der Schweiz mit besonderer Berücksichtigung der Diglossie in der Deutschschweiz. Bulletin VALS/ASLA 79: 2–30. Wetzel Kranz, B. 2000. Mehrsprachige fachsprachliche Kommunikation am Beispiel der trinationalen Ingenieurausbildung am Oberrhein. Lecture given at the colloquium “Kommuni kation in einer mehrsprachigen Arbeitswelt”, Lugano, 14–16 Sept. 2000.

chapter 7

Receptive multilingualism in business discourses Bettina Dresemann Universität Erfurt

English is the number one language of business communication today. However, sometimes various languages are applied in one discourse, which may lead to opaque sequences. Thus comprehension of utterances in a language that was not learnt by one of the participants is necessary in order to achieve non-linguistic goals. Participants in international business encounters cannot rely on their linguistic knowledge alone, but have to make use of different forms of knowledge. This article discusses examples of situations in which comprehension has to be achieved although the participants’ linguistic knowledge alone is not sufficient for this task. A qualitative-interpretative approach is used to explain how participants achieve comprehension and the capacity to act in linguistically opaque situations. The examples discussed are taken from a corpus of authentic business negotiations. Keywords: business negotiations, English as lingua franca, professional knowledge

1. Introduction Today’s globalised world with its increasingly international markets is constantly setting new challenges for communication. Communication can be carried out in one language only or in various languages, with or without the help of an interpreter. However, not only individuals can be multilingual, discourses also tend to be multilingual because often more than one language is used. These forms of multilingualism can be organised and use agreed forms, or they can develop spontaneously, sometimes without a participant’s agreement. Thus multilingual discourse can be problematic if the interlocutors do not have the same mastery of the language(s) used. In situations where languages are used which are not mastered by one or more interlocutors, opaque situations (cf. Müller 1989) may arise with the result that the communicative efforts become increasingly demanding. Thus participants need to make use of various strategies to cope with these situations. In forms of professional discourse like business negotiations, comprehension is paramount. In most instances of business encounters today, the language chosen for communication is English. Yet situations arise in which other languages play a decisive role.

180 Bettina Dresemann

There are instances in which the speaker’s mother tongue (L1) or the business partner’s national language is used. Although speakers sometimes do not know their interlocutor’s L1, they are still able to react or interact in these situations. In this paper examples from business encounters will be used to discuss how participants succeed in their non-linguistic aims even though they may not know the language that is being used. There will be further discussion as to how participants compensate for a lack of linguistic (i.e. mainly lexical and semantic) knowledge and how communication is made possible. After a short overview of some aspects of multilingualism and multilingual discourse, this paper will discuss two different situations in which languages are used that the participants are not familiar with. It will discuss forms and functions of multilingual discourse as well as the ways in which communicators cope with the situation and achieve their goals.

2. Aspects and functions of multilingualism Although some researchers differentiate between multilingualism as a societal phenomenon and bilingualism as individual behaviour, in this paper the term multilingualism is used to cover the behaviour of individuals in communicative situations in which two or more languages are applied. That is because in the situations discussed below two languages are spoken — English as lingua franca (ELF) plus the speaker’s L1, which is not known to the hearer. So at least three languages are involved, including the hearer’s L1, which influence the communicative situation. Thus, the abbreviation L2 is not only used for languages that have been acquired or learned by the user — if this differentiation is at all necessary — but for any language that is not the L1 of the language user in question. Most research in the field of multilingualism has been carried out on “active” multilingualism, i.e. speech production. The main aspect of interest has been grammar, i.e. morpho-syntax, and the lexicon (cf. Myers-Scotton 1997, 2002, Lüdi 1996, Clyne 1997). Here, the “compatibility” of two language systems has attracted most of the interest. On the pragmatic level, code-switching, as a realisation of multilingualism, has mostly been described as a means of establishing a social identity (cf. e.g. Heller and Pfaff 1997, Holmes and Stubbe 2004) or achieving a specific effect, e.g. being dramatic (cf. Clyne 1997: 309). However, what is required when discussing the ability and functions of multilingual communication is to focus on the interaction between participants (an approach that is in line with Firth and Wagners’ demand for a new perspective on communication; cf. Firth and Wagner 1997) and between participants together with their different kinds of knowledge (e.g. general, linguistic, professional, institutional knowledge). The pragmatic functions of different (speech) actions need to be considered when interpreting both their production and their reception.

Receptive multilingualism in business discourses 181

Few researchers have actually tried to explain the processes and conditions of receptive multilingualism.1 An exception is Zeevaert (2004, this volume), who describes the ability of Scandinavian speakers to interpret discourse sequences in other Scandinavian languages. Yet in Zeevaert’s data, the hearers were actually prepared to understand other Scandinavian languages, which were not their L1, but were linguistically related to it. According to Zeevaert (this volume) the ability to understand the interlocutor’s language is either dependent on the linguistic knowledge (A knows B’s L1) or on the closeness of the language, i.e. their typological similarity. There are situations, though, in which the languages are less closely related and yet one speaker uses his/her L1. Thus it is quite difficult to actually describe multilingualism as a personal ability, especially when it comes to less proficient speakers. It might be more fruitful to shift the perspective from linguistic knowledge to the interaction and to both the non-linguistic and linguistic procedures enabling communication. Thus a communicative situation can be defined — and even more objectively than an individual — as multilingual if more than one language either is applied or plays a role in speech production or in comprehension. As House and Rehbein (2004: 1) point out, multilingual communication is characterised by the following features: • • • •

The use of several languages for the common purposes of participants Multilingual individuals who use language(s) to realize these purposes The different language systems which interact for these purposes Multilingual communication structures, whose purposes make individuals use several languages

Interestingly, according to House and Rehbein, the use of different languages is determined by the “common purposes of participants”. The question is, however, whether the use of more than one language in discourse is necessarily determined by the common purpose or whether participants try to take advantage of using one language in preference to another. So far, encounters in which people had to use languages different to their L1 or different languages in one situation have usually been described as intercultural communication. Most research in the field of intercultural communication takes a critical view and regards communicative problems as arising from linguistic deficits or different cultural backgrounds. However, Loos (1999) e.g. shows that communication can be successful despite the fact that the participants’ linguistic and cultural knowledge may not be sufficient. According to his findings, interlocutors can create a common ground for their interactions by using cues.2 Although these contextualisation cues can, according to Gumperz (1982), lead to misunderstandings as hearers from different cultural backgrounds may interpret them differently, according to Loos, linguistic references to other oral or written texts are interpreted as cues indicating in which social context the speaker is acting, e.g. a business negotiation or organisational procedures for staying at a bungalow park (cf. Loos 1999). Thus communication can be successful because speakers refer to the situational


context to indicate this reference to the hearer. Ehlich and Rehbein (1986) even assume patterns of action in institutionalised contexts which structure the actions of speakers and hearers and at the same time enable the interpretation of actions on the grounds of predetermined steps of communication. Zeevaert (this volume) states that besides the linguistic and situational knowledge, the hearer can also rely on his general knowledge, body language, prosody and his general communicative competence to understand the situation. However, Zeevaert does not define “general communicative competence” in detail. Thus it is hard to differentiate between prosody and general communicative competence as interpreting prosody might be included in the latter category. In my opinion, prosody forms part of pragmatic competence, at least in typologically related languages. All in all, hearers can rely not only on their linguistic knowledge in communicative situations, but also on their general and professional knowledge to make sense of the ongoing discourse.

3. Business negotiations One type of discourse that is — perhaps more than many others — characterised by aims and conditions which lie outside the discourse itself are business negotiations. There is no clear-cut theoretical definition of what a negotiation is (cf. Wagner 1995). However, what most definitions of negotiations have in common is the awareness of the interlocutors that they either have interrelated goals which they want to combine for the profit of both, or that at least one of them has control over the other one’s goal (Wagner 1995, Rehbein 2001, 1995). Wagner and Petersen call this specific situation, in which cooperation in discourse is required to reach competitive goals, cooperation under competitive conditions (Wagner and Petersen 1993: 271). While Wagner (1995) and Wagner and Petersen (1993) define negotiations as a discourse type in its own right, Rehbein (1995) defines negotiations as an auxiliary device that can be applied within patterns of action, e.g. in the “pattern of buying and selling”, if the participants cannot proceed in the pattern of action because they have to achieve agreement first (Rehbein 1995). Wagner and Petersen (1993), on the other hand, define business negotiations as a whole process consisting of various phases of which only one phase is carried out in face-to-face interaction. They also include the preparation phase and the ratification phase. These are distinguished from the negotiation phase as such, in which details such as conditions of delivery are clarified argumentatively (Wagner and Petersen 1993: 272). Whether negotiations are regarded as an auxiliary device or a speech pattern, pattern of action, language game or the like — the decisive aspect is that they are characterized by cooperation as well as competition. Cooperation exists first and foremost on the communicative level; otherwise the whole interaction would not be possible. Especially if the interlocutors use an L2, cooperation is required to reach communicative goals, e.g. through a greater tolerance of non-standard forms or through negotiation of meaning.


However, the interlocutors’ aims might still be competitive. And thus the way they use language(s) to reach their non-linguistic goals might be less cooperative.

4. Language choice If the interlocutors do not share an L1, they might choose to rely on an interpreter or intermediator. In most encounters, especially in business, however, no interpreter is made use of. Participants usually choose to use a lingua franca instead. The language they choose in most cases today is English, i.e. English as a Lingua Franca (ELF), which means English used by non-native speakers, who only apply this language for a specific purpose (cf. Meierkord and Knapp 2002: 10). Although participants do not have complete mastery of the language (Knapp 2003), they are usually capable of achieving understanding and carrying out negotiations in English. Thus if other languages are used, the question is why this is so and what the communicative purpose or the speaker’s motivation is. Another question is how interlocutors react and how they make sense of the situation, i.e. which form of background knowledge is activated and enables their (inter)actions. To investigate this, it is necessary to go beyond the participants’ observable verbal and non-verbal behaviour. One approach that takes the participants’ knowledge as well as their institutional roles into consideration is the approach of Functional Pragmatics (cf. Ehlich and Rehbein 1986). It allows for conclusions about the different procedures and forms of knowledge which enable the continuation of communication. Thus some conclusions will be made about the way participants in multilingual business discourses process information and make use of various sources of knowledge to take part in the ongoing discourse even though their semantic knowledge alone does not enable them to do so.

5. The data The data in this paper were recorded for a PhD project on the features of English as a Lingua Franca (Dresemann forthcoming), which concentrates on the use of ELF in business negotiations. The examples do not form part of the project itself. Rather, they are an interesting by-product as they reveal that the language actually used in the negotiations is not only English. The use of other languages besides English was quite unexpected, especially as the interlocutors did not know their business partners’ respective L1’s. The examples are taken from two different business encounters. In both encounters, a label manager of a German record company, S1mGr, is involved in negotiations with different (potential) business partners about distributing music titles in Germany. All speakers are non-native speakers of English and have different L1’s, namely German, Norwegian, and French as indicated by the abbreviations Gr, Nr, Fr in the speaker’s re-


spective abbreviation. The encounters were audio-recorded at two international music trade fairs in Germany. They were transcribed in a relation of approximately 1: 20 using the conventions of HIAT 23 with some additions (see appendix).

6. Multilingual discourse Lexical inferences According to Wagner and Petersen’s definition, the preparation of a negotiation forms part of the negotiation process (Wagner and Petersen 1993: 272f.). So searching for contact data of possible business partners on a website is part of the preparation phase and thus of the negotiation in the wider sense. And the owner of the website can be expected to provide the necessary information for the potential business partner. Thus, the website can be regarded as forming part of the discourse — although in a wider sense — between potential business partners. However, web pages are not always in English. Sometimes they are created in the national language. In these situations there are no communicative hints, which the interlocutor usually gives in interaction, such as prosody or non-verbal behaviour. Thus the communicative problems have to be solved solely by the recipient’s reliance on his or her own knowledge. Example 1 shows how S2fFr, a native speaker (NS) of French and head of distribution of a French record company, tells S1mGr how she tried to contact him. She used his company’s website although it was only in German at that time, a language she does not know. During their meeting at the trade fair, she explains her problems by identifying the correct contact person. Although she was able to identify the correct department, she did not send her first email to S1mGr, but to his boss, Achim. Example 1. Inference distribution (665) S1mGr I mean/ you, you approached Achim, my/ – he is actually the head of the department – you approached him, and he forwarded your email to me. (666) S1mGr Yeah· ... (673) S2fFr Ehm ((1s)) and eh/ well, ehm I approached him… (674) S2fFr Honestly, I just picked up, you know, the name Meyer and… (675) S1mGr I know´. (676) S1mGr Yeah, I receive a lot of those mails, ( )… (677) S2fFr And it was written [“Vertrieb”], and I thought: “Hm¯ [Vertrieb, probably distribution.” [vεrtrɪeb [vεrtrɪeb (678) S1mGr Yeah¯ ((lacht)) (679) S2fFr So… (680) S1mGr Yeah, but it was perfect.


S2fFr contacted S1mGr’s company. As S1mGr mentions in segment (se) 665, she did not contact the right person. Here, the communicative problem in the preparation phase shows: S2fFr had identified the correct department of the company, but not the person in charge. However, she established contact with S1mGr through his boss and offered him products of one of her music labels. S1mGr explains the misunderstanding to her, obviously to let her know why she has to deal with another person than the one she originally contacted. After S1mGr has mentioned the problem, she explains what had happened, starting with some hesitation (se 673) but does not complete her utterance. Instead, she starts a new one in se 674 with “Honestly” and explains her searching the website and picking out the name. Thus her reaction can be seen as an apology for not contacting S1mGr. S1m Gr’s reaction, the backchannel “I know” with a rising intonation (se 675) and the statement “I receive a lot of those emails” (se 676), is an acceptance of her apology. It also indicates that many users of the website face the same problem and do not contact the correct person. It also shows that the way the information was presented may not be the most logical one as it obviously leads to misunderstandings. However, all in all S2fFr had achieved her aim as she organised a meeting with a representative of the correct department. In the rest of the short excerpt, she tells S1mGr how she looked for the information and then contacted his company, explaining her communicative problems and efforts: She read the German website and found the German term “Vertrieb” (Germ. pronunciation: /feɑtrip) (se 677). Although she did not know the term, as can be seen by her French pronunciation of it and by the passive construction “it was written”. She does not use any active voice here, by which she distances herself from the German term. At the same time the verb “written” only indicates the appearance of the term, but gives no indication of any form of understanding it on her side. In the next part of the utterance she indicates that she had to guess what its meaning was: the verb “thought” shows that she tried to process the information but was not sure about its meaning; so does the “Hm”, by which she indicated her insecurity. She then repeats the term and mentions her supposition as to its possible meaning, distribution. Her statement is weakened by the adverb “probably”, which makes her hesitation and insecurity explicit. The German noun Vertrieb is of Germanic origin and not related to the Latin distributio and thus not related to the English form distribution, a homograph of the French noun. So the form cannot have been of much help to her. However, she found the correct department. As the linguistic form did not indicate the department she was looking for, she must have used her professional knowledge to solve the problem. As a professional she knows that enterprises indicate the departments of external relations on their websites. So she can assume that the (virtual) interlocutor, the company owning the website, provides the information that is required. So one of the units of S1mGr’s company given on a website had to be the distribution department. Although she does not give any hint as to how she identified it — be it, e.g. , by exclusion of other organisational units given or by the order in which they appeared — it was not the linguistic form which enabled her to identify the department she was looking for.


Rather, a combination of two types of knowledge helped her identify the necessary information. Firstly, her professional knowledge about the organisation of companies and secondly her discourse knowledge that companies usually provide the contact person and email address on a website for possible business partners. So here, communication is a form of combining professional and discourse knowledge, i.e. the culturally and professionally determined use of websites, and inferring the meaning of the foreign language data she found. Besides, she had a high tolerance of ambiguity because she was not sure that her inference was correct, but decided to run the risk of contacting the wrong department. And although she sent the email to the head of the department and not to the person in charge, her communicative efforts were successful because she reached her overall goal, meeting a representative of the company to establish a new business contact.

7. Code Switching 7.1 Code Switching for Small Talk Code switching is a phenomenon that may also appear in business conversation, for example if more than two speakers are involved and two of them share the same L1. Examples 2–4 are taken from a negotiation in which three speakers, S1mGr, S2mNr, Sm3Gr are involved. S1mGr and S2mNr, the owner of a Norwegian record label and promotion manager at a Norwegian record company, are involved in a negotiation about the distribution of S2mNr’s products in Germany. In contrast to Rehbein’s as well as to Wagner and Petersen’s model, the ratification has already been done, i.e. the contracts have been signed, but S2mNr and S2mGr meet to negotiate specific details, such as how to import the goods to Germany. They have met at a trade fair and have been joined by another label manager, S3mGr, one of S1mGr’s colleagues, who is not involved in the deal. During their encounter, sequences of negotiation alternate with sequences of small talk. Before the sequence shown in example 2, S1mGr and S2mNr discussed a way of transporting the Norwegian discs to Germany, e.g. with the help of transportation companies, but did not reach an agreement. In example 2, S3mGr starts his leave-taking whilst still using English. Suddenly he switches to German and opens a side discourse with his German colleague S1mGr about working at the company’s fair stand. Starting his leave-taking with his announcement in se 827, S3mGr changes the topic. In se 828 he gives the reason for his leaving: having to “work behind the counter”, referring to a stand at the trade fair. Although S2mNr takes up the new topic in se 831 with his utterance: “Yeah, you have a stand”, S3mGr switches to German and asks S1mGr whether he has already worked at the stand (se 832). S1mGr offers a brief answer to the question in German (se 834), only confirming that he has done so already.


Example 2. stand (827) (828) (829) (830) (831) (832)

S3mGr S3mGr S2mNr S3mGr S2mNr S3mGr

(833) (834) (835) (836) (837)

S2mNr S1mGr S3mGr S3mGr S1mGr

(838) (839) (840)

S1mGr S3mGr S2mNr

I think I have to go, in a couple of minutes. I have to work behind the counter. Yeah. For one hour. Yeah, you have a stand. {Have you done your stand duty?} Ja, ja. {Yes, yes}. Ja· Ist das okay? {Is it okay?} Ja, ich hab da [ne ganz nette • • Frau von Nata kennengelernt]. [lachend (laughing) {Yeah, well I met quite a nice woman from Nata}. ((talking to S3mGr in German; small talk, leave taking; 1min, 15sec)) Nice to meet you. Yeah it was really nice meeting you.

In se 835, S3mGr gives a backchannel to S1mGr, and then he asks him in German how he liked it (se 836). S1mGr takes up the conversation in German and tells S3mGr about his meeting a woman attached to one of their partner companies during his stand duty. The two continue the small talk topic, and then S3mGr says good-bye to S1mGr (short conversation in German, not given here), before turning to S2mNr again to say good-bye. However, he actually does not leave at this point, but switches to English again and stays with the two. He starts another side discourse in German with S1mGr later (see below). Although the conversation between the three men originally takes place in English and the topic has been introduced in English and taken up by S2mNr, S3mGr switches to German. S2mNr, however, does not speak German. Müller (1989) describes sequences where one party is not capable of understanding the ongoing discourse without the help of an interpreter as “opaque” sequences. Here, the sequence is opaque to S2mNr, and none of the two Germans offers a translation. Thus S2mNr is consciously excluded from the ongoing discourse. However, the topic seems rather irrelevant, and the sequence can be regarded as small talk. Yet it obviously is an interesting small talk topic to the two men, which might have been of interest to S2mNr, too. Here, code switching is used to exclude S2mNr although there seems to be no reason for doing so. Yet later on in the encounter, S3mGr switches to German again for a relevant matter. Thus the side sequence here, in which S2mNr does not involve, can be regarded as a kind of “testing situation”, where S3mGr tries to find out whether S2mNr knows German. As he obviously does not, S3mGr can use German for a strategic side discourse with his colleague S1mGr and keep the talk opaque to the other party in the negotiation (see example 3).


7.2 Strategical Code Switching In example 3, code switching is used in a situation where the topic is of relevance to S2mNr, who is again excluded through language choice. It is again S3mGr who switches to German, but this time to address a strategically relevant topic. S1mGr and S2mNr are discussing the distribution of the Norwegian goods via another country. S3mGr addresses S1mGr in German and suggests using existing business connections to import the products from Norway. The idea was brought up by S1mGr earlier in the encounter. However, S2mNr did not favour this option. They did not decide on the matter, but changed the topic then. Here, the topic is on the schedule again. S3mGr’s suggestion is to use their business relation to the distribution company Righthand to organise for the transport of the Norwegian CD’s to Germany. Righthand imports goods from S2mNr’s company besides dealing with S1mGr and S3mGr’s company, too. However, S2mNr is not satisfied with Righthand’s work, as became apparent earlier in the discourse. S2mNr and S1mGr did not decide on that aspect although S3mGr argued in favour of Righthand at that particular time. Here S3mGr brings the topic up again, but this time in German: Example 3. Righthand (846) (847) (848) (849) (850) (851) (852)

S1mGr So we have to, we have to see how we can… S3mGr Kannst ja mit Righthand anfangen zu dealen. {Why don’t you start dealing with Righthand.} S1mGr Nee, das ist… {No, that is...} S3mGr Das ist ja kein Problem. {That’s no problem.} S3mGr Die haben ja alle ( ). {They have all the ( ) anyway.} S1mGr Ja, müssen wir mal gucken, wie wir das machen. {We’ll see how we’re gonna do it.} S2mNr It’s up, it’s up…

In 846 S1mGr tries to postpone the decision by showing his willingness to find a compromise with S2mNr as is indicated by the use of “we” in the formulation “we have to see how we can...”. He is about to start a new topic when S3mGr interrupts him in se 346 to suggest in German that he use Righthand for organising the transport. S1mGr does not favour the idea as can be seen by his negative, though incomplete, answer “No, that is…” (se 848). However, S3mGr does not give up and supports his own suggestion by stating it was no problem (se 849). He seems to be quite convinced of his idea and gives a reason in the next segment. However, S1mGr is not willing to discuss the matter as his noncommittal answer “We’ll see…” (se 851) shows. At this point, S2mNr tries to engage in the conversation, but his start “It’s up…” is uttered simultaneously with S1mGr’s next utterance. So his two interlocutors seem not to hear him, and there is no reaction to his attempt. He will repeat his utterance later on, leaving the decision to S1mGr (cf. se 876 below). However, at this point, the two German speakers continue to discuss the possibility. S3mGr suggests S1mGr pass the


orders on to him. He discusses the internal procedures with S1mGr until they are interrupted by S2mNr, who asks S3mGr about his responsibilities concerning music labels (se 862): Example 4. Procedures (852) (853) (854) (855) (856) (857) (858) (859)

(860) (861) (862)

S2mNr … S1mGr <Es geht ja auch> dann auch nur um die/ um den Kontakt. {It’s only about the contact anyway.} S1mGr Es ist ja das eine, die Warenbestellung. {That is the one thing, the orders.} S1mGr ( )… S1mGr Ja, das ginge schon. {Yeah, it would be possible}. S3mGr Ja, du sagst uns, was bestellt werden muß. {Yeah, you tell us what has to be ordered.} S3mGr Ich mein, der Guido bestellt sowas fast immer bei Righthand. {I mean, Guido almost always orders those things at Righthand}. S3mGr Der bestellt dann auch meinen Kram mit, aber ich hab trotzdem sehr viel Kontakt direkt zu Y C K, weil Y C K mir dann auch die (die Labels ), die Infos rüberschickt. {He also orders my stuff, but I still have a lot of direct contact to Y C K because Y C K then sends me the, (the labels ), the information.} S3mGr Weil anlegen und so tu ich alles selbst. {Because the creating and so on I do myself}. S1mGr Das würde ich dann halt . {I would do that myself, too}. S2mNr , you are not doing all the label, right?

S3mGr tries to convince S1mGr of his idea of using Righthand for transporting the goods to Germany. His interest in the matter is not quite clear as he is not involved in the deal with the Norwegian partner and S1mGr is in charge of the matter. However, he is in favour of using that contact and of combining S1mGr’s orders with his own (se 857). He himself passes his orders on to another colleague, namely Guido (se 859). He seems to convince S1mGr, who agrees he would administer the labels himself, as S3mGr does. And their other colleague, Guido, would send the orders to Righthand (se 859). At the point where S1mGr seems to give in and agree to S3mGr’s proposal, S2mNr asks S3mGr whether he takes care of all the labels (se 862), correctly assuming a negative answer, as the negative verb form shows. At this point he interrupts the conversation. He indicates that he knows what is going on by drawing a conclusion about S3mGr’s tasks (se 862). He runs the risk of having misinterpreted the discourse. However, by his interrupting the talk at the decisive point, he regains his influence in the ongoing decision process. He shifts the topic from the cooperation to the different labels S3mGr is in charge of (following segments; not given here). Then he ends the topic in se 875 with a “Yeah.” and turns to S1mGr. He ends the discussion in se 876. Instead of starting another “round” in the negotiation process, he leaves the decision to S1mGr. However, he regains his influence by interrupting the two and by handing the decision over to S1mGr.


Example 5. (876)

S2mNr ((2s)) Yeah, yeah, yeah, it’s eh • it’s up to you.

Although he is not able to express himself in German, S2mNr is able to follow the conversation and draw the conclusion that S3mGr is only in charge of part of the labels. Certainly the explicit mentioning of the company in se 859 proved to have been useful in this case. It must have served as a tying (cf. Loos 1999, Firth 1990), which connected the German sequence to the negotiation of import procedures mentioned earlier in the discussion. Another aspect was the fact that the problem had been discussed, but no agreement had been found. S2mNr knows that it is one of the important aspects of the negotiation still to be settled. So his discourse knowledge helped him to interpret the other participants’ actions. However, the exact topic — using Righthand as a link between Norway and Germany — was established in S3mGr’s utterance (se 847; example 3) in German. So S2mNr must have been able to conclude the topic before being able to follow the conversation and interrupt it at a point at which a decision is about to be taken without him (se 861). The utterance itself contains the two following hints serving as tyings on the lexical level: firstly, the name Righthand and secondly, the German verb dealen. Dealen is a loan word in German, here carrying the ending of the German infinitive (-en) because it appears after the modal kannst (second person singular, indicative, present tense, active of können = can). As it contains the English verb deal as its root, it can be identified if it is segmented from the German ending. Thus it is likely that S2mNr identified the topic with the help of the two cues and concluded that a negotiation was going on about the proceeding of importing his goods to Germany. Yet at that time the negotiation was taking place between the two German colleagues without him. However, the lexical hints might not have been the only aspects helping S2mNr identify S3mGr’s intentions. S3mGr’s utterance “Kannst ja mit Righthand anfangen zu dealen.” (se 847) is a proposal. Thus, S2mNr must have identified the illocution, presumably by the prosody of the utterance. In two typologically related languages like Norwegian and German, the prosodic patterns and the forms of speech acts are often quite similar. So presumably S3mGr’s utterance was comprehensible to S2mNr because he combined the cues with the speech act, a proposal, which he identified through its prosody. He used his professional knowledge about the different companies and the procedures. Adding his discourse knowledge — the fact that no decision about the delivery of the products to Germany had been taken — he ascertained what the content might be and, what is more, he came to a correct conclusion as to the pragmatic function of the side discourse carried out in German. Thus the combination of the different types of knowledge compensated for the lack of knowledge of German.


And his ability to engage in a conversation without mastering the language in use enabled S2mNr to stand his ground and regain the floor, and thereby his influence in the decision making process that was going on.

8. Conclusion Although English certainly is the main language in most international business encounters nowadays, mastery of this language is not sufficient to achieve non-linguistic goals in business. There are instances in which the abilities to understand another language and to interpret the ongoing situation are important. These instances are of different kinds. They may be unintended or they may serve strategic purposes, e.g. if one participant wants to exclude another participant form the ongoing discourse. As shown in the analyses above, the hearer can compensate for a lack of linguistic, i.e. mainly lexical and semantic knowledge, by using elements of the talk as tyings and cues, which enable him or her to relate the utterance to the situation, e.g. a business negotiation. Furthermore, discourse knowledge may enable the interpretation of the utterance in combination with institutional, professional and general knowledge. In order to interpret problematic situations, it is first and foremost the concluding of the pragmatic function of an utterance or behaviour that is decisive. For the comprehension of pragmatic functions and speaker’s intentions, linguistic knowledge of other languages can be helpful. However, if this knowledge is not given or not sufficient, the interpretation of a speaker’s intention with the help of other forms of knowledge, e.g. pragmatic knowledge, is an indispensable means in business encounters. Thus the term “passive language use”, as receptive language use is often called, is perhaps not helpful to describe comprehension of languages in multilingual communication. It might actually be misleading as the hearer actively needs to decode messages, purposes and intentions to be able to interpret an ongoing interaction. And the less a hearer knows the language that is used, i.e. its lexis, syntax and semantics, the more he/she has to rely on other forms of knowledge, such as pragmatic knowledge, general knowledge or professional knowledge to interpret the situation. So the underlying mental processes are more demanding when it is not possible for the hearer to decode the message on its lexical or semantic level. All in all, the combination of different types of knowledge and the trust in one’s own ability to interpret the situation accordingly lead to communicative competence. This communicative competence is essential, especially when the linguistic means are not enough to handle the situation. Thus the ability to draw conclusions from linguistic and non-linguistic cues and to combine them with other forms of knowledge, such as pragmatic and institutional knowledge, make for a successful communicator in multilingual (business) discourses.


Notes 1. For an overview of research of semicommunication and receptive multilingualism see Braunmüller and Zeevaert (2001). 2. Cf. contextualisation cues in Gumperz (1982). 3. Cf. Ehlich and Rehbein (1979).

Signs and abbreviations in transcriptions / self correction ( ) not understandable; English sentences in { }: translations of German original ´ rising intonation ` falling intonation ¯ midlevel intonation < > simultaneous talk S speaker f female m male Fr French Gr German Nr Norwegian se 413 segment 413

References Braunmüller, K. and Zeevaert, L. 2001. Semikommunikation, rezeptive Mehrsprachigkeit und verwandte Phänomene [Arbeiten zur Mehrsprachigkeit 19]. Hamburg: Universität Hamburg. Clyne, M. 1997. Multilingualism. In The Handbook of Sociolinguistics, F. Coulmas (ed.), 301–14. Oxford: Blackwell. Dresemann, B. Forthcoming. Features of English as Lingua Franca Communication in Professional Contexts. Ehlich, K. and Rehbein, J. 1979. Erweiterte halbinterpretative Arbeitstranskriptionen (HIAT 2): Intonation. Linguistische Berichte 55: 51–75. Ehlich, K. and Rehbein, J. 1986. Muster und Institution. Untersuchungen zur schulischen Kommunikation. Tübingen: Narr. Firth, A. 1990. Lingua franca negotiations: Towards an interactional approach. World Englishes 9(3): 269–80. Firth, A. and Wagner, J. 1997. On discourse, communication, and (some) fundamental concepts in SLA research. Modern Language Journal 81(3): 285–300. Gumperz, J. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Heller, M. and Pfaff, C. 1997. Code-Switching. In Kontaktlinguistik. Ein internationales Handbuch zeitgenössischer Forschung. 1. Halbband [Handbücher zur Sprach- und Kommunikations


wissenschaft 12.1], H. Goebl et al. (eds), 594–609. Berlin: Mouton de Gruyter. Holmes, J. and Stubbe, M. 2004. Strategic code-switching in New Zealand workplaces: Scaf folding, solidarity and identity construction. In Multilingual Communication, J. House and J. Rehbein (eds), 155–78. Amsterdam: John Benjamins. House, J. and Rehbein, J. 2004. What is ‘multilingual communication’? In Multilingual Com munication, J. House and J. Rehbein (eds), 1–17. Amsterdam: John Benjamins. Knapp, K. 2003. Interpersonale und interkulturelle Kommunikation. In Interkulturelles Management, N. Bergemann and A. Sourisseaux (eds), 3rd ed., 109–35. Berlin: Springer. Loos, E. 1999. Erfolgreiche deutsch–niederländische Betriebskommunikation: Wahrheit oder Dichtung? WirtschaftsDeutsch International: 142–50. Lüdi, G. 1996. Mehrsprachigkeit. In Kontaktlinguistik. Ein internationales Handbuch zeitgenössischer Forschung. 1. Halbband [Handbücher zur Sprach- und Kommunikationswissen schaft 12.1], H. Goebl et al. (eds), 1. Halbband, 233–45. Berlin: Mouton de Gruyter. Meierkord, C. and Knapp, K. 2002. Approaching lingua franca communication. In Lingua Franca Communication, K. Knapp and C. Meierkord (eds), 9–28. Frankfurt/Main: Lang. Müller, F. 1989. Translation in bilingual conversation: Pragmatic aspects of translatory interaction. Journal of Pragmatics 13: 713–39. Myers-Scotton, C. 1997. Code-switching. In The Handbook of Sociolinguistics, F. Coulmas (ed.), 217–37. Oxford: Blackwell. Myers-Scotton, C. 2002. Contact Linguistics. Bilingual encounters and grammatical outcomes. Oxford: Oxford University Press. Rehbein, J. 1995. International Sales Talk. In The Discourse of Business Negotiation, K. Ehlich and J. Wagner (eds), 67–102. Berlin: Mouton de Gruyter. Rehbein, J. 2001. Intercultural negotiation. In Culture in Communication. Analyses of intercultural situations, A. di Luzio, S. Günthner and F. Orletti (eds), 173–207. Amsterdam: John Benjamins. Wagner, J. 1995. What makes a discourse a negotiation? In The Discourse of Business Negotiation, K. Ehlich and J. Wagner (eds), 9–36. Berlin: Mouton de Gruyter. Wagner, J. and Petersen, U. 1993. Zur Definition von Verhandeln. In Interkulturelle Wirtschafts kommunikation [Studium Deutsch als Fremdsprache — Sprachdidaktik 9], B.-D. Müller (ed.), 2nd ed., 261–76. Munich: Iudicium. Zeevaert. L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Ver ständigung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.

chapter 8

Speaker stances in native and non-native English conversation* I + verb constructions Nicole Baumgarten and Juliane House Universität Hamburg

In lingua franca communication the participants operate under the assumption of mutual intelligibility despite the fact that they are often unable to know whether their interlocutors’ variety of English in fact provides the same repertoire of linguistic expression and hence the context to decode utterances in the sense intended by the speaker. The article investigates the expression of speaker stance in English L1 and English as a lingua franca (ELF) discourse, examining in particular the use of I+verb constructions in order to establish whether EFL (English as a foreign language) speakers’ talk patterns of subjectivity typically differ from those in L1 English discourse. Findings suggest that ELF discourse differs indeed from English L1 discourse in the use of verb types, speaker-specific patterns of the expression of stance and speakers’ preference for expressing prototypical rather than grammaticalized and pragmaticalized meanings. Keywords: lingua franca, mutual intelligibility, English, speaker stance, subjectivity

1. Introduction This article explores subjectivity in discourse from the point of view of the construction of speaker stances by native and non-native speakers of English in L1 and English as Lingua Franca (ELF) conversation. We will suggest on the basis of our results that subjectivity in ELF communication — despite the assumption of mutual intelligibility — is often expressed in such a way that it leads to communicative incongruities caused by speakers’ different L1s, their respective learner variety, or their sociolinguistic and culture-specific backgrounds, of which individual EFL speakers may not be aware. Linguistic subjectivity refers to the ways in which speakers use language to express their perceptions, feelings, opinions and evaluations in discourse. Benveniste (1966/ 1971), who first introduced the notion of subjectivity in language, claims that the expression of subjectivity in discourse is the linguistic reflection of the ability of speakers to view themselves as ‘subjects’, and that it is especially — though not exclusively — the grammatical category of pronouns which enables the expressive capacity of speakers,

196 Nicole Baumgarten and Juliane House

i.e. , their ability to talk about themselves by positing themselves as the grammatical subject of the utterance and the topic of the talk. The starting point for the present investigation is the first person singular pronoun I in subject position and its co-occurrence patterns with semantic verb types and lexical verbs. According to Scheibman (2002) the most common features of spoken discourse are subject-predicate combinations which allow the speakers to personalize their talk, to mark attitude, evaluation and empathy, and the first person pronoun is the most basic (i.e. prototypical) source of subjectivity in language because it always explicitly refers to the speaker thereby automatically introducing an explicit argumentational perspective to the discourse. “I signifies ‘the person who is uttering the present instance of the discourse containing I’” (Benveniste 1966/1971: 218) such that the proposition expressed will always be directly related to the person of the speaker by the hearer(s). The expression of subjectivity in discourse is then, as it were, the closest the interlocutors can get to the speakers and how they see themselves. A concept closely related to subjectivity is “stance”. A speaker’s stance is his or her attitude towards what he or she is saying (Conrad and Biber 2001). That is, in addition to propositional content speakers express — very often in the same utterance — meanings which convey their personal feelings, attitudes, values, judgements, and assessments (Biber et al. 1999). The expression of stance is associated with a number of particular linguistic structures. Beyond pronouns and temporal and spatial deictic elements — the categories which were initially suggested by Benveniste, and extended in subsequent research to include aspect, modality and sentence structure as building blocks of a subjective perspective in discourse (cf. e.g. Lyons 1977; Palmer 1986; Iwasaki 1993; Dorgeloh 1996; Smith 2002) — the concept of stance encompasses larger units of discourse, i.e. , grammatical structures, as well as lexical means. Among these, complement clause constructions (1), so-called comment clauses (finite adverbial clauses such as utterance-final I think or I guess) (2), lexical verbs (3) and adjectives (4) with evaluative meaning overtly attribute a stance to the speaker.

(1) (2) (3) (4)

I hope that I got it right. I got it right, I guess. I love TV drama. I’m surprised.

Many other linguistic structures mark speaker stance in discourse (cf. Biber et al. 1999). In the present context, however, only those which systematically co-occur with the first person pronoun I — the speaker’s prototypical resource for the expression of a subjective perspective — are of interest. Grammatical and lexical means, (1), (2) and (3), (4), respectively, encode stance differently. The grammatical marking of stance always involves two structural components that can be said to be in a frame relation to each other: the first component presents the attitude of the speaker and frames the second, the proposition. In the case of complement clause constructions, for example, the main clause expresses the

Speaker stances 197

stance and serves as an attitudinal frame for the complement clause which expresses the proposition. The lexical marking of stance, on the other hand, does not provide an attitudinal frame for another proposition; the speaker’s attitude is inferred directly from the use of the evaluative lexical item. In recent years there has been an upsurge of scientific interest in the linguistic sources of the expression of subjectivity, the historic semantic-pragmatic process whereby subjective meanings evolve and how and to what purpose subjective meanings are expressed in communicative interaction (Traugott 1989, 1995; Smith 2002; Hunston and Thompson 2000; Scheibman 2001, 2002). Other work has focused on register- or genre-specific patterns of stance-marking, on differences in stance-marking in spoken and written registers, the diachronic development of the linguistic marking of stance in texts (Biber 2004; Biber and Finegan 1994; Hyland 1999, 2005; Kärkkäinen 2003), language-specific patterns of stance-marking (Biber 1995) and the differences in stance-taking and stance-marking by native speakers of English and non-native (EFL) speakers (Hyland 2004; Mauranen 1993). The latter investigations have almost exclusively focused on written communication, in particular written academic prose (see, however, Mauranen 2006). It has also been observed (Hohenstein 2004) that even advanced L2 speakers in L1–L2 communication (in Japanese, in this case) have difficulties in the situationally and pragmatically appropriate use of stance-taking devices. This “pragmatic failure” (Thomas 1983; House and Kasper 1981; House 1996a, 2003), however, does not seem to be the result of deficiencies in the speakers’ grammatical knowledge but could rather be explained as the shining through (Teich 2003) of L1 communicative conventions or the non-availability of certain types of stance-marking devices in the learner varieties. On the whole, with few exceptions, existent research has focused on subjectivity and stance in L1 American and British English. There are, to our knowledge, no investigations of the linguistic construction of stance-marking and stance-taking in ELF conversation. The goal of the present investigation is to describe the construction of speaker stances through I + verb-constructions in L1 and ELF conversation. This specific linguistic structure was chosen because direct personal self-reference through the first person pronoun is the most basic/prototypical and at the same time the most explicit type of encoding the speaker’s subjective perspective in the discourse. At the same time, even though all known languages have the category of ‘first person’, the actual use of the first person pronoun in discourse is more often than not highly constrained. Given that stance-marking is register-specific, language-specific, diachronically changing and differently marked in native and non-native discourse, it is reasonable to assume that the observed differences are the expression of superordinate preferences for communicative styles in the different speech and writing communities involved. In other words, because the linguistic expression of feelings, opinions and evaluations is inherently bound up with considerations of the speaker’s and his or her interlocutor’s face and the degrees of politeness appropriate in the communicative encounter — which, in turn, are governed by community- and culture-specific patterns of interaction — the


communicative conventions regulate the type (i.e. the linguistic pattern) and the frequency of stance-taking in discourse. In ELF communication, speakers of different L1 with potentially differing conventionalized patterns of stance-marking and stance-taking and differently diversified (learner) varieties of English interact. At the same time, the communicative conventions of the speakers’ L1 may or may not coincide with the expected communicative styles in comparable English L1 interactions. Which communicative style, including the linguistic expression of speaker stance, we may ask, do EFL speakers in ELF conversation then adopt? One that reflects the conventions of stance-marking and stance-taking in their L1? An imitation of native English conventions even though no L1 speaker is present? Or an altogether different type, which is always negotiated anew in the course of each ELF interaction (cf. Hüllen 1992)? As a first step towards answering these questions we will try, in the present article, to find out whether the linguistic expression of stance through the first person pronoun in ELF conversation is different from L1 English — and thus possibly typical of ELF conversation. In the following sections, we will first describe the database for our investigation (Section 2). Secondly, we will present the global frequencies and syntactic co-occurrence patterns of the first person pronoun I in the native and ELF discourse (Section 3). More detailed descriptions will then be given for the three most frequent stance-expressing collocations I think, I don’t know and I mean. Section 4 concludes the article with a discussion of the L1 and EFL construction of speaker stance in discourse, first, in the light of variation between the English L1 and the learner varieties in our corpus, which may be the result of a combination of factors, ranging from transfer of the EFL speakers’ L1 or Lx communicative styles into English, negative transfer into English and the influence of formal instruction to different degrees of involvement in or exposure to grammaticalization processes. Secondly, we will discuss the results with respect to the question of how to relate EFL speakers’ individual properties to the properties of the ELF discourse in which they take part. In other words, we will ask whether our results can be said to be characteristic of ELF communication or characteristic of the individual EFL speakers in each interaction.

2. The database The data consists of three audio-taped elicited conversations in English — one with a group of English L1 speakers and two with EFL speakers. The participants in each conversation talk about a specific topic of general interest provided by the researchers, for example “the role of English in the world” or “men and women in the contemporary arts scene”. The topics were chosen to elicit the linguistic expression of opinions, beliefs, feelings and personal experiences as well as an overall argumentative discourse. The participants were not informed about the research design. They were led to believe that they were taped because of a general, unspecified interest in their views on the

Speaker stances 199

Table 1. Contextual data of the speakers Speaker

Sex

L1

L2

Lx

L1

C M W F

male male female female

English (US) English (US) English (US) English (US)

German German German German

Hebrew

EFL1

H B L JH

male female male female

Indonesian German Chinese Korean

English English English English

German, Japanese German German

AS MB DG PD

female female female male

German French Nepali Gujrati

English English English English

German German German

EFL2

topic under discussion. The researchers themselves were not present during the conversation. Even though the discussions were in principle open-ended, each came to a close after about 30 minutes. The conversations were recorded in university settings in Germany, i.e. , the language of the environment is German and all speakers (including the native speakers of English) are proficient in German. Thus, the speakers have, in theory, the possibility to fall back on another shared language. All participants are university students (either visiting or full-time) in their twenties, and apart from possibly hierarchical gender relationships, the conversations are characterized by symmetrical role relationships. Table 1 shows the main relevant contextual data for the 3 groups of participants. The conversations were transcribed according to the HIAT transcription conventions (Rehbein et al. 2004).1 For the present investigation, all utterances with the first person pronoun I in subject position were analyzed with respect to their co-occurrence with semantic verb types, with lexical verbs, the grammatical structure of the I + verb combinations (i.e. complement clause construction, simple clause, finite adverbial (verbal routine)), their syntactic position (initial, medial, final) and function in the utterance (e.g. main process, evidential, hedge), and the function of the complete utterance for the discourse (e.g. agreeing, contradicting, elaborating, introducing a new topic). In order to assess the I + verb combinations in terms of the expression of subjectivity, the main verbs of the I-utterances had to be classified by semantic type. Here we followed the taxonomy of verbal process types introduced by Halliday (Halliday 1994). Halliday posits a set of process types which model the linguistic expression of human experience along the general processes of “doing”, “sensing” and “being”. Verbal processes and the participant in subject position are part of the “experiential center” of the clause (Halliday and Matthiessen 2004: 176). Table 2 provides a summary of the four semantic classes used in this study.


Table 2. Main verb types Verb type

Description

Examples

Material

doing, happening, creating, changing

do, go, take

Mental

cognition/cognitive activity, emotion, desideration, perception

know, think, like, feel, want, hear

Relational

possession (x has a), being (x is a; x is at, on, in… a)

have, be

Verbal

saying

say, talk, mean, be like (quotative), go (quotative)

3. Co-occurrence patterns of I in English L1 and ELF conversation Table 3 presents the overall frequency of utterances with I. Complete utterances minimally consist of a finite clause with I in subject position and a main verb. Table 3. Utterances with I L1 ELF1 ELF2

Incomplete

Complete

Total

Total utterances

5 1 8

120 80 188

125 81 196

489 418 1209

The co-occurrence of I with verb types in the L1 and ELF discourses is displayed in Table 4. Table 4. I-utterances by verb type in L1 and ELF discourses Verb type

L1 (n=120)

ELF1 (n=80)

ELF2 (n=188)

Material Mental Relational Verbal

16 77 15 12

8 59 3 10

10 113 7 58

(13.4%) (64.1%) (12.5%) (10.0%)

(10.0%) (73.8%) (3.7%) (12.5%)

(5.3%) (60.1%) (3.7%) (30.9%)

The combination of I with mental verbs is by far the most frequent choice in both the L1 and the ELF discourses. They are at least twice as frequent as the next frequent verb type. The distribution of verb types in the L1 discourse suggests that the expression of subjectivity and stance is systematically parcelled out across the different experiential domains. The two ELF discourses show only partially overlapping distributions, both however differ from the L1 discourse. The following sections present the collocation of I with each of the four verb types, starting with the most frequent type of mental verbs.

Speaker stances 201

Table 5. Mental verbs Subtype

Mental verbs

Cognition

believe figure find forget guess know realize remember think

Emotion

Perception

Desideration

L1 1

ELF2

1

29 1

7

1 3 1 27

22

46

2 54

2

4

2

feel like love need

3

hear look make out see

3 1

want/wanna

ELF1

1

1 10

11

2

1 5

3

1

4

Total

77

59

113

Type–Token ratio

14.2%

10.1%

10.6%

3.1 Mental verbs Among the verb types, mental processes are the most prototypical sites of the expression of a subjective perspective in discourse because they refer directly to the speakers’ cognitive processes, feelings and desires. Table 5 presents the distribution of mental verbs across the L1 and the ELF discourses. The type–token ratio for the mental verbs is somewhat lower in the ELF discourses than in the L1 data. This might indicate that the EFL speakers rely on a more restricted, less diversified repertoire of what they perceive as standardized means of encoding subjective attitudes by means of mental verbs, although the use of mental verbs clusters around the verbs of cognition think and know in both the L1 and the ELF discourses. In all three discourses, I think and I don’t know are the most frequent collocations. I don’t know is the single most frequently used collocation in the L1 discourse, followed by I think. They account for 21 of the 22 (95.4%) occurrences of think and 24 of the 29 (82.7%) occurrences of know. I think is the single most frequent mental verb in both ELF discourses, while I don’t know is comparatively rare.2 I think makes up 43 of 46 (93.4%) and 44 of 54 (81.4%) occurrences of think in ELF1 and ELF2, respectively. I don’t know accounts for 4 of 7 (57.1%) and 14 of 27 (51.8%) occurrences of know.


In our data, I think and I don’t know occur in three formal structures: Simple clause construction (5) I don’t know. (L1) (6) I think about the Chinese people … (ELF1) Main clause in complement clause construction (7) I don’t know how that works. (L1) (8) I think it’s just gonna happen. (L1) Verbal routine (Coulmas 1981; Edmondson 1989, 1999; Edmondson and House 1981; House 1996a,b) (9) And so especially the younger people, I think, go for that, the new stuff. (L1) (10) And that’s just äh/ I think that’s just äh well crazy äh disgusting and äh I don’t know äh idiotic äh sick? (ELF2) I think in British and American English has been studied from the perspectives of its polysemous semantic meaning and its functional diversification in discourse. Aijmer (1997, 1998) distinguishes between ‘cogitation’ (‘thinking’) as the prototypical meaning of I think — evident in the single clause construction — and three other epistemic meanings, namely ‘belief ’, ‘opinion’ and ‘subjective evaluation’, which derive from the linguistic and situational context of occurrence and the hearer’s inferencing. The epistemic meanings of I think are the result of a process of grammaticalization in English, in the course of which the selection of first person subjects for think outnumbered all other combinations and the referential meaning of the collocation I think was bleached and gradually replaced with subjective meanings (Traugott 1995; Thompson and Mulac 1991; Hopper and Traugott 1993). We distinguish uses of I think as main clause in complement clause constructions with or without complementizer (“epistemic main clause”) on the one hand, and verbal routine-like uses of I think in utterance-medial and clause-medial and utterance-final and clause-final position on the other hand, which are more detached from the grammatical structure of the utterance and thus could be said to represent a more grammaticalized and pragmaticalized form of the expression. Both structural forms, however, express the same meaning of belief, opinion and subjective evaluation. Table 6 displays the distribution of I think over the structural forms in the L1 and ELF discourses. Table 6. I think: Distribution over formal structures L1

ELF1

ELF2

Simple clause construction Complement clause construction Verbal routine

0 10 (62.5%) 6 (37.5%)

6 (13.9%) 26 (60.4%) 11 (25.6%)

2 (4.5%) 30 (68.2%) 12 (27.3%)

Total

16

43

44

Speaker stances 203

Both in the L1 and the ELF discourses I think is predominantly used as main clause in complement clause constructions, where it serves to express the speaker’s belief, opinion and subjective evaluation, and his or her attempt to hedge or boost his or her stance towards the upcoming proposition in the complement clause. The majority of the cases show the omission of the complementizer that. However, with only 7.6% percent of full structures, the EFL speakers show a much higher ratio of that-omission than the L1 speakers (25% full structures), which might be interpreted as an indicator of hypergeneralization on their part. The EFL speakers seem to be less aware of the structural variability of the collocation, the concomitant meaning differences with respect to the ‘tentativeness’ and ‘deliberativeness’ of the speaker’s stance, and the concomitant communicative effects. L1 and ELF discourses further differ in that single clause constructions of the type I think so, which overtly signal the speakers’ agreement or their belief, opinion and subjective evaluation in the context of corroborating a preceding proposition, and I think about, which expresses the prototypical meaning ‘cogitation’, only occur in the ELF discourses. Conversely, I think as a verbal routine is more frequent in the L1 data than in the ELF data. The reduced frequency of the “pragmaticalized” (Hopper and Traugott 1993) verbal routine form of I think and also the comparatively unvaried and inflexible use of I think without complementizer in complement clause constructions in the ELF discourses might be related to differences in the repertoire of markers of subjective meanings of the L1 and the EFL speakers, or the wish to resort to the most basic and prototypical meanings of the structure in order to ‘keep it simple’ and to avoid miscommunication. In its prototypical meaning, I don’t know is a marker of the speaker’s insufficient knowledge about the topic of the discourse, i.e., the expression is a declaration of the inability to supply information. Tsui (1991) distinguishes six pragmatic functions of I don’t know which derive from this core meaning of insufficient knowledge. Depending on its sequential placement in discourse (e.g. in a reply to a question, assessment, or request) I don’t know is used whenever the speaker wants to avoid assessment, explicit disagreement or commitment, and intends to preface disagreement, to minimize impoliteness, or to indicate uncertainty. It also serves as a hearer-oriented element of discourse organization opening up the conversational floor to the other participants (cf. Beach and Metzger 1997; Östman 1981), and it can be used as a ‘cajoler’ to make the communicative act that is transmitted more palatable to the hearer, to appeal for sympathy (Edmondson and House 1981; House 1996a). Table 7 presents the distribution of I don’t know over structural forms in the L1 and the ELF discourses. The L1 and the ELF discourses differ with respect to the distribution over structural forms: The L1 speakers use I don’t know predominantly as a verbal routine (11), whereas the EFL speakers prefer the simple clause constructions and the use of I don’t know as main clause in complement clause constructions (12, 13).

204 Nicole Baumgarten and Juliane House Table 7. I don’t know: Distribution over formal structures L1

ELF1

Simple clause construction Complement clause construction Verbal routine

9 (37.5%) 5 (20.8%) 10 (41.7%)

1 3 0

Total

24

4

(25.0%) (75.0%)

ELF2 7 (50.0%) 3 (21.4%) 4 (28.6%) 14

(11) It’s like, I don’t know, if you look like and think of (this?), it goes to what works best especially in business and things like that because I don’t know if that’s people get lazier but they wanna be more effective. (L1) (12) I don’t know so much about Japanese. (ELF1) (13) I don’t know why it’s not, not pre... I, I really ask myself why there is no LAW against such THINGS. (ELF2) Among the nine instances of simple clause constructions in the L1 discourse, three can be characterized as utterance-like verbal routines (14). Unlike utterances, they do not in themselves constitute a move in the interactional structure of the talk in the sense of forwarding the conversation towards a potential outcome (Edmondson and House 1981). These single clause constructions do not denote insufficient knowledge on the part of the speaker, rather they express the pragmatic meanings of the speaker’s uncertainty, and avoidance of full commitment to the upcoming or preceding proposition. (14) L1 C: And so it’s like, it’s already spread so far, it’s like you’re g/ like you’re trying to change it and make another world-language. That’ll be really hard to do though just because nobody will wanna do it. F: Hm. C: And I don’t know. Like I look here and just like going on in (??) and stuff when I lived in Hannover F: Hm. C: it was like… And it’s funny because everything h/ every third word was like an American word. F: Hm. The distribution across formal structures shows that the expression of subjectivity by means of I don’t know differs between the L1 and the EFL speakers. The EFL speakers predominantly encode the subjective meaning ‘insufficient knowledge’.3 In contrast, the L1 speakers commonly use I don’t know to encode the pragmatic meanings, which do not denote a real deficit in knowledge. It may be possible that the distribution of I don’t know across the L1 and the ELF data is a reflection of the different stages of the grammaticalization of the expression in the L1 and the EFL varieties: In the variety of English in the L1 data we find both

Speaker stances 205

the prototypical (single clause constructions and main clause in complement clause constructions expressing the meaning ‘insufficient knowledge’) and the verbal routine variants, each fulfilling different functions in the discourse. In the EFL varieties in the ELF discourses, we find a frequency and distribution over formal structures which suggests that I don’t know is not as functionally diversified in the EFL varieties as in the L1 variety. This is evident in the fact that the EFL speakers clearly prefer to use I don’t know prototypically, in single clause constructions and as main clause in complement clause constructions, i.e. as an expression of insufficient knowledge, than in the verbal routine form, expressing pragmatic and discourse organizational meanings. Moreover, the verbal routine form, which exclusively expresses on-line planning difficulties in the ELF discourses, is only used by three speakers. In two of these, speakers’ L1 pragmatic markers or verbal routines are available which can be translated by I don’t know. The meaning of German keine Ahnung (‘[I have] no idea’) und French (je) (ne) sais pas (moi) (‘I don’t know’) at least partially overlaps with English I don’t know so that the speakers may employ I don’t know in their L2–English in a way that mirrors the use of their L1 expression. In contrast, at least in Chinese similarly corresponding lexical expressions do not seem to be used in these contexts.4

3.2 Material verbs Material verbs may encode a subjective perspective in the discourse when they are used to convey propositional information about the speakers’ experiences — in the sense of activities carried out (15) or events taken part in (cf. Scheibman 2001). Example (16) presents a use of material verbs in which the speakers cast themselves in hypothetical roles or fictional dialogical exchanges in order to illustrate their point. (15) Today, I read a äh ähm book äh a famous äh women in Hamburg äh some ähm • • female artists. (ELF2). (16) Here it’s like and it’s like what you’re taking for foreign language or French or or whatever and it was never like: oh, I’m taking English ’cause that was just part of it. (L1). Material verbs are the second most frequent verb type in the L1 discourse but rank only in third place in the ELF discourses. Table 8 (see p. 206) presents the distribution of material verbs across the discourses.

3.3 Relational verb types Relational verbs are about three times as frequent in the L1 discourse as in the ELF discourses. Relational verbs can be direct expressions of subjectivity in the collocation I am + evaluative adjective (e.g. I’m serious.) or when they — like material verbs — encode the speaker’s subjective perspective by referencing propositional information about the speaker as in example (17).


Table 8. Material verbs Material verbs

L1

do fit go grow up have a look learn live make move back paint read see (=activity) sit back start study take

1 1 1 2

Total

ELF1

ELF2 1

1

2 1

6

1

1

1 5 1

1 1

1

1 1

3 16

7

10

(17) I started when I was ten. How old were you? (ELF1) The distribution of relational verb types across the L1 and ELF discourses is displayed in Table 9. Table 9. Relational verb type Relational I am in/at/on x I am x I have x Total

L1

ELF1

3 8 4

1 2

15

3

ELF2 5 2 7

3.4 Verbal verbs Similar to material and relational verbs, verbal verbs (e.g. say, talk and the quotatives be like and go) can also express propositional information which implicates the speakers’ involvement in the communicative task and their subjective evaluation of the subject matter (18, 19). (18) So I was like hey, that looks cool, I take German. (L1) (19) Cause they see the movies and they’re like: Wow, that’s so America. And they… I’ve talked to people — often people thought that. (L1)

Speaker stances 207

Table 10. Verbal verbs Verbal verbs agree ask be like go mean say talk Total

L1

ELF1

1 4

3 3 (4)* 2 3 11

ELF2

8 2 10

1 49 1 1 57

* One occurrence of I mean is inside the routine formula you know what I mean. Although I mean does express subjectivity in this collocation, it is not included in the present context because the present discussion focuses on the single occurrences of the collocation I mean.

The explicit expression of speaker stance through verbal verbs, however, is achieved by the use of I mean. I mean is the single most frequent collocation in the group of verbal verbs. Table 10 presents the distribution of verbal verbs across the L1 and the ELF discourses. According to Edmondson and House (1981), I mean is used as a so-called ‘gambit’ used in pre-, mid-, and post-position to express the speaker’s attempt to appeal to his or her interactant’s agreement and cooperation, particularly at instances in the discourse where the speaker cannot be sure that his or her message is welcome to the hearer. This afffective-interactional meaning was found by Kasper (1981) to be so strongly underrepresented in her learner data that she diagnosed zero-occurrence. Schiffrin (1987) identifies three functional aspects for I mean, which bear on the ideational structure of the discourse, on the participant framework (i.e. the relationship between the speaker and the hearer(s)), and on the information status of the proposition. With respect to the ideational structure, I mean marks the upcoming modification of prior propositional information. With respect to the participant framework, speakers use I mean, first, to express their orientation toward the meanings of their own talk and, secondly, to focus attention on themselves. Schiffrin suggests that I mean is especially frequently used when speakers present opinions or disclose personal information. In these contexts, I mean serves to put additional focus on the speakers’ self, i.e. to make the speakers’ involvement in the subject matter and the communicative task more salient. With respect to the information state of the proposition, finally, I mean maintains the speaker’s and the hearers’ focus on prior material. The hearer is instructed to continue to attend to the prior information in order to hear how it will be modified by the speaker. The occurrences of I mean in our data largely mirror the uses described by Edmond son and House and Schiffrin. I mean occurs as a marker of subjective perspective in turn-initial position (20, 21), as a turn-internal means of reformulation and clarifica-


tion within or between utterances (22, 23), and as emphatic evaluative tag in turn-final position (24). (20) I mean, for Americans it’s exciting to go to Hawaii. (L1) (21) Well I mean, • • it’s also a very, very delicate question because… (ELF2) (22) And then I saw ahm a documentary about a VERY SMALL kid. I mean, she is like FOUR YEARS old. (ELF2) (23) F/ for us, we have, we don’t have problem, I mean, Asia, Asian people. (ELF1) (24) I don’t know why, why people consider THIS as art and why people go for it and why he gets a forum in a nice MUSEUM! Why? I mean... (ELF2) In all three discourses, the reformulation/clarification type of I mean is most frequent. I mean occurs more than twice as often in the ELF data as in the L1 data. At first sight, this seems to indicate that the speakers in ELF conversation are particularly anxious to ensure hearer comprehension because I mean serves the speakers to structure their turns by keeping prior information in focus and focusing attention on themselves. This seems to be true for the ELF1 data. However, the contexts of occurrence for I mean in ELF2 suggest that I mean can have, on the basis of the reformulation/clarification property, yet another, related function already described by Edmondson and House (1981). In ELF2 data, the reformulation/clarification use of I mean always seems to include the expression of a subjective evaluation, i.e. I mean is usually accompanied by (strongly) evaluative elements (lexical and prosodic prominence) in the preceding or following utterances or clauses (or in both) (see again examples 20 to 22 and 24 above).5 Where I mean connects two evaluative utterances we usually find the stronger, unequivocal evaluation after I mean, in the second utterance. In all these cases I mean seems to function not as a device for a genuine ‘reformulation’ and ‘clarification’ in the sense of an alternative wording for the preceding information — as we find it in the L1 and the ELF1 data — but rather as a focalizing device in the speaker’s contribution which serves as the point of departure for an explicit expression of a subjective evaluation. Hence, I mean in the ELF2 data appears to be a marker of speaker’s affective involvement in the topic and in the discourse; it signals that the speaker has an opinion on the topic and focuses the hearer(s) on the verbalization of this opinion. As a caveat to this explanation, it has to be said that in ELF2 81.2% percent of all instances of I mean are produced by the Nepali EFL speaker. And even though this particular use of I mean is not restricted to her, the results are very heavily skewed in her direction. We will return to the possible speaker-specificity of ELF talk in Section 5. To sum up, in all discourses, the first person pronoun-mental verb combination is the construction of choice for the expression of the speaker’s subjective perspective. The L1 and ELF discourses differ, however, in the use of the other verb types for the expression of subjective meanings in discourse. In the L1 data, the expression of subjective meanings is almost evenly distributed across the material, relational and verbal verb types. It appears as if the expression of speaker stance is systematically parcelled

Speaker stances 209

out over the different experiential domains (doing, feeling, being). The ELF discourses do not show a comparable pattern. EFL speakers handle the expression of stance differently. Stance-taking appears to be predominantly and firmly associated with the prototypical means of expression, i.e. mental verbs. The remainder of the expressions of subjectivity by means of I + verb constructions in the discourses are unevenly distributed across the other verb types. What is more, we do not find the same patterning in ELF1 and ELF2, which might indicate that the conventions of expressing stance may be speaker-specific — in the sense of an idiosyncratic style in their L2 English — or specific of the speakers’ L1. This is particularly evident in the case of the collocation I mean in ELF2. Furthermore, the speakers in ELF2 only seldom use material verbs to express their personal experiences in the discourse. In both ELF discourses the speakers rarely use relational processes to ascribe themselves attributes, which entails an overt expression of subjective attitudes and the disclosure of personal information. The fact that both ELF discourses show almost identical values for the relational verb types suggests that this construction type may not be established as a viable stance-taking device in EFL varieties of speakers’ or is judged as an inappropriate or ineffective means of expression in ELF talk.

4. Conclusion L1 speakers in L1-only conversation and EFL speakers in ELF conversation show only partially overlapping patterns of constructing speaker stance by means of I+verb constructions. Apart from a general preference for the use of mental verb types, L1 and EFL speakers differ in their use of verb types for the expression of a subjective perspective in the discourse. They differ also with respect to the most frequently used collocations: While I think and I don’t know are the two most frequent mental verb collocations in all discourses, L1 speakers use the expression I don’t know three times as often as EFL speakers and EFL speakers use the expression I think at least 50 percent more often that the L1 speakers. In addition, the L1 and the EFL speakers use the two expressions differently. The L1 and the EFL speakers differ both in their preferences for particular structural forms and their preferences for particular meanings. With respect to both I think and I don’t know, the EFL speakers seem to prefer on the whole the more prototypical meanings and formal structures over the more grammaticalized structures and pragmaticalized meanings as they are expressed in the verbal routine forms of the two expressions. With respect to the use of the verbal verb I mean, we find the expression of strongly evaluative meanings which are not found in the L1 data. The differences in the construction of speaker stance in the L1 and the ELF discourses may be the result of a variety of factors which need further investigation: Firstly, the EFL speakers may show a specific pattern of stance-marking and stance-taking because they verbalize in English the communicative style appropriate in comparable L1 communicative situations. This could, for instance, explain the different distribution of I +


verb constructions across verb types in the L1 and the ELF discourses; secondly, as we have seen in the case of I don’t know, the EFL speakers may transfer the function and meaning of semantically related expressions in their L1 into their L2 English; thirdly, because for the majority of EFL speakers the use of English is domain-specific (i.e. restricted to certain settings, times, topics and interlocutors), the repertoire of stancemarking in the L2 English varieties may be restricted to what is conventionalized in the particular domains; fourthly and finally, the EFL varieties may differ systematically from the L1 English varieties because they are not involved in the same processes of linguistic evolution. New linguistic items and new functions of established items in the L1 varieties will reach the EFL varieties inevitably later — i.e. , only after they are introduced to the EFL varieties by a ‘stylistic leader’ — which in turn will result in an uneven and domain-specific distribution of the innovation across the EFL varieties – or through formal instruction, if the innovation reaches the standard variety in the L1 and is included in EFL textbooks. As a consequence, an innovation may not progress very far or very easily through the EFL varieties or the ELF communities (if one can legitimately speak of such phenomena). This picture leaves us with the question of whether ELF interaction can in fact be said to have characteristic features, apart from such very general ones as an overall enhanced cooperativeness (House 2002a,b), or whether each interaction is not rather the sum of the participants’ EFL varieties — crucially including their L1s. Hence, although all speakers in ELF communication are certainly competent in English, their varieties may differ with respect to their repertoire of subjectivity markers and the expression of speaker stances, which might be mutually unintelligible and thus a potential trouble spot in ELF interaction. One characteristic feature of ELF talk is the phenomenon of ‘Let-it-pass’ (Firth 1996; House 1999, 2002a,b). ‘Let-it-pass’ refers to the participants’ willingly ignoring grammatically incorrect, incomprehensible or dubious, i.e. incompatible with the overall goal of the talk, contributions of their interactants. ‘Let-it-pass’ thus highlights the general goal-orientedness of ELF talk and presumably also of interactions which are characterized by receptive multilingualism; it tends to reduce the talk to the referential, transactional dimension of communication by focusing speakers’ attention on the retrieval of the informational content of a contribution to the detriment of grammatical correctness and interactional and interpersonal appropriateness. Under such circumstances, inappropriate and unexpected expressions of subjectivity and constructions of speaker stances may not be overtly detectable on the surface of the discourse unless they represent serious breaches of the communicative norms and evoke reactions by the other participants. For these reasons, ELF talk has also been described as particularly “robust” (Firth 1996). In a way then, the interactional dimension of communication, i.e. , the exchange of interpersonal subjective and affective meanings may be less important for the outcome of ELF interactions. Our analysis shows, however, that EFL speakers endeavour to express speaker stances although it is at present not clear whether their use of the more prototypical linguistic structures is a motivated choice resulting from an attempt to communicate speaker

Speaker stances

stances which can be decoded with comparative ease, regardless of the degree of proficiency of the hearers, or whether it is caused by the level of their English competence and is presumably beyond speaker’s conscious choice. It remains to be seen whether the linguistic make-up of ELF talk might be related to the fact that EFL speakers operate under the assumption of mutual intelligibility just because they all use the ‘same’ language.

Transcription symbols ( ) (??) / … • CAPS

unclear incomprehensible repair aborted utterance pause of 0.25 seconds prosodic prominence

Notes * We would like to thank two anonymous reviewers for their helpful comments. Any remaining shortcomings are, of course, our own. 1. The transcription conventions are given at the end of the article. 2. The high frequency of the collocation I think in EFL speakers’ student writing has been reported on by Ringbom (1998) and Aijmer (2001). 3. The speakers in ELF1 use I don’t know rarely and exclusively to express the meaning ‘insufficient knowledge’. 4. We would like to thank Nicole Bartel and Yingpin Wang for the information on French and Chinese. 5. As opposed to Edmondson and House (1981), Schiffrin (1987) does not mention a co-occurrence of evaluation and I mean.

References Aijmer, K. 1997. I think — An English modal particle. In Modality in the Germanic Languages, T. Swan and O. Jansen Westvik (eds), 1–47. Berlin: Mouton de Gruyter. Aijmer, K. 1998. Epistemic predicates in contrast. In Corpora and Cross-Linguistic Research: Theory method and case studies, S. Johansson and S. Oksefjell (eds), 278–95. Amsterdam: Rodopi. Aijmer, K. 2001. I think as a marker of discourse style in argumentative Swedish student writing. In A Wealth of English. Studies in honor of Göran Kjellmer, K. Aijmer (ed.), 247–57. Göteborg: Acta Universitatis Gothoburgensis.

211


Beach, W. A. and Metzger, T. R. 1997. Claiming insufficient knowledge. Human Communication Research 23: 560–85. Benveniste, E. 1966/1971. Subjectivity in language. In Problems in General Linguistics, E. Benveniste, 223–30. Coral Gables FL: University of Miami Press. Biber, D. 1995. Dimensions of Register Variation: A cross-linguistic comparison. Cambridge: Cam bridge University Press. Biber, D. 2004. Historical patterns for the grammatical marking of stance: A cross-register comparison. Journal of Historical Pragmatics 5 (1): 107–36. Biber, D. and Finegan, E. (eds). 1994. Sociolinguistic Perspectives on Register. Oxford: Oxford University Press. Biber, D. , Johansson, S. , Leech, G. , Conrad, S. and Finegan, E. 1999. Grammar of Spoken and Written English. London: Longman. Conrad, S. and Biber, D. 2001. Adverbial marking of stance in speech and wrting. In Evaluation in Text: Authorial stance and the construction of discourse, S. Hunston and G. Thompson (eds), 56–73. Oxford: Oxford University Press Coulmas, F. (ed). 1981 Conversational Routine: Explorations in standardized communication situations and prepatterned speech. The Hague: Mouton. Dorgeloh, H. 1996. Viewpoint and subjectivity in English inversion. In The Construal of Space in Language and Thought, R. Dirven and M. Pütz (eds), 509–26. Berlin: Mouton de Gryuter. Edmondson, W. 1989. Discourse production, routines and language learning. In Englisch als Zweitsprache, B. Kettemann, P. Bierbaumer, A. Fill and A. Karpf (eds), 287–302. Tübingen: Narr. Edmondson, W. 1999. Twelve Lectures in Second-language acquisition. Tübingen: Narr. Edmondson, W. and House, J. 1981. Let’s Talk and Talk about It. A pedagogic interactional grammar of English. München: Urban & Schwarzenberg. Firth, A. 1996. The discursive accomplishment of normality. On ‘lingua franca’ English and Conversational Analysis. Journal of Pragmatics 26: 237–60. Halliday, M. A. K. 1994. An Introduction to Functional Grammar. London: Arnold. Halliday, M. A.K and Matthiessen, C. M. I. M. 2004. An Introduction to Functional Grammar. London: Arnold. Hohenstein, C. 2004. A comparative analysis of Japanese and German complement constructions with matrix verbs of thinking and believing: ‘to omou’ and ‘ich glaub(e)’. In Multilingual Communication, J. House and J. Rehbein (eds), 303–41. Amsterdam: John Benjamins. Hopper, P. J. , and Traugott, E. C. 1993. Grammaticalization. Cambridge: Cambridge University Press. House, J. 1996a. Developing pragmatic fluency in English as a foreign language: Routines and metapragmatic awareness. Studies in Second-language acquisition 18: 225–52. House, J. 1996b. Contrastive discourse analysis and misunderstanding: The case of German and English. In Contrastive Sociolinguistics, M. Hellinger and U. Ammon (eds), 345–61. Berlin: Mouton. House, J. 1999. Misunderstanding in intercultural communication: Interactions in English as lingua franca and the myth of mutual intelligibility. In Teaching and Learning English as a Global Language, C. Gnutzmann (ed.), 73–93. Tübingen: Stauffenburg. House, J. 2002a. Communicating in English as a lingua franca. In EUROSLA Yearbook 2, 243–62. Amsterdam: John Benjamins.

Speaker stances 213

House, J. 2002b. Developing pragmatic competence in English as a lingua franca. In Lingua Franca Communication, K. Knapp and C. Meierkord (eds), 245–69. Frankfurt: Peter Lang. House, J. 2003. Misunderstanding in intercultural university encounters. In Misunderstanding in Social Life. Discourse approaches to problematic talk, J. House, G. Kasper and S. Ross (eds), 22–56. London: Longman. House, J. and Kasper, G. 1981. Politeness markers in English and German. In Conversational Routine: Explorations in standardized communication situations and prepatterned speech, F. Coulmas (ed.), 157–86. The Hague: Mouton. Hüllen, W. 1992. Identifikationssprachen und Kommunikationssprachen. Zeitschrift für germanistische Linguistik 20: 298–315. Hunston, S. and Thompson, G. (eds) 2000. Evaluation in Text: Authorial stance and the construction of discourse. Oxford: Oxford University Press. Hyland, K. 1999. Disciplinary discourses: Writer stance in research articles. In Writing: Texts, processes and practices, C. Candlin and K. Hyland (eds), 99–121. London: Longman. Hyland, K 2004. Disciplinary interactions: Metadiscourse in L2 postgraduate writing. Journal of Second Language Writing 13: 133–51. Hyland, K. 2005. Stance and engagement: A model of interaction in academic discourse. Dis course Studies 7: 173–92. Iwasaki, S. 1993. Subjectivity in Grammar and Discourse: Theoretical considerations and a case study of Japanese spoken discourse. Amsterdam: John Benjamins. Kärkkäinen, E. 2003. Epistemic Stance in English Conversation. Amsterdam: John Benjamins. Kasper, G. 1981. Pragmatische Aspekte in der Interimsprache. Eine Untersuchung des Englischen fortgeschrittener deutscher Lerner. Tübingen: Narr. Lyons, J. 1977. Semantics. Vol. 2. Cambridge: Cambridge University Press. Mauranen, A. 2006. Spoken discourse, academics and global English. In Spoken English, TESOL, and Applied Linguistics: Challenges for theory and practice, R. Hughes (ed.). Basingstoke: Palgrave. Mauranen, A. 1993. Contrastive ESP rhetoric: Metatext in Finnish‑English economic texts. English for Specific Purposes 12: 3–22. Östman, J. 1981. You Know: A discourse-functional approach. Amsterdam: John Benjamins. Palmer, F. R. 1986. Mood and Modality. Cambridge: Cambridge University Press. Rehbein, J. , Schmidt, T. , Meyer, B. , Watzke, F. and Herkenrath, A. 2004. Handbuch für das computergestützte Transkribieren nach HIAT. [Arbeiten zur Mehrsprachigkeit 56]. Hamburg: Universität Hamburg. Ringbom, H. 1998. High-frequency verbs in the ICLE corpus. In Explorations in Corpus Linguistics, A. Renouf (ed.), 191–200. Amsterdam: Rodopi. Scheibman, J. 2001. Local patterns of subjectivity in person and verb type in American English conversation. In Frequency and the Emergence of Linguistic Structure, J. Bybee and P. Hopper (eds), 61–89. Amsterdam: John Benjamins. Scheibman, J. 2002. Point of View in Grammar: Structural patterns of subjectivity in American English conversation. Amsterdam: John Benjamins. Schiffrin, D. 1987. Discourse Markers. Amsterdam: John Benjamins. Smith, C. 2002. Accounting for subjectivity. In The Legacy of Zellig Harris, B. Nevin (ed.), 137– 63. Amsterdam: John Benjamins.


Teich, E. 2003. Cross-linguistic Variation in System and Text: A methodology for the investigation of translations and comparable texts. Berlin: Mouton de Gruyter. Thomas, J. 1983. Cross-cultural pragmatic failure. Applied Linguistics 4(2): 91–112. Thompson, S. and Mulac, A. 1991. A quantitative perspective on the grammaticization of epistemic parentheticals in English. In Approaches to Grammaticalization, Vol. II, E. C. Traugott and B. Heine (eds), 313–29. Amsterdam: John Benjamins. Traugott, E. C. 1989. On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65(1): 31–55. Traugott, E. C. 1995. Subjectification in grammaticalisation. In Subjectivity and Subjectivisation: Linguistic perspectives, D. Stein and S. Wright (eds), 31–54. Cambridge: Cambridge University Press. Tsui, A. B. 1991. The pragmatic functions of I don’t know. Text 11: 607–22.

part 3

Testing mutual understanding in receptive multilingual communication

chapter 9

Understanding differences in interScandinavian language understanding Gerke Doetjes Høgskolen i Østfold*

Linguistically seen, the overlap between the mainland Scandinavian languages Danish, Norwegian and Swedish is relatively large. To a large extent, the languages are mutually intelligible and communication within Scandinavia therefore normally takes place in the speakers’ respective first languages. However, this so called semi-communication is not always unproblematic. This article presents a critical overview of the studies that have been conducted on inter-Scandinavian language understanding, pointing out that none of them really seems to be able to provide a secure answer to the question of how well Scandinavians can understand each other’s languages. This mainly depends on methodological problems. Moreover, the article presents a survey of different methods that have been used or possibly could be used to investigate interScandinavian language understanding. The survey makes clear that test results are strongly effected by the method chosen and that these effects ought to be taken into account when the above-mentioned investigations into understanding between Danes, Norwegians and Swedes are used for further research. Keywords: Danish, Nowegian, Swedish, mutual intelligibility, semi-communication

1. Introduction Within the Scandinavian branch1 of the Germanic language family, two separate groups of languages can be defined: (1) the insular Scandinavian languages Faroese and Icelandic, and (2) the mainland Scandinavian languages Danish, Norwegian and Swedish.2 Mutual intelligibility is possible within, but not between the language groups. Faroese and Icelandic are both small traditional languages that have managed to hold their posts in the Scandinavian periphery. They have approximately 60.000 and 300.000 speakers respectively. Both languages have conservative lexicons and grammars, and still have many aspects in common with Old Norse. Compared to the mainland Scandinavian languages, the percentage of loan words is very low. Language distance between Faroese and Icelandic is modest, thus making mutual understanding possible. For centuries, the Faroe Islands and Iceland were part of the Danish territory. Both economical and political connections between Denmark and its former territor-

218 Gerke Doetjes

ies have since remained tight. As a consequence of the strong relations to Denmark, most speakers of Faroese and Icelandic show a very high command of Danish. Large parts of the populations even have Danish as a first language. Many inhabitants of the Faroe Islands and Iceland speak English as a L2 to a very high standard. The mainland Scandinavian languages Danish, Norwegian and Swedish have a significantly greater number of mother tongue speakers (approximately 5.500.000 in Denmark, 4.500.000 in Norway and 8.500.000 in Sweden). They are also less traditional than the insular Scandinavian languages. Throughout the past 1000 years, the mainland Scandinavian languages have gone through a number of language change processes, e.g. loss of case markers, reduction of the verb paradigm accompanied by a strong loan influence from Low-German, Latin, French, and English (cf. Bergman 2001). Those changes have over time resulted in a widening language gap between insular and mainland Scandinavian languages. This gap means that mutual intelligence between Faroese and Icelandic on the one side, and Danish, Norwegian and Swedish on the other side is no longer possible. To overcome the understanding gap between both parts of the Scandinavian language branch, the first group needs to make use of their knowledge of Danish when communicating with mainland Scandinavians.3 Language drift processes have also effected mutual distances between Danish, Norwegian and Swedish. Nevertheless, they should still be seen as very closely related languages. Neither on the pragmatic nor on the syntactic level can any significant differences be found. Even on the lexical and the phonetic level, following centuries of separation, similarities are still relatively large and rather remind of differences between dialects than between languages. From a linguistic point of view, the differences between Danish, Norwegian and Swedish can be compared to interdialectal differences, e.g. those between the dialects of German spoken in different parts of Germany, Austria and Switzerland. Following this view, Braunmüller (2002: 6–9) describes the mainland Scandinavian languages as mutual dialects placed on a dialectal continuum, roofed by a non-existing pan-Scandinavian language. The fact that no roofing language exists is compensated for by what once started as a pan-Scandinavian student movement and throughout the years has grown into a common idea of Scandinavian togetherness. It is not that such a subjective pan-Scandinavian idea just exists, cooperation is also institutionalised. The Scandinavian countries reached far-ranging free trade agreements shortly after World War II and the governments have since been working together tightly within the Nordic Council (Nordisk ministerråd).4 Within the inter-Scandinavian context, Danish, Norwegian and Swedish are usually referred to as being mutually understandable, i.e. Scandinavians should be able to make use of their own language when communicating with fellow Scandinavians (cf. Zeevaert this volume). The necessity of speaking English within inter-Scandinavian businesses or universities, on holidays or in contacts via the internet, is thus ruled out. In practice, mainly international businesses, e.g. the Scandinavian airline SAS (cf. Bruntse 2004) or the internationally operating bank Nordea, have an English language policy. Yet in most other inter-Scandinavian relationships, speakers normally speak

Differences in inter-Scandinavian understanding 219

Danish, Norwegian or Swedish respectively, expecting to be understood by the remaining Scandinavians (including those from the Faroe Islands, Iceland and Finland capable of speaking and understanding one of the mainland Scandinavian languages). Einar Haugen, a pioneer in the field of inter-Scandinavian research, labeled this type of L2-communication semi-communication5 (Haugen 1966: 153). Semi-communication was shown to be relatively unproblematic whilst largely functioning like normal L1-communication. In their dissertations, both Ulla Börestam Uhlmann and Ludger Zeevaert have found inter-Scandinavian communication not to be strongly affected by the fact of two or more different languages being involved (Börestam Uhlmann 1994, Zeevaert 2004). As shown by Zeevaert, even speakers of Faroese and Icelandic as well as Finnish speaking Finns seem to be able to successfully take part in the Scandinavian semi-communication process. Some differences appear, but generally, semi-communication seems to function quite well. As pointed out before, the similarities between the Scandinavian languages are, looked at from a linguistic point of view, comparable with the overlaps between other closely related languages in Europe, such as Low- and High-German and Dutch, Polish and Czech, or even Spanish, Portuguese, Italian and French. None of those language pairs though seem to play a role in binding together two or more countries politically and economically though, as it is the case in Scandinavia. This leads to the conclusion that semi-communication is not only a linguistic phenomenon, but that it also depends on cultural and political factors (cf. Lüdi this volume, Werlen this volume). The specialty of the inter-Scandinavian type of communication lies in the political sphere. It is not only that people have the possibility of speaking to each other in their own languages, it is far more the fact that they are urged to do so. Within Scandinavia, it is often not really appreciated when one speaks English in a context in which there is no necessity to do so (cf. Grünbaum and Reuter 1997: 4). Both the dialect-like language distance and the relatively well-functioning communication in practice indicate that there are only few problems on the language reception level. Although, this must not always be true, seen as language differences exist, understanding problems are reported frequently from practice and special language guides for inter-Scandinavian communication have been published (e.g. Grünbaum and Reuter 1997). It might be the case that some of the communication problems emanating from the differences between the languages are actually tackled by intensive use of language external strategies, such as body language or world knowledge. In this context, Kurt Braunmüller refers to the possibility of using a let it pass-strategy and other non-elaborating strategies (Braunmüller 2006; Zeevaert this volume). But still, too little is known about the way language distance within Scandinavia is challenged in semi-communication. To be able to give a better description of how semi-communication works, we first have to answer the question of how well mainland Scandinavians actually understand each other’s languages — regardless of the seemingly well functioning interScandinavian communication. The goal of this article is to investigate what is already

220 Gerke Doetjes

known about mutual language understanding between Danes, Norwegians and Swedes. As will be shown in Chapter 2 of the article, the question of how well people in Scandinavia understand each other has been a key part of semi-communication research for a long time. The question though has apparently been very difficult to answer. Therefore, in Chapter 3, it will be shown that these difficulties mainly depend on the way primary investigations were carried out. In order to put this in a broader perspective, an investigation of language understanding based on six different test types will be presented. Finally, it will be concluded that a straight out answer to the above question is hard to provide, but that other questions are more relevant and that there are ways to answer those questions.

2. Differences in understanding Empirical investigation of mutual understanding between mainland Scandinavians began in 1953, when Einar Haugen sent out a questionnaire to over 300 inhabitants of Denmark, Norway and Sweden (Haugen 1953). He asked them to report on how well they thought they could understand their neighbouring tongues. The questionnaire was answered and sent back by 28% of the recipients. As a result (see Table 1), Haugen found that Norwegians tend to understand Swedes (89%) quite well. Understanding Danish seemed to be at a somewhat lower level (83%). Norwegian is seen as the easiest language to be understood by both Danes (86%) and Swedes (94%). Furthermore, the results clearly indicate that there is an understanding problem between Danes and Swedes (Danish in Sweden: 54%; Swedish in Denmark: 56%). Haugen’s questionnaire was largely reproduced as part of a Gallup/SIFO investigation carried out in the 1970ties (Nordiska rådet 1973, see Table 2). All in all, the numbers in the Haugen investigation were at a somewhat lower level than in the Gallup/ SIFO poll, a fact most likely caused by differences in the answering categories between both studies. Anyhow, in general, their findings were comparable, apart from the apparently low estimation of the understanding of Swedish in Denmark in the Haugen study. Table 1. Haugen’s (1953) questionnaire Denmark Norway Sweden

Danish

Norwegian

Swedish

– 83% 54%

86% – 94%

56% 89% –

Note: Figures shown are combined results in for two out of five answering categories: understanding almost everything and understanding everything. Other answering categories were pointing in the direction of participants not or hardly being able to understand the neighbouring languages. Questions not answered were left out here.


Table 2. Gallup/SIFO investigation initiated by the Nordic Council (1973) Denmark Norway Sweden

Danish

Norwegian

Swedish

– 79% 47%

85% – 95%

78% 97%

With the first two investigations being of a socio-linguistic type, Øyvind Maurud carried out the first large-based empirical language understanding test, investigating both written and spoken language. His participants were young recruits from the respective countries capitals. The test covered Denmark, Norway and Sweden and was complemented by questions relating to the understanding of the participants own language (Maurud 1976, see Table 3; cf. also Delsing this volume). This step is necessary in order to interpret mutual language understanding test results in a way that accounts for the fact that L1-understanding can be fully measured in such experiments. The Danes’ understanding of Swedish has therefore to be linked to the Swedes’ understanding of the same text material in their own language within the same test context. Other tests often fail in this point, but neither did Maurud go all the way. The figures for L1 written language understanding are no real test results; seen as Maurud just set the figure at 100%, awaiting his participants to be able to read their own language without any problems. Maurud’s findings seem to be highly compatible with those of Haugen’s investigation. Again, Norwegians reach the highest level of understanding and Norwegian is the language best understood (written language: 81% in Denmark and 76% in Sweden; spoken language: 65% in Denmark and 41% in Sweden). Danes and Swedes again reveal problems in understanding each other (spoken language: 21% for Danish in Sweden and 40% for Swedish in Denmark). Around the same time, Inge Bø carried out a test similar to that of Maurud (Bø 1978). Whilst Maurud took a general view on inter-Scandinavian understanding, Bø focused on one specific factor, namely the possibility of receiving TV from another Scandinavian country. A TV-dependent difference was observed, though the differTable 3. Maurud’s (1976) all-round test. The left column shows the results of the written language (W) part, the figures to the right are the spoken (S) language results. Figures are median results Danish Denmark Norway Sweden

Norwegian

Swedish

W

S

W

S

W

S

100% 94% 68%

93% 72% 21%

81% 100% 76%

65% 97% 41%

60% 87% 100%

40% 81% 93%

222 Gerke Doetjes

Table 4. Bø’s (1978) investigation. Results in % for the TV-condition for each language are situated in the left column. The figures in the right column show the results in places without Scandinavian TV. Results are separated for the written (W) and the spoken (S) language part. Danish

Norwegian

+TV Denmark Norway Sweden

–TV

+TV

Swedish –TV

+TV

–TV

W

S

W

S

W

S

W

S

W

S

W

S

– 72 48

– 58 39

– 80 29

– 67 19

51 – 59

67 – 76

45 – 50

58 – 71

53 76 –

63 84 –

35 75 –

35 75 –

ence was only minimal in most cases (see Table 4). In the Norwegian case, participants even understood Danish better without having the possibility to watch Danish television (with TV: 72%; without TV: 80%). Bø did not investigate L1-understanding. Up to the year 2000, only a few smaller case studies were carried out (e.g. Börestam 1984, 1985, 1987). In 2000, a combined railway and road bridge connecting Copen hagen (Denmark) with Malmö (southern Sweden) was opened. In connection with this occasion a new, widespread investigation of mutual understanding within Scandinavia was carried out (Delsing and Lundin Åkesson 2005; cf. also Delsing in this volume, Jörgensen and Kärrlander 2001, Lundin and Zola Kristensen 2001). The test consisted of a written and a spoken language part and covered the understanding of Danish, Norwegian, Swedish and, in order to compare inter-Scandinavian communication alternatives, even English in Denmark, Norway, Sweden and five further Scandinavian areas. L1 language understanding was also tested (see Table 5). In accordance with the previous studies it was found that Norwegians are most capable of successfully mastering inter-Scandinavian communication (written language: 77% for Danish, 74% for Swedish; spoken language: 41% and 69%). According to the results, communication between Swedes and Danes could be problematic. Norwegians’ understanding of English is approximately at the same level as their inter-Scandinavian language understanding (written language: 75%; spoken language: 70%). This is not the case for Denmark and Sweden. In both countries, English is in most cases unTable 5. Delsing and Lundin Åkesson’s (2005) investigation of mainland Scandinavian and English understanding in Scandinavia. Figures are given in % and calculated from the points given on a 1-to-10 scale for each test part (one written language part, W; two spoken language parts, S: video sequence, V; audio sequence, A) Danish

Norwegian

Swedish

W

S V

A

W

S V

A

W

S V

English A

Denmark 61 Norway 77 Sweden 55

92 64 40

78 41 19

73 74 55

18 76 38

35 87 56

63 74 75

15 43 83

27 69 81

W

S V

A

58 72 77

61 75 73

63 70 76


derstood better than the remaining Scandinavian languages. At first sight, this would strengthen the assumption of English functioning as a lingua franca for communication between Danes and Swedes. This, however, is only a linguistic supposition, not taking the political and ideological background of Scandinavian semi-communication into consideration. The investigations introduced above reveal several problems. For most of them, it is impossible to say how well Scandinavians are actually capable of understanding each other. To be able to provide a satisfactory answer to this question, it is necessary to measure L1-understanding as a level of normal understanding, in order to compare the tests for the other languages with these figures. Only Maurud’s spoken language test (Maurud 1976) results in figures for both L1 understanding and interScandinavian understanding. In this case, the problem is a relatively wide range of understanding (between 93% and 97% correct answers), which opens up the possibility of the test groups from the three countries not being comparable. Delsing and Lundin Åkesson (2005) also present L1 understanding figures, but those differ even more than Maurud’s. In this case, even the texts tested and the questions asked were partly different. Another problem is that language understanding has so far only been investigated by testing one language at a time, whereas in real semi-communication, more than one language is used. For example, in a situation in which two Danes and one Swede communicate with each other, the Danes only have to act semi-communicatively in part of the conversation. This should have a positive effect on the overall understanding within such communication situations. Ridell (2000), who analysed this kind of mixed (Danish/Swedish) understanding on a small scale in her MA-thesis, found that it functioned relatively well. But her test also fails in establishing the difference between L1 understanding and understanding in the semi-communicative test situation. Furthermore, it is also impossible to establish a long time comparison between the different tests in order to find out whether mutual understanding in Scandinavia has improved or deteriorated over the years. This is partly due to the differences between the test groups. Maurud tested recruits, all men at the age of 18, from the countries’ capitals. Bø’s test persons were high school students aged around 14–15, mainly from smaller villages. Delsing and Lundin Åkesson also tested high school students, between 15–19 years of age, from the capitals, larger cities and smaller towns. The latter compared their results to those of Maurud in order to find out whether mutual language understanding has developed in a positive or a negative manner. They came to the conclusion that, even if Maurud’s results are on a generally higher level, it cannot subsequently be claimed that the understanding has deteriorated seen as the problem exists of both tests being incomparable (Delsing and Lundin Åkesson 2005: 87, cf. also Delsing in this volume). In addition to that, the recent investigation makes use of a new test type in which the stimulus material is presented in video sequences. Yet the differences between the tests, both technically and principally, are simply too large to reach reliable conclusions on long-term developments in understanding.

224 Gerke Doetjes

Therefore, it still remains open as to how well Scandinavians do actually understand each other. In fact, the only conclusion that can be drawn so far, is that Norwegians are generally best suitable for taking part in inter-Scandinavian communication, seen as they understand the other languages better and are generally understood slightly better than the Danes and Swedes respectively. Between Denmark and Sweden, the understanding gap seems to be fairly large. This conclusion, again, is mainly in line with Haugen’s results.

3. Testing understanding in different ways In Chapter 2, it was stated that the language tests carried out so far are not sufficient in order to find an accurate answer to the question of how well Scandinavians understand each other. This leads us to the question as to which extent varying difficulty levels and different test types can influence test results. In the following, the effect of six different test types on the measurement of inter-Scandinavian language understanding will be presented (cf. Golinski and Doetjes 2005 for a preliminary description of the investigation and its results). The six test types were defined as follows: (A) open questions; (B) true–false questions; (C) multiple choice questions; (D) word translation; (E) summary; and (F) short summary. In order to keep the test conditions stable otherwise, all six tests were based on the same short text6 written in Swedish about tourism on the Baltic Sea island of Gotland and were carried out in test groups from high schools in two small province towns (Odense and Kolding) in western Denmark (total number of students: N = 282; age 15–19; results in Table 6). The first test (A) was carried out in two school classes with a total of 54 students. The text was read out loud, followed by four open answer questions, each of them relating to a different part of the text. The students could answer as much as they wished, i.e. there were no space or time limits. The answers were then coded with 0 (wrong answer), 0.5 (partly correct answer) and 1 point (correct answer). In total, 180.5 out of a possible total of 216 points were reached, thus resulting in a mean understanding level of 83.6%. The number of points achieved by the students individually ranged between 1.5 and 4.0. For the second test condition (B), 43 students from two classes were tested. The questions asked in this test related to the same text parts as in test A, but now demanded a true–false answer. Correct answers were again awarded with 1 point, wrong answers with 0 points. The test type excludes partly correct answers. With 172 points being the possible total for this group, 160 correct answers resulted in an understanding level of 93.0%. The lowest number of correct answers was 2, the highest 4. The third test type (C) also used a fixed questioning method, namely multiple choice. Again, four questions were asked, relating to the same four text parts as in the test conditions above. All four questions had four possible answers, of which only one answer


resulted in 1 point. All other answers were incorrect and coded with 0 points. None of the possible answers were partly correct and in that way distracting from the correct answer. Under the true–false condition, 55 students took part. Of 220 possible points, 91.4% were gained. This corresponds to 201 correct answers, ranging between 2 and 4 points per participant. From the perspective of language assessment, the fourth test (D) was principally different from the test conditions applied in A–C, as a translation task was used to measure Danish understanding of Swedish written language. Within the text, 16 words were underlined. Those words had to be translated into the test persons L1, in this case Danish. Like in test A, the answers were divided into three groups: 0 points for wrong (or no) translations, 0.5 points for partly right translations and 1 point for correct translations. 50 test persons took part in this test, leading to a maximum of 800 points. All participants together reached 659 points, corresponding to an understanding level of 82.4%. The lowest number of correct translations was 9, three participants scored the full 16 points. A further test type was applied in test condition (E). The students were asked to write a summary of the Swedish text in Danish. They could write as much as they wanted, no restrictions were given. The summaries were supposed to correctly refer to the four points which the questions in test conditions A-C referred to. To guarantee that the assessment covered the whole text and not just parts of it, each test person’s summary was assessed in a comparable and transparent way making it possible to compare the results from all test conditions with each other. Again, fully and correctly summarised text parts were rewarded with 1 point, partly correct answers scored 0.5 points and wrong or left out parts 0 points. 36 students took part in this test, making up 144 possible points. In total, the understanding measured in this test condition totalled 76.4% (110 points, results ranging between 1 and 4 points). In the sixth test condition (F), the procedure applied in (E) was repeated, but now the instructions for the test were complemented by the test person being asked to keep the summary short and clear. This restriction in general resulted in shorter answers and, in accordance with the expectations, the results of the text understanding were clearly lower than those in (A) to (E). Nevertheless, with 44 persons tested and a maxi-

Table 6. Understanding of Swedish in Denmark; tested in six different test conditions Test condition

Understanding

Overall result (A) Open questions (B) True/false questions (C) Multiple choice questions (D) Word translation (E) Summary (F) Short summary

82.7% 83.6% 93.0% 91.4% 82.4% 76.4% 66.2%

226 Gerke Doetjes

Table 7. Index figures for the different test conditions (mean = 100) Test condition

Understanding

(A) Open questions (B) True/false questions (C) Multiple choice questions (D) Word translation (E) Summary (F) Short summary

101.2 112.5 110.5 99.6 92.4 80.1

mum of 176 points achievable, even here 116.5 points were reached (66.2%). The maximum of points achieved was 4, one participant reached no points. The overall result was 82.7%,7 with the results of three test conditions situated under this level, and three above. However, this does not mean that the understanding was worse in some cases, and better in others. It rather points to the fact that some test conditions lead to higher result levels than others. This mainly is a test effect. Some test types automatically lead to higher results, with the differing possibilities of answering correctly playing a huge role (true–false and multiple choice tests resulting in an automatic 50% and 25% chance respectively of getting a correct answer). Furthermore, pre-formulated answers play a strong role. In these cases participants can rely both on the text and on the answering possibilities to reach the correct answer. In more open test methods, e.g. open questions or summary tests, merely advanced knowledge of the texts used could play a role with some participants. Such test types are also more complicated to assess and evaluate. Both factors could have a slightly negative effect on the results. Results from different test conditions can compensate for such test and rating effects by recalculating the mean result of 82.7% to a base level of 100 (see Table 7). The index figure for the translation task, as well as both summary conditions stay under this level (99.6, 92.4 and 80.1 respectively); the other test conditions reach index figures over 100 (112.5 for the true–false condition, 110.5 for the multiple choice test and 101.2 for the open question test type).

4. Conclusions Based on research done so far, it is virtually impossible to give a straight out answer to the question of how well Scandinavians can understand each other’s languages. The applied tests do not provide clear answers, seen as the results are always effected by the test type being used and by several other factors. The investigations carried out by Haugen (1953) and Nordiska rådet (1973) as well as Maurud (1976), Bø (1978) and Delsing and Lundin Åkesson (2005) all manage to answer parts of the main question posed in this article, but fail to cover all aspects of it. Furthermore, most tests fail to control the L2 language understanding results (e.g. of Danish by Swedes) neglecting to test the same text material and using the same methods in comparable L1


test groups (e.g. Danish language understanding tested in a Danish-speaking control group). Only when the level of understanding in this group is known, can the results achieved by groups of speakers for the other languages be interpreted referring to that level. Maurud (1976) reaches this prerequisite, but only for the spoken language part of his test. Even his figures reveal differences that could make the interpretation of inter-Scandinavian language test results difficult. Other researchers only present figures for the L2 condition or L1 test results that vary between the languages. In Chapter 3, it was shown that different test conditions, all other things being equal, present us with figures of understanding varying between 66.2% for the short summary condition and 93.0% for the true–false test type. Short summaries lead to relatively low results, leading to an understating of the inter-Scandinavian understanding. In contrast, results in the true–false condition point in the other direction. This is important to keep in mind when interpreting test results in inter-Scandinavian language understanding investigations. It is often stated that no testing method is perfect (cf. e.g. Alderson 2000: 203). This is no less true for the test types presented in Chapter 3 of this article as well as for the test types used in the primary research on the topic as presented in Chapter 2. Therefore, it might be a good idea to use a combination of several test methods in order to be able to adequately answer the above-mentioned question. For now, we can conclude that a thorough investigation of language understanding between the mainland Scandinavian languages is still missing and that we therefore are not able to precisely describe under what exact conditions Danes, Norwegians and Swedes take part in the inter-Scandinavian semi-communication. The Scandinavian countries are intensively working together, there is a strong desire for this information. A next step in inter-Scandinavian language testing could be to focus more on language understanding in special situations and under certain conditions, e.g. mixed Danish– Norwegian–Swedish constellations, or in situations in which speakers use accommodation strategies (Grünbaum and Reuter 1997, Giles and Smith 1979) in order to make it easier to understand dialect-like neighbouring languages. But not only within Scandinavia is there use for more information on the possibilities of understanding neighbouring languages, as studies on other combinations of related languages have been carried out focusing on the receptive use of L1 languages (e.g. van Bezooijen and Gooskens 2005 for Dutch, Frisian and Afrikaans, cf. also van Bezooijen and Gooskens in this volume, as well as Klein and Stegmann 2000 for the Roman languages).

Notes * This article is partly based on a test the author designed and carried out within the research project Semicommunication and Receptive Multilingualism in Scandinavia in 2004/2005. From 1999 until 2005, this project was part of the Collaborative Research Center 538 Multilingualism, located at the University of Hamburg. The test results presented in Section 3 of this article were

228 Gerke Doetjes initially discussed in a working paper from the research center named above (cf. Chapter 3 of Golinski and Doetjes 2005). This article aims to deepen the theses presented there and to put them in a broader perspective. I would like to thank Kurt Braunmüller, Bernadette Golinski and Ludger Zeevaert for their help. 1. Cf. Vikør (1995) for comprehensive descriptions of the Scandinavian languages. 2. Cf. Dahlstedt (1974) for a discussion on the separation of the Scandinavian language branch into two groups, or as he puts it, a first (mainland) and a second (peripheral) language community. 3. This principle is also used by speakers of other, non-Scandinavian languages in Scandinavia wanting to take part in inter-Scandinavian communication, e.g. Finnish speaking Finns using Swedish. 4. The Nordic Council has a website, http://www.norden.org, which is aimed to provide information about the official cooperation between its members (including Denmark, Norway and Sweden as well as Finland, Iceland, Greenland and the Faroe Islands). In addition, the nonofficial association Föreningen Norden (http://www.norden.se) has actively stimulated panScandinavian cooperation since it was founded 1919. 5. This term might cause some confusion, with semi- not being used in a negative way (as in e.g. semi-bilingualism). Haugen rather wanted to point out that semi-communication combines receptive bilingualism with productive monolingualism (cf. Braunmüller and Zeevaert 2001: 4–5). 6. The test text was taken from the Swedish newspaper Gotlands Allehanda (http://www. gotlandsallehanda.se/) in December 2003 and is presented in the Appendix. 7. Calculated as the mean result of all participants. The mean result weighted for test conditions was 82%.

References Alderson, J. C. 2000. Assessing Reading. Cambridge: Cambridge University Press. Bergman, G. 2001. Kortfattad svensk språkhistoria. Stockholm: Norstedts akademiska förlag. Bezooijen, R. van and Gooskens, C. 2005. How easy is it for speakers of Dutch to understand Frisian and Afrikaans, and why? Linguistics in the Netherlands 22: 13–24. Bø, I. 1978. Ungdom og naboland. En undersøkelse av skolens og fjernsynets betydning for nabo språksforståelsen [Rogalandsforskning 1978 4]. Stavanger: Rogalandsforskning. Börestam, U. 1984. Språkförståelse och språkpreferenser i internordisk kommunikation på Island [FUMS 120]. Uppsala: Uppsala universitet, Institutionen för nordiska språk. Börestam, U. 1985. Dansk-svensk språkförståelse på Jyllands västkust. En punktstudie [FUMS 126]. Uppsala: Uppsala universitet, Institutionen för nordiska språk. Börestam, U. 1987. Dansk-svensk språkgemenskap på undantag. Nordisk språkförståelse i nutids historiskt och regionalt perspektiv belyst av svenska gymnasieungdomars förståelse av äldre och nutida talad danska [FUMS 137]. Uppsala: Uppsala universitet, Institutionen för nordiska språk. Börestam Uhlmann, U. 1994. Skandinaver samtalar. Språkliga och interaktionella strategier i samtal mellan danskar, norrmän och svenskar [Skrifter utgivna av Institutionen för Nordiska Språk vid Uppsala Universitet 38]. Uppsala: Institutionen för Nordiska Språk.


Braunmüller, K. 2002. Semicommunication and accommodation: Observations from the linguistic situation in Scandinavia. International Journal of Applied Linguistics 12: 1–23. Braunmüller, K. 2006. Vorbild Skandinavien? Zur Relevanz der rezeptiven Mehrsprachigkeit in Europa. In Praxen der Mehrsprachigkeit [Reihe Mehrsprachigkeit 20], K. Ehlich and A. Hornung (eds), 11–29. Münster: Waxmann. Braunmüller, K. and Zeevaert, L. 2001. Semikommunikation, rezeptive Mehrsprachigkeit und verwandte Phänomene. Eine bibliographische Bestandsaufnahme [Arbeiten zur Mehrspra chigkeit, Folge B 19]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehr sprachigkeit. Bruntse, J. 2004. It’s Scandinavian. Dansk-svensk kommunikation i SAS. Copenhagen: Køben havns universitet. MA-thesis. Dahlstedt, K.-H. 1974. Den nordiska språkgemenskapen. Språkvårdsstudier 53: 171–87. Delsing, L.-O. and Lundin Åkesson, K. 2005. Håller språket ihop Norden? En forskningsrapport om ungdomars förståelse av danska, svenska och norska [TemaNord 2005 573]. Copenhagen: Nordiska ministerrådet. Giles, H. and Smith, P. 1979. Accommodation theory: Optimal levels of convergence. In Language and Social Psychology, H. Giles and R. St. Clair (eds), 45–65. Oxford: Basil Blackwell. Golinski, B. and Doetjes, G. 2005. Sprachverstehensuntersuchungen im semikommunikativen Kontext [Arbeiten zur Mehrsprachigkeit, Folge B 64]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehrsprachigkeit. Grünbaum, C. and Reuter, M. 1997. Att förstå varandra i Norden — språkråd till nordbor i nor diskt samarbete. En handledning utgiven av Nordiska rådet, Nordiska språksekretariatet och Nordiska språk- och informationscentret. Oslo: Nordiska språksekretariatet. Haugen, E. 1953. Nordiske språkproblemer — en opinionsundersökelse. Nordisk tidskrift ör vetenskap, konst och industri 29: 225–49. Haugen, E. 1966. Semicommunication: The language gap in Scandinavia. Sociological Inquiry 36: 280–97. Jörgensen, N. and Kärrlander, E. 2001. Grannspråksförståelse i Öresundsregionen år 2000. Gym nasisters hörförståelse [NordLund 22: 1]. Lund: Institutionen för nordiska språk. Klein, H. G. and Stegmann, T. D. 2000. EuroComRom — Die sieben Siebe: Romanische Sprachen sofort lesen können, 2nd edn. [Editiones EuroCom 1]. Aachen: Shaker. Lundin, K. and Zola Kristensen, R. 2001. Grannspråksförståelse i Öresundsregionen år 2000. Gymnasisters läsförståelse [NordLund 22: 2]. Lund: Institutionen för nordiska språk. Maurud, Ø. 1976. Reciprocal comprehension of neighbour languages in Scandinavia. An investigation of how well people in Denmark, Norway and Sweden understand each other’s written and spoken languages. Scandinavian Journal of Educational Research 20: 49–72. Nordiska rådet 1973. Den nordiska allmänheten och det nordiska samarbetet 1973. Attitydunder sökning [Nordisk utredningsserie 4 73]. Stockholm: Nordiska rådet. Ridell, K. 2000. Språkbroar. En studie av dansk-svenska radiosamtal i Öresundsregionen. Uppsala: Uppsala universitet, Institutionen för nordiska språk (unpublished MA-thesis). Vikør, L. 1995. The Nordic Languages. Their status and interrelations [Nordiska språksekretariatet 14]. Oslo: Novus. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verständi gung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.

230 Gerke Doetjes

Appendix The following text was used for the test presented in Chapter 3. The text was taken from the online news paper Gotlands Allehanda in December 2003. Sommarens turister höll hårt i plånboken Sommaren 2003 blev en rekordsommar för Gotland, aldrig förr har antalet besökare varit så högt. Ändå nådde inte omsättningen några rekordnivåer. Anders Murat, turistchef på Gotlands turistförening, tror att lågkonjunkturen är orsaken till att det höga antalet turister inte lett till större intäkter. –Folk reste mer, men höll hårdare i plånboken, säger Anders Murat. Framför allt konsumerades det mindre. Besökarna åt mindre på restaurang, roade sig billigare och handlade mer sparsamt. Trenden ser likadan ut nationellt. Turistdelegationen har genomfört en undersökning om hur turistsommaren 2003 såg ut för hela riket och kommit fram till att volymen har höjts medan intäkterna sjunkit. Som förklaring uppges att fler väljer billigare alternativ för sitt resande. Många kör egen bil på bekostnad av flyg, tåg och buss eller bor i sommarstugan eller hos vänner istället för på hotell och vandrarhem. En annan tydlig trend i riket såväl som på Gotland är att resandet förskjutits till den senare delen av sommaren. –Många tar ut sin semester i augusti och jag tror detta beror på att vi anpassat oss efter Europa. Att turismen fördelas över hela sommaren är bra för oss i turistnäringen, säger Anders Murat.

Chapter 10

Scandinavian intercomprehension today Lars-Olof Delsing Lunds universitet

The article presents the results from a study on the understanding of written and spoken Danish, Norwegian and Swedish. The study was carried out in the years 2003– 2005 within the INS-project and involves pupils and adults from the Scandinavian countries Denmark, Faroe Islands, Finland, Greenland, Iceland, Norway and Sweden. The results are compared to those from a study published by Maurud in 1976. The comparison of the two studies reveals significant changes in the mutual understanding in Scandinavia that can be attributed to political and economical internationalisation and globalisation, changes in the media landscape and an increasing migration rate into Scandinavia, but also between the Scandinavian countries. The study is completed by a simultaneous investigation of the comprehension of English. Keywords: Danish, Norwegian, Swedish, mutual understanding, globalisation, English

1. Introduction The three Mainland Scandinavian languages, Danish, Swedish and Norwegian, are to a great extent mutually intelligible. It seems safe to state that most communication between Danes, Swedes and Norwegians is carried out in the Scandinavian languages, where each person speaks his/her own language and tries to understand the other’s, when they are speaking theirs. Often, also Finns, Icelanders, Faeroese, and Greenlanders participate in communication in the Scandinavian languages, since many of them are sufficiently fluent in one of the Scandinavian languages to do so (many Finns know Swedish and many Icelanders, Faeroese and Greenlanders know Danish, c.f. also Doetjes this volume, Zeevaert this volume). The linguistic situation is not unique; it can, for instance, be found in communication between speakers of various Romance, Slavic, North Indian or Bantu languages. Politically the situation is however somewhat unusual, since there has been a political consensus in the Scandinavian countries in particular, and in the Nordic countries in general to promote and strengthen this Scandinavian linguistic fellowship. Danish has long been the language of education in Greenland and the Faeroe Islands (together with Greenlandic and Faeroese, respectively), and it has been taught as a foreign language in Icelandic schools also after the independence from Denmark in 1946. Similarly, Swedish has been taught as a foreign language to Finns in Finland. Only recently has English taken the place as the first foreign language taught in Icelandic and Finnish schools.1

232 Lars-Olof Delsing

In Scandinavia proper (Denmark, Sweden and Norway), there is an agreement to teach the other Scandinavian languages in school. This has been done since the 1950s, but with varying degree of enthusiasm amongst the teachers. Public service television in the Nordic countries make an effort to cast TV-series and films from the neighbouring countries.2 The knowledge of the Scandinavian neighbouring languages was tested amongst military recruits in an investigation from 1972 (Maurud 1976).3 In this paper, I present a new investigation — the INS investigation, carried out with funding from Nordiska kulturfonden (the Nordic Cultural Fund) from 2003–2005. There were several reasons to investigate the intercomprehension of Scandinavian languages again, since the Nordic countries and their interaction have changed considerably in the last thirty years. First, the three Scandinavian countries, Denmark, Sweden and Norway, have seen the entry of a large number of immigrants and refugees. The percentage of foreign born persons from outside the Nordic countries in Sweden was 2.9% in 1975 and 8.4% in 2001.4 The second generation immigrants have increased even more, but we lack reliable statistics. The situation is similar in Denmark and Norway, but on a slightly lower level. In Finland and Iceland the immigration has been almost insignificant. Second, the supply and use of mass media has changed considerably. In 1972, the Nordic countries had one public service channel each, except for Sweden where a second channel was started in 1969. The only ones that were able to watch two television channels before that were those living near the boarder with a neighbouring country. Nowadays, there are many commercial channels available to most people. Through cables and parabolic aerials, there are now national commercial channels, as well as international channels, and many immigrants are able to watch television from their native countries. English dominates the television channels that people actually view. Third, the school systems have changed a great deal, especially in Sweden and Norway. There, the old gymnasium is, in practice, transformed into a compulsory school form, from where everyone is supposed to gain access to higher education. English is taught from the third or fourth class, and nowadays everyone below 65 years of age knows English to some extent. Fourth, the political situation has changed since Denmark’s membership in the EEC in 1974, which was followed by Sweden and Finland who became members of the EU in 1995. Norway and Iceland are outside the EU, but they are members of the EES and Schengen. Thus, Europe is much more in focus for Scandinavians now than 30 years ago. The traditional foreign languages taught in Scandinavian schools were English, German and French. English has been strengthened, but the other languages have received competition from languages such as Spanish and Italian and other subjects. Simultaneously, pupils have tended to choose other subjects than foreign languages. The result of this is that English is stronger and is now very dominant. Fifth, general economic globalisation has lead to the creation of several common Scandinavian companies, especially in industry, banking and media. Some companies have tried to use Scandinavian as their common language, but most have turned to

Scandinavian intercomprehension today 233

English, at least in official circumstances. Frequently, employees find themselves in a diglossia-like situation, speaking English at the board meeting, but Scandinavian over lunch. Sixth, the pattern of commuting and migration between the Nordic countries has changed. Due to varying states of the market, the exchange of labour between the Nordic countries is considerable. A bridge between Copenhagen in Denmark and Malmö in Sweden, the Öresund bridge, was opened in the year 2000, and the new Svinesund bridge between Sweden and Norway was completed in 2005. Some of the changes mentioned above should lead to more interaction in Scan dinavia, but other will lead to new problems for communication. Some are clearly favouring English as an alternative to the Scandinavian linguistic fellowship. In the rest of this paper, I will present some of the results of our new investigation on Scandinavian intercomprehension, the INS-investigation. As a consequence of the increasing importance of English in the Nordic countries, this investigation was also accompanied by a parallel investigation of the comprehension of English.5

2. The investigation The investigation was initiated and funded by Nordiska Kulturfonden (The Nordic Cultural Fund). One objective was to make this new investigation comparable to the investigation of Maurud, carried out in 1972, which is the only previous major investigation of this type. Our investigation was carried out during 2003 and 2004. Our main group of informants were pupils in the gymnasium and other schools for the same age group (15–19 years of age). We tested these pupils in Danish, Swedish, Norwegian and English. First we tested some 1200 pupils in the two Scandinavian languages that they were supposed to know the least, so Swedes and Finns were tested in Danish and Norwegian, Danes, Faeroese, Icelanders and Greenlanders in Norwegian and Swedish, and Norwegians in Danish and Swedish. In a second investigation we tested roughly a third of the classes in English and one Scandinavian language. The investigation was carried out in the capitals of the countries and semi-independent areas (partly in order to make a comparison with Maurud’s investigation possible) and in seven other cities. Our choice of cities was partly guided by our knowledge of contact patterns. Stockholm, for instance, has quite little contact with Denmark or Table 1. Locations of the investigation Denmark

Sweden

Copenhagen Stockholm Århus Malmö

Norway

Finland

Iceland

Oslo Bergen

Helsinki Reykjavik Jyväskylä Akureyri Vaasa Mariehamn

Faeroe Isl.

Greenland

Torshavn

Nuuk


Norway, and therefor we chose Malmö as the second city of investigation; Malmö has extensive contacts with Denmark, especially since the Öresund bridge was opened. Similar considerations apply to the choice of other Scandinavian cities. In Finland we were also guided by the size of the Swedish speaking population, Helsinki having roughly 8% Swedish speaking population, Vasa 28%, Mariehamn 95% and Jyväskylä practically no Swedish speaking population at all. This investigation of pupils was followed up by a smaller investigation of some of their parents. We contacted the parents of some 20% of the pupils in Denmark, Sweden and Norway. The intention was to compare two generations with regard to comprehension of the neighbouring languages. The last part of the investigation was done on the web, and the public was invited to do parts of the test on the web. Their answers have been gathered in a database. The intention was that these answers will be able to answer some questions that the main investigation will not, since the web investigation gets a broader range of informants, with regard to age and geography. The web investigation is not analysed yet, and is not presented here.

2.1 The test All the informants were presented with a questionnaire. They filled in data about themselves, such as sex, age, home language, attitudes towards the neighbouring countries and languages. After that they were tested in two languages, for instance one test on Swedish and one on Norwegian for Danish consultants. Every language test is constituted by three parts, one video sequence, one tape recording and one written text. The video sequence is a short extract from the TV-program “Who wants to be a millionaire”, which has been broadcast in all the three Scandinavian countries and in Great Britain. The general concept and procedure of the program is the same in all countries, but the persons participating in the quiz and the questions are of course different. In between the questions, there are intermissions of small talk, where the interviewer asks the participants about their everyday life, their occupation and their hobbies. A sequence of this kind has been chosen for each of the four languages (Danish, Swedish, Norwegian and English). In connection to these sequences the questionnaire contains five questions about the content of the small talk, such as “The woman has a pet, what kind of a pet is that?” A correct answer gives two points (a partly correct answer gives 1 point), and the informant may thus get 10 points on this part of the test. The tape recording contains a piece of a radio news broadcast. Originally the text is a newspaper article in Danish. This has been translated into the other three languages, and it has been read by professional news reporters or actors having the respective languages as their mother tongues (female for Danish, Swedish and Norwegian, but male for English). The text is thus identical for all the four languages, and the questionnaire contains five identical questions about the content, such as “How fast


did the kangaroo run?” A correct answer gives two points (a partly correct answer gets 1 point), and the informant may thus get 10 points also on this part of the test. The written text is an article from a newspaper. The original article was written in Danish and then translated into the other three languages. This is a vocabulary test. Ten words are underlined in the questionnaire, and for every word there are four alternatives, of which only one is a correct translation. The texts are identical in the four languages, but the underlined words are not always the same. The alternatives are sometimes very similar in the Scandinavian languages, but not always. Every correct answer gives 1 point, and the informant may thus get 10 points on this part of the test. To summarise, there are three parts of the test, where the first one has different content and questions, the second has identical content and questions, and the third has identical content but only partly identical questions. The latter two have identical content, and in order to be able to test two languages at the same occasion, we produced two different radio news broadcasts and two different newspaper articles. The two radio sequences consist of light news. One is about a runaway kangaroo in Copenhagen, the other about people in the USA who count frogs by listening for their croaking. The two newspaper articles are more serious. One is about new legislation about synthetic drugs within the EU, and the other one is about tinnitus and compensation in connection with this disease. In this way we avoid playing the same radio broadcast, for instance first in Danish and then in Norwegian for our Swedish informants, and likewise the informants get two different texts to read. Roughly half of the pupils have been tested on each variant of the radio news broadcast and the newspaper article. The video sequences have different content from the beginning and have not called for two versions in this way.

2.2 The investigation The investigation has been carried out by 20 field workers. Twelve of them were trained during a weekend course in Copenhagen in February 2002. They learned about the overall purpose and design of the investigation, and they got a thorough survey of the questionnaire and the video and tape recordings. They were also instructed about how to contact schools and parents etc. Since then eight new field workers have been trained by the original field workers. The field workers were instructed to contact classes attending both theoretical and practical education. In the three Scandinavian countries, we also tried to find classes with immigrants in each city, so that approximately 25% of the informants should have a foreign home language. The test was carried out in 40 minutes in each class. The pupils first filled in their personal data, and were then tested in two languages. Most classes were tested in the languages of the two neighbouring countries, but some were tested in the national language and English. The investigation has always been carried out in the ordinary language of education, Norwegian in the Norwegian classes, Finnish in the Finnish classes in Finland etc. The

236 Lars-Olof Delsing Table 2. Number of language tests for each language and country Danish Denmark 113 Sweden 256 Norway 303 Finland (Sw.) 92 103 Finland (Fi.) 47 Faeroe Isl. Iceland 144 Greenland 31 Total 1089

Swedish

Norwegian

English

Total

267 78 303 56 140 24 107 20 995

267 256 78 92 152 39 107 11 1002

110 75 78 56 115 16 70 0 520

757 665 762 296 510 126 428 62 3606

Greenlanders and the Faeroese get their education in both Danish and the native language. In these classes the investigation was carried out in Danish. When we tested the pupils’ parents, we only did so in Scandinavia, but otherwise they were tested in the same way as the pupils. Some of the parents were gathered in the pupils schools, but mostly the field workers visited them in their homes. Thus the test situation was normally less noisy for the parents. The field workers have entered the questionnaire results into our data base. The answers from Iceland and the Finnish classes in Finland have been translated into Danish and Swedish, respectively, by the field workers. A few questionnaires have been omitted from the database. Only people 25 years of age or younger have been entered into the database, in order to keep the age group homogenous and separate from the parental group. A few persons did not take the investigation seriously, and some have been omitted because of their home language; we have excluded immigrants outside Scandinavia and a few Scandinavians living in another Scandinavian country (Danes in Norway etc.). Swedish speakers in Finland and Danish speakers in Greenland are considered more stable minorities and are not seen as immigrants. Finally persons with English as their home language were excluded from the English test. With these omissions we have 1806 questionnaires in the data base. All of them except six, contain two language tests.6 Thus we have 3606 language tests, distributed as in Table 2. For Finland, Finnish speaking classes and Swedish speaking classes are shown separately.7 Below, I give a short overview of the results. Further details may be found in Delsing & Lundin Åkesson (2005).

2.3 General results All figures are calculated as averages in the range between 0 and 10.8 Those who were tested in their own mother tongue usually reach an average of 8 to 8.5 points, but it should be pointed out that the reading comprehension test probably was more difficult when the native language was tested than when English or neighbouring languages


Table 3. Comprehension of neighbouring Scandinavian languages for countries/areas Country/area

Danish

Swedish

Norwegian

Neighbouring languages

Denmark Sweden Norway Finland (Sw.) Finland (Fi.) Faeroe Isl. Iceland Greenland

– 3.80 6.07 3.64 1.54 8.28 5.36 6.61

3.53 – 6.21 – 3.24 5.75 3.34 2.23

4.18 4.98 – 4.77 1.63 7.00 3.40 3.73

3.86 4.39 6.14 4.20 2.14 7.01 4.03 4.19

were tested. In Table 3 the results for the neighbouring languages are given for the different countries/areas. In the rightmost column, the average for the neighbouring Scandinavian languages are given; it is based on two languages for Denmark, Sweden, Norway, and Swedish classes in Finland, whereas it is based on all the three languages for the other groups. Table 3 shows that the Faeroese and the Norwegians understand the neighbouring languages better than other groups. The comparison between Faeroese and Norwegians is a bit misleading, since the figures are based on three and two languages respectively. It is however clear that the Norwegians are considerably better than the Swedes and the Danes, and that the Faeroese are clearly better than the Icelanders, Finns and Greenlanders. According to Table 2, the Swedes seem to manage a bit better than the Danes, but that is entirely due to the video recording, where the Danish version happened to be easier than the Swedish (and the Norwegian) one. In the other tests, which are more comparable, the Danes score 4.55 points in Swedish and 5.39 in Norwegian, whereas the Swedes scored 3.72 in Danish and 5.55 in Norwegian. This is in accordance with Maurud’s (1976) investigation, i.e. the Swedes are worse than the Danes, although the Danes are more close to the Swedes in our investigation.

2.4 Scandinavian as a second language In the former section, we saw that the Swedish speaking classes in Finland managed much better than the Finnish speaking classes, two to three times better. To have a Scandinavian mother tongue thus seems to be a clear advantage, when it comes to understand the neighbouring languages. As for the Faeroese and the Icelanders, they have a North Germanic language as their mother tongue, and the former also have Danish as their language of education to a great extent. Both the Faeroese and the Icelanders are of course better in Danish, but they manage to understand Swedish and Norwegian astonishingly well. The Greenlanders on the other hand (just like the Finns) have a non-related language as their mother tongue. They have Danish as their language of education to a great extent, and so they manage Danish quite well, but are

238 Lars-Olof Delsing Table 4. Comprehension of neighbouring Scandinavian languages for natives and immigrants in Scandinavia Danish

Swedish

Norwegian

Country

Natives

Immigrants

Natives

Immigrants

Natives

Immigrants

Denmark Sweden Norway

– 4.26 6.55

– 2.87 4.61

3.67 – 6.75

3.04 – 4.57

4.35 5.24 –

3.63 4.39 –

poor in the other two.9 To have a Scandinavian language as a mother tongue is thus clearly favourable for the comprehension of the related languages. In Scandinavia, we distinguish natives from immigrants by means of their home language. Persons who claim to speak only the national language (Danish, Swedish or Norwegian, respectively) at home are classified as native and all others (speaking only a non-Scandinavian language or such a language together with the national language) are classified as immigrants. With this (approximate, but practical) classification we arrive at the results in Table 4. Table 4 clearly shows that natives understand the neighbouring languages better than immigrants. This is of course to some extent a class issue; as a group, the immigrants live in socially and economically less favourable circumstances than native Scandinavians, which is also witnessed by the fact that they do not manage the English test as well as the natives (see below). To some extent, the worse results for the immigrants will, however, have to be accounted for linguistically. Significantly, native Scandinavians have a larger vocabulary, which helps considerably when confronted with related languages. For instance, native Swedish youngsters will be more likely to know low frequency vocabulary (often antiquated words) such as icke, vansklig, upprinnelse, and spörsmål (meaning ‘not’, ‘difficult’, ‘origin’ and ‘question’), than their immigrant classmates. The youngsters almost certainly use the Modern Swedish equivalents inte, svår, ursprung and fråga in their own language, but knowing the less frequent variants will be very useful for them when trying to understand Danish and Norwegian, where these words are frequent. Thus, having a Scandinavian home language is clearly an advantage. The immigrants that we have tested are of course not a homogeneous group, and their proportion among the informants deviates more or less from their proportion of the general population. They are for instance over-represented in Stockholm and underrepresented in Bergen. In the following sections, I will only give results for the native Scandinavians.

2.5 Comprehension of neighbouring languages in the different cities We chose to investigate two cities in each country, partly in order to be able to check what role the contact with the neighbouring countries plays. We assumed, for instance, that people in Malmö would be better at Danish than people in Stockholm, because of


Table 5. Comprehension of neighbouring languages (by natives) in the different cities of Scandinavia City

Danish

Swedish

Norwegian

Neighbouring language

Århus Copenhagen Malmö Stockholm Bergen Oslo

– – 5.08 3.46 6.50 6.57

3.74 3.60 – – 6.15 7.12

4.68 4.13 4.97 5.56 – –

4.21 3.87 5.02 4.51 6.32 6.85

the closeness of Malmö to Denmark and Copenhagen. Our assumptions were, by and large, confirmed, which can be seen in Table 5. From Table 5, we may conclude that the people of Bergen and Oslo are equally good at Danish, whereas the people in Oslo understand Swedish better than the people in Bergen. This should of course be explained by the closeness of Oslo to Sweden. In Sweden, we find that the informants in Malmö understand Danish much better than those from Stockholm. Danish is normally harder to understand for Swedes than Norwegian, but Malmö’s closeness to Denmark is significant enough to make Danish and Norwegian equally well understood by our Malmö informants (quite in contrary to the situation in Stockholm). In Denmark the results are surprising; the youngsters in Copenhagen do worse in the Swedish test than those in Århus. Even though Copenhagen is close to Sweden, and Swedish television has been available to Danes in the Copenhagen area since the childhood of television, the youngsters of Copenhagen seem to manage less well than those of Århus. This might be due to lack of interest among the Copenhageners. Our field workers reported more interest from the pupils in Århus, and it seems as if part of the bad result for Copenhagen should be accounted for by attitudes.10

2.6 Age differences In Scandinavia, we returned to the classes we had tested in the spring 2003 and tested some of their parents. We aimed at testing the parents of one third of the pupils. In Stockholm, we had problems finding parents who would do the test, but in the other five Scandinavian cities we were able to carry out the investigation of the parents as planned. For some pupils, both parents have been tested, and for others only one, depending on whether the pupil has two parents at home, and whether they were both at home at the time of our appointment. We have tested some 170 parents, and they are to be compared to their own children, not to the whole group of pupils. The advantage of this testing procedure is that we may compare a group of pupils with just their parents, thus avoiding problems that might occur where upper middle class parents tend to participate in tests like these to a greater extent than parents from the lower middle class and the working class. The comparison of the parents and their children is given in

240 Lars-Olof Delsing Table 6. Comprehension of neighbouring languages in Scandinavia for pupils and their parents Country

Number of informants

Danish Swedish

Norwegian Neighbouring languages

Parents Denmark Sweden Norway

64 47 59

– 7.26 7.92

5.93 – 7.98

6.62 6.86 –

6.28 7.06 7.95

Pupils Denmark Sweden Norway

36 32 48

– 4.96 6.73

3.59 – 6.88

4.54 5.55 –

4.06 5.25 6.81

Table 6, where the mean values for the parents are given first, and then the mean values for the pupils (the children of the tested parents). Table 6 must be interpreted with care. As mentioned above, this kind of investigation normally attracts the parents of students who manage quite well in school; this is by and large true about our investigation as well. Thus, we should only compare the results of the pupils and their parents, not for instance the parents/pupils across countries. According to Table 6, parents are unambiguously better at understanding the neigh bouring Scandinavian languages than their children. This may to some extent be accounted for by the fact that parents are more mature language users generally; they have greater experience of communication with people speaking other languages, they have a larger and more advanced vocabulary etc. Furthermore, the pupils were tested in classes of 20 to 30, whereas the parents were tested in smaller groups, often in their own homes. Nevertheless, the differences are striking; Table 6 reflects two language tests in six cities, each with five separate tests (video, two listening tests, and two reading tests), which amounts to 60 separate language tests. In 58 out of these 60 tests, the parents managed better than their children; in the last two, the children were only slightly better than their parents. Even if a part of the difference between the parents’ and the pupils’ results should be explained by the fact that the parents are more mature language users and were tested in smaller groups (often individually), I interpret the results as a decline in comprehension over time. The decline seems especially significant in Denmark, but also quite clear in Sweden. In Norway it is harder to tell.

2.7 Comparison with Maurud Our investigation should also be compared to Maurud’s from 1972 (published in 1976). The two investigations are partly different; Maurud tested military recruits (only men) in the capital areas of the three Scandinavian countries, whereas we tested high school pupils (both men and women) in two cities in each country. Furthermore Maurud


Table 7. Comparison of Maurud’s and our investigations Country Maurud (1972) Denmark Sweden Norway INS (2003–2004) Denmark Sweden Norway

Danish

Swedish

Norwegian


– 5.17 8.33

6.13 – 9.10

7.90 7.42 –

7.01 6.29 8.72

– 3.87 6.62

3.41 – 8.05

4.64 6.43 –

4.02 5.15 7.33

tested both listening and reading for content and word comprehension. We tested content comprehension for listening, only, and word comprehension for reading, only. Maurud’s investigation involves fewer informants (and covers fewer cities and areas), but his tests are more extensive for each informant. To make the two tests comparable, we have compared only boys in the capital areas for comprehension of content in listening, and comprehension of words in reading. Furthermore we count only native Scandinavians in our comprehension tests, since Maurud’s investigation included practically no immigrants.11 With these qualifications, the comparison between Maurud’s investigation and ours can be summarised as in Table 7, where Maurud’s results (medians given as percentages) have been recalculated to be comparable to ours (averages given in points). In Table 7, we see that, for instance, the Danes have 6.13 (recalculated) points for Swedish in Maurud’s investigation, as compared to 3.87 in the INS-investigation. We should, however, immediately stress that we cannot control for the differences in difficulty between the two tests (calculations of the difficulty of the texts that we used shows that our texts were more difficult than Maurud’s; see Delsing and Lundin Åkesson 2005: 34–8). Thus it is more useful to compare the differences of country-language pairs between the two investigations. This is done in Table 8, where Maurud’s result is the reference point (the base of an index of 100). Table 8 thus shows the results of our investigation (INS) relative to those of Maurud’s. An index of 75 means that he result of our investigation is three fourths of that of Maurud’s. Table 8 shows that the Danes tested for Swedish in the INS-investigation managed to get 55.7% of the points that the Danes got in Maurud’s investigation. Comparing the three Scandinavian groups, Table 8 shows that the Danes in our study differ more from Table 8. Results of the INS investigation relative to Maurud’s Country

Danish

Swedish

Norwegian


Denmark Sweden Norway

– 74.8 79.5

55.7 – 88.4

58.6 86.6 –

57.3 81.8 84.1


Maurud’s than the Swedes and the Norwegians do. It also shows that the results of the Swedes and the Norwegians in our investigation differ more from Maurud’s when it comes to Danish than when they are tested in Norwegian and Swedish, respectively. Keep in mind that the figures do not say anything about deterioration in general, since we cannot control for the differences in difficulty between the two tests, but the difference between the Danes on the one hand, and the Swedes and Norwegians on the other says something. This comparison thus confirms the comparison of pupils and their parents. The comprehension of neighbouring languages has deteriorated in Denmark. It also indicates that the Norwegians are the ones that are best at Scandinavian comprehension, and that they might be at the same level as 30 years ago. Admittedly, the parent investigation indicates that the Swedes are closer to the Danes in deteriorating, whereas the comparison with Maurud indicates that the Swedes almost match the Norwegians. A very clear result is that the Swedish–Norwegian intercomprehension is similar to the situation 30 years ago, whereas there have been more changes where Danes or Danish are involved.

2.8 Comprehension of English In our investigation, we also tested some of the pupils in English. We returned to the classes where we tested the neighbouring languages, and carried out the same test in English (normally some six months later). The result is, not surprisingly, that Scandinavian youngsters are very good at English. Most of the tested pupils have normally learned English in school for six to nine years. In Table 9, the results of the English tests are shown together with the tests of neighbouring languages. The table shows only native Scandinavians; immigrants are thus excluded.12 Before we interpret the results above, it should be stressed that the English video test obviously was simpler than the Swedish and Norwegian video test (the Danish one was also rather simple). It should further be noted that some of the pupils doing the Table 9. Results for comprehension of neighbouring languages and English in the Nordic countries Denmark Sweden Norway Finland (Sw.) Finland (Fi.) Faeroe Isl. Iceland Greenland

Danish

Swedish

Norwegian

English

– 4.26 6.55 3.64 1.54 8.28 5.36 6.61

3.67 – 6.76 – 3.24 5.75 3.34 2.23

4.35 5.24 – 4.76 1.63 7.00 3.40 3.73

6.08 7.55 7.22 7.66 5.94 7.60 7.17 –

Note: Unfortunately, the investigation of English comprehension was not carried out in Greenland.


English test had been tested for the same news recording and news paper article in one of the Scandinavian languages before (normally at least 6 months before the English test). This means that they might remember some of the content from that test. Nevertheless, the results are clear; the result of the English comprehension test is better than that of the Scandinavian neighbouring languages in all countries/areas. Only in the Faeroe Islands and Norway do we find test results for neighbouring languages that may compete with the English results. The comprehension of English is similar in all the Nordic countries, even though the Finnish classes in Finland and the Danish get a lower score than the others. The fact that the Finnish manage less well is hardly surprising, since they have a non-Germanic mother tongue, differing from English in vocabulary, morphology and syntax much more than the Scandinavian languages.13 The lower figures for the Danes are however surprising; there is no obvious reason that the Danes should have significantly worse results than the Swedes and the Norwegians.

2.9 Analysis When we interpret the investigation, we begin with some clear results. One of these is that Norwegians have clearly better results than Swedes and Danes. We would like to put forward the following suggestions as an explanation. First, Norwegian (especially Standard Norwegian, i.e. written bokmål and the spoken Oslo variety) takes an intermediate position between Danish and Swedish in pure linguistic terms. Norwegian shares in many ways its phonological system with Swedish, and the differences between written and spoken language are mostly the same in these two languages. Danish, on the other hand, is phonologically more conservative (primarily when it comes to initial consonants and consonant clusters) and more radical (when it comes to post-vocalic consonants and prosody). If we turn to the vocabulary, Norwegian is clearly much more similar to Danish than to Swedish. Norway was a part of Denmark from the late fourteenth century to the early nineteenth century. Also after the separation of Norway and Denmark in 1814, the Danish language continued to be the only written language, used by writers like Henrik Ibsen in the late nineteenth century. Norwegian was by and large replaced by Danish as the written language for more than 300 years. When Norwegian reappears as a written language, or rather as two written languages, it has a heavy Danish heritage. Bokmål is in principle a Norwegianised continuation of written Danish, whereas Nynorsk was created as a new written language based on Norwegian dialects in the nineteenth century. Also after Norway’s independence from Sweden in 1905, Denmark and Norway have had similar institutions and similarities in their political history (such as the German occupation and the membership in Nato) and their institutions (such as higher education), which have caused similar innovations and changes of the vocabulary. Thus, Norwegian is close to Swedish phonologically and close to Danish in its vocabulary. This intermediate position of Norwegian is helpful, and it helps to explain not only


why the Norwegians understand the neighbouring languages better than Danes and Swedes, but also why Swedes and Danes understand Norwegian better than Danish and Swedish, respectively (compare Table 3). Second, there is a clear difference between the attitudes to, and the interest for, the neighbouring countries. Norwegians sometimes consult Swedish and Danish journals, magazines, television etc. for their special interests. We also find that the cultural interest for the neighbouring countries among Norwegians is more intense than that of Danes and Swedes. It is easier to find a book in Danish or Swedish in a Norwegian book store than vice versa. Maybe, the Norwegian interest for Sweden is most evident; Norwegians normally know what is going on in Swedish popular culture in a way that is unheard of in Sweden. Swedish singers and musicians may be popular in Norway in a way that Norwegian singers and musicians are not in Sweden. Whether this is due to the fact that Norwegians, who understand Swedish well, have less trouble to access Swedish media, or that the Norwegian interest for Swedish culture generates good language comprehension is hard to tell. Third, Norwegians are used to linguistic variation, both in spoken and written language. Norway has two written languages (Bokmål and Nynorsk) and both are more liberal in orthographic and morphological variation than Danish and Swedish. The spoken Norwegian language is also more varied; dialectal forms and pronunciation in public situations are clearly much more accepted in Norway than in Denmark and Sweden. This makes the Norwegian eye and ear used to variation. Norwegians thus have strategies to decode and understand spoken and written language that is slightly different from their own language. This is of course an enormous advantage when it comes to understanding Swedish and Danish. Norwegians may, in effect, consider Danish and Swedish as deviant Norwegian dialects. A last kind of explanation is a matter of size. Generally multilingualism is more common in small language communities than in large ones. Norway has slightly more than four million inhabitants, whereas Denmark has more than five and Sweden nine million. Additionally, Denmark and Sweden have a long history of independence and independent cultural prestige. As a result of this, Norwegians are more likely to turn to Swedish or Danish journals, television etc. when they have a special interest that cannot be satisfied on the Norwegian market. This argument should however not be exaggerated; the size difference between Norway and Denmark is small. Even if the explanations above do not give the whole answer to the Norwegian head start, they are clearly part of the answer.

3. Final remark When we communicate, there are two requirements for success. We have to be able to express ourselves and we have to be able to understand what is said. In other words, we have to produce and consume language. The results given in our investigation are


confined to the consumption part. Thus, it only says something about comprehension; our tests investigate one way communication, where our informants try to understand what is said or written. When we communicate in real life, there is a lot of information of non linguistic character, gestures, facial expressions, etc. , which help the speaker to adjust his language (to accommodate) to the listener (c.f. also Dresemann this volume). Our investigation says nothing about how such phenomena help interlocutors to understand each other. The second half of the communication, production, is not investigated at all in our study.14 The Scandinavian linguistic fellowship has precisely that advantage, namely that the speakers may speak their own mother tongue when communicating with neighbours. We do not have to search for words or expressions, at least no more than we do when we speak to our countrymen.

Notes 1. For a special study of Icelanders using Danish, see Hauksdottir (2001). 2. For a short survey of the history of the Scandinavian linguistic fellowship and modern language policy, see Lund (2006). 3. Apart from Maurud’s investigation, some less extensive investigations have also been made. The Öresund Bridge project investigated Swedish–Danish intercomprehension in relation to the opening of the Öresund Bridge between Malmö and Copenhagen; see Delsing and Kärrlander (2002) and references cited there. 4. Source: Statistiska Centralbyrån, ‘Central Office of Statistics in Sweden’. 5. The investigation and its results is presented in detail in Delsing and Lundin Åkesson (2005). For ease of reference to the tables there, it should be pointed out that the Swedish name of Helsinki is Helsingfors. 6. These six questionnaires are those testing English and another language, where the informants claimed English as their home language. In these cases the English test has been excluded. 7. For Finland, Swedish speaking classes and Finnish speaking classes are shown separately. 8. The figures are calculated in the following way: For the video test an average was calculated for all consultants in the group. For the two listening comprehension tests (the kangaroo and the frog texts), a separate average was calculated for each test, and then the average of these two averages was calculated. For the reading comprehension tests (the tinnitus and drug tests), the same procedure was followed, calculating an average of two separate averages. In this way we can eliminate the difference in difficulty of the different tests. Finally an average of the three averages (for video, listening and reading tests) was calculated. 9. The Greenlanders manage well in the Norwegian test, especially in comparison with the Icelanders. This is however probably due to the fact that they are very few, and that several of them have Danish as their home language. 10. The pupils were asked whether they were willing to live and study abroad. The Danish youngsters were significantly less interested in going abroad than other youngsters in the Nordic coun-


tries. In this connection, there is however no clear difference between Århus and Copenhagen. 11. Recruits have to be citizens, and in 1972 very few 18 year old citizens had a non-Scandinavian language as their home language. For further details on the comparison, see Delsing and Lundin Åkesson (2005). 12. Immigrants are excluded, first, because their proportion of the population is not the same as their proportion in our investigation, and, second, because they constitute a heterogeneous group, where some are very well integrated into the general population, whereas others live in segregated areas. Furthermore, immigrants have different origins in the different Scandinavian countries; for instance, Denmark and Norway have large Pakistani groups, whereas Sweden has very few Pakistani immigrants. 13. For a special study comparing Finns and Finland Swedes learning English, see Ringbom (1987). 14. For some studies on the actual communication we refer to Börestam Uhlman (1994), Zee vaert (2004) and the proceedings from the Symposium on Nordic intercomprehension 1980 (see Elert 1981).

References Börestam Uhlmann, U. 1994. Skandinaver samtalar. Språkliga och interaktionella strategier i samtal mellan danskar, norrmän och svenskar [Skrifter utgivna av Institutionen för Nordiska Språk vid Uppsala Universitet 38]. Uppsala: Institutionen för Nordiska Språk. Delsing, L.-O. and Kärrlander, E. 2002. The Öresund Bridge Project. Swedish Danish Inter comprehension. In Mehrsprachiges Europa durch Interkomprehension in Sprachfamilien, G. Kischel (ed.), Hagen: Fernuniversität. Delsing, L.-O. and Lundin Åkesson, K. 2005. Håller språket ihop Norden. En forskningsrapport om ungdomars förståelse av danska, svenska och norska [TemaNord 2005 573]. Copenhagen: Nordiska ministerrådet. Elert, C.-Chr. (ed.) (1981): Internordisk språkförståelse. Föredrag och diskussioner vid ett symposium på Rungstedsgaard utanför Köpenhamn den 24–6 mars 1980, anordnat av Sekretariatet för nordiskt kulturellt samarbete vid Nordiska ministerrådet [Acta Universitatis Umensis 33]. Umeå: Universitetet i Umeå. Hauksdottir, A. 2001. Lærerens strategier — elevernes dansk. Dansk som fremmedsprog i Island. [TemaNord 2001 6]. Copenhagen: Nordiska ministerrådet. Lund, J. 2006. Norden, de nordiske sprog og nordisk sprogpolitik. In Nabosprogsdidaktik, L. Madsen (ed.), 19–38. Copenhagen: Dansklærerforeningen. Maurud, Ø. 1976. Nabospråksforståelse i Skandinavia. En undersøkelse om gjensidig forståelse av tale- og skriftspråk i Danmark, Norge og Sverige [Nordisk utredningsserie 13]. Stockholm: Nordiska rådet. Ringbom, H. 1987. The Role of the First Language in Foreign Language Learning [Multilingual Matters 34]. Clevedon: Multilingual Matters Ltd. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verständi gung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.

part 4

Determining the possibilities of reading comprehension in related languages

chapter 11

Interlingual text comprehension Linguistic and extralinguistic determinants* Renée van Bezooijen and Charlotte Gooskens Radboud Universiteit Nijmegen/Rijksuniversiteit Groningen

The three West-Germanic languages Dutch, Frisian and Afrikaans are so closely related that they can be expected to be mutually intelligible to a large extent. In the present investigation, we established the intelligibility of written Afrikaans and Frisian by Dutch-speaking subjects. It appeared that it is easier for speakers of Dutch to understand Afrikaans than Frisian. In order to explain the results, attitudes as well as linguistic distances were assessed. There was no evidence of a relationship between attitude and intelligibility. Three linguistic distances did show a relationship with reading comprehension, namely the number of non-cognates, the transparency of the lexical relatedness of cognates, and the Levenshtein distance, which calculates the similarity between the written forms of words. Keywords: Dutch, Frisian, Afrikaans, mutual intelligibility, attitudes, linguistic distance, Levenshtein distance

1. Introduction 1.1 Aim of the research When two persons speaking different languages meet, there are three possibilities. One speaker switches to the language of the other, both persons take recourse to a third language, or both persons keep using their own language. The third type of communication, which we will refer to with the term ‘receptive multilingualism’ (Braunmüller and Zeevaert 2001), can only be used when the languages involved are sufficiently similar. Receptive multilingualism offers many advantages, especially on the production side. People can express themselves more easily and more precisely in their mother tongue than in a later acquired language. The question is, of course, what the effect is on the reception side. How much of an effort is needed to understand the spoken or written message? What is the risk of communication breaking down? What exactly are the causes of communication failure?

250 Renée van Bezooijen and Charlotte Gooskens

1.2 Frisian and Afrikaans in relation to Dutch Research into receptive multilingualism has a long tradition. In the 1950s the mutual intelligibility of American Indian languages was studied (Hickerton, Turner and Hickerton 1952, Pierce 1952, Wolff 1959). More recently much attention has been paid to the communication among Scandinavians (e.g. Haugen 1966, Maurud 1976, Bø 1978, Börestam Uhlmann 1994, Zeevaert 2004). Other languages that have been investigated include Spanish and Portuguese (Jensen 1989) and Slovakian and Czech (Budovičová 1987). The present study focuses on the intelligibility of Frisian and Afrikaans for speakers of Dutch. There have been two previous studies of the intelligibility of Frisian by Dutch-speaking subjects (Van Bezooijen and Van den Berg 1999a, 1999b), but only in their spoken form, not in the written mode. The intelligibility of Afrikaans for speakers of Dutch has never been investigated experimentally. Here we focus on the relative intelligibility of written Frisian and Afrikaans. Both Frisian and Afrikaans are related to Dutch. However, the historical backgrounds of the relationship are different (see Section 1.2). As a consequence, the present-day linguistic relationship between Dutch and Frisian deviates in many respects from that between Dutch and Afrikaans. These linguistic differences can be expected to affect the relative intelligibility of the two languages. Therefore, the first factor we will consider in order to explain intelligibility is linguistic distance. In addition, we will look at the possible role of attitudes. It is often contended in the literature that a positive attitude towards a language will motivate people to try and understand that language, whereas a negative attitude will hinder intelligibility. So, the present study addresses the following three questions: 1. Which language is more difficult to understand for Dutch-speaking readers, Frisian or Afrikaans? 2. Can the difference in intelligibility, if any, be explained by different attitudes towards the two languages? 3. Can the difference in intelligibility, if any, be explained by differences in the linguistic distances to the two languages? Before going into the research method and the results, we will first provide some background information on Frisian and Afrikaans.

1.2 Frisian and Afrikaans The present study deals with the variety of Frisian as spoken in the province of Friesland in the Netherlands. In the literature, this variety is sometimes referred to as West-Frisian, to distinguish it from the North-Frisian and East-Frisian varieties spoken in Germany. In this paper we will simply use the term ‘Frisian’. Frisian is the second official language of the Netherlands, in addition to Dutch. It is spoken by about 350.000 people. Historically, Frisian is closely related to English. In the course of time,

Interlingual text comprehension 251

however, Frisian has become increasingly similar to Dutch. Frisian is now linguistically more distant from English than for example Swedish (Gooskens and Heeringa 2004). For centuries, Frisian was almost exclusively a spoken language. From the middle of the nineteenth century the status of Frisian slowly started to rise. This was largely due to the efforts of the Frisian movement, which strived for the acceptance of Frisian as a fully fledged language, fit to be used for all communicative purposes, including cultural expression. This goal was never reached completely, due to the pervasiveness of Dutch in Frisian society. Nowadays, Frisian is used more often in the lower than in the higher social strata, more often in the countryside than in the towns, and more often in the informal than in the formal domains. Frisian is taught to a limited degree in education. Only a small minority of the inhabitants of Friesland (17%) can write Frisian, but most (64%) report that they are able to read Frisian (Gorter and Jonkman 1995). Due to the dominance of Dutch in the media, education, and administration, Frisian loses more and more of its typical characteristics. Originally, Afrikaans was a dialect that developed among a small group of Dutch colonists who settled in South Africa at the beginning of the seventeenth century.1 In the course of time its nature changed, among others because it was largely used by nonnative speakers with an insufficient command of Dutch. In the beginning, Afrikaans was mainly a spoken language, but in the nineteenth century it started to be used in writing. In 1921 Afrikaans was acknowledged as a separate language from Dutch. At present, there are ten additional languages in South Africa with an official status. According to the last census of 1996, Afrikaans is spoken by about 6 million people, both black and white. This is 14% of the South African population. The majority lives in the Western Cape (39%), Gauteng (21%), the Northern Cape (10%) and the Eastern Cape (10%). Afrikaans is the mother tongue of people of all social classes. Most speakers can be found in the urbanized areas. Afrikaans is taught and used at all educational levels. However, English has more prestige and is therefore gaining ground. To be able to answer the research questions formulated in 1.1, three types of data are needed. First, the intelligibility of Frisian and Afrikaans texts by speakers of Dutch has to be assessed (Section 2). Second, the attitudes of speakers of Dutch towards Frisian and Afrikaans have to determined (Section 3). Third, the linguistic distances between Dutch and Frisian and between Dutch and Afrikaans have to be measured (Section 4).

2. Intelligibility 2.1 Method Subjects Twenty native Dutch language students (two men and 18 women) from the Radboud University Nijmegen and the University of Groningen were selected as subjects. They had no active knowledge (speaking or writing) of Frisian and Afrikaans and no or only very limited passive knowledge (hearing or reading). Their mean age was 23 years.


The youngest subject was 18 and the oldest was 34. They all had Dutch as their mother tongue. Two subjects spoke a dialect of Dutch at home. In both cases this was the Limburg dialect, which is not related to Frisian nor to Afrikaans. Task Intelligibility was assessed by means of a variant of the cloze test.2 As a basis we used two Dutch newspaper articles with an average level of difficulty.3 One article (‘the dating text’) was about dating agencies and comprised 329 words; the other (‘the feminist text’) dealt with the image of women created by modern music stations and consisted of 317 words. In either article, five nouns, five adverbs, five adjectives, and five verbs were selected at random. These were placed in alphabetic order above the text and replaced by blanks in the text. Next, the two texts were translated into Frisian and Afrikaans. The Frisian translation was made by a native Frisian-speaking student of Frisian and checked by a language specialist from the Fryske Akademy in Leeuwarden (Friesland). The Afrikaans translation was made by a native Afrikaans-speaking student of English and checked by a linguist from South Africa. In the Frisian and Afrikaans translations, the same words were removed and placed above the texts as in the original Dutch texts. These are the test versions which were used for the present investigation. The subjects were given ten minutes to put the 20 words back in the right place in the texts. The percentage of words placed back correctly was taken as a measure of text comprehension. In Appendix A, the Afrikaans version of the dating text is given as an example. Design All subjects were tested in both languages. In order to avoid order effects, subjects that got the Frisian feminist text got the Afrikaans dating text, and the other way around. Furthermore, half of the subjects started with Frisian, whereas the other half started with Afrikaans. So there were four test versions that were each administered to five subjects. All subjects first filled in a form enquiring about their origin, home language, age, and sex. Then they were given two blocks, with respect to the first and the second language tested, respectively. Both blocks contained a set of questions about the subject’s attitudes towards the language at hand (see 3.1) and the intelligibility test.

2.2 Which language is more difficult to understand, Frisian or Afrikaans? To determine the relative intelligibility of Frisian and Afrikaans we counted the number of words that were placed correctly in the original sentence context. For greater stability we summed the data for the two texts. The results are shown in Table 1. The mean percentage of correct responses appears to be considerably higher for Afrikaans (81.8%) than for Frisian (50.3%). All but one of the subjects had a higher score for Afrikaans than for Frisian. The difference, tested by a paired-samples t-test, is significant (t=-7,991, df= 1,19, p=.00). So, the answer to the first question is clear, written Frisian is more difficult for Dutch-speaking readers than Afrikaans.


Table 1. Mean, minimum and maximum percentages of correct responses for Frisian and Afrikaans. Language

mean

minimum

maximum

Frisian Afrikaans

50.3 81.8

20.0 45.0

90.0 100.0

3. Attitudes 3.1 Method Before carrying out the intelligibility test, subjects first answered some questions that aimed at gaining insight into their attitudes towards the two languages. Not only the attitudes towards the languages themselves were probed, but also the attitudes towards the speakers of the languages and the countries where the languages are spoken. The idea behind this is that language attitudes may be influenced by social connotations (Trudgill and Giles 1978). Subjects noted down their responses on six five-point scales. As an example we present the questionnaire related to Frisian; the same questionnaire had to be filled in for Afrikaans.

Does Friesland appeal to you as a vacation destination? very much

□ □ □ □ □ not at all

What is your impression of Frisians? likeable intelligent reliable

□ □ □ □ □ not likeable □ □ □ □ □ not intelligent □ □ □ □ □ not reliable

What do you think of the Frisian language? beautiful

□ □ □ □ □ ugly

Would you like to learn Frisian? very much

□ □ □ □ □ not at all

Diagram 1. Frisian questionnaire

3.2 Can the difference in intelligibility between Frisian and Afrikaans be explained by differences in language attitudes? The mean ratings on the attitude scales are presented in Table 2. In a number of cases the difference between Frisian and Afrikaans is significant. The subjects • find South Africa a significantly more attractive vacation destination than Friesland, • judge South Africans to be significantly more intelligent than Frisians,

254 Renée van Bezooijen and Charlotte Gooskens Table 2. Mean attitude scores for Frisian and Afrikaans with the minimum and maximum scores in parentheses. A score of 3 is neutral, lower than 3 is positive and higher than 3 is negative t-test

Mean Scale

Frisian

Does Friesland/South Africa appeal to you as a vacation destination? What is your impression of Frisians/South Africans? likeable–not likeable intelligent–not intelligent reliable–not reliable

t

df

sig.

3.3 (2–5) 2.4 (1–4)

−2.70

19

.01

2.6 (1–4) 2.4 (1–4) 2.8 (1–4) 2.6 (1–3) 2.6 (1–3) 3.1 (2–4)

−1.07 −2.18 3.25

19 19 19

.30 .04 .00

Afrikaans

What do you think of the Frisian/Afrikaans language? beautiful–ugly

3.2 (2–4) 2.9 (2–4)

−1.67

19

.11

Would you like to learn Frisian/Afrikaans?

3.8 (2–5) 2.9 (1–5)

−4.50

19

.00

• are significantly more motivated to learn Afrikaans than Frisian, • think Frisians to be significantly more reliable than South Africans. Averaged over subjects, the attitudes towards Afrikaans are more positive than those towards Frisian. As the subjects generally had fewer problems with Afrikaans than with Frisian, this outcome suggests that at the level of the group the two types of data may be related. To examine the relationship between attitudes and comprehension at the individual level, we correlated the two types of data of each subject. None of the correlation coefficients proved to be significant. So, it is not the case that subjects who expressed a relatively positive attitude towards Frisian (or Afrikaans) were also relatively successful in understanding Frisian (or Afrikaans). This outcome pertains to the subjects’ performance for each of the two languages separately. Finally, we also compared relative attitudes and comprehension for the two languages combined at the level of the subject. To do this we subtracted for each subject the attitude scores for Frisian from those for Afrikaans and the intelligibility scores for Frisian from those for Afrikaans. We then correlated the two types of difference scores. Again, no significant correlations were found. So, it is not the case that subjects who were more positive towards Frisian than towards Afrikaans (or the other way round) also performed more successfully for Frisian than for Afrikaans (or the other way round). We think that the last measure is the most convincing one and therefore conclude that attitudes and text comprehension are not related.


4. Linguistic distances 4.1 Method To calculate the linguistic distances, we first aligned the Dutch texts with the Frisian translations and with the Afrikaans translations. In a few cases we adapted the word order to obtain a better alignment. The aligned word pairs formed the basis for seven distance measures. Measures A, B, and C express the nature of the relationship between the Dutch and the Frisian/Afrikaans words, whether they are cognates or not. Measures D, E, and F express the transparency of the relationship from the viewpoint of the Dutch reader. And measure G, the so-called Levenshtein distance, expresses the degree of orthographic similarity between corresponding words. Because we assumed that lexical words (nouns, adjectives, numerals, main verbs) are more important for intelligibility than function words (articles, conjunctions, prepositions, pronouns, auxiliaries, modals, particles, adverbs), distances were calculated separately for these two word categories. The distance measures will now be explained in further detail. A. Percentage of cognates A large proportion of cognates, i.e. words in the two corresponding texts with a common root, can be expected to facilitate comprehension. However, a direct relationship is not a necessary condition for mutual intelligibility. In some cases, the meaning of a word can be deduced via a cognate synonym. For example, the Dutch word samenleving (‘society’) in the original newspaper article was translated in the Frisian text with maatskippij. These two words are non-cognates. Nevertheless, the Dutch reader can easily understand the Frisian word maatskippij because of the existence of the Dutch synonym maatschappij. The original Dutch text could just as well have contained this word. The percentage of cognates, either related directly or via a synonym, constitutes the first linguistic distance measure. B. Percentage of cognates related via a paradigm It is also possible to deduce the meaning of a word paradigmatically. For example, the Frisian translation of the Dutch word zijn (third person plural present tense of the verb to be) is binne. These two words are not related, not directly and not via a synonym. However, a Dutch reader may nevertheless understand the meaning of the Frisian word because it is related to the Dutch word ben (first person singular present tense of the verb to be), which belongs to the same paradigm as zijn. The percentage of cognates related via a paradigm constitutes the second distance measure. Most of the words in this category are function words. C. Percentage of non-cognates It should be impossible to deduce the meaning of a word in an unknown language if it bears no formal relationship with the corresponding word in the mother tongue. The


percentage of non-cognates should therefore be an important indicator of (the impossibility of) mutual intelligibility. It is the complement of the first two measures, but for the sake of completeness we present all three. D. Transparency of the lexical relatedness As stated above, it is impossible to deduce the meaning of a non-cognate (measure C). In the case of cognates, however, deducibility varies. In the course of time, the phonology or spelling of cognate words may have changed to such an extent that to a reader with no historical-linguistic background the relationship is no longer transparent. We therefore included a measure expressing the transparency of lexical relatedness. The scoring was done by the first author and compared to the independent scoring of a second linguist. The percentage of identical scores was 82.2% (Pearson’s r = .92) for the two Frisian texts and 90.1% (Pearson’s r = .92) for the two Afrikaans texts. On the basis of these results we concluded that the scores were sufficiently reliable. In the scoring the following four grades of transparency were distinguished. • Relatedness completely transparent (0 points) When two cognates have an identical form, there is no recognition problem. In this case a score of 0 points is assigned. Example: Afr. uitbuiting vs. Du. uitbuiting ‘exploitation’. Cognates with a small spelling difference not leading to a difference in pronunciation were also assigned a score of 0 points. Example: Fri. buro’s vs. Du. bureaus ‘bureaus’. • Relatedness fairly transparent (1 point) A score of 1 point is assigned whenever two cognates are so similar that the reader can be assumed to recognise the relationship fairly easily. In most cases there is a difference in only one letter. Example: Afr. sewentig vs. Du. zeventig ’seventy’. • Relatedness rather untransparent (2 points) A score of 2 points is assigned whenever two cognates have so little in common that it will be quite difficult for a Dutch reader to recognize the relationship. Usually several letters will be different. Example: Fri. jierren vs. Du. jaren ‘years’. • Relatedness completely untransparent (3 points) A score of 3 points is assigned if the two cognates bear so little resemblance that it must be (virtually) impossible for a Dutch reader to see the relationship. In most cases the majority of letters will differ. Example: Afr. hê vs. Du. hebben ‘have’. Distance measure D was calculated by averaging the total number of points over the total number of word pairs; it can vary between 0 and 3. E. Transparency of the grammatical relatedness In addition to the lexical meaning, a correct interpretation of the grammatical meaning of a word is also a prerequisite for text comprehension. Is it clear which word category (noun, verb, etc.) is involved? Is it clear what tense, number, gender, person, etc.


is involved? An example of the first type of problem is Afr. die, which Dutch readers will be inclined to interpret as a demonstrative pronoun rather than as a definite article. An example of the second type of problem is Afr. is, which will be interpreted by Dutch readers as a singular form of the verb ‘to be’, whereas in Afrikaans the form is also used for the plural. If the grammatical relatedness was transparent, 0 point was assigned; if it was not transparent 1 point was assigned. The two types of grammatical transparency were scored separately and summed. The scores thus range between 0 and 2. F. Transparency of total relatedness To determine the total transparency of the word relatedness, measures D and E were summed. The scores can range between 0 and 5. G. Levenshtein distance Measures D, E, and F are to some extent subjective and also very time consuming to compute, as the transparency has to be estimated separately for each pair of corresponding words. The advantage of the Levenshtein distance is that similarity between word forms can be computed objectively and automatically. In the present study, the simplest variant of the Levenshtein measure was opted for. As the texts were presented to the subjects in a written form rather than in an auditory form, we based the distance measure on the orthographic representation of the words rather than on the phonemic representation. Distance was determined on the basis of the minimum number of letters that need to be inserted, deleted, or substituted in order to transform the word in the one language into the other. All three operations were given an equal weight of 1 point. Word length was compensated for by dividing the total sum of costs by the number of alignments of letters (see Heeringa (2004) for an extensive explanation of the nature and application of Levenshtein distances). As an example we present the calculation of the distance between the Dutch word zoeken and the Frisian word sykje ‘search’ (see Table 3). The sum of costs (1+1+1+1+1=5) is divided by the number of alignments (7). The result is a distance of 0.71 or 71%. The total distance between two texts is the mean distance over all word pairs. The Levenshtein distances were calculated on the basis of the texts in which the original source word had been replaced by a synonym if a synonym was available (see Measure A). It should be noted that measure G and measure D are related. The larger the orthographic distance as measured by G, the less transparent the lexical relatedness as measured by D. Table 3. Levenshtein distance for Dutch zoeken and Frisian sykje ‘search’ Alignment

1

2

3

4

5

6

7

Dutch Frisian Cost

z s 1

o y 1

e – 1

k k

– j 1

e e

n 1


4.2 Can the difference in intelligibility between Frisian and Afrikaans be explained on the basis of the linguistic distance? Assuming that intelligibility is inversely related to linguistic distance, we expected the linguistic distances between the Dutch and Frisian reading texts to be larger than between the Dutch and Afrikaans reading texts. In Table 4, the results of the first three linguistic distance measures are presented. We see that for the content words the percentage of cognates (related directly or via a synonym) is almost identical (94.1% versus 94.6%). However, there is a marked difference in the function words. Whereas in Frisian almost all function words (93.4%) are related directly to their Dutch counterparts, Afrikaans has relatively many function words (23.7%) that are related to Dutch via a paradigm. Furthermore (not shown in the table), we want to mention the fact that the nature of the cognates differs between Frisian and Afrikaans, in the sense that the Afrikaans texts have a larger proportion of words that are related via a synonym. The meaning of these synonyms sometimes slightly deviates from that of the original Dutch word. For example, the Dutch word opeens ‘suddenly’ was translated by Afrikaans skielik. Dutch readers probably interpret this word correctly via the Dutch synonym schielijk. However, Dutch schielijk is less frequent and slightly archaic compared to opeens. In quite a few cases the relationship with the Dutch counterpart appears to be more direct in Frisian than in Afrikaans. However, we do not know to what extent this affects intelligibility. What we do know for sure is that a large proportion of non-cognates must have a negative influence on intelligibility. In this case the term non-cognates is used in the strict sense. It refers to those words in the Afrikaans or Frisian text which are not related to the corresponding words in the Dutch text and which have no related synonyms in Dutch. The percentage of non-cognates is higher for Frisian than for Afrikaans, both for function words (2.0% and 0.8%) and for content words (5.9% and 3.7%). This does not only hold for the number of tokens (24 and 14, respectively) but also for the number of types (20 and 10, respectively). These differences between Afrikaans and Frisian might seem small, but it should be recalled that one single unintelligible word can make a whole sentence or even a complete paragraph unintelligible. As stated above, not all cognates are equally transparent for the Dutch reader. In Table 5 the transparency scores are presented, separately for the lexical relatedness Table 4. Percentage of Dutch-Frisian and Dutch-Afrikaans cognates, cognates via a paradigm, and non-cognates, for function words and content words separately and total Frisian

Afrikaans

Function Content A. Cognates 93.4 B. Cognates via a paradigm 4.6 C. Non-cognates 2.0

94.1 – 5.9

Total

Function Content

Total

93.8 2.5 3.7

75.5 23.7 0.8

84.1 13.8 2.1

94.6 1.7 3.7


Table 5. Mean transparency of the relatedness of cognates, split up for function words and content words. Higher values denote lower transparency. Between brackets the scores converted to percentages. All differences between Frisian and Afrikaans are significant at the 1% level. Frisian Transparency

Afrikaans

Function Content

Total

Function Content

Total

1.24 (41.3)

1.45 (48.3)

1.33 (44.3)

0.40 (13.3)

0.63 (21.0)

0.50 (16.7)

E. Grammatical relatedness 0.03 (range 0–2) (1.5)

0.07 (3.5)

0.05 (2.5)

0.35 (17.5)

0.21 (10.5)

0.29 (14.5)

F. Total relatedness (range 0–5)

1.52 (30.4)

1.39 (27.8)

0.77 (15.4)

0.82 (16.4)

0.79 (15.8)

D. Lexical relatedness (range 0–3)

1.28 (22.7)

(measure D), the grammatical relatedness (measure E) and the complete relatedness (F). The scores have also been converted to percentages to facilitate comparison among the three measures. For example, the deducibility of the lexical relatedness of Frisian function words is (1.24/3)100−41.3%. It can be seen in Table 5 that the lexical relatedness of the Frisian cognates (1.33) is more difficult to deduce than of the Afrikaans cognates (0.50). This holds both for content words and function words. On the other hand, it is easier to deduce the grammatical relatedness of Frisian words (0.05) than of Afrikaans words (0.29). However, the latter difference is smaller than the former. Also, it has to be assumed that the transparency of the lexical relatedness contributes more to text comprehension than the transparency of the grammatical relatedness. We therefore think that measure D forms part of the explanation of why Frisian is more difficult to understand than Afrikaans. Finally, we also calculated the linguistic distance of Afrikaans and Frisian to Dutch by means of the Levenshtein distance (measure G). The Levenshtein measure expresses how many letters the Afrikaans and Frisian words differ from the corresponding Dutch words as a proportion of the total number of alignments (see 4.1). We only show the results for the cognates that are related directly or via a synonym (measure A). Any correspondence between the cognates that are related via a paradigm (measure B) and between the non-cognates (measure C) is largely based on chance and therefore not interesting. It appears from Table 6 that the Levenshtein distance (total distance) is higher for Frisian (34.2%) than for Afrikaans (20.9%). The difference is significant at the 1% level. This means that with respect to cognates the form of the Frisian words deviates significantly more from Dutch than the form of the Afrikaans words. Table 6. Mean Levenshtein distance between Frisian–Dutch and Afrikaans–Dutch cognates (%) Frisian G.

Afrikaans

Function

Content

Total

Function

Content

Total

37.6

30.1

34.2

22.5

19.4

20.9


5. Conclusion and discussion This study of interlingual text comprehension reveals that Dutch-speaking readers with no previous exposure to Frisian and Afrikaans are able to some extent to read Frisian and Afrikaans newspaper articles. There is, however, a considerable difference between the two languages. Reading Afrikaans presents few problems. In a cloze test, four out of five words were placed back correctly in their original sentence context. Reading Frisian appears to be much more difficult, as not more than half of the words are placed back correctly. The results of this study suggest that in the written communication with Dutch-speaking people, South Africans can use (to a large extent) their own language, whereas this is not possible for Frisians. This pertains to informative texts of average difficulty read by linguistically trained students. Recent research has shown that the results can be generalized to non-linguistically trained subjects (Van Bezooijen and Gooskens 2005b, Gooskens and Van Bezooijen 2006). In the literature, problems in interlingual communication are often attributed to the attitude of the receiver (listener or reader). It is assumed that the reported or measured comprehension problems are not so much due to a lack of transparency of the relatedness of the language, but rather to a lack of motivation (see Wolff (1959) for various examples from West-Africa). Unfortunately, it is very difficult to determine experimentally whether it is a question of lack of ability or lack of willingness. In the present study the attitudes towards Afrikaans tended to be more positive than towards Frisian. It is unlikely, however, that this has led to greater success in the cloze tests. First, there were no significant correlations at the individual level. So, subjects with a relatively positive attitude did not perform any better in the intelligibility test than subjects with a relatively negative attitude. Second, the subjects had had (virtually) no personal contact with (speakers of) Frisian and Afrikaans. Their reported attitudes probably reflect general, non-emotionally based stereotypes with no effect on performance. Third, the test was administered as part of a course in sociolinguistics. The students saw it as an interesting and challenging assignment. We observed no signs of an aversion towards the task. As mentioned in the introduction, receptive multilingualism has been studied for many language pairs in the past. The most systematic investigations have been on the Scandinavian languages. In the Scandinavian studies, differences in mutual comprehensibility between speakers of Norwegian, Swedish and Danish have been attributed to differences in attitudes towards the neighboring languages. For example, language tests have shown that Danes have a better understanding of spoken Swedish than Swedes have of spoken Danish. Also, Danes are more positive towards Swedes and their language than Swedes are towards Danes and their language (cf. Delsing this volume, Doetjes this volume). However, measurements performed by Gooskens (2006) indicate that the asymmetric comprehension scores can to a large extent be attributed to asymmetric phonetic relationships between the subjects’ language and the test language in the two countries. Additionally, the asymmetric comprehension scores


can be explained by the rapid development of the Danish spoken language, which has resulted in a large distance between phonetic characteristics of the spoken language and the rather conservative spelling of the written language. This difference is much smaller in the case of Swedish. When listening to Swedish, Danes can therefore understand many Swedish words via the Danish written form. This is much less the case for Swedes when listening to Danish. Finally, in contrast with oral comprehension, an asymmetric relationship has not been found for Swedish–Danish text comprehension, which again shows that, like in our investigation, attitude probably plays a minor role for the test results. We think that, just like for Danish, Swedish and Norwegian, the difference in intelligibility which we found between Frisian and Afrikaans by Dutch readers has also mainly been caused by linguistic factors. A first factor is the number of ‘true’ non-cognates, i.e. Frisian or Afrikaans words that have no relation to the corresponding word in the Dutch text nor to a Dutch synonym. These occur more often in the Frisian texts than in the Afrikaans texts. Not understanding a single, central term may have disastrous effects on text comprehension. A second factor which may have played a role is the fact that the Frisian cognates diverge more from Dutch than the Afrikaans cognates. This appears both from the objective Levenshtein measurements and from the subjective transparency scores. Finally, a third factor which may have worked to the advantage of Afrikaans is the spelling. The Afrikaans spelling is more similar to the Dutch spelling than the Frisian spelling. Especially, the spelling of the Frisian vowels, with a great number of unfamiliar diacritics and letter combinations (e.g. ii, û, ú, ô, â, ê, ea, eo, oa, ue, iu, oai, uoi) may be confusing to the Dutch reader (for more examples of differences between the Frisian, Afrikaans, and Dutch spelling systems, see Van Bezooijen and Gooskens 2005a). There are many morphological differences between Afrikaans and Dutch, but these are mainly simplifications. For a Dutch reader the simplified morphological system may be unusual, but in the end it may have little effect on text comprehension. A text generally contains so much redundancy grammatically, that the absence of explicit marking of, for example, number in the verb system presents no problems. After all, it does not present any problems to the Afrikaans-speaking readers either. The morphological differences between Frisian and Dutch are of a different nature. Here the Dutch readers are confronted with meaningful endings that they have never seen. Both the Frisian and Afrikaans texts contained few syntactical differences with Dutch, so in this respect both languages must have presented few problems. In future research, we plan to make a more detailed analysis of communicatively relevant differences between closely related languages at different linguistic levels. For example, few investigations have been carried out so far on the role of graphemic differences between languages for the receptive comprehensibility of closely related languages. An exception is formed by Möller (this volume), who used statistical techniques to estimate the importance of various graphemic correspondences between German and Dutch for the recognition of Dutch words by German readers. However,


his findings still have to be tested experimentally. We also wish to extend our analysis to spoken language comprehension. In Van Bezooijen and Gooskens (2005b) and Gooskens (2006) the first steps in this direction have been taken.

Notes * We would like to thank Vincent van Heuven, Sulette Bruwer, Jelske Dijkstra and Bart Ale wijnse for their help in various stages of the investigation. 1. The data for Afrikaans have been taken from Webb (2002). 2. The cloze test was developed in 1953 in the United States by William Taylor and has been used extensively for measuring text comprehension. Sometimes the words are placed above the text, like in the present study. Alternatively, it may be left to the subjects to think of suitable words. 3. We determined the readability of the text by means of the so-called LIX-index (Björnsson 1968). This index consists of the mean number of words per sentence plus the percentage of words exceeding seven letters. Texts with a value of between 35 and 44 have a mean degree of difficulty. The mean LIX-value for the two Dutch texts was 42.

References Björnsson, C. H. 1968. Läsbarhet. Stockholm: Liber. Bø, I. 1978. Ungdom og naboland [report 4]. Stavanger: Rogalandsforskning. Börestam Uhlmann, U. 1994. Skandinaver samtalar. Språkliga och interaktionella strategier i samtal mellan danskar, norrmän och svenskar [Skrifter utgivna av Institutionen för Nordiska Språk vid Uppsala Universitet 38]. Uppsala: Institutionen för Nordiska Språk. Braunmüller, K. and Zeevaert, L. 2001. Semikommunikation, rezeptive Mehrsprachigkeit und verwandte Phänomene. Eine bibliographische Bestandsaufnahme [Arbeiten zur Mehrspra chigkeit, Folge B 19]. Hamburg: Universität Hamburg, Sonderforschungsbereich Mehr sprachigkeit. Budovičová, V. 1987. Literary languages in contact (A sociolinguistic approach to the relation between Slovak & Czech today). In Reader in Czech Sociolinguistics, J. Chloupek, J. Nekvapil et al. (eds), 156–75. Amsterdam: John Benjamins. Gooskens, C. 2006. Linguistic and extra-linguistic predictors of Inter-Scandinavian intelligibility. In Linguistics in the Netherlands 23, J. van de Weijer and B. Los (eds), 101–13. Amsterdam: John Benjamins. Gooskens, C. and Bezooijen, R. van 2006. Mutual comprehensibility of written Afrikaans and Dutch: symmetrical or asymmetrical? Literary and Linguistic Computing 23: 543–57. Gooskens, C. and Heeringa, W. 2004. The Position of Frisian in the Germanic Language Area. In On the Boundaries of Phonology and Phonetics, D. Gilbers, M. Schreuder and N. Knevel (eds), 61–87. Groningen: University of Groningen. Gorter, D. and Jonkman, R. 1995. Taal yn Fryslân op ‚e nij besjoen. Leeuwarden: Fryske Aka demy.


Haugen, E. 1966. Semicommunication: The language gap in Scandinavia. Sociological Inquiry 36: 280–97. Heeringa, W. 2004. Measuring Dialect Pronunciation Differences Using Levenshtein Distances. Groningen [Groningen dissertations in linguistics 46]. Groningen. Hickerton, H. , Turner, G. D. and Hickerton, N. P. 1952. Testing procedures for estimating transfer of information among Iroquois dialects and languages. International Journal of American Linguistics 18: 1–8. Jensen, J. B. 1989. On the mutual intelligibility of Spanish and Portuguese. Hispania 72: 849–52. Maurud, Ø. 1976. Nabospråksforståelse i Skandinavia. En undersøkelse om gjensidig forståelse av tale- og skriftspråk i Danmark, Norge og Sverige. Stockholm: Nordiska rådet. Pierce, J. E. 1952. Dialect distance testing in Algonquian. International Journal of American Linguistics 18: 203–10. Trudgill, P. and Giles, H. 1978. Sociolinguistic and linguistic value judgments: Correctness, adequacy, and aesthetics. In Functional Studies in Language and Literature, F. Coppieters and D. L. Goyvaerts (eds), 167–90. Gent: Story-Scientia. Van Bezooijen, R. and van den Berg, R. 1999a. Word intelligibility of language varieties in the Netherlands and Flanders under minimal conditions. In Linguistics in the Netherlands 1999, R. van Bezooijen and R. Kager (eds), 1–12. Amsterdam: John Benjamins. Van Bezooijen, R. and Van den Berg, R. 1999b. Hoe verstaanbaar is het Fries voor nietFriestaligen? Philologia Frisica anno 1999: 9–26. Ljouwert: Fryske Akademy. Van Bezooijen, R. and Gooskens, C. 2005a. Intertalig tekstbegrip. De begrijpelijkheid van Afrikaanse en Friese teksten voor Nederlandse lezers. Nederlandse Taalkunde 10: 129–52. Van Bezooijen, R. and Gooskens, C. 2005b. How easy is it for speakers of Dutch to understand spoken and written Frisian and Afrikaans, and why? In Linguistics in the Netherlands 22, J. Doetjes and J. van de Weijer (eds), 13–24. Amsterdam: John Benjamins. Webb, V. 2002. Language in South Africa. The role of language in national transformation, reconstruction and development. Amsterdam: John Benjamins. Wolff, H. 1959. Intelligibility and inter-ethnic attitudes. Anthropological Linguistics 1: 34–41. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verstän digung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.


Appendix A: The Afrikaans version of the dating text Fill in the 20 words on the empty lines in the Afrikaans text below. aantreklike fantasties moet stel

al internet nooit teleurgesteld

blyk lank onbetroubare verhale

doen misbruik paar werk

eksplosief moeilike soms woorde

Nie lank en blond nie maar ’n klein, kaalkop lamsak Jaarliks soek tienduisende alleenlopers via ’n verhoudingsbemiddelingsagentskap of op die __________ na ’n metgesel. __________ met sukses, maar ook dikwels daarsonder. __________ mans blyk in werklikheid oud en lelik, en __________ agentskappe verdwyn skoonveld sodra die inskryfgeld betaal is. Dit het __________ geklink in die kletskamer op die net. Die ou met wie Annemarie op die internet in kontak gekom het, het aan al haar vereistes voldoen: hy was __________, blond en atleties. Het hy geskryf. “Maar tydens die eerste afspraak in die kroeg het daar ’n lamsak op my gesit en wag,” sê die 27-jarige vrou. “Hy was klein, kaalkop en het ook nog tien jaar ouer gelyk as die ouderdom wat hy op my webtuiste opgegee het.” Annemarie het drie __________ gewissel met die man en het kwaad en teleurgesteld vertrek. Die sprokie was verby. Vir sulke kliënte is Joke Pronk aangestel. Sy __________ vir die Algemene Vereniging Verhoudingsagentskappe (AVV) en bemiddel in geskille tussen kliënt en verhoudingsagentskap. ’n __________ saak. “Die mense is __________,” aldus Pronk, ’n __________ honderd euro terugbetaling bied dikwels min troos, hulle wil veral stoom afblaas.” Pronk __________ vas watter agentskappe nie hul afsprake nagekom het nie en deel ‘geel kaarte’ uit. Die aantal verhoudingsagentskappe het die afgelope jare — met name op die internet __________ gegroei. Daar is __________ vier — à vyfhonderd (op die net). Met die groei neem die __________ ook toe. Baie internetagentskappe blyk te verdwyn sodra die inskryfgeld betaal is. “Skielik bestaan die webwerwe nie meer nie. Of hulle __________ ’n kantoor in Kazachstan te hê.” Pronk ken die __________-, maar dit kom volgens haar nie voor by die vyftien agentskappe met ’n AVV-keurmerk, waarvoor sy werk, nie. Sy raai kliënte aan om __________ ’n eie e-posadres te gee nie, maar alleen e-posse te stuur via die webwerf. “As mense afsprake wil maak, adviseer ons nadruklik om dit te __________ op ’n openbare plek. Dit is regtig nie net psigopate wat op die internet na ’n vrou soek soos sommiges beweer nie, maar jy __________ wel versigtig wees.

chapter 12

Processing levels in foreign-language reading Madeline Lutjeharms Vrije Universiteit Brussel

Reading is a highly complex activity that can be described according to different interacting processing levels. These levels correspond to the use of specific kinds of knowledge. This paper presents a survey of reseach data on reading, with the focus on reading a foreign language, especially when related to previously acquired languages. Most of the data on the reading process come from cognitive psychology and deal with reading L1 English. Data and hypotheses on word recognition and sentence processing both in other languages and in foreign language reading are considered where possible. Keywords: processing levels, foreign language reading, word recognition, sentence processing

1. Introduction Reading comprehension requires the use of several kinds of knowledge and abilities, just as any use of language does. The description of the reading process is normally structured according to the different kinds of knowledge required for processing. The use of a specific kind of knowledge is called processing level. These levels interact during processing, but for description purposes processing levels have to be separated and are normally ranked from lower to higher level processing, i.e. starting from eye movements and lexical access to semantic processing. This bottom-up approach corresponds approximately to fluent receptive reading, a linear style of reading for getting to the gist of the text. Until about 1980, two contrasting models of the reading process had been proposed: a data-driven or bottom-up model and an expectancy-based or top-down model. Currently, there is consensus that both data-driven and expectancy-based processing interact during reading. A description of reading as a bottom-up process is to some extent artificial for, in perception, knowledge about what we perceive always interacts with perceived information. However, it is clearer to describe reading as a serial rather than as a parallel process. In what follows, special focus will be put on aspects of reading research relevant for reading a foreign language related to previously acquired languages. An elaborate didactic method for learning to read related languages is the EuroCom-method,

266 Madeline Lutjeharms

which was originally developed for Romance languages (Klein and Stegmann 2000, cf. McCann et al. 2003, Hufeisen and Marx this volume). To facilitate transfer from acquired languages, “sieves” to support the search for similarities — or: to look for familiar linguistic material — have been worked out for this method. The seven sieves are (1) international vocabulary, (2) common vocabulary typical of the language family such as “Pan-Romance vocabulary”, (3) sound correspondences or systematic sound shifts, (4) spelling and pronunciation, (5) syntactic structures common to the language family, (6) morphosyntactic elements and (7) prefixes and suffixes.

2. Research on reading In cognitive psychology the reading process is described as a form of information processing. Aspects like motivation or reading purpose are in general ignored, although they largely determine how we read. They probably influence higher level processing more than lower level processing. The purpose for reading determines reading styles (like scanning, search reading, skimming, receptive reading, intensive reading . . .), whereas motivation influences selective attention control and thus efficiency and depth of processing. Positive feelings enhance the ability to organize information more creatively, which implies a stronger representation in long term memory (Ashby et al. 2002: 247). However, the introduction of aspects such as motivation that have begun to gain attention in research only recently, would complicate even more the already complex research design. Moreover, most research is on single word recognition and on sentence parsing. It is not clear, whether motivation and reading purpose have an influence on those processing levels and if so, how. Research data on the reading process mainly come from laboratory experiments in cognitive psychology. Comprehension processes can hardly be observed and many aspects are not open to introspection. This requires the design of often complex experiments starting from a hypothesis on (aspects of) the reading process. Details of the reading process are isolated and controlled under certain conditions. Data collected during an experiment allow for verification or falsification of the hypothesis and mostly lead to new research questions. Participants have to solve a task, e.g. to decide whether a visually presented string of letters is a word or not. This is an aspect of word recognition. Reaction times, error rate and sometimes eye movements are measured under different conditions. Experiments on word recognition often use a kind of priming to influence conditions. Priming means the presentation of a somehow related word before presenting the target, which can facilitate, i.e. prime, or inhibit the task. Often masked priming is used, which means the presentation of the prime is so short (subliminal), that it cannot be processed consciously, as perception happens without awareness. For syntactic analysis, manipulated sentences, such as garden path and filler-gap sentences are used, often with ambiguity resolution, but also computer simulations. For reading L1 many such research data are available for all processing

Processing levels in foreign-language reading 267

levels, although far less so for higher processing levels like semantic processing than for lower, form based levels. With newer methods like the ERP-method (event-related potentials) measurement of electrical activity in the brain related to particular tasks is possible. Another imaging technique of neural activity in the brain is MRI (magnetic resonance imaging, Francheschini et al. 2004). Such data give a better insight into ongoing processes, whereas traditional data mostly focus on the end product of a processing task. Although more research on bilinguals and even multilinguals has been published recently, experimental and empirical data for foreign language reading are still rather scarce. Therefore, a description of interaction models and of different processing levels during L1 reading is the starting point for a discussion of reading in a foreign language. Reading in a foreign language that is not completely unfamiliar because of typological proximity to a language or to languages previously acquired and/ or common vocabulary will be the subject of special focus.

3. Models of interaction and processing levels during reading The two types of processing in working memory that are important for reading are automatic processing and attentional or controlled processing. Working memory is a term used to refer to systems that store information temporarily in order to perform cognitive tasks on this incoming information. Working memory regulates incoming information and maintains it shortly, while executive processes manipulate the information (Traxler et al. 2005). During this manipulation information retrieved from long-term memory is used to prepare the new information for integration into longtime memory. Automatic processing is fast, parallel and without effort. This does not imply that it is unintentional. There are no capacity constraints. Attentional processing demands effort, is serial, has capacity limitations and is often a conscious process or one that is open to consciousness. Attentional processing is required for new, unfamiliar or not fully mastered information as well as for unexpected and inconsistent information and especially for semantic processing or comprehension. Non-semantic aspects of processing can be automatized with practice (Rawson 2004, cf. Paradis 2004: 31). Lower level processing — eye movements, word recognition and syntactic analysis — is an automatic process in fluent reading. On perceiving, expectations based on experience interact with incoming stimuli. Expectations regarding text continuation facilitate further processing when confirmed, but expectations are also adapted to the incoming data as an ongoing process. To explain automatic lower level processing two hypotheses play a role in current research: modular and connectionist models. Fodor (1983) explains processing as a modular process. A module is an autonomous device that only allows for domain-specific, encapsulated processing units like lexical access, word recognition or syntactical analysis. Modular processing is described as fast and data driven, without influence of context and of higher processing levels on lower modules. The information proc-


essed by one module is passed on to the next module. Central processes are required for comprehension and for processing information that is experienced as illogical or as too difficult. Difficulty may be due to knowledge deficits that can be compensated for by inferring. Such central processes are non-modular and slow, as they require attention. A modular model could explain why, in the case of good linguistic competence, reception is mostly error free. Connectionist models are more recent and “describe the physiological substrate of memory better than other models” (McClelland and Rumelhart 1985: 101). In such highly interactive models many units or nodes of a network function in parallel. This causes activation spreading and inhibition, both topdown and bottom-up. Interactive models explain why — as has been observed — higher level information like meaning speeds up lower level processing. Although McClelland and Rumelhart applied their interactive model to explain word recognition (Ellis and Humphreys 1999: 315ff.), only quite recently connectionist models have been adopted more generally as a model for research on word recognition. For a long time, modular approach dominated the field, whereas connectionist models were adopted more frequently to explain parsing. It is conceivable that the two approaches are not mutually exclusive (cf. Paradis 2004: 122–50; Lutjeharms 1994: 54f.)

4. Processing levels of the reading process The following presentation of processing levels is not meant to be a model of the reading process. The purpose is to introduce the kinds of knowledge and processes or abilities that control reading. Levels and processes could be split up into more or into fewer units and there may be parallel processing. Whereas lower level processing normally functions automatically in fluent reading, when reading a language not yet well acquired, attentional resources are often required for processing form-based linguistic information. Because of the limited capacity of working memory for attentional processing, avoidance or guessing strategies are frequently used, which involve skipping parts of text to make (some) sense of its meaning. This is a typical strategy for beginning and weak readers in L1 because of the lack of automaticity in word recognition and syntactic analysis (Lutjeharms 1997). In didactic reading research on beginning readers, the idea of reading as a process of predicting and confirming was launched in the late sixties — an idea that became quite popular in foreign language teaching and was for some time propagated as an ideal strategy, which it certainly is not (Bernhardt 1993). When the form level requires attention, not enough processing capacity is available for content. Avoidance and guessing strategies are then inevitable, but they also entail misunderstanding, esp. when the reader is not familiar with the content and when cues for guessing are insufficient. Fluent L1 readers need to resort to the strategies of weak L1 readers when acquiring a new language for as long as their foreign language competence is too low because of lacking linguistic knowledge or insufficient automaticity


in using the knowledge. Focus on form like in the EuroCom-method, where cognates are used systematically for faster recognition and acquisition, is useful both for foreign language reading and acquisition. For reading it supports guessing or rather inferring; for acquisition the possibility of connecting parts of the input to previous knowledge supports memorizing. Moreover, reading is a useful source for linguistic input, including repeated input. Fluent reading implies automatic processing of linguistic form so that all attentional resources can be focused on processing the semantic level, i.e. on reading for content.

4.1 Graphemic level: eye movements, visual word recognition and phonological recoding Eye movements during reading have been studied for over a century. Eye movements are the only directly observable component of the reading process. We do not move our eyes evenly along the text when reading. Eye movements consist of fast jumps in between short fixations. These jumps are called saccades. Fluent readers fixate only some words or parts of words. Together with these fixations peripheric sight guides the saccade to the next fixation. When reading a related language with the same script and the same kind of word order, eye movement habits can be transferred from the L1 or another acquired language to the target language. However, when linguistic competence is low, more fixations will be required, the reader might be less efficient in choosing the next fixation point and many regressions can be expected. Regressions also occur more often when reading semantically difficult text. Text difficulty is, of course, largely a relative, reader-dependent variable. Words are read faster than a string of single letters; this is called the word superiority effect. We recognize orthographic patterns and/or morphemes more easily than a string of single letters. Frequency is an important factor in speeding up recognition, because frequently activated words or parts of words are recognized faster. Word beginnings are important for visual recognition (Deutsch and Rayner 1999: 416 “readers of Hebrew and English tend to land initially about halfway between the beginning of the word and the centre of the word”). This has been observed both for isolated words and for words in context (ibid.). When reading a foreign language, this habit can lead to the use of prefix for meaning search without attention being paid to the stem morpheme, even if the prefix as such should seem familiar from acquired languages (Lutjeharms 2001, Müller-Lancé 2002). Phonological recoding is a — probably very abstract — conversion of the visually recognized form in sound. Evidence as to whether such a phonological recoding is obligatory for lexical access, whether it is a by-product or not (always) occurring is conflicting (Folk 1999, Kim and Davis 2002: 570); certainly, both orthographic and phonological features appear to be activated at some stage when reading a word. Experimental results could be language specific. Subvocalization is often observed for processing linguistically or semantically difficult text, but this conversion is a post-lexical process


necessary for rehearsal in working memory, for which the acoustic code is required. This implies that knowledge of the pronunciation of a language is important not only for production and listening comprehension, but even for reading to avoid the acquisition of defective pronunciations. Rehearsal is also required for looking up a word in a dictionary.

4.2 Word recognition with lexical access and recognition in context Mental lexicon is the designation for the representation of linguistic knowledge in memory. A visual and an auditory in- and output system are considered to be separate from the mental lexicon (Jescheniak and Schriefers 2001: 372, also for further literature). The graphemic level of the reading process would then correspond to the visual input lexicon. The next phase in word recognition is lexical access, the access to a word’s entry in the mental lexicon. With lexical access, all representation levels of a word in the mental lexicon are available for further processing. Several hypothetical models have been proposed for lexical access. Best known is the Logogen Model (Morton 1969). A logogen is “an evidence collecting device . . . which becomes increasingly activated the more features of the incoming stimulus resemble those of the word that it represents” (Taft 1991: 4). According to serial-discrete models, only selected entries are activated (Jescheniak et al. 2001); according to cascaded activation models, however, all activated lexical nodes cause a weaker or stronger activation, even when they are not selected (Costa et al. 2000). Serial-discrete models are modular, cascaded activation models connectionist. In the literature the possible role of syntax and discourse level semantic constraints during lexical access is controversial and might be task-dependent (Balota et al. 1999: 36ff. , 46ff.; cf. McQueen and Cutler 2001: 485f.). Following an overview of the literature Sanford concludes that evidence suggests “that at least some of the meaning of a word is accessed as soon as it is encountered in a sentence”. Apparently a “selective (or partial) availability of semantic information” occurs as an early process, but semantic processing can be shallow or incomplete (Sandford 1999: 329, cf. Bowers et al. 2005, who observed some form of semantic access). Proofreading is a case of recognition with focus on form and no or shallow semantic processing as a controlled process. But even if access to a word’s entry in the mental lexicon does not imply immediate access to its syntactic and semantic use, the access to such information must be a consequence of lexical access in as far as the word has been acquired. It is generally accepted that words have representations on several levels in the mental lexicon. These include phonological, visual and morphological form or syntactical information and use (including word class or — for verbs — valency), but also linguistic code. When reading a foreign language such word representation levels tend to be transferred from the native translation equivalent to the foreign language target word (cf. Chen et al. 1997, Jiang and Forster 2001). Especially in the case of related languages, this often proves a successful strategy. Meaning (denotation and connota-


tion) could be seen as part of semantic memory, but how this is linked to the mental lexicon is not clear. The generally proposed organizational form for all this knowledge in the mental lexicon is a kind of network of interconnected units or features. “Processing flows across a series of representation levels” (Peterson et al. 2001: 1223). Balota et al. (1999: 47ff.) stress “the importance of attentional selection of appropriate processing pathways” with different patterns found “depending upon the types of goals engaged by a given task” and “argue for a flexible lexical processing system in which attentional systems modulate the importance of the numerous codes available for a given word”. Evidence is conflicting as to whether whole words are the basis of lexical representation or, in the case of morphologically complex words, morphemes — which would imply such words are then accessed via the base form (Marslen-Wilson 1999: 102f.); the access may also be language specific (Marslen-Wilson 2001, Longtin et al. 2003: 331). However, many data have been found for the facilitating role of morphemes during lexical access (Feldman 2000), and in some models parallel processing of whole word and morphemic units is proposed (Longtin et al. 2003, Longtin and Meunier 2005, Melinger 2003, Moscoso et al. 2005). Morphemes, in many languages at least, appear to be one of several organizational principles in the mental lexicon. This implies that attention to morphemes, whether they are affixes, grammatical morphemes or stem morphemes, is useful for improving lexical access. For inflectional morphemes and stem, separated morphological representation seems robust. Data for a number of languages show that the type of morpheme influences morphological processing, whereby several properties (such as word formation type, i.e. inflection vs. derivation, productivity of the morpheme and affixal homonymy) may influence processing in combined form (Bertram et al. 2000b). In foreign language reading (related languages, L1 Dutch, L4 German), I found that words with familiar looking morphemes are often accessed on the form level only, if the morphemes do not support understanding (as in German Vorgang or entsprechen, Lutjeharms 1988: 220ff.). The familiarity may depend on inter or intralinguistic similarity. Such words are then recognized on the form level and this recognition interrupts further processing (Lutjeharms 2004: 17), probably also as a kind of avoidance strategy during a demanding task. For the acquisition of foreign language reading, recognition exercises focusing on specific types of morphemes can be recommended because of the important role of morphemes during lexical access (cf. sieve 7 in the EuroCom-method). Lexical access can only occur if there is a representation for this word or its morphemes in the mental lexicon. New, already encountered, but not yet acquired words can be represented in episodic memory. Episodic memory is a memory system for specific episodes, i.e. for information linked to a special place and time. We might remember that we have already seen the word and where and when, maybe even remember we have already looked it up, without having access to meaning or use. Access to the meaning of a newly acquired word form is often brought about by way of a translation in a more strongly fixed linguistic code. In the case of morphologic-


al similarity between the target word and the translation equivalent, lexical access via the equivalent seems plausible. For reception, the processing of target language words can in such a case be transferred from a stronger language. For use in production, the correct word form should also be recallable. Here, sound correspondences are possibly helpful (cf. sieve 3 in the EuroCom-method). It has often been observed that words in context are processed faster than words in isolation (e.g. Kutas 1993). This is considered to be a form of priming: When accessing a word’s representation, a whole network of links on several levels is activated. This activation spreading causes faster processing of appropriate information. It has also been observed for morphological similarity, even without semantic transparency (Zwitserlood 1995, Bertram et al. 2000a, cf. Feldman 2000), and on the sentence level (Kutas 1993). In research on bi- or multilingual participants, priming experiments are generally used as a research method. After presenting a form and/or meaning related prime in one language, reaction times for the recognition of a target word in another language are measured to see whether this prime speeds up reaction time or slows it down. Most of this research is on single content words in Indo-European languages. The extent to which these results can be generalized is not yet clear. However, even if sentence information and semantic context influence word recognition as early as lexical access, such experimental data do have relevance for non-experimental processing situations as they show whether certain types of prime activate adequate or non-adequate networks. On the whole, research data related to lexical access in the case of bilingualism produce a rather complex picture. Because the mental lexicon originally develops together with L1-acquisition, the native language has a prominent role in shaping it. In the case of low foreign language proficiency, a target language word is often connected to the concept by an activation of the translation equivalent (Chen et al. 1997). This corresponds to a subordinate model of bilingualism. But even in case of high L2 or L3 proficiency, the connection between the L1-word and the concept is stronger than that between the target language word and the concept (ibid.: 279ff.). And even with near-native proficiency, translation priming effects appear earlier than phonological priming effects (Grainger and Frenck-Mestre 1998). This means that activation of an L1 translation equivalent influences lexical access earlier than the activation of phonologically similar morphemes. Are several languages represented in a language specific way, i.e. separate or shared? And if so, is this a feature of the mental lexicon, of input and output systems or both? There are no clear answers to these questions. A general answer might not even be possible, as word qualities, L1- reading skills and linguistic proximity between the languages concerned may influence research data. Thomas and Allport (2000) observed language specific word recognition procedures for listening comprehension but not for reading comprehension, although in reading some influence of typical aspects of spelling were found. In reading, the non-relevant language could not be de-


activated. Dijkstra et al. (1999) found cumulative support in the case of orthographic and semantic overlap, whereas phonological overlap resulted in inhibition. In speaking, language choice is under control; in reading, there is less control (Dijkstra et al. 2000: 460ff. , Paradis 2004: 230). Perhaps auditory input and output systems are rather language specific (cf. Friel and Kennison 2001: 250), visual systems far less separate — maybe only via phonological recoding. Roelofs’ data (2003, L1 Dutch, L2 English), however, seem to indicate a shared auditory output system for production (cf. also Costa et al. 2000: 1285 and Dijkstra et al.1998: 178). The mental lexicon might be seen as shared or, as Ameel et al. (2005: 63) summarize the evidence, it may be the case that “the representations of the bilingual’s two languages may be readily and broadly permeated by one another”. For access to cognates, strong priming effects were observed, probably because form and meaning related activation strengthen each other (Grainger and Frenck-Mestre 1998, cf. Costa et al 2000: 1290). De Groot (1993) assumes common representation for cognate translation equivalents, but language specific representation for non-cognate equivalents (for more data: Friel and Kennison 2001). Research with cognates supports “a view of language non-selective access implying all languages known to an individual may affect word activation and recognition” (Lemhöfer et al. 2004: 586). The data of Lemhöfer et al. (2004: 601, 603) for Dutch(L1)-English(L2)-German(L3) trilinguals demonstrate for cognates “that co-activation of three lexicons occurs even within the same words” and that “not only the dominant language (Dutch) exerts an influence on a word recognition task carried out in a weaker language (German), but that a second, non-target language (English) can also affect the recognition process on top of the strong facilitation caused by the mother tongue”. In teaching reading skills in German as an L4 to Dutch speakers, I often observed that conscious knowledge of the correct meaning of a deceptive cognate can only be used in automatic lexical access after frequent exercising (cf. Friel and Kennison 2001: 253); the strong connection to the L1-meaning (often strengthened by the L2/3 English meaning, as in German weil vs. Dutch terwijl and English while) first has to be inhibited, esp. when the deceptive L1-cognate is frequently activated. The possibly shared word representation of cognates justifies the conscious connecting of cognate words in the EuroCom-method. For internationalisms, shared representation may be assumed as well.

4.3 Sentence processing Syntax is expressed by words, by their order, their morphology, their word class and even by their meaning. To a large extent, syntactic markers are word-bound. It seems probable that a lexical entry consists not only of the word form(s), but also contains possible uses of this form. Just as in linguistic theories on syntactic processing (Lutjeharms 1998), a tendency to allocate always more importance to the lexicon can be observed in parsing models in cognitive psychology. Bates and Goodman (1997: 568) present “evidence from language development, language disorders and real


time language processing in support of a lexicalist approach to grammar”. Information about word class, morphological structure, subcategorisation and argument structure are connected to the lexical entry. Such cues guide syntactic analysis. There are, however, two types of information and, accordingly, two types of parsing models. In lexical entry driven models, properties linked directly to the word are emphasized. These properties might be summarized as inflectional markers and valency or “attachment biases of verbs, prepositions and nouns” (Snedeker and Trueswell 2004: 275). Shapiro et al. (1993: 110) found data for a lexical entry driven parser: “When the verb is encountered, the various argument structure possibilities are exhaustively activated. . . ... The preferred structure is then projected as a possible initial parse”. In other models, phrase structure driven rules are prominent, with sentence related properties, mainly word order, being used. In cognitive psychology most of the research is on English as L1. Syntactically, English is a special case as word order is the most important cue for sentence processing. MacWhinney and Bates (1989: XIV) even call it an exotic language, because of the overriding importance of word order. In most languages far more cues are used for parsing. In English for instance the assignment of the relative pronoun to a subject or object function can be decided immediately after the pronoun has occurred for the subject is followed by a verb, the object by a subject, as in (Schriefers et al. 1995: 503): The senator that the reporter attacked . . . . Ambiguous relative pronouns in Dutch and German tend to be assigned a subject function. If necessary, a reinterpretation follows, but only at the end of the clause after the auxiliary. In English, however, semantic information influences processing immediately (ibid.). From such data it may be concluded that it is dangerous to generalize data found for English. Dutch and German are considered to be SOV-languages due to the underlying structure (Haegeman 1991: 520ff. , Suchsland in Borsley 1997), but all combinations of S (subject), O (object) and V (verb) are possible, although OSV and VOS are rare. In SOV-sentences, such as subordinate clauses, there may be no expectations about slots to fill, as complements precede the verb. That is why argument structure may play a role that is different from that in SVO-sentences (Konieczny et al. 1997). Experimental evidence has been found for the assumption “that case marking languages such as German may employ non-syntactical processing routes to determine the thematic interpretation of a sentence. . . . .This finding supports proposals assuming that morphological case reflects the thematic relations between the arguments” (Bornkessel et al. 2003: 270, 295). In Dutch, case marking is, as in English, only a matter of pronouns. Case marking cues are therefore very often not available. As a result, word order cues appear to be more widely used by Dutch-speakers than by German-speakers. I found evidence of this strategy in errors by Dutch-speakers reading German LSP and avoiding case markers (Lutjeharms 1998: 140f.). An interesting theory of syntactic analysis — a functional theory — is MacWhinney and Bates’s Competition Model (1989). This connectionist model shows what kind of elements guide syntactic processing in a specific language. Several surface cues such as


word order, lexemes with their properties, e.g. noun animacy, congruency and morphological indicators, e.g. subject-verb agreement, are used simultaneously in different combinations and with changing cue strength to reach correct role assignment (see also Traxler et al. 2005). Surface indicators may be syntactic cues (word order, inflectional morphology), phonological cues (prosodic information and intonation) for listening comprehension, and semantic cues (animacy). They are used simultaneously in complex configurations of parallel activation, i.e. in competition. Cues can reinforce each other. Cue validity of a surface indicator develops during language acquisition from availability, i.e. frequency, and from reliability. Salience, i.e. detectability, also plays a role. Two levels map in the model, a functional level (representation of meaning) and a formal level (surface forms), which are processed in complete interaction. Form-function mappings are made as directly as possible. A single form can map onto several functions and a single function can map onto several forms. The surface indicators are language specific. Within the Competition Model experiments on syntactic analysis in several languages have been carried out, e.g. on the role of word order for the assignment of the subject. In English, word order is the main cue; in Italian, it is verb morphology (Li et al. 1993, Liu et al. 1992). In Dutch, word order is also the main cue, but the case form of a personal pronoun has higher conflict validity (McDonald and MacWhinney 1991: 409, see also Lutjeharms 1998). If animacy and word order conflict, word order is stronger in English; in Italian it is animacy (McDonald and MacWhinney 1989). In German, morphological cues are prominent as soon as they are complete (Bates and MacWhinney 1989: 56). So, depending on the language, different cues have higher conflict validity. Syntactic cues appear not to be as universal as often proposed (but see Universal Grammar research or Clifton 1995). Most of the data on syntactic analysis in a foreign language have been collected within the framework of the Competition Model. It could be shown that learners of English with several L1s (Spanish, Italian, Dutch, Japanese, Chinese) use L1 strategies, but also some of the typical English word order strategies. L1 strategies do not necessarily interfere, even if not used by native speakers of English. With growing proficiency, increasingly more target language strategies are adopted (for a survey of the literature see McDonald and Heilenman 1991, Liu et al. 1992, Koda 1993). But noninterfering L1 strategies might still be applied (Kilborn and Cooreman 1987). FrenckMestre and Pynte (1997: 121, 134) demonstrated that transfer of syntactic cues from more proficient languages depends on perceived structural similarity of the languages concerned. Readers often do not perceive target language cues if they have not been prepared to react to them. But even after some preparation, non-automated cues tend to be avoided. As syntactic analysis is normally automatic, influencing it is not easy. For Dutch speakers reading German, I found that case markers are avoided when this morphology is not familiar from L1 (Lutjeharms 1988: 272ff.). Familiar looking case markers are not avoided, but are mostly understood in the L1-function. Genitive ‑s, in German


a marker of singular, is thus systematically processed as a plural marker or the ‑er-case marker of an adjective (as in ein großer Tisch, mit großer Mühe) is understood as a comparative, which it would be in Dutch. The fact that knowledge — as evidenced by productive tasks — is often not used in reception is probably due to lack of control, i.e. lack of automatic processing of the foreign language cue. Explicit knowledge has not yet developed into implicit knowledge; no routine has as yet been developed. Knowledge representation does not imply automatic knowledge retrieval.

4.4 Form, content and the role of prior knowledge So far, mainly processing of linguistic form — also called decoding or lower level processing — has been discussed. Lower level processing in fluent reading is a highly automated process, i.e. fast, parallel and non-attentional. Working memory capacity is then completely available for semantic processing. When reading a still insufficiently acquired language, decoding will be at least partially an attentional process. Besides, the use of resources such as a dictionary or the help of linguistically more proficient readers may be required. Inferring and avoidance strategies are inevitable to arrive at some — not necessarily correct — interpretation, when all working memory capacity is used for lower processing. As such conscious inferring strategies - sometimes called intelligent guessing or predicting — are too demanding for extensive use. We need prior knowledge on all processing levels: script, spelling pattern, words with rules for their use, syntactic patterns, but also knowledge of the world for higher level processing and inferring. When reading a new language, we will try to transfer prior knowledge as soon as this seems possible. We need interpretable forms for the transfer of knowledge. Content knowledge can support the transfer of knowledge to the decoding level in combination with recognizable linguistic forms (internationalisms, cognates, expectations for word order etc.). However, the transfer of non-suitable knowledge can be very misleading, and in trying to attain some coherence one misinterpretation can lead to more errors. The best strategy to avoid such problems is to make greater use of resources, and especially to achieve more language acquisition and to increase language use automaticity. More proficiency means a better chance to infer correctly on all levels.

4.5 Semantic analysis Text comprehension is brought about by an interaction between decoding results and prior knowledge of content. Lexical access and syntactic analysis result in the incremental construction of a propositional representation of sentence content — a process that in fluent reading and with for the reader appropriate content probably occurs nonattentionally. Some evidence has been found that in reading for the gist of the message readers may be satisfied with underspecified representation of aspects of text information. This has been found for anaphoric inferences (Klin et al. 2006).


Content-based previous knowledge is mostly called knowledge of the world or schema-based knowledge. A schema “is an organized packet of knowledge that enables us to make sense of new knowledge”; it “can be conceptualized as a series of slots that can be filled with particular values” (Harley 2001: 329). Schema knowledge develops as a kind of abstraction based on previous experience. This prior knowledge supports anticipation and allows for the integration of the newly acquired information. To activate suitable prior knowledge titles, illustrations, abstracts etc. are used for written texts as a means of foregrounding. Schema-based knowledge of the subject matter of a text is a prerequisite for content-based inferring during reading. This prior knowledge is what makes texts coherent. On the sentence level, the usual topic-comment structure of texts, i.e. known information before new information, follows the same principle of foregrounding. In our use of language much information is implicit. If we read Mary is happy. She won the lottery we infer, that she is happy because she won the lottery and not because e.g. her holiday has just started. Prior knowledge about text types, text patterns and about the functions of texts is also used during semantic analysis. All these kinds of prior knowledge elicit expectations about text purpose and content that may support decoding. About how exactly and at which level this occurs in fluent reading is not very clear. In the case of knowledge gaps on the decoding level, all these kinds of prior knowledge can support attentional search of meaning. When reading incoherent information, bridging inferences have to be made in order to make sense. Together with internal text properties (such as style, sentence length, word choice etc.) the prior knowledge that readers bring to the text determines text difficulty. Today, the result of semantic processing — when text comprehension is reached — is usually called a mental model (Johnson-Laird 1983). Mental models “are multidimensional and represent five kinds of information”, such as spatial, causal and temporal information as well as information about people’s goals and about “the characteristics of people and objects” (Harley 2001: 332). There are other models like story grammars, Van Dijk’s macrostructures (1980) or Kintsch’ (1988) construction-integration model. In constructing a model of text information we infer with the help of prior knowledge. The extent to which (some) inferences may be processed automatically is a controversial point in research (Rawson 2004). Inferences can be intended, i.e. the author thinks the target group will be able to make them, or elaborated, which means they go beyond what is required for comprehending text content. Decoding is of course the first step to be able to construct a mental model, but other subskills and memory representations such as use of prior knowledge about text topic and text structure, richness of semantic representations for word meanings, inference skills and metacognitive monitoring to integrate text information are also required (Oakhill et al 2003: 463f.). Mental models are constructed incrementally, i.e. gradually using incoming data. They only contain text content, not the linguistic form of a text. Memory representation of a text consists of condensed information that is taken from the text, combined with prior knowledge on the topic. This information is integrated into long term memory. Such representations may vary individually according to prior knowledge


and reading purpose. Motivation for text content implies more attention is given to it, which leads to deeper processing. Reading purpose influences which parts of content are focused on. Surface text form is not retrievable after reading. However, traces of surface forms are encoded somehow, for even when readers no longer remember having seen the text before, they read a once already processed text faster than a text about the same subject, but in a different form (Levy 1993: 54ff. , Moscovitch and Bentin 1993). This observation explains why even a text that is read without attention for content supports language acquisition. How different is conceptual knowledge across languages? At least two aspects are relevant in answering the question: general human perception and naming. For listening comprehension, language specific effects in perception have been observed for “speakers of different languages become sensitised to different prosodic pattern and phonetic constraints” (Green 1998: 253). Consequences of this phenomenon have been observed for foreign-language acquisition as well. But are such language specific differences also found for concepts? Here, the data are not conclusive (ibid.), but differences in naming patterns have been found across languages (Malt et al. 2003). There is “the notion that all languages use the same set of semantic primitives identical to a subset of conceptual primitives in which thinking takes place”. But these primitives can be combined “into specific concepts of a particular language” (Green 1998: 253). There may be more or less difference according to the kind of information, such as more difference in the domain of abstract knowledge, less for visual, auditory and kinaesthetic systems. Related languages may present less difference in naming of conceptual knowledge than languages with little proximity.

5. Code-switching In general, code-switching is seen as a phenomenon of discourse. Franceschini et al. (2004) have studied code-switching as a neural activity, when reading a continuous coherent story in two languages (in which after approximately three short sentences the language switches again). In their research, code-switching occurs on the decoding level. Whether and how this might influence semantic processing is another question. However, code-switching on the level of semantic processing can be observed as well. When making notes when reading (or listening to) English or French LSP-texts I mix three languages: the language being decoded, Dutch (L1) and German (the language I use for teaching and mostly for writing LSP-texts). A colleague with the same linguistic skills confirms that she does exactly the same. Other colleagues have told me they use their L1 when taking note while reading or listening to LSP-texts in less strongly acquired languages. This means that the text content has been processed, but that the surface form level is no longer present in working memory, and that content is rendered in a stronger, more accessible code — in a cognitively demanding situation. It might be seen as a special case of multilingual communication — in the case of reading


“distributed over two situations” (House and Rehbein 2004: 3), that of reception and that of production.

6. Processing levels, reading and acquiring related languages In the EuroCom-method, a foreign language is acquired with the help of an L1 and/or L2, L3 etc. that is related to one or more of the acquired languages. From a didactic perspective this makes sense for several reasons. For the intake of new information, we have to organize it and connect it to activated knowledge in long-term memory. The situation is similar for content and for linguistic information. Activated entries in the mental lexicon that belong to already acquired languages can function as a connection or “transfer base” (Meißner 2003). As cognates probably share a representation in the lexicon, connecting related words and internationalisms (sieves 1 and 2) is a sensible process from a learning perspective. By activating a network, such connections can also strengthen already acquired linguistic knowledge. Improving recognition of cognates by using interlingual sound correspondences also facilitates recall of meaning. Not all learners are good at discovering such correspondences. I have observed this with my students, who with L1 or L2 Dutch have to learn L4 German. German and Dutch are closely related languages, but German had a second sound shift, which distanced it from all other Germanic languages. Some learners need help and attention to discover the relationship of cognate words, while others see it automatically. To see the correspondences seems to be easier when the related language is the L1. This might be due to the prominent role of L1 in developing a mental lexicon. Whatever the case may be, “recent studies have shown that language learners find it easier to acquire cognate translations” (Friel and Kennison 2001: 249). New graphemes and their pronunciation (sieve 4) have to be acquired, also to facilitate recognition (e.g. in English book, German Buch and Dutch boek the vowel is rendered with three different graphemes, but the pronunciation is quite similar). Moreover, as mentioned above, pronunciation is required even if the learning goal is reading comprehension only, both for rehearsal in working memory and memorizing, as for possible future acquisition of other skills. Because of the importance of morphemes as a probable representational level in the mental lexicon, emphasis on affixes (sieve 7) is useful for learning. Moreover, linguistic knowledge can be applied more flexibly when some insight into derivation and compounding has been acquired. For syntactic analysis (sieve 5, main sentence structures), new routines for syntactic processing only have to be developed in the case of new structures. German or Dutch speakers will not need new routines for processing SVOstructures, but English speakers have to develop new routines to cope with non-SVOstructures when they are first encountered in new languages. At the level of syntax there is more typological distance between, on the one hand, German and Dutch, with their very flexible word order, and English and the Scandinavian languages, with their


fixed word order, on the other, than between the latter group and French. Typological distance between languages can differ according to the processing level. To acquire a new language means especially to acquire new forms or perhaps only partially new forms as it is often the case in related languages. For the conceptual level, shared representations seem most plausible (cf. Zeelenberg & Pecher 2003, Francis 1999: 322f.). In semantic memory, concepts and the networks they belong to may have to be altered slightly or to be developed, but the learning load is mainly due to the form level. Differences in thought traditions and in knowledge systems may occur, but this can also be an intralinguistic phenomenon; it is not typical of foreign-language acquisition. Different cultures can share a language and different thought patterns can occur inside a linguistic community depending on factors such as social group or scientific community. The seven sieves use focus on form for all decoding levels and thus support the acquisition of form-based processing levels — often using inferring for the recognition of possible transfer bases (Meißner 2003: 36). This implies that both datadriven and expectancy-based processes are used.

References Ameel, E. , Storms, G. , Malt, B. C. and Sloman, S. A. 2005. How bilinguals solve the naming problem. Journal of Memory and Language 53: 60–80. Ashby, F. G. , Valentin, V. V. and Turken, A. U. 2002. The effects of positive affect and arousal on working memory and executive attention. In Emotional Cognition, S. Moore and M. Oaksford (eds), 245–87. Amsterdam: John Benjamins. Balota, D. A. , Paul, S. T. and Spieler, D. H. 1999. Attentional control of lexical processing pathways during word recognition and reading. In Language Processing, S. Garrod and M. Pickering (eds), 15–57. Hove: Psychology Press. Bates, E. and Goodman, J. C. 1997. On the inseparability of grammar and the lexicon: Evidence from acquisition, aphasia and real-time processing. Language and Cognitive Processes 1: 507–84. Bates, E. and MacWhinney, B. 1989. Functionalism and the competition model. In The CrossLinguistic Study of Sentence Processing, B. MacWhinney and E. Bates (eds), 3–73. Cambridge: Cambridge University Press. Bernhardt, E. B. 1993. Reading Development in a Second Language: Theoretical, empirical and classroom perpectives. Norwood NJ: Ablex. Bertram, R. , Baayen, R. H. and Schreuder, R. 2000a. Effects of family size for complex words. Journal of Memory and Language 42: 390–405. Bertram, R. , Schreuder, R. and Baayen, R. H. 2000b. The balance of storage and computation in morphological processing: The role of word formation type, affixal homonymy, and productivity. Journal of Experimental Psychology: Learning, Memory, and Cognition 26: 489– 511. Bornkessel, I. , Schlesewsky, M. and Friederici, A. D. 2003. Eliciting thematic reanalysis effects: The role of syntax-independent information during parsing. Language and Cognitive Processes 18: 269–98.


Borsley, R. D. 1997. Syntax-Theorie. Ein zusammengefaßter Zugang. Deutsche Bearbeitung von Peter Suchsland. Tübingen: Max Niemeyer. Bowers, J. S. , Davis, C. J. and Hanley, D. A. 2005. Automatic semantic activation of embedded words: Is there a ‘hat’ in ‘that’? Journal of Memory and Language 52: 131–43. Chen, H.-C. , Cheung, H. and Lan, S. 1997. Examining and reexamining the structure of Chinese– English bilingual memory. Psychological Research 60: 270–83. Clifton, C. 1995. Thematic roles in sentence parsing. In Reading and Language Processing, J. M. Henderson, M. Singer and F. Ferreira (eds), 94–118. Mahwah NJ: Lawrence Erlbaum. Costa, A. , Caramazza, A. and Sebastian-Galles, N. 2000. The cognate facilitation effect: Implications for models of lexical access. Journal of Experimental Psychology: Learning, Memory, and Cognition 26: 1283–1296. de Groot, A. 1993. Word type effects in bilingual processing tasks: Support for a mixed-representational system. In The Bilingual Lexicon, R. Schreuder and B.Weltens (eds), 27–51. Amsterdam: John Benjamins. Deutsch, A. and Rayner, K. 1999. Initial fixation location effects in reading Hebrew words. Language and Cognitive Processes 14: 393–421. Dijkstra, T. , van Heuven, W. J. B. and Grainger, J. 1998. Simulating cross-language competition with the Bilingual Interactive Activation model. Psychologica Belgica 38: 177–96. Dijkstra, T. , Grainger, J. and van Heuven, W. J. B. 1999. Recognition of cognates and interlingual homographs: The neglected role of phonology. Journal of Memory and Language 41: 496– 518. Dijkstra, T. , Timmermans, M. and Schriefers, H. 2000. On being blinded by your other language: Effects of task demands on interlingual homograph recognition. Journal of Memory and Language 42: 445–64. Ellis, R. and Humphreys, G. 1999. Connectionist Psychology. A Text with readings. Hove: Psychology Press. Feldman, L. B. 2000. Are morphological effects distinguishable from the effects of shared meaning and shared form? Journal of Experimental Psychology: Learning, Memory, and Cognition 26: 1431–1444. Fodor, J. A. 1983. The Modularity of Mind. Cambridge MA: MIT Press. Folk, J. R. 1999. Phonological codes are used to access the lexicon during silent reading. Journal of Experimental Psychology: Learning, Memory, and Cognition 25: 892–906. Franceschini, R. , Krick, C. M. , Behrent, S. and Reith, W. 2004. The neurobiology of code-switching. Inter-sentential code-switching in an fMRI-study. In Multilingual Communication, J. House and J. Rehbein (eds), 179–93. Amsterdam: John Benjamins. Francis, W. 1999. Analogical transfer of problem solutions within and between languages in Spanish–English bilinguals. Journal of Memory and Language 40: 301–29. Frenck-Mestre, C. and Pynte, J. 1997. Syntactic ambiguity resolution while reading in a second and native languages. The Quarterly Journal of Experimental Psychology 50(A): 119–48. Friel, B. M. and Kennison, S. M. 2001. Identifying German–English cognates, false cognates, and non-cognates: methodological issues and descriptive norms. Bilingualism: Language and Cognition 4: 249–74. Garrod, S. and Pickering, M. (eds). 1999. Language Processing. Hove: Psychology Press. Grainger, J. and Frenck-Mestre, C. 1998. Masked priming by translation equivalents in proficient bilinguals. Language and Cognitive Processes 13: 601–23. Green, D. W. 1998. Bilingualism and thought. Psychologica Belgica 38: 251–76.


Haegeman, L. 1991. Introduction to Government & Binding Theory. Oxford: Blackwell. Harley, T. 2001. The Psychology of Language. Hove: Psychology Press. House, J. and Rehbein, J. (eds). 2004. Multilingual Communication. Amsterdam: John Benja mins. House, J. and Rehbein, J. 2004. What is ‘multilingual communication’? In Multilingual Communication, J. House and J. Rehbein (eds), 1–17. Amsterdam: John Benjamins. Jescheniak, J. D. and Schriefers, H. 2001. Priming effects from phonologically related distractors in picture-word interference. The Quarterly Journal of Experimental Psychology 54(A): 371–82. Jescheniak, J. D. , Schriefers, H. and Hantsch, A. 2001. Semantic and phonological activation in noun and pronoun production. Journal of Experimental Psychology: Learning, Memory, and Cognition 27: 1058–1078. Jiang, N. and Forster, K. 2001. Cross-language priming asymmetries in lexical decision and episodic recognition. Journal of Memory and Language 44: 32–51. Johnson-Laird, P. N. 1983. Mental Models: Towards a cognitive science of language, inference and consciousness. Cambridge: Cambridge University Press. Kilborn, K. and Cooreman, A. 1987. Sentence interpretation strategies in adult Dutch–English bilinguals. Applied Psycholinguistics 8: 415–31. Kim, J. and Davis, C. 2002. Using Korean to investigate phonological priming effects without the influence of orthography. Language and Cognitive Processes 17: 569–91. Kintsch, W. 1988. The role of knowledge in discourse comprehension: A construction-integration model. Psychological Review 95: 163–82. Klein, H. G. and Stegmann, T. D. 2000. EuroComRom — Die sieben Siebe: Romanische Sprachen sofort lesen können. Aachen: Shaker Verlag. Klin, C. M. , Guzman, A. E. , Weingartner, K. M. and Ralano, A. S. 2006. When anaphor resolution fails: Partial encoding of anaphoric inferences. Journal of Memory and Language 54: 131–43. Koda, K. 1993. Transferred L1 strategies and L2 syntactic structure in L2 sentence comprehension. The Modern Language Journal 77: 490–500. Konieczny, L. , B. Hemforth, C. Scheepers and Strube, G. 1997. The role of lexical heads in parsing: Evidence from German. Language and Cognitive Processes 12: 307–48. Kutas, M. 1993. In the company of other words: Electrophysiological evidence for single-word and sentence context effects. Language and Cognitive Processes 8: 533–72. Lemhöfer, K. , Dijkstra, T. and Michel, M. C. 2004. Three languages, one ECHO: Cognate effects in trilingual word recognition. Language and Cognitive Processes 19: 585–611. Levy, B. A. 1993. Fluent reading: An implicit indicator of reading skill development. In Implicit Memory. New directions in cognition, development, and neuropsychology, P. Graf and M. E. J. Masson (eds), 49–73. Hillsdale NJ: Lawrence Erlbaum. Li, P. , Bates, E. and MacWhinney, B. 1993. Processing a language without inflections: A reaction time study of sentence interpretation. Journal of Memory and Language 32: 169–92. Liu, H. , Bates, E. and Li, P. 1992. Sentence interpretation in bilingual speakers of English and Chinese. Applied Psycholinguistics 13: 451–84. Longtin, C.-M. and Meunier, F. 2005. Morphological decomposition in early visual word processing. Journal of Memory and Language 53: 26–41. Longtin, C.-M. , Segui, J. and Hallé, P. A. 2003. Morphological priming without morphological relationship. Language and Cognitive Processes 18: 313–34.


Lutjeharms, M. 1988. Lesen in der Fremdsprache. Versuch einer psycholinguistischen Deutung am Beispiel Deutsch als Fremdsprache. Bochum: AKS-Verlag. Lutjeharms, M. 1994. Lesen in der Fremdsprache: Zum Leseprozess und zum Einsatz der Lesefertigkeit im Fremdsprachenunterricht. Zeitschrift für Fremdsprachenforschung 5(2): 36–77. Lutjeharms, M. 1997. Reading skills for weak learners. In Languages for Work and Life: The Council of Europe and vocational oriented language learning, G. Egloff and A. Fitzpatrick (eds), 164–77. Strasbourg: Council of Europe Publishing. Lutjeharms, M. 1998. Die syntaktische Verarbeitung bei der Rezeption von Sprache. In Betrach tungen zum Wort. Lexik im Spannungsfeld von Syntax, Semantik und Pragmatik, E. Klein and S. J. Schierholz (eds), 117–51. Tübingen: Stauffenburg. Lutjeharms, M. 2001. Reading comprehension and the C-test from a psycholinguistic viewpoint: Aspects of the lower processing levels. In Tests and Translation. Papers in memory of Christine Klein-Braley, H. Pürschel and U. Raatz (eds), 21–37. Bochum: AKS-Verlag. Lutjeharms, M. 2004. Der Zugriff auf das mentale Lexikon und der Wortschatzerwerb in der Fremdsprache. Fremdsprachen Lehren und Lernen 33: 10–26. MacWhinney, B. and Bates, E. (eds) 1989. The Cross-linguistic Study of Sentence Processing. Cambridge: Cambridge University Press. Malt, B. C. , Sloman, S. A. and Gennari, S. P. 2003. Universality and language specificity in object naming. Journal of Memory and Language 49: 20–42. Marslen-Wilson, W. D. 1999. Abstractness and combination: The morphemic lexicon. In Lan guage Processing, S. Garrod and M. J. Pickering (eds), 101–19. Hove: Psychology Press. Marslen-Wilson, W. D. 2001. Access to lexical representation: Cross-linguistic issues. Language and Cognitive Processes 16: 699–708. McCann, W. J. , Klein, H. G. and Stegmann, T. D. 2003. EuroComRom — The Seven Sieves: How to read all the Romance languages right away. Aachen: Shaker Verlag. McClelland, J. L. and Rumelhart, D. E. 1985. Distributed memory and the representation of general and specific information. Journal of Experimental Psychology: General 14: 159–88 (quoted after the reprint in R. Ellis and G. Hymphreys 1999, 75–105). McDonald, J. L. and Heilenman L. K. 1991. Determinants of cue strength in adult first and second language speakers of French. Applied Psycholinguistics 12: 313–48. McDonald, J. L. and MacWhinney, B. 1989. Maximum likelyhood models for sentence processing. In The Cross-linguistic Study of Sentence Processing, B. MacWhinney and E. Bates (eds), 397–421. Cambridge: Cambridge University Press. McDonald, J. L. and MacWhinney, B. 1991. Levels of learning: A comparison of concept formation and language acquisition. Journal of Memory and Language 30: 407–30. McQueen, J. M. and Cutler, A. 2001. Spoken word access processes: An introduction. Language and Cognitive Processes 16: 469–90 (Special issue spoken word access processes). Meißner, F.-J. 2003. EuroComDidact: Learning and teaching plurilingual comprehension. In Sprachkompetenz — Mehrsprachigkeit — Translation. Akten des 35. Linguistischen Kolloqui ums, L. N. Zybatow (ed.), 36–46. Tübingen: Gunter Narr. Melinger, A. 2003. Morphological structure in the lexical representation of prefixed words: Evidence from speech errors. Language and Cognitive Processes 18: 335–62. Morton, J. 1969. The interaction of information in word recognition. Psychological Review 76: 165–78.


Moscoso del Prado Martin, F. , Deutsch, A. , Frost, R. , Schreuder, R. , De Jong, N. H. and Baayen, R. H. 2005. Changing places: A cross-language perspective on frequency and family size in Dutch and Hebrew. Journal of Memory and Language 53: 496–512. Moscovitch, M. and Bentin, S. 1993. The fate of repetition effects when recognition approaches chance. Journal of Experimental Psychology: Learning, Memory, and Cognition 19: 148–58. Müller-Lancé, J. 2002. La corrélation entre la resemblance morphologique des mots et la probabilité du transfer interlinguistique. In Eurocom. Mehrsprachiges Europa durch Interkom prehension in Sprachfamilien. Tagungsband des internationalen Fachkongresses im Euro päischen Jahr der Sprachen 2001, Hagen, 9.-10. November 2001, G. Kischel (ed.), 141–59. Hagen: Fernuniversität. Oakhill, J. V. , Cain, K. and Bryant, P. E. 2003. The dissociation of word reading and text comprehension: Evidence from component skills. Language and Cognitive Processes 18: 443–68. Paradis, M. 2004. A Neurolinguistic Theory of Bilingualism. [Studies in Bilingualism 18]. Amster dam: John Benjamins. Peterson, R. R. , Burgess, C. , Dell, G. S. and Eberhard, K. M. 2001. Dissociation between syntactic and semantic processing during idiom comprehension. Journal of Experimental Psychology: Learning, Memory, and Cognition 27: 1223–37. Rawson, K. A. 2004. Exploring automaticity in text processing: Syntactic ambiguity as a test case. Cognitive Psychology 49: 333–69. Roelofs, A. 2003. Shared phonological encoding processes and representation of languages in bilingual speakers. Language and Cognitive Processes 18: 175–204. Sanford, A. J. 1999. Word meaning and discourse processing: A tutorial review. In Language Processing, S. Garrod and M. Pickering (eds), 301–34. Hove: Psychology Press. Schriefers, H. , Friederici, A. D. and Kühn, K. 1995. The processing of locally ambiguous relative clauses in German. Journal of Memory and Language 34: 499–520. Shapiro, L. P. , Nagel, H. N. and Levine, B. A. 1993. Preferences for a verb’s complements and their use in sentence processing. Journal of Memory and Language 32: 96–114. Snedeker, J. and Trueswell, J. C. 2004. The development constraints on parsing decisions: The role of lexical-biases and referential scenes in child and adult sentence processing. Cognitive Psychology 49: 238–99. Taft, M. 1991. Reading and the Mental Lexicon. Hillsdale NJ: Lawrence Erlbaum. Thomas, M. S. C. and Allport, A. 2000. Language switching costs in bilingual visual word recognition. Journal of Memory and Language 43: 44–66. Traxler, M. J. , Williams, R. S. , Blozis, S. A. and Morris, R. K. 2005. Working memory, animacy, and verb class in the processing of relative clauses. Journal of Memory and Language 53: 204–24. van Dijk, T. A. 1980. Macrostructures. Hillsdale NJ: Lawrence Erlbaum. Zeelenberg, R. and Pecher, D. 2003. Evidence for long-term cross-language repetition priming in conceptual implicit memory tasks. Journal of Memory and Language 49: 80–94. Zwitserlood, P. 1995. Prozesse und lexikalische Repräsentation bei der visuellen Worterkennung. In Sprache: Verstehen und Verständlichkeit. Kongreßbeiträge zur 25. Jahrestagung der Gesellschaft für Angewandte Linguistik e.V. , B. Spillner (ed.), 115–18. Frankfurt a.M.: Peter Lang.

chapter 13

A computer-based exploration of the lexical possibilities of intercomprehension Finding German cognates of Dutch words* Robert Möller Université de Liège

Group relations between languages, especially cognate words, provide an excellent opportunity to develop receptive competence (“intercomprehension”). This paper presents a computerized approach to the investigation of the extent of Dutch–German cognates in Dutch and the difficulties a German reader might have in recognizing them. The main procedure used to finding the most similar German counterparts of Dutch words is based on the Levenshtein algorithm, in which findings and assumptions on general similarity perception have been integrated along with a set of statistically important sound correspondences. Results show that about 75% of the tested frequency list of 5,000 Dutch word forms can be decoded with the help of German. Keywords: Dutch, German, cognates, Levenshtein algorithm, intercomprehension

1. The role of cognates in reading a closely related language The attitude of Germans towards written Dutch primarily falls — roughly speaking — into two categories: a. “Dutch is a foreign language which I haven’t learnt and consequently cannot read.” b. “Dutch is so similar to German that it is easy to understand even if you have never learnt it.” The difference is, of course, partly a matter of optimism and self-confidence in dealing with foreign languages. But if you explain the systematic relationships between German and Dutch words to German students and let them try to decode small texts in Dutch, you can often observe an “aha!” reaction which shifts the attitude from a) to b). There seems to be a threshold which can (or could) be surmounted rather easily (at least by persons with experience and knowledge of languages, such as students of philology), but which nonetheless blocks spontaneous access so that many people will not attempt to overcome it (cf. also Ház 2005: 38, 110). In addition, customary methods of language learning normally do not concentrate on receptive competence; this is especially true for “small” languages. The market is dominated by a series of learning

286 Robert Möller

aids and courses which in fact suggest that learning Dutch in 30 lessons and learning Chinese in 30 lessons involves the same processes and effort invested. Dutch and German certainly cannot be regarded as easily mutually intelligible. But taking only written language into consideration and leaving aside the problems of pronunciation, we find that syntactical differences are unimportant and morphological differences asymmetrical: it is often difficult to decode a German sentence without being familiar with the German inflectional system, while the more analytic Dutch structures in general also exist in German, mostly as less formal alternatives. The only major obstacles in decoding Dutch with help from German (as a mother tongue or as a well-mastered foreign language) are the lexical differences,1 more precisely: real lexical differences on the one hand and phonological differences between cognates (which hinder recognition) on the other.

2. The project In order to assess the actual elevation of the recognition threshold, the most important things to know are the number of cognates in the lexicon, the possibilities for recognizing them, and whether many are misleading. The project that will be presented here aims to investigate the possibilities for Dutch–German reading comprehension by exploring the number of recognizable cognates and identifying the necessary knowledge and the risks of this approach with a computer program. Beyond the general interest in the synchronic relationship between the languages, the investigation also has a practical benefit: The results and the computer procedures are to be used for developing learning aids for intercomprehension in which cognates and non-cognates are treated in a different way and regular and non-regular sound correspondences are indicated in each case. Certainly, the recognition of isolated cognates is a far cry from real text understanding (see Lutjeharms 1988, 2004), but it defines the minimum that can be understood (and the maximum that could be misunderstood). Decoding strategies, such as inferencing from textual and general context, which play a major role in real foreign language comprehension, rely on familiar elements as a starting point. In reading an unknown language, the only familiar elements are names, international words and cognates. If a sufficient number of these is lacking, and inferencing can therefore only be based upon a few clues, the general context may perhaps help the reader to guess that a text contains certain information he or she already knows; if the context is restricted, the reader may even guess which of several possible statements is being made (X is elected president or Y is) — but a transmission of more extensive new information will generally not take place. The textual context, in turn, consists of words which are either decodable as cognates or need a helping textual context as well. Thus, the number as well as the transparency of cognates are crucial for intelligibility (see also

Finding German cognates of Dutch words 287

van Bezooijen and Gooskens this volume). In real text reading, the context may help to bridge a certain number of lexical gaps, but this depends very much on chance and on the individual text, whereas the starting point, the number of potential gaps and bridge piers, can be determined for a given constellation of languages. As mentioned earlier, this does not mean that all individual readers make use of existing possibilities in the same way. If the number of cognates is invariable, their recognition depends very much on the reader’s proficiency in detecting cognates in cases where the phonological relation is not so obvious (and in making use of them in cases where they are not semantically identical, cf. below). This is a special skill that can (and must) be trained (cf. Lutjeharms 2004: 78). The question of under which conditions the relationship between phonologically/orthographically different cognates is easily recognizable has been to date little investigated,2 although results from research in reading comprehension, acquisition of reading and writing and (receptive) language processing3 provide some hints (cf. Lutjeharms 1988: 249, Scheerer-Neumann 1996, Schade and Barattelli 2003). It appears that for word recognition, the beginning of a word is most important. Moreover, consonants seem to be more salient than vowels, as can be seen from consonant “skeleton writings” used by young children (cf. also the different role of consonants and vowels in the ablaut system). Additionally, the amount of (accumulated) differences in a word, in relation to its length, is likely to be important. This could explain why cognates between German and other Germanic languages are often not recognized: due to the Old High Germanic (OHG) Consonant Shift, consonants in particular are not the same. This is often the case in the initial position, and even for the closely related German and Dutch it is not unusual to find pairs of cognates in which nearly all segments differ, although according to general correspondence regularities (Du. duik-en–G. tauch-en ‘to dive’). Without some knowledge of Dutch–German sound correspondences, such words can be decoded only with difficulty. Thus, a catalogue of relevant rules of sound correspondences could be a valuable aid for readers of German. One of the questions that will be pursued here also is how to find the most relevant instances from the standpoint of intercomprehension. The linguistic preoccupation with sound correspondences in related languages has nearly always been dominated by a historical perspective.4 Referring to a historical system such as (reconstructed) Germanic makes it possible to formulate unidirectional phonological “laws”. The sound correspondences between related systems, however, can rarely be formulated into laws, since relationships are generally obscured by different sound mergers on both sides. Things are further complicated by loan words, which in historical phonology are simply excluded or treated separately, but synchronically often cannot be recognized as such. However, for receptive competence, the unambiguity of phonological correspondences is not imperative so long as assumed correspondences do not lead to incorrect assignments.5 Moreover, highly differentiated systems of rules and conditions of sound context are in any case not practicable as an easy access route to intercomprehension. Thus it is important to find those rules of

288 Robert Möller

sound correspondence which are truely necessary for decoding and to test the extent to which the remaining ambiguities may result in misunderstandings. Finding etymological equivalents in one’s own language does not, however, guarantee correct decoding of unknown words. The fear of false cognates (“false friends”) is in fact very often mentioned by Germans as a main reason for distrusting an intercomprehensive approach to Dutch. In language production, this poses a certain problem (cf. Arntz 1997: 170) — even if the real risk is smaller than many people fear, since in most cases, the context helps to avoid misunderstandings. In addition, it is far too simplistic to differentiate only between “true” and “false” friends. In Section 6, a further differentiation, which focuses particularly on reception, will be proposed. This differentiation cannot be done automatically. For the rest, the particular advantage of using a computer program is that it differs from human intelligence in that it does not know anything it has not learnt, but never forgets. In that respect, it is quite different from real human behaviour; however, if the aim is not to test what humans do, but rather what can be attained in relation to a particular input, a program will provide much more precise answers. Furthermore, it permits testing the outcome under different conditions — without secretly extending its skills during the test.

3. Which “previous knowledge” is necessary? 3.1 The main sound correspondences and their statistical importance The first necessary step is to identify the relevant sound correspondences. From a diachronic point of view, the relevant laws of sound change are well-known. Here, however, it is important to assess the usefulness of rules which have to be formulated on a strictly synchronic basis, referring only to the results of these changes in modern Dutch. First of all, the usefulness depends on how many words are affected by a certain rule. To establish this, a list of 2490 cognates was examined by a program that detects and records the differences. (This list is an extract of the bilingual word list which is also used later.6) It contains only one instance (preferably a simple form) of each (indigenous) lexical morpheme, because here the question is the frequency of certain correspondences in different pairs of cognate morphemes and not the frequency of certain cognate morphemes in different words. In contrast to the procedure presented below, which searches the cognates themselves as well, for this first step the pairs of cognates are given. No distinction is made between true cognates and misleading ones, since only phonologic relationships are of interest. The central procedure, however, is the same: The Levenshtein algorithm (cf. Heeringa 2004) is used to determine the operations (substitution, deletion, insertion of segments) that are necessary to transform the Dutch word into the German one. Prior to this, the differences caused by the respective orthographical conventions are harmonized (a potential reader is supposed to know, for example, that the Dutch pronunciation of is [z], not [ts] as in German,

f(f) → pf offer–Opfer f → b (af–ab) (ff → b gaffel–Gabel) v → b geven–geben w→b erwt–Erbse

p → f(f) slapen–schlafen p(p) → pf pijp–Pfeife p(p) → b knaap–Knabe b(b) → p(p) rib–Rippe f → p(p) hoofd–Haupt

s, z → r bes–Beere vriezen– frieren

57

3

40 3

2

2

10

2

56

t → s(s), ß eten–essen t(t) → (t)z tijd–Zeit t→d slenteren–schlendern d→t dag–Tag s(s) → (t)z pels–Pelz, klos–Klotz s(s), z → sch zwan–Schwan (s)s → chs vos–Fuchs

51

3 ch[t] → f[t] kracht–Kraft (-achtig–-haft)

147 j → g jicht–Gicht 9

5 k→g klinken–klingen 178 g(g) → (c)k brug–Brücke 35

75 (c)k → ch maken–machen 98

Table 1. Sound correspondences in Dutch-German cognates (consonants)

-el → -er wandelen–wandern 7 -er → -en / -el ijzer–Eisen, donker- dunkel

3

11 [Vok]r → r[Vok] vorst–Frost

109 [l,n]d → [l,n]0 kelder–Keller [Vok]0[Vok] → d,t leer–Leder 4

2/1

3

4

14

8

ie → e(e) ziel - Seele ie→ ü,eu liegen - lügen vriend - Freund i(e) → ei kiem - Keim

i → u,ü kil - kühl slikken - schlucken [n. verw.] u→i hulp - Hilfe ie → ei

ee → ei been - Bein e → ei emmer - Eimer ei → e weinig - wenig i→e blik - Blech e(e) → i berk - Birke

oo → au ook - auch o → au stof - Staub

4

u → o, ö kurk - Kork

o(o) → eu loochenen - leugnen o → u,ü worst - Wurst oo → u noot - Nuss

5 7 7 2 6

uu → au,eu vuur - Feuer

e→ö lepel - Löffel o→e worden - werden oe → au kroes - kraus oe → o,ö roest - rost

79

13

2

2

45

Table 2. Sound correspondences in Dutch-German cognates (vowels)

38

32

e → a fles - Flasche 10

a(a) → e aarde - Erde

a(a) → o halen - holen 10

6

458

2

o(o) → a oksel - Achsel

u → a snurken - schnarchen

0 → e [ɵ] vlag - Flagge e [ɵ] → 0 woede - Wut [gb] e [ɵ] [+ln] → 0 geloven - glauben

26

12

ui → au, äu, eu buik — Bauch

ij → i(e) krijgen - kriegen

eu → a, e, ie neus -> Nase

8

73

21

3

6

ui → ie sluiten - schließen

23

eu → ü vleugel - Flügel eu → u breuk - bruch

10

3

2

11

ui → u,ü duin - Düne kruim - Krume ui → ö fluit - Flöte

3

2

4

ui → ei sluier - Schleier

13

12

15

104

8

ou[d,t] → al,ol,ul [d,t,t] zout - Salz

13


and the program knows also that postvocalic in German only indicates that the vowel is long and that <sch>and in both languages indicate single segments, not clusters). In the comparison, diphthongs and affricates are treated as single segments. Whereas the Levenshtein algorithm will later be used to calculate the overall difference between the compared forms (“Levenshtein distance”), here it serves to detect and register the individual differences, i.e. the corresponding sounds, and generate a sorted list of them. The result of this procedure was that, of all 2490 cognates, 542 (22%) are (graphically) identical in both languages, a further 61 differ only according to different orthographical conventions, and the others exhibit the relationships listed in Tables 1 and 2. The difference between regular relationships and exceptional relationships due to detours such as borrowing or prevailing of dialectal forms is rather obvious, but nevertheless the overall impression becomes a little fuzzy.

3.2 Correspondence and similarity Not all these correspondences have to be learnt as rules. Not only will a great number of rules discourage people from attempting to read Dutch with the help of German, but more importantly, many of these differences would probably never block the decoding of a word because corresponding elements are so similar. So long as an empirical investigation remains to be undertaken, one has to rely on theoretical considerations concerning similarity here. In this case, similarity means both orthographic similarity and phonetic similarity: the relationship between the Dutch and German (meloen–Melone) is obvious graphically and will dominate the knowledge of the (different) pronunciation. For goed–gut it is the other way around. The question of perceptual similarity of sounds is still being discussed, for example, in foreign accent research or in measuring dialect differences (cf. Archibald 1998, Heeringa 2004). For written language, the situation is once again different. In any case, partial identity in phonological and/or in graphical features is probably not the only factor; the connections between sounds in the L1 also play a role in the recognition of correspondences. German speakers, for example, are familiar with the neutralization of the opposition between voiced and unvoiced obstruents from the regular process of final devoicing in German. As well, in many German regions the phonetic distinction between the Standard German voiced and unvoiced plosives is very reduced, so that people are accustomed to phenomena such as the correpondence Dutch baviaan–German Pavian from their familiarity with regional and stylistic variation as well. In the same way, the most frequent difference by far between Dutch and German cognates, the absence of German final -e [ə] in Dutch (bloem–Blume), should not impede recognition for several reasons: it is not phonetically salient (unstressed syllable, reduced vowel), common in German allomorphy (Wolle–Woll|decke), and omission of final schwa is normal in most regional German varieties as well as in supraregional allegro speech. The correspondence Dutch <s>/[s]–German <sch>/[ʃ], in particular

292 Robert Möller

before consonants (snoer–Schnur), is supported by phonetic similarity and also by the fact that German orthography only partly reflects the (Early New High German) change s > sch, so that <s> before and
regularly has to be pronounced as [ʃ]. Moreover, as a consequence of this change, the combinations sn, sl, sm, sw are excluded in indigenous German words, and thus loan words have been assimilated to this regularity (in Schlips ‘tie’ < Engl. slip(s), this assimilation is even reflected by orthography). Taking these considerations into account, the number of necessary correspondence rules which should be given to a human reader can be limited to those which not only concern a certain number of words/morphemes, but also go beyond the correspondences covered by similarity. One could argue that in investigating the possibilities for intercomprehension between two languages, correspondence rules should be excluded. Ház (2005) and Wenzel (in press) test spontaneous reading comprehension and do not present even basic information about Dutch orthography to their subjects. From a technical point of view, correspondence rules are not necessary to make a program find similar words: the Levenshtein algorithm (see 3.3) is able to determine the degree of difference for any given pair of strings that differ in a regular or irregular way. Even if similarity relations are not taken into consideration (cf. van Bezooijen and Gooskens this volume), the results of this procedure are mostly close to human impressions of difference. Including similarity in the calculation (see 4) makes them rather realistic. But if possibilities for intercomprehension can be considerably improved by applying a small set of correspondence rules, this little effort seems reasonable, and it is just as interesting to find out what can be reached in the end as it is to test what is possible without any previous knowledge. Thus, both similarity and correspondence rules are integrated in the program presented here, but it is important to keep a fundamental difference in mind: which correspondence rules are given is a variable decision and the rules depend on the relation between the languages, while which similarities are perceived without such help is an invariable starting point that only depends on phonetic and perceptual universals and probably on the native language of the readers. In order to explore the lexical possibilities of intercomprehension between two other languages (e.g. Dutch and Swedish or Portuguese and Italian), both sets of rules must be replaced (and the spelling adaptation as well — the rest of the program can remain as is); if, however, the potential readers are Germans, only the set of correspondence rules must be changed.

3.3 Correspondence rules The most relevant of the obscure sound correspondences are those which go back to the OHG Consonant Shift: they play a statistically important role, are not covered by (strong) similarity and concern consonants, which seem to be more “recognition-sensitive” than vowels. The disagreeable thing about the OHG Consonant Shift is the different outcome according to position and sound context:


– t, p 7 in initial position: G. z (tijd–Zeit), pf (pijl–Pfeil) – t before r, p and t in combination with fricatives: no difference (trap–Treppe, steen– Stein) – t, p intervocalic: G. ss/ß (laten–lassen, heten–heißen), f (slapen–schlafen, open–offen) – tt, pp intervocalic: G. tz (ketter–Ketzer), pf (appel–Apfel) – t, p postconsonantic: G. z (hart–Herz), pf, f (damp–Dampf, helpen–helfen) – final, postvocalic t, p: G. s,ß (dat–das, daß), f (op–auf) In the tradition of historical philology, the OHG Consonant Shift is usually presented in this way, but is it necessary, or at least helpful, to make all these differentiations to facilitate access to written Dutch, or only some of them, or none? (Even if [ts] and [s] are similar, the question remains, because the combination correspondence rule + similarity would require double effort.) The answer is easy in cases where there is no change: even knowing a correspondence rule such as t→ z, a reader will always find a German cognate with t instead of z (otherwise all loan words from Latin/Greek, French or English would be unrecognizable, too), because identity makes a stronger immediate impression than correspondence via a well-known rule (cf. also Lutjeharms 1997: 157). For the rest, things are less evident. In general, the simplest rule should be the preferable one. Here, this would be: t→(t)z, ss/ß, s. The benefit of deterrent differentiations according to position and surrounding depends on the number of incorrect assignments they help to prevent. Even if a particular correspondence (as part of a more generally formulated correspondence rule) is etymologically completely false, here it would do no harm if the respective “cognates” do not exist in the lexicon. Therefore it is useful to let the computer test different formulations of these correspondence rules, A. B. C. D.

not differentiating at all differentiating between initial position and others differentiating between initial position, gemination and others differentiating between all positions as above — but with the full range of possibilities in the final position, because in modern Dutch former geminates also appear here as simple consonants (kat < Middle Dutch catte) and it is not always possible to rely on vowel length as a clue.

In turn, the program is given these different versions of the OHG Consonant Shift correspondence rules (together with the other correspondence rules and similarity assumptions) and has to find possible cognates for every item concerned in the Dutch list in an extensive list of German words8 (including inflected forms). The different correspondence rules are taken into account in the procedure, which compares the Dutch word with similar German forms and determines the degree of difference. To accomplish this, the Levenshtein algorithm is used. The Levenshtein algorithm sums up the “costs” of the most straightforward way to transform string A into string B by

294 Robert Möller

substituting, inserting or eliding single segments.9 The “costs” that are caused by these operations (whereas identity of segments is “free”) can now be differentiated: a substitution which reflects a relationship given as a correspondence rule or as a similarity relationship is “cheaper” than other substitutions. Here, “costs” vary between 0.1 for correspondences, 0.2–0.8 for different degrees of similarity, 0.9 resp. 1 for “normal” substitutions/insertions/deletions of vowels resp. consonants and 2 for substitutions of consonants by vowels or the other way round. To date, these values have been fixed more or less ad hoc according to similarity assessments (even a more systematic application of phonological feature systems apparently does not automatically agree with perception, cf. Heeringa 2004: 185). The distribution of correct and incorrect assignments (from an etymological viewpoint; the difference between “good” and “misleading” cognates does not depend on the form of correspondence rules) for the different versions A–D of the OHG Consonant Shift rules now results in the following (see Table 3). It is evident that the results of the versions C and D are hardly better than those of B. Thus, these differentiations do not make much sense. Between B and A, however, the difference is somewhat more important. (Certainly the majority of the incorrect assignments for A are German words with a voiced s which could be generally excluded here, but if the reader is instructed to replace with , this difference in pronunciation could easily be overlooked.) It seems useful to choose B, which only Table 3. Assignments resulting from different rule versions rule version

correct assignments

incorrect assignments, same word class

incorrect assignments, different word class

A

279

B C D

277* 276 276

62 (e.g. tand → Sand, pers → Ferse) 42 (e.g. heten → hetzen) 38 38 (e.g. hert → Herz, krant → Kranz, krijt → Kreis, hout → Haus, griep → Griff, koper → Koffer, doop → Topf)

61 (e.g. tegen → Segen, pijn → fein) 39 36 36 (e.g. taart → zart , biet → bis, grot → groß)

(The results for t and p are parallel; moreover it would not make sense to obscure the systematic parallel by giving different formulations for t and p.) * If the etymologically less precise versions B and C do not only lessen the number of incorrect assignments, but also slightly the number of “right” ones, this is due to some “cognates” that are actually not real cognates, but nevertheless correspond in form and meaning. For plek–Fleck and pluis –Fluse the background of this phenomenon is not clear, perhaps it has to do with concurrence between an inherited and a borrowed form. For verwijten–verweisen a fusion of different etyma has taken place in German (Middle High German still verwîzen with unvoiced s).


differentiates between the particular prominent initial position and all other positions, but even A would not lead to too disastrous a result. Another case in which testing the outcome can help determine the usefulness of different rules is Dutch ui. This mostly represents a former Germanic *û which in German has become au (äu, eu in case of umlaut), or Germanic *eu (G. äu, eu as well). As can be seen in Table 1, this is reflected by the majority of cognates. However, there are also German cognates with u and ü, o and ö, ei and ie corresponding to the Dutch ui. The correspondence between the Dutch and German / (which goes back to borrowing from Dutch or Low German in German) is covered by the graphemic similarity. The cases with the German , <ei>are very rare and not systematic.10 Those with , however, are evidenced in a certain number of verbs with the same development (sluiten–schließen).11 Thus, the rule –produces 27 (14) incorrect assignments as opposed to six correct ones, and testing the result without this rule shows that even if, in that case, some more incorrect assignments appear as the most similar forms, the correct ones are still the second best — so it is preferable to give only one rule for : ui→ au, äu, eu. The program is now given the rules and similarities in Table 3 as basic knowledge. For very frequent correspondences which often appear in salient positions (such as d → t), rules are given despite similarity correspondence. If in further empirical research it turns out that a reader’s intuitions about similarity differ from what has been Table 4. “costs” fixed for the calculation of Levenshtein distance Correspondence rules initial p → pf; t → z p → pf/f; t → (t)z/s(s)/ß; k → ch d→t v, f → b ou [+d/t] → al/ol [+ d/t/z]

0.1 0.1 0.1 0.1

a(a) ↔ e e(e) → i(e), o → u ee → ei; oo → au/äu ui → au, äu, eu; uu → au, äu, eu

0.1 0.1 0.1 0.1

Substitutions (in both directions) s–sch

0.2

long vowel–short vowel a–ä; o–ö; u–ü (umlaut) aai–ai/ä; ui–u/ü ei–e/i; ou/au–o/u

0.2 0.2 0.2 0.3

Consonants: voiced–unvoiced

0.4

e–ö; i–ü

0.4

Other consonant substitutions without 0.8 change of place of articulation

a–e; a–o; e–i; o- u

0.5

Other consonant substitutions

1.0

other vowel substitutions

0.9

Insertion/deletion Insertion/deletion of final e (schwa)

0.2

ld–ll, nd–nn (assimilations)

0.4

0.2

other deletions/insertions

1.0

Differences in connecting elements of compounds

296 Robert Möller

assumed here, the respective adaptations can easily be made. And if the lexical basis for intercomprehension between two other languages is to be explored, only this list must be replaced by another list — if the reader’s language is German, only the first part, the correspondence rules.

4. The program NL-D-KOG The job of the program NL-D-KOG12 is to find, for a given list of Dutch lemmas, the most similar German lemmas, to determine the difference between the two forms and whether or not they are cognates (taking into account the given correspondence rules and instructions about similarity), and finally to check if the/a real German equivalent is among the words that have been found and if other words have been found and determined as similar which are not equivalents but incorrect assignments or misleading cognates. The basis for this investigation are a “top 10,000” (frequency) list of Dutch words13 to which the German equivalents have been added using the already mentioned bilingual Dutch–German list, and the German–Dutch version of the latter, reduced to the German forms. All lists also contain inflected word forms, so it can be tested if the right assignments can be made without knowledge of Dutch morphology. The method of finding the possible German cognates (see Fig. 1) is structured to be both as efficient as possible and as close to real human cognate processing as possible (which actually is identical — only the latter’s flexibility is unattainable). A first selection from the German list is made according to the first segment of the Dutch word, also taking into account any correspondent or similar segments the program knows (cf. newer versions of the cohort model of lexical access, see Schade and Barattelli 2003: 85f., cf. also Dijkstra 2005: 1723). An important second step consists of focussing on a more abstract pattern of the consonantic structure of the word: This ‘structure pattern’ encodes the sequence of consonants and vowels as a sequence of numbers in which obstruents are grouped according to their place of articulation, nasals and liquids form their own groups and all vowels only appear as “V”. Sequences of two identical consonants, affricates and diphthongs appear as one single position, and endings (-en, -el) are left out (cf. also Lutjeharms 1997: 156). Affixes are generally not included, but automatically marked as such and replaced by the German equivalent.14 Thus we get as a structure pattern for Dutch duiken: |2V3| (dental–vowel–palatal/velar [ending left out]), and for German tauchen as well (similarly for German dick, taugen, Tag, zeigen, . . .), for Dutch pijp we get |1V1| as in German Pfeife (and Pfaffe, Fabel . . .). Whether or not this really mimics human cognate processing remains to be tested, but it does correlate to the saliency of consonants and other findings about similarity perception (see Archibald 1998: 102, cf. above). There are a few relationships between Dutch and German cognates with which this approach has difficulty coping: the results of the Dutch r-metathesis (Dutch borst– German Brust), the loss of the intervocalic d (weer–Wetter, leer–Leder), the Dutch

Figure 1. Finding German cognates


298 Robert Möller

change ft > cht (kracht–Kraft) and the vocalization of l in al/ol before d/t (oud–alt, goud–Gold). However, this may be realistic in so far as these relationships seem to be more difficult to recognize for a human reader, too. On the other hand, these phenomena can rather easily be integrated via correspondence rules which generate alternative structure patterns to be looked for. So the correspondence ou/al/ol (before d/t), which concerns a certain number of morphemes, is included in the list of rules, while the others are not. To speed up the procedure, the structure patterns are made up only for the Dutch words while the program is running, whereas the German list already contains them. Thus, the German forms to be compared with the Dutch word can be found very quickly. These are then compared with the Dutch lemma by calculating the Levenshtein distance modified by the correspondence rules (“cost” 0.1) and similarity assumptions (“cost” 0.2–0.9), e.g.: – duiken–tauchen: d → t (0.1), ui → au (0.1), k → ch (0.1) ⇒ sum 0.3 – duiken–taugen: d → t (0.1), ui → au (0.1), k → g (similar, 0.4) ⇒ sum 0.6 – duiken–decken: ui → e (0.9) ⇒ sum 0.9 – duiken–zeigen: d → z (both dentals, 0.8), ui -> ei (0.9), k → g (similar, 0.4) ⇒ sum 2.1 – duiken–dick: ui → i (0.9), e → 0 (0.6–if not in final position or in a context where e is systematically deleted in German as in -elen, -eren), n → 0 (1) ⇒ sum 2.5. This shows once more the different role of rules and similarity: the “natural” order by similarity would be 1. decken 2. taugen 3. tauchen (perhaps it could be argued that ui — as a grapheme and as a phoneme — is more similar to a diphthong than to a short vowel, but still taugen would be preferable to tauchen). With the additional knowledge introduced by the correspondence rules — reflected in the program by lower costs in spite of inferior similarity — the right cognate moves to the first place. For the human reader, the difference between spontaneous associations and rule-governed associations would be analogous, even if (as already mentioned) the quantification of similarity by the program is only based on ad hoc assumptions about the perception of similarity by German speakers. Compound and derived words cause additional complications. If they exist in parallel in both languages, in many cases only the different connecting element may cause problems (zonnebloem–Sonnenblume). As this also belongs to German allomorphy and is therefore easily recognizable, these differences do not “cost” much. If a direct counterpart does not exist, the question is if the morphemes can be recognized. Here, the problem of segmentation has to be resolved first, a problem human readers often have in a similar way: if the parts of the word are not obvious, it can only be segmented following phonotactical regularities which the readers knows, that is, the sonority scale and the restrictions of consonant clusters in German. But there are often several possibilities and the right segmentation is not always the phonologically optimal one, so the first guess may be wrong in which case a second try is to be made. The main problem for a computerized approach, however, is that it cannot be tested auto-


matically if the meaning of the whole word can be deduced from the cognate parts. The result of the whole procedure is a list of the “best” German cognate candidates, sorted according to their “quality”, i.e. the Levenshtein distance to the Dutch word (see the example duiken above). Finally, this list is compared to the translation equivalents for the Dutch word found in the bilingual Dutch–German list. If the translation equivalent (or one of several equivalents) is among the candidates that have been found, the Dutch word is marked by “!”, otherwise by “-”. If the equivalent is the candidate with the smallest Levenshtein distance, and thus the “best” candidate, this is indicated by “!!”; “≠” denotes “false friends”, namely candidates which are “better” or “as good” as the real equivalent or, in case the latter is not among the candidates found, whose Levenshtein distance to the Dutch word does not exceed a certain level (here set to 1.2).

5. Results To date, the procedure has been applied to the first 5000 Dutch word forms of the “top 10,000” list (excluding the numerous names and abbreviations). It would be easy to test a more extensive corpus, but the errors and gaps in the bilingual list still necessitate a lot of manual revision. (Moreover, the profit of the partial assignments of compound or derived words cannot be examined by a program.) However, tests with parts of the Dutch–German bilingual list as a corpus suggest that for lower frequency words, too, the overall share of German cognates hardly differs from the results presented here. For the “top 5,000” (see Fig. 2), it results that more than half of the word forms are immediately decodable via German cognates: the German cognates of these forms are the assignments with the smallest Levenshtein distance (“!!”, e.g. bezoek ‘visit’ : Besuch). The share of recognizable cognates is further augmented by those which are not the most similar German form but one of several similar forms (“!” — here the context must help to chose the right assignment, e.g. to assign koper ‘buyer’/‘copper’ to Käufer or Kupfer and not to Koffer ‘suitcase’). Together with the complex words whose meaning can be deduced from the meaning of cognate morphemes (“→ !”, e.g. gebruiker

Figure 2. German cognates of the “top 5,000” Dutch word forms (%)

300 Robert Möller

‘user’ : gebrauchen ‘to use’ + suffix -er ‘agent’), we arrive at a share of about 77% cognates. The remaining 23% mainly consist of two types of word forms: one for which no assignment to a German word is possible (“0”) — the complex forms whose structure is transparent but whose meaning cannot be deduced from cognate resp. similar German morphemes (“→ 0”, e.g. ontslagen ‘dismissed’ or uitdaging ‘challenge’: *entschlagen or *Aus-tag-ung do not make sense) can be added here — and one group of misleading cognates resp. incorrect assignments (“≠”, e.g. wet ‘law’ : Wette ‘bet’ or vervelen ‘to bore’–verfehlen ‘to miss’). The incorrect assignments where the Dutch word and the similar German word do not belong to the same word class ((“≠ ~ 0”), e.g. omlaag ‘down’ : Umlage ‘division of costs’) form a further small group in between. Here, the risk of being misled exists, but is rather low.

6. Misleading cognates The amount of “misleading” cognates is not so large as it is often assumed. Moreover, not all “misleading” cognates are equally so. The different typologies of “false friends” that have been proposed (cf. Kroschewski 2000) normally refer to production and include types which are absolutely irrelevant for reception (such as cognates with different spellings) or hardly relevant, such as cognates with a different stylistic value (e.g. Dutch advocaat–German Rechtsanwalt/Advokat). For general comprehension this is unimportant, and for a better comprehension more familiarity with the language is indispensable. For comprehension, it seems reasonable to differentiate misleading cognates in the following way: 1. How misleading are they really (do they present a danger of misunderstanding)?, and 2. If being misled can be avoided, can they somehow point in the right direction instead? Both aspects are independent of each other: there are “false friends” which are “harmless” but can hardly help (normally homonyms such as Dutch bellen ‘to ring’ — German bellen ‘to bark’–the difference between real etymological cognates and homonyms does not matter for the reader) and others which are “dangerous” but helpful if the danger can be avoided (Dutch vuil ‘dirty’–German faul ‘rotten’). The danger of misunderstanding is often negligible in most contexts due to syntactical-semantic reasons. Even if the word class is the same, misleading cognates often differ with regard to the complements they demand. Certainly it cannot be taken for granted that “good” cognates always demand the same complements in both languages, but at least it constitutes grounds for distrust if a presumed cognate has a different valence (cf. Lutjeharms 1988: 262, 286). If the reader has recognized that a cognate does not fit into the semantic and/or syntactical context, it is possible that it may, on the contrary, lead to a better equivalent via associations. These may vary, but mainly concern three different aspects. First, semantic change (as the main source of “misleading” cognates) often does not affect all contexts equally in which a word or morpheme appears; an older meaning or a common area of meaning may remain in a set expression or may


become visible in another member of the same word family. Second, languages often use similar metaphors; a “misleading” cognate, which in German lacks the metaphorical use it has in Dutch, may lead to a synonym that is an appropriate equivalent. (Thus in Dutch zakken also means ‘to fail (an exam)’ whereas in German sacken only means ‘to sink’. However, the idea of a downwards movement is the same as in German durchfallen, meaning ‘to fail (an exam)’.) Third, emancipation from Latin (and other loan influence) has brought forth many semantic loans and calques in Dutch and German (as in other Germanic languages). The results of this often correspond, but sometimes different but more or less synonymous indigenous words have been chosen, or the Latin word has not been replaced in one language. So, in the two latter cases, synonyms of the German cognates can lead to a correct translation. Especially the first case is quite frequent, so the reader who often has to deal with this type of cognate will develop intuitions about possible semantic change instead of only being misled.

7. A “Cloze test” That fact that results presented above confirm that German cognates are so numerous that they can provide access to written Dutch, shall be demonstrated in a short text example, in which the categories “0” , “→0” and “≠” of Figure 2 are marked (grey/underlined/italics). The text can be conceived as a gap-filling task (“Cloze test”) in which the marked words have been deleted. De internationale en nationale ontwikkelingen van het afgelopen jaar hebben de onzekerheden in ons dagelijks bestaan doen toenemen. Burgeroorlogen, aanslagen en andere vormen van geweld treffen iedere dag weer vele onschuldige mensen. Dit roept de vraag op hoe Nederland kan bijdragen aan duurzame vrede, veiligheid en armoedebestrijding. Aardbevingen, overstromingen en droogte confronteren overal in de wereld de mens met zijn beperkingen. Ook ons land is niet gespaard gebleven voor de gevolgen van langdurige droogte, ondanks de grote aandacht voor het waterbeheer. De teruggang van de economie is in Nederland in alle scherpte voelbaar geworden. Na jaren van voorspoed leidt dit voor veel burgers onverwacht tot onzekerheid. Dagelijks worden honderden mensen werkloos. De problemen van onze economie zijn niet alleen conjunctureel van aard. Om tot duurzaam herstel te komen is het noodzakelijk de economische structuur te versterken en de sociale zekerheid grondig te herzien. De regering beseft dat dit in eerste instantie voor veel mensen ingrijpende gevolgen zal hebben. Er zijn ook grote zorgen om de cohesie in onze samen leving. De waarden van verschillende bevolkingsgroepen blijken soms ver uit elkaar te liggen en de integratie verloopt niet voorspoedig. Voorts zijn de onveiligheid en overlast op straat en de aantasting van de leefomgeving verontrustend. De grote steden zien zich geplaatst voor een opeenhoping van problemen. (From: Queen Beatrix, Speech from the Throne 200315)

302 Robert Möller

The words in italics illustrate what has been said about “misleading” cognates: As ‘devotion’ (G. Andacht) does not fit into a non-religious context, the adjective andächtig in the expression andächtig zuhören (‘to listen very attentively’) can provide an association which leads to ‘attention’ (Du. aandacht). As it cannot be the duty of the Netherlands to deny (G. bestreiten) poverty, the meaning of bestrijding must be something else. Streiten means ‘argue, quarrel’, but e.g. in Streitkräfte ‘armed forces’, Streitaxt ‘battleaxe’ it is completely synonym to kämpfen ‘fight’ (Du. bestrijding = G. Bekämpfung ‘fight against’). However, one final caution has to be made: the frequency of the German cognates themselves has not been taken into account thus far, although this is not only an important factor in word recognition (see Dijkstra 2005: 1723). Moreover, it is not uncommon that two cognate words differ considerably in frequency in both languages. If the phonological correspondence between the Du. ziek ‘ill’ and G. siech is obvious and the meaning is the same, the former is the normal word and the latter is completely obsolete. Even for educated native speakers of German, the knowledge of less frequent or peripheral vocabulary (archaisms etc.) should not be overestimated, but vague reminiscences can often be activated (here, for example, dahinsiechen ‘to languish’ as an elevated, but still-used word may support the receptive survival of siech). By contrast, many existing cognates will not be of any help for most readers who have learnt or are learning German as a foreign language.16 These readers may, however, have some other cues from their own language. This distinction should not be forgotten in connection with the concept of a “bridge language” (cf. Klein 2002: 43), in order to avoid frustration due to overly optimistic expectations.

8. Conclusion Using a computer program makes it possible to explore the lexical basis of intercomprehension between two languages in a detailed and extensive way, arriving at quantifiable results. The present study suggests that by far the majority of Dutch vocabulary is accessible for German readers, if they are familiar with a small set of sound correspondences. Using the same procedures, one could test the lexical distance between any other languages and compare the results. The Levenshtein procedure can provide a realistic measure of lexical accessibility by integrating specific factors of human cognate processing such as similarity or saliency of differences in the calculation (one such factor, frequency/currency of the cognates, has been neglected in this study, but a frequency list of German words would be sufficient to include this, too). In addition, the possibility for improving cognate recognition by applying sound correspondence rules can be integrated in the procedure — as demonstrated here — in order to test what can be reached under this condition, or left out, if the basis for spontaneous understanding is to be tested.


To what extent the recognition of isolated words can really represent the possibilities of intercomprehension between two languages in general can only be examined by comparing the results of this approach with those of empirical studies of text comprehension.17 However, the studies by van Bezooien and Gooskens (this volume) and Ház (2005) corroborate the assumption that the amount and the transparency of cognates play a decisive role in real text understanding, and the results reported by Ház lead to the same conclusion as the results of the present study: understanding Dutch against the background of German is not completely trivial, but absolutely possible. Considering the fact that Ház’ test persons were not given any information about Dutch orthography and sound correspondences, it is clear that the results cannot be as good as they would be under the conditions the present study supposes as a starting point. Normally, given the same previous knowledge, the results of real text understanding should always be a little better than those of decoding isolated words, but the former will have a direct relation to the latter. Thus, a computer-based search for cognates seems to be a practicable way to achieve a realistic and comparable measure of the feasibility of intercomprehension, at least if there are no major differences in grammar, and furthermore to register in detail the specific differences and problems in order to provide a systematic help for this route to receptive competence.

Notes * I would like to thank the editors and referees, Nicole Marx and Veronika Wenzel for helpful comments 1. According to Zeevaert (2004: 273, 275), such central problems arise from lexical differences in inter-Scandinavian communication as well. 2. Ház (2005: 130–46) presents some experiments concerning the understanding of spoken words in context (decoding of Dutch words by Germans and vice versa), but she only tests a few words and in addition to phonological correspondences, other factors vary (the context is more or less clear, in some cases identical English cognates exist and the test persons certainly know them), so it is impossible to draw detailed conclusions. Lutjeharms (1988) is a comprehensive study of problems and strategies used by Dutch reading German, but here the readers have already had German instruction. Cf. also Wenzel (in press). 3. Here too, many problems are still unresolved, see Dijkstra (2005: 1724). 4. The overview given by Kuen (1993) is partly an exception, but here the perspective is reverse (German → Dutch) and oriented on production, which makes a considerable difference. 5. For reasons of tradition, the term cognate will here normally be used to indicate etymologically related words, but in fact the difference between a — truely — misleading cognate and an incorrect assignment based on generally existing sound correspondences does not matter in this context. 6. http://werners-index.de/niederlande.htm (April 6th, 2004). This list is not perfect, but it is difficult to find freely available bilingual word lists.

304 Robert Möller

7. k becomes ch or remains as is, which is no problem here. Moreover, German words with initial ch do not exist. 8. Based on the German–Dutch version of the bilingual word list (see above). 9. For an application of Levenshtein distance without differentiation of “costs” see also van Bezooien and Gooskens (this volume). 10. Some correspondences are only a coincidence, e.g. the relationship between Du. ruiter and G. reiter ‘rider’. 11. Here, early umlaut in certain inflectional forms and later analogical adjustment within the paradigm has taken place in both languages, but the latter in a different way. 12. Programmed in VBA to date. 13. http://wortschatz.uni-leipzig.de/html/wliste.html (March 2004). 14. Nearly all the Dutch affixes that are here included in a list could be associated with their cognate German equivalent in the same way as the other cognates, but this procedure will not be repeated for every word concerned, as these cognate affixes will quickly be stored by a human reader (cf. Lutjeharms 1988: 244f.). 15. http://www.koninklijkhuis.nl/content.jsp?objectid=4127 (September 2004). 16. Lutjeharms (1988: 160 and 1997: 159) reports an important difference between the success of native and non-native speakers of Dutch in reading German. The respective knowledge of peripheral Dutch vocabulary is not among the possible reasons she proposes, but the same fundamental difference can be noted in the comprehension of Middle High German texts by German students, where the vocabulary problem is obviously dominant. 17. Such as the studies by Ház (2005) and Wenzel (in press), but using more than only one short written text.

References Archibald, J. 1998. Second Language Phonology. Amsterdam: John Benjamins. Arntz, R. 1997. Passive Mehrsprachigkeit — eine Chance für die ‘kleinen’ Sprachen Europas. In Einsprachigkeit ist heilbar. Überlegungen zur neuen Mehrsprachigkeit Europas [Socio linguistica 11], P. H. Nelde (ed.), 166–77. Tübingen: Niemeyer. Dijkstra, T. 2005. Word recognition and lexical access: Connectionist approaches. In Lexikologie/ Lexicology: Ein internationales Handbuch zur Natur und Struktur von Wörtern und Wort schätzen [Handbücher zur Sprach- und Kommunikationswissenschaft, 21], D. A. Cruse, F. Hundsnurscher, M. Job, Michael and P. Rolf (eds), vol. 2, 1722–1730. Berlin: de Gruyter. Ház, É. 2005. Deutsche und Niederländer. Untersuchungen zur Möglichkeit einer unmittelbaren Verständigung [Philologia 68]. Hamburg: Dr. Kovač. Heeringa, W. 2004. Measuring Dialect Pronunciation Differences using Levenshtein Distance [Groningen Dissertations in Linguistics 46]. Groningen. Klein, H. G. 2002. Entwicklungsstand der Eurocomprehensionsforschung. In EuroCom — Mehr sprachiges Europa durch Interkomprehension in Sprachfamilien [Editiones EuroCom 8], G. Kischel (ed.), 35–45. Aachen: Shaker.


Kroschewski, A. 2000. False Friends and True Friends. Ein Beitrag zur Klassifizierung des Phäno mens der intersprachlich-heterogenen Referenz. Frankfurt: Peter Lang. Kuen, E. 1993. Deutsch–Niederländisches Lernwörterbuch. Hamburg: Buske. Lutjeharms, M. 1988. Lesen in der Fremdsprache. Versuch einer psycholinguistischen Deutung am Beispiel Deutsch als Fremdsprache. Bochum: AKS. Lutjeharms, M. 1997. Worterkennen beim Lesen einer Fremdsprache. In Kognitive Linguistik und Fremdspracherwerb: das mentale Lexikon, W. Börner and K. Vogel (eds), 149–67. Tübingen: Narr. Lutjeharms, M. 2004. Verarbeitungsebenen beim Lesen in der Fremdsprache. In Neuere For schungen zur Europäischen Interkomprehension [Editiones EuroCom 21], H. G. Klein and D. Rutke (eds), 67–82. Aachen: Shaker. Schade, U. and Barattelli, S. 2003. Kognitionswissenschaftliche Beiträge zu Sprachproduktion und Sprachrezeption. In Psycholinguistik — Psycholinguistics. Ein internationales Handbuch [Handbücher zur Sprach- und Kommunikationswissenschaft 24], G. Rickheit et al. (eds), 80–91. Berlin: de Gruyter. Scheerer-Neumann, G. 1996. Der Erwerb der basalen Lese- und Schreibfähigkeiten. In Schrift und Schriftlichkeit. Ein interdisziplinäres Handbuch internationaler Forschung [Handbücher zur Sprach- und Kommunikationswissenschaft 10], H. Günther and O. Ludwig (eds), 1153–1169. Berlin: de Gruyter. Wenzel, V. In press. Nederlands voor iedereen. In Akten des IVG-Kongresses 2005, Paris. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verstän digung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.

chapter 14

How can DaFnE and EuroComGerm contribute to the concept of receptive multilingualism? Theoretical and practical considerations Britta Hufeisen and Nicole Marx TU Darmstadt and Westfälische Wilhelms-Universität Münster

If the goal of receptive multilingualism is to be attained in the European context, new pedagogical concepts must be developed, tested and implemented. The following discussion has as its focus the learning of a new, tertiary and possibly related foreign language (FL), employing two models, Meißner’s Spontaneous Learner Grammar and Hufeisen’s Factor model, as a theoretical framework. It is shown how these models have been used to investigate tertiary language learning and how these concepts apply to improving the reception of new FLs. Two projects, DaFnE and EuroComGerm, are considered, both of which aim to help students use their knowledge of a related, previously learned FL in order to achieve receptive competencies in a new FL. Keywords: Spontaneous Learner Grammar, Factor model, DaFnE, EuroComGerm, receptive competence

1. Introduction According to the White Book of the Council of Europe, every EU citizen should have the opportunity to learn at least two foreign1 EU languages (Council of Europe 2002). In reality, the situation looks rather different (Krumm 2005, Besters-Dilger, de Cilia, Krumm and Rindler-Schjerve 2001): countries whose principle language is used mainly or only within its own boundaries, like Finnish in Finland or Dutch in the Netherlands, have long fulfilled this requirement. Finnish students learn — in addition to Swedish, their second official language — English and at least one further language. On the other hand, speakers of global languages, like English or Spanish, often learn at the most one foreign language. Although this is understandable from a pragmatic point of view, it undermines the concept of a European Union with mutual understanding and respect between citizens. Various pedagogical concepts have been introduced to make foreign language learning easier and more interesting, the latest being the communicative, the inter-

308 Britta Hufeisen and Nicole Marx

cultural, the task-based and the multilingual approaches. Another interesting idea not specifically developed for teaching purposes is the concept of the polyglot dialogue (Posner 1991), or pluri-lingual discourse (Clyne 2003). In the Scandinavian context, this phenomenon is dubbed interscandinavian semicommunication (see Zeevaert’s and Braunmüller’s contributions in this volume and Zeevaert 2004): each communication partner speaks his/her first or native language, perhaps with some accommodation to ease comprehension, and all others understand. For example, one speaker mentions something in Swedish, which the Dutch person understands and replies to in Dutch. The French speaker — having a receptive knowledge of both languages — adds something in French which both earlier speakers do not speak actively but understand. In this volume, this phenomenon is called receptive multilingualism. Researchers in DaFnE (Deutsch nach Englisch = German L3 after English L2, thirdlanguage acquisition/tertiary language learning) and specifically in EuroComGerm combine the concept of receptive multilingualism with the idea of the polyglot dialogue or plurilingual discourse in order to assist communication among EU citizens. They study linguistic processes which underlie such communication patterns and search for methods to prepare and enable speakers/hearers to tackle this challenge. It is important to note here that receptive knowledge of one or more languages is not at all passive. It includes a variety of complex processes, competencies and techniques that only a few speakers develop instinctively. It is a highly challenging cognitive and linguistic task and must be specifically trained, as will be shown below. Therefore, persons trying to gain receptive knowledge in various languages often need instruction and help. Receptive multilingualism focuses on the two classical skills listening and reading and on multiple language processing. In order to accept this new type of communicative goal, the former pedagogical aim of native-like competence has been replaced by a concept of a set of competencies which comprise (meta)linguistic and intercultural understanding, domain specificity, and cooperative communicative competence (Sprachhandlungskompetenz). The concept of receptive multilingualism, which extends far beyond minimal linguistic knowledge, is neither a simple pidgin nor incomplete language learning. Instead, it represents both the acquisition of receptive competencies in more than one given target language, and includes a set of specific foreign language learning strategies in order to fill in the receptive side of the polyglot dialogue. While the linguistic focus of this new direction rests on what happens during these learning and communication processes, the pedagogical focus is the enhancement of specific learning processes. The potential political and educational consequences of such a novel approach to language learning include the enhancement of Europeans’ mobility. Citizens may then spend time abroad regardless of their previous competencies in the target language. The aim is to equip them with the tools to function in the target language as soon as possible.

DaFnE, EuroComGerm and receptive multilingualism 309

2. Theoretical background During the past ten years, L3 researchers have begun to develop and evaluate their own models on the process of learning and acquiring more than one second language. These researchers argue that it is neither sufficient simply to use the established L2 acquisition models for the considerably more complex setting of learning an L3 or Lx (x≥3), nor is it appropriate to extend existing L2 models as these cannot explain the differences that are evident between L2 learning and L3 learning. The speech production models of Williams and Hammarberg (1998), de Bot (2004) and Clyne (2003) show that the mechanisms of language choice are based mainly on modified L2 models (chiefly Levelt 1989). These feature a relevant element that is not the focus of other models, namely speech production (see Table 1). Table 1. Multilingual speech production models Researcher(s)

Name of model

Main characteristics

Sarah Williams and Björn Hammarberg (1998)

Languages Switches Model

Psycholinguistic speech production model; languages assume different roles/ functions in communication become apparent through certain switches

Kees de Bot (2004)

Selection and Control Model

Psycholinguistic speech production model; automaticity, selection and control determine the choice of speech units

Michael Clyne (2003)

Model of Plurilingual Processing

Sociolinguistic model; includes sociocultural aspects

In the following, we would like to demonstrate the various models that have been recently introduced to explicate the differences between L2 and L3 learning and to illustrate how the general notion of multilingualism or — as Aronin & Ó Laoire (2004; see Table 2) phrase it, individual multilinguality — is described. This model is applied to multilingual societies such as Israel or Ireland in order to explain multilingual settings and dominant language situations (Aronin 2006). It comprises a set of nine characteristics: complexity, interrelatedness, fluctuation, variation and inconsistency, multifunctionality, inequality of function, self-balance, selfTable 2. Sociolinguistic multilingual acquisition model Researcher(s)

Name of model


Larissa Aronin & Muiris Ó Laoire (2004)

Biotic Model of Multilinguality

Sociolinguistic model that tries to explain multilingual acquisition in multilingual societies

310 Britta Hufeisen and Nicole Marx Table 3. Psycholinguistic multilingual acquisition model Researcher(s)

Name of model


Philip Herdina and Ulrike Jessner (2002)

Dynamic Model of Multilingualism

Psycholinguistic model that includes aspects such as language maintenance and loss

extension, and non-replication. They control the individual language behaviour and setting. As can be seen in the Dynamic Model of Multilingualism (Table 3), which explains individual conditions of the learning process, language maintenance and language loss play as important a role as the learning process itself. We will discuss two of the three applied models in detail in order to lay the foundation for the subsequent discussion. We are convinced that the acquisition/learning of an L3 differs greatly from the acquisition/learning of an L2 because of changes in the individual learner’s background (see Table 3). Groseva was the first researcher to highlight the special role that the L2 plays in the L3 and Lx (x≥3) learning process, as it acts as a reference base for future language learning. New items and rules are constantly revised and compared with the L2 and the L2-learning process. Meißner’s and Hufeisen’s models, developed within the frameworks of EuroCom and DaFnE, portray a similar conception of the L2, which is called the bridge language (Hufeisen 1991) or transfer base (Meißner 2002) (see Table 4). Table 4. Models in applied linguistic settings Researcher(s)

Name of Model

Main Characteristics

Maria Groseva (2000)

Foreign-language acquisition Model

L2 becomes the comparative and contrastive basis for all further languages

Franz-Josef Meißner (2004)

Multilingual Processing Spontaneous Model

Development of a Learner Grammar by zero beginners

Britta Hufeisen (Hufeisen/ Gibson 2003)

Factor Model

Describes the factors that control and influence the learning process, it emphasizes difference between L2 and L3

2.1 Multilingual processing model: Spontaneous learner grammar From within the context of constructivist learning theory and multilingualism pedagogy (EuroComDidact, cf. Meißner 2002), the Spontaneous Learner Grammar model was proposed. Franz-Joseph Meißner’s model demonstrates the processes involved during the reception — both written and oral — of texts in an unknown foreign language. The main focus is on how a learner with a high level of competence in one foreign language (such as French) deciphers a text in an unknown but closely related

DaFnE, EuroComGerm and receptive multilingualism

language (such as Spanish), and what processes are involved when attempting to understand the structure of this new language. Meißner suspects that learners continuously turn for help to at least one known, related language, while at the same time formulating hypotheses about the new language. In this way, learners build a new spontaneous (or hypothetical) grammar for this system. In contrast to Selinker’s concept of interlanguage, this spontaneous grammar “covers the linguistic knowledge of a target language still ‘unknown’ to the learner, even with the first encounter of the lexical, morphematical and syntactical structures” (Meißner 2003: 40); it is characterized by its systematicity and dynamics and “is constructed less upon the target language than upon the patterns of one or more pre-learned closely related foreign languages” (ibid.). With the assistance of all the other languages that the learner already knows, the spontaneous grammar is at first composed less of the actual structures of the target language, but rather of the sum of all the learner’s language knowledge. Over time, it is then adjusted and re-formulated to become ever closer to the actual structure of the new language system. During the process of hypothetical grammar formation, at least one helper or bridge language is always present, allowing the learner to compare structures of the known and the new language(s), and will be activated during the first contact with the new target language if certain conditions are met. These include an etymological relationship between the languages involved, a certain proficiency in the bridge language, and finally, a previous amount of preparation of the bridge language for its new roll. Only when these three conditions are met is a continuous comparison of the two languages possible (Meißner 1998), such that the development of the spontaneous grammar is optimised. In our context, it is interesting to see how Meißner’s model explains the process of grammar building, especially since Meißner is concerned primarily with receptive multilingual contexts. The procedure contains four steps (Meißner and Senger 2001: 41–3): 1. First, a spontaneous grammar is built in the target language. This ensues through the acquaintance with a bridge language (for example, the L2 English provides a structure for the initial understanding of the L3 German) and is highly dynamic. This grammar recognizes interlingual regularities via the process of systemising and generalising input. 2. An interlingual correspondence grammar is then constructed through the spontaneous grammar. Here, the spontaneous grammar develops interlingual correspondence rules which lie somewhere between the previous linguistic knowledge (i.e. of the bridge language) and the growing knowledge of the target language system. 3. Next, a plurilingual inter-system is built. This system saves all successful (as well as some unsuccessful) interlingual transfer processes and is composed of transfer bases which provide a general framework for decoding the new language. In this model, six such transfer bases are of interest: communicative strategy transfer, transfer of

311


interlingual processing procedures, transfer of cognitive principles, transfer as proor retroactive overlapping (Überlagerung), learning strategy transfer, and finally transfer of learning experiences (in this way, the model provides many parallels to the model developed by Britta Hufeisen, below). 4. Finally, a metacognitive strategy collection develops, in which learning experiences in the target language are saved. The spontaneous grammar is thus a subset of the interlanguage, but not interchangeable with it, as it is in a state of continual change, and is produced without explicit target language knowledge, which is, however, necessary for the development of interlanguage. In addition to formulating hypotheses about the target language, a learner gradually develops plurilingual system knowledge, which contains both positive and negative correspondence rules, and thus more information in total than the spontaneous grammar alone. This knowledge can then be re-applied when reading or listening to texts in the target language, or be extended and modified through contact with another language system. However, Meißner’s model also has limits for application to our situation, since it is primarily concerned with how zero beginners decode a language — but not how learners continue to learn, and not how they learn when guidance or instruction is available. This is where Hufeisen’s learning model becomes relevant.

2.2 Hufeisen’s factor model Hufeisen’s factor model consists of four parts, each describing one stage of language learning/acquisition, namely the acquisition of an L1 or various L1s, the stages of the learning of an L2, the learning of an L3, and the learning of an L4. Each separate stage refers to the initial learning steps taken by the learner of the respective language, and accounts for the factors that influence and control the learning process. The groups of factors responsible for influencing language learning and acquisition include the following: • Neurophysiological factors are the basis of general language learning/acquisition, production and reception capability. If one of these factors is hindered, language acquisition fails or is flawed. • Learner external factors include socio-economic and sociocultural surroundings such as the learning tradtitions, and the type and the amount of input the learner receives. If, for instance, sufficient or qualitatively adequate input is lacking, acquisition/learning is more difficult or even impossible. • Emotional factors such as anxiety, motivation, or acceptance of the new target language are highly influential in the learning process. If a learner is for example very tense or afraid to speak the target language, this emotional state slows or even hinders the learning process and success (see Dewaele 2002).


• Cognitive factors include language awareness, linguistic and metalinguistic awareness, learning awareness, knowledge of one’s own learner type and the ability to employ learning strategies and techniques (see Jessner 2004). • Linguistic factors comprise the learner’s L1(s). Schematically, L2-learning could be described as shown in Figure 1. Hufeisen’s model emphasizes the differences between learning an L2 and learning an L3. Whereas at the beginning of the L2 learning process the learner is a complete novice in the learning process of a second language, the L3 learner already knows what it feels like to approach a new language. She has developed (consciously or unconsciously) certain techniques of learning new words. She knows that a new text is often unclear, and is able to cope with the insecurity of having knowledge gaps. The learner may have previously explored what foreign language learner type she is and will be able to employ suitable strategies2 and techniques accordingly. In comparison to the beginning L2 learning process, the learning of an L3 therefore must recognize a new set of learning factors: • Foreign/second language learning-specific factors, such as individual second language learning experiences, interlanguages of other learned languages, and foreign language learning strategies. The linguistic factors must be extended from L1(s) to L1(s) and L2, which can function as the bridge language on the way to the L3 (Hufeisen, 1991). Thus, the L3 learner is able to rely on language specific knowledge and competencies that an L2 leaner has not yet gained. In summary, the two stages of the learning process are shown in Figures 1 and 2.

Figure 1. Learning of an L2


Figure 2. Learning of an L3

Besides describing the prototypical language learning process, the model can be applied to explain specific learning situations. It has to be kept in mind that each factor interacts with the others and that the change in one complex of factors might be followed by a change in other factor complexes, and that the importance of single factors and their relevance for the success of the learning process change from learner to learner. The model provides an explanation for the fact that native speakers of Dutch are not necessarily good language learners of German only due to linguistic factors (i.e. Dutch and German belong to the same language family and are — in comparison to other language constellations — fairly similar). Linguistic factors are merely one component of these complexes, and as they influence each other, some factors can become predominant and or unimportant in certain situations. In spite of the high amount of similarity between Dutch and German, the Dutch learner might still find it difficult to learn German because he is not motivated to learn this language. In projects such as German after English and English after German (see below), the starting point for the learning process is the relative similarity of languages (linguistic factors), and certain specific strategies and techniques are employed to exploit these similarities in order to get a grasp of the target language as quickly as possible. Therefore, the foreign language specific factors are used to enhance the learning process. The above-mentioned projects show that learning strategies can be taught, learned and successfully employed (this was also shown by Marx 2005 and Neuner 2005 in another study). Recent studies indicate that in different cultures, this method is not necessarily successful. Ouédroago (2005) reports that the learning traditions in Burkina Faso (Africa) do not emphasise cognitive factors but rather other learning methods such as condi-

DaFnE, EuroComGerm and receptive multilingualism

tioning. In these cultures, teaching and learning approaches that emphasize language awareness, learning strategies or learning awareness might not be as helpful for language learners who grow up in a learning tradition that does not foster these cognitive aspects. Other factors such as the existence of teaching and learning materials will influence the learning success to a greater degree.

3. Recent examples of linguistic research and good applied practice 3.1 Multiple-language acquisition: Tertiärsprachenkonzept — German after English (DaFnE) German is today a typical L3, which means that it is rarely chosen as a first foreign language at school, where English takes priority. If an L3 is learned in Europe, German, Spanish or French are the most common choices. Thus, beginning learners of German already have some knowledge of English, and these earlier experiences might ease the learning process of the new foreign language, as proposed by the models mentioned above. On a more general level, this means that students not only speak another language besides their L1, but that they also have foreign language learning experience (see the Factor Model). These are experienced language learners who can apply a large range of linguistic, meta-linguistic and strategic skills to the new language learning process through continual transfer of declarative and procedural knowledge.3 The long-term Plurilingualism Project (DaFnE = Deutsch nach English/German after English), supported by the European Center of Modern Languages in Graz, has explored these issues further within linguistic, educational, and pedagogical contexts, and has had several piloting phases in various European countries (Hufeisen and Neuner 2004). This project was not directed primarily at receptive multilingualism, but did consider features such as employing receptive skills in the beginning phases of learning, drawing on earlier languages and language learning experiences, and using learning techniques and strategies to exploit earlier competencies when undertaking the learning of new languages. Thus, in this approach, the concept of receptive multilingualism is extended and eventually comes to encompass the productive skills of speaking and writing. Learners tend not to make use of their previous knowledge on a systematic basis, however, as only the lexicon is evident to them as a transfer base (see Cenoz, Hufeisen and Jessner 2003). Moreover, it seems that learners do not use their previous language and strategy knowledge automatically, but rather have to be made aware of parallels and transfer possibilities between languages, as well as be introduced to potentially useful techniques of how to use and employ previous foreign language knowledge and language learning strategies. This was shown by Marx (2005) in her doctoral dissertation, which focused on two balanced groups of zero beginners of German with knowledge of

315


English as an L2. In the study, an experimental group was exposed to a specific DaFnEmethodology involving a sensitization to the learning situation and the specific benefits that multilingual learners possess; the course involved assisting learners recognize similarities between their two foreign languages and use these bridges to aid learning. Both groups of learners were given weekly listening comprehension tasks and were tested on their LC abilities over a six-week intensive language course. Compared to the control group, the experimental group achieved significantly higher scores on listening tasks at the language level of the learners. As well, sensitized learners performed significantly better on difficult language tasks, so long as texts contained similarities between the two languages (i.e. a higher rate of cognates). On these texts, the main advantage shown by the sensitized learners reflected their ability to quickly, often automatically recognize English–German cognates and use this knowledge to build an understanding of the text as a whole. Learners in the control group, who had not practiced interlingual strategies, were less able to make use of their previous knowledge of English and thus had fewer footholds on which they could build text comprehension. Thus, training in interlingual strategies seems to be an important factor in whether declarative knowledge from other languages is, in fact, drawn upon when a task involves comprehension of an L3. Various exercises to train such interlingual strategies have been developed, and an important outcome of the DaFnE project mentioned above was the development of specific learning/teaching materials which are designed to take into account the previous language knowledge and language learning experiences of students. Smaller European countries like Bulgaria or Sweden have introduced specific curricula (Dikova, Mavrodieva and Stankulowa 2001) and textbooks (Elfving-Vogel, Rydén and Mertens 1998). Distance education material targeted at specific language groups also tries to introduce tertiary language learning aspects at a more general level using as illustration the succession of English as L2 and then German as L3 (Neuner, Hufeisen, Erlenwein, Koithan, Kuršisa and Marx, in press).

3.2 EuroComGerm and eag While the DaFnE concept has been applied in German language courses and emphasises all four competencies in learning, other programmes concentrate chiefly on receptive skills, with the goal of enabling students to understand multiple languages, even if they do not (yet) speak or write them. Such is the philosophy of the EuroCom (European (Inter-) Comprehension) concept, which strives for intercomprehension between members of language families (i.e. receptive multilingualism).4 In the field of Romance language intercomprehension, parallels between various key modern Romance languages are drawn and on this basis, the ability to understand a group of languages with a common background is developed — thus, a German native speaker with a solid basic knowledge of French will learn to read and then aurally comprehend Spanish, Italian, Portuguese, Rumanian and Catalan within a relatively short time on


the basis of their knowledge of French (and possibly English). Since the Romance languages are closely related to each other, and more than a third of European citizens speak one as their mother tongue, it is easy to understand why the pioneering work in this project was completed in this field (Klein and Stegmann 2000). The EuroCom strategy is to optimize inferencing techniques for students learning (i.e. developing receptive competencies in) new, etymologically related foreign languages. This inferencing is made possible through the comparison of languages at seven levels — by “sieving” them in seven ways. These levels include applying previous and contextual knowledge to reading texts through the search for international and pan-Romance vocabulary, for sound correspondences, for common graphemes and pronunciations, for syntactic and morphosyntanctic regularities in the language family, and finally, for common prefixes and suffixes. With practice, this process can become automated (see Stoye 2000), so that understanding texts is possible within a language family. This method, which has been taught at the Johann Wolfgang von Goethe University in Frankfurt since 1991, has been used successfully at both universities and schools. While EuroComRom is the oldest and thus most advanced of the projects, two further ones focus on receptive competencies in European language groups: EuroComSlav concentrates on the Slavic family, and EuroComGerm 5 considers European Germanic languages (Hufeisen and Marx 2007). In the latter project, the EuroCom concept has been applied to intercomprehension in eight languages, using German and English as the linguistic transfer bases or “bridges”. Although most coursework in these programmes is done in a regular language classroom under the guidance of well-informed instructors, the EuroCom concept also acknowledges the fact that there are many people who have already finished school and who either do not have access to such programmes, or are not able to invest the time required in order to develop receptive strategies in many languages at once. For this reason, on-line courses have been introduced with the goal of providing learners with opportunities to develop reading competencies in just one (related) language. These courses require the learner to work autonomously and remain motivated by doing the necessary coursework, as there is no on-line tutor available to answer questions. They have the advantage of allowing the learner to choose when she wishes to work on her language learning, how intensively, to what degree she reviews material and also which material she chooses from the collection. Further, such a course is not tied to a specific profession or course of study. One project employing the EuroCom on-line platform is eag 6 (=English after German), a course designed to enable students to learn to read English on the basis of their previously learned German (in the case of, for example, Eastern Europeans who learned Russian and German in school but not English) or who have acquired German language skills (German native speakers). Because reading English texts has become a necessity during university studies and professional life, many people are now realising that their lack of English knowledge poses a hindrance. However, these (future) pro-


fessionals might not have the time, the desire or the need to achieve productive competencies in English and might therefore wish to begin with receptive knowledge of this language of wider communication. This step is made possible via their previous knowledge of German, a language which (as noted above) is related to English and thus is able to provide many bridges to new learners of English. German is a rich source of international and pan-Germanic vocabulary, as well as syntactic structural aids. The on-line materials follow a somewhat different structure than the classroom work, moving in a structured, step-by-step progression through the seven sieves and the languages involved. After a brief introduction to the principles of the Seven Sieves for the Germanic languages, the learner works through a series of authentic texts, utilizing her previous and developing knowledge of Germanic language systems to develop and further improve her reading potential in English. The programme incorporates various levels of analysis, including lexical and semantic (each lexeme is linked to hints — but not to a translation), morphological (tips divide words into morphemes, and separate explanations of morphological rules can be viewed) and grammatical (again, explanations are provided for various phenomena appearing in the texts). Once the learner has worked through a particular text — preferably without the aid of a dictionary — she can continue to answer multiple-choice questions both on content as well as on language structure, check her understanding against a parallel translation of the text, and in the final step, see which new, “non-sievable” words (profile words) and which new structures appear in the text under study. In this way only those structures and lexemes necessary for text comprehension are explicitly analyzed, and only at the junctures they become important. A grammatical or lexical progression in the traditional sense is missing from this concept, and the learner will not, like for example within traditional communicative approaches, be taught to introduce himself before he is confronted with newspaper articles, tourist information or job adverts. The on-line eag course underwent a trial phase in 2005, in which students of the Technical University of Darmstadt, Germany, with no knowledge of English familiarized themselves with the programme before continuing autonomously, either within a university language laboratory or at home, the variety chosen by most of the learners. Students were requested to read an English text in the first session and then to answer questions about it before the programme was explained to them. At this point they were allowed to start working on the texts. Responses to the programme were quite positive, with many students commenting that they would continue to work with the materials during the summer months. However, motivation for autonomous learning seems to be a key factor (as in most on-line learning programmes), as many students found that, as the semester wore on, they had less time to devote to their English reading comprehension. Most people will need an understanding of other languages for communication purposes, and the EuroCom strategy is a method holding a great deal of promise for reaching the goal of interlingual comprehension within one language family. Whether


in a traditional language classroom or on-line, students have the opportunity to familiarize themselves with linguistic structures and similarities in many languages at once, thereby moving closer to the real possibility of a polyglot dialogue.

4. Conclusion Results to date of the aforementioned projects lend much support to previous data from the Romance, Slavic and Scandinavian language arenas, providing clear evidence that receptive multilingualism within the German language constellation is a real possibility, not only linguistically, but also with respect to the political aims of the European White Book. It is imperative that further research within specific language constellations (such as German, English, Swedish; Dutch, English, German; or German, English, Icelandic) be conducted in order not only to validate the models and their explanatory value, but as well, to foster learning and, specifically, receptive competencies in related foreign languages. In effect, the term “foreign languages”, viewed from the concept of receptive multilingualism, is actually a misnomer. No language — especially from within the same language family — can truly be considered unknown or “foreign”. Instead, we should begin to consider languages being learned for receptive purposes to be systems already well understood. The crucial step for the learner is to find the proper key to fully unlock their mysteries.

Notes 1. We are aware that the concepts of acquisition and learning share a number of common features and that these two types of acquisition rarely occur separately. Still, the learning process involves an approach to language that does not usually happen during the natural, undirected acquisition of a given language. Therefore, we use learning to refer to the formal setting of instruction and concede that acquisition takes also place during the learning process. 2. Dörnyei purports that language learning strategies do not exist because research cannot agree on a common definition (2005a, b). In contrast, we argue that when a learning activity has been consciously directed, purposefully employed and specifically used in a certain setting to reach a certain goal, and is used at further points in order to reach a similar learning goal, it is well justifiable to describe this activity as learning strategy. 3. We will not explore interference here as it is minimal compared to positive transfer, especially in the case of receptive competencies. 4. See http://www.eurocom-frankfurt.de/ for more information on all programmes. 5. See http://www.eurocomgerm.de/ for more information on the EuroComGerm project. 6. See http://dmz02.kom.e-technik.tu-darmstadt.de/eurocomgerm/index.php for the eag course.


References Aronin, L. 2006. Dominant language constellation: An approach to multilingualism studies. In Multilingualism in Educationals Settings, M. Ó Laoire (ed.), 142–61. Baltmannsweiler: Schneider Hohengehren. Aronin, L. and Ó Laoire, M. 2004. Exploring multilingualism in cultural contexts: Towards a notion of multilinguality. In Trilingualism in Family, School and Community, Ch. Hoffmann and J. Ytsma (eds), 11–29. Clevedon: Multilingual Matters. Besters-Dilger, J. , de Cillia, R. , Krumm, H.-J. , and Rindler-Schjerve, R. (eds). 2001. Mehr sprachigkeit in der erweiterten Europäischen Union. Klagenfurt: Drava. Cenoz, J. , Hufeisen, B. and Jessner, U. (eds) 2003. The Multilingual Lexicon. Dordrecht: Kluwer. Clyne, M. 2003. Dynamics of Language Contact. Cambridge: Cambridge University Press. Council of Europe, 800 Million Europeans. Strasbourg: Council of Europe. www.coe.int/T/E/ Com/About_Coe/Brocures/800millionsE.asp [2 Mar. 2007]. De Bot, K. 2004. The multilingual lexicon: Modeling selection and control. International Journal of Multilingualism 1 (1): 17–32. Dewaele, J.-M. 2002. Psychological and sociodemographic correlates of communicative anxiety in L2 and L3 production. International Journal of Bilingualism 6 (1): 23–8. Dikova, V. , Mavrodieva, L. and Stankulowa, K. 2001. Curriculum für Deutsch als zweite Fremd sprache in der bulgarischen allgemeinbildenden Oberschule. Zeitschrift für Interkulturellen Fremdsprachenunterricht [Online] 5 (3), 24pp. http://www.spz.tu-darmstadt.de/projekt_ ejournal/jg_05_3/beitrag/dikova.htm [15 Apr. 2005]. Dörnyei, Z. 2005a. The Psychology of the Language Learner: Individual differences in second-language acquisition. Mahwah, NJ: Lawrence Erlbaum. Dörnyei, Z. 2005b. Individual differences in second-language acquisition. Paper presented at the AILA conference in 14 Madison, WI, 25–9 July 2005. Duke, J. , Hufeisen, B. and Lutjeharms, M. 2004. Die sieben Siebe des EuroCom für den multilingualen Einstieg in die Welt der germanischen Sprachen. In Neuere Forschungen zur Europäischen Interkomprehension, H. Klein and D. Rutke (eds), 109–34. Aachen: Shaker. Elfving-Vogel, M. , Rydén, K. and Mertens, H. 1998. Lust auf Deutsch. Stockholm: Bonnier. Groseva, M. 2000. Dient das L2-System als ein Fremdsprachenlernmodell? In Tertiärsprachen. Theorien. Modelle. Methoden, B. Hufeisen and B. Lindemann (eds), 21–30. Tübingen: Stauf fenburg. Herdina, P. and Jessner, U. 2002. A Dynamic Model of Multilingualism. Perspectives of change in psycholinguistics. Clevedon: Multilingual Matters. Hufeisen, B. 1991. Englisch als erste und Deutsch als zweite Fremdsprache — Empirische Unter suchung zur zwischensprachlichen Interaktion. Frankfurt: Peter Lang. Hufeisen, B. and Gibson, M. 2003. Zur Interdependenz emotionaler und kognitiver Faktoren im Rahmen eines Modells zur Beschreibung sukzessiven multiplen Sprachenlernens. In Gehirn und Sprache: Psycho- und neurolinguistische Ansätze. Bulletin suisse de linguistique appliquée 78, R. Francheschini, B. Hufeisen, U. Jessner and G. Lüdi (eds), 13–33. Hufeisen, B. and Marx, N. (eds) 2007. EuroComGerm — Die sieben Siebe: germanische Sprachen lesen lernen. Aachen: Shaker. Hufeisen, B. and Neuner, G. (eds) 2004. The Plurilingualism Project: Tertiary language learning — German after English. Strasbourg: Council of Europe Publishing.


Jessner, U. 2004. Zur Rolle des metalinguistischen Bewusstseins in der Mehrsprachigkeitsfor schung. In Beim Schwedischlernen sind Englisch und Deutsch ganz ‘hilfsvoll’. Untersuchungen zum multiplen Sprachenlernen, B. Hufeisen and N. Marx (eds), 17–32. Frankfurt: Lang. Klein, H. and Stegmann, T. 2000. EuroComRom — Die sieben Siebe: Romanische Sprachen sofort lesen können. Aachen: Shaker. Krumm, H.-J. 2005. Von der additiven zur curricularen Mehrsprachigkeit: Über die Notwendig keit der Einbeziehung von Minderheiten-, Migranten- und Nachbarsprachen. In Gesamt sprachencurriculum — Integrierte Sprachendidaktik — Common Curriculum. Theoretische Überlegungen und Beispiele der Umsetzung, B. Hufeisen and M. Lutjeharms (eds), 27–36. Tübingen, Gunter Narr (Giessener Beiträge zur Fremdsprachenforschung). Levelt, W. J. M. 1989. Speaking: From intention to articulation. Cambridge MA: MIT Press. Marx, N. 2005. Hörverstehensleistungen im Deutschen als Tertiärsprache: zum Nutzen eines Sensi bilisierungsunterrichts in „DaFnE“. Baltmannsweiler: Schneider Hohengeren. Meißner, F.-J. 1998. Gymnasiasten der Sekundarstufe I lernen den interlingualen Transfer. In Mehrsprachigkeitsdidaktik. Konzepte, Analysen, Lehrerfahrungen mit romanischen Fremd sprachen, F.-J. Meißner and M. Reinfried (eds), 217–37. Tübingen: Narr. Meißner, F.-J. 2002. EuroComDidact. In Europäische Mehrsprachigkeit. Analysen — Konzepte — Dokumente, D. Rutke (ed.), 45–64. Aachen: Shaker. Meißner, F.-J. 2003. EuroComDidact: Learning and teaching plurilingual comprehension. In Sprachkompetenz — Mehrsprachigkeit — Translation. Akten des 35. Linguistischen Kolloqui ums. Innsbruck, 20.-22. September 2000, L. Zybatow (ed.), 33–46. Tübingen: Narr. Meißner, F.-J. 2004. Transfer und Transferieren. Anleitungen zum Interkomprehensionsunter richt. In Neuere Forschungen zur Europäischen Interkomprehension, H. Klein and D. Rutke (eds), 39–66. Aachen: Shaker. Meißner, F.-J. and Senger, U. 2001. Vom induktiven zum konstruktiven Lehr- und Lernparadigma. Methodische Folgerungen aus der mehrsprachigkeitsdidaktischen Forschung. In Bausteine für einen neukommunikativen Französischunterricht, F.-J. Meißner and M. Reinfried (eds), 21–50. Tübingen: Narr Neuner, St. 2005. Fremdsprachenlernen und Lernerautonomie. Sprachlernbewusstsein, Lern prozessorganisation und Lernstrategien zum Wortschatzlernen in Deutsch als Fremdsprache. Baltmannsweiler: Schneider Hohengehren. Neuner, G. , Hufeisen, B., Erlenwein, S., Koithan, U. , Kuršisa, A. and Marx, N. In press. Deutsch als zweite Fremdsprache nach Englisch. Berlin: Langenscheidt (Fernstudieneinheit). Ouédraogo, D. 2005. Mehrsprachigkeit und Deutschunterricht in Burkina Faso. Welche Strate gien zur Verbesserung der Lehr- und Lernmethoden. Vortrag auf der internationalen Deutschlehrertagung im August 2005 in Graz in der Sektion Deutsch als zweite oder weitere Fremdsprache. In Begegnungssprache Deutsch. Motivation – Herausforderung – Perspektiven. Thesenband der XIII. Internationalen Tagung der Deutschlehrerinnen und Deutschlehrer in Graz, 1.-6.8.2005, B. Sorger (ed.), 336, Wien: Internationaler Deutschlehrerverband. Posner, R. 1991. Der polyglotte Dialog. Sprachreport 3: 6–10. Stoye, S. 2000: Eurocomprehension. Der romanistische Beitrag für eine europäische Mehrsprachigkeit [Editiones Eurocom 2]. Aachen: Shaker. Williams, S. and Hammarberg, B. 1998. Language switches in L3 production: Implications for a polyglot speaking model. Applied Linguistics 19 (3): 295–333. Zeevaert, L. 2004. Interskandinavische Kommunikation. Strategien zur Etablierung von Verständi gung zwischen Skandinaviern im Diskurs [Philologia 64]. Hamburg: Dr. Kovač.

Index of names A Acklin Muji 137 Aijmer 202 Alber 164 Alderson 227 Allport 272 Ameel 273 Archibald 291, 296 Arntz 96, 288 Aronin 309 Ashby 266 Aspeslagh 76 Auer 79 B Balota 270, 271 Barattelli 287, 296 Barðdal 1 Bates 166, 273–5 Baumgarten 4, 6, 13 Beach 203 Bentin 278 Benveniste 195, 196 Bergman 218 Bernhardt 268 Berthoud 166 Bertram 271, 272 Besters-Dilger 307 Beuerle 34 Bhatia 2 Biber 196, 197 Bø 111, 128, 221–3, 226, 250 Börestam 109,112, 219, 222, 246, 250 Bornkessel 274 Borsley 274 Bourdieu 76, 164, 173 Bowers 270 Brattegard 37 Braunmüller 7, 8, 32, 34, 38, 42, 78, 105, 106, 218, 219, 249, 308 Brown 174 Bruntse 218 Budovičová 105, 250 Bührig 10, 81, 92 Burger 57, 69 Burke 33 C Canale 165 Cenoz 315 Chen 270, 272 Chirita 39 Chomsky 129, 165

Clifton 275 Clyne 52, 76, 96, 180, 308, 309 Conrad 145–7, 151, 196 Cooreman 275 Coseriu 4, 106, 128–9 Costa 270, 273 Coulmas 202 Cutler 270 Czeitschner 54, 62, 64 D Dahlstedt 109, 130, 228 De Bot 77, 309 De Groot 273 De Jong 79 De Jongste 77 Delsen 79 Delsing 8, 10, 13, 15, 33, 43, 128– 30, 221–3, 226, 236, 241, 245–6, 260 De Stefani 170 De Swaan 76 Deutsch 269 Dewaele 312 Diercks 34, 43–4 Dieth 155 Dijkstra 273, 296, 302–3 Dikova 316 Doehler 166, 173 Doetjes 1–4, 8, 10, 14, 19, 224, 228, 231, 260 Dorgeloh 196 Dorian 26 Dresemann 4, 6, 13, 183, 245 Duke 110 Dürmüller 140 E Eco 107, 129, 174 Edmondson 202–4, 207–8, 211 Ehlich 81–2, 95, 97–8, 182, 183 Elfving-Vogel 316 Ellis 268 Elmiger 144, 145 Ervin-Tripp 107 F Faerch 165 Feitsma 106 Feldman 271, 272 Fellerer 54, 60 Finegan 197 Finkenstaedt 14, 76

Firth 180, 190, 210 Fischel 61, 62 Fodor 267 Folk 269 Forster 270 Franceschini 166, 173, 278 Francis 280 Fremdling 96 Frenck-Mestre 272–3 Friel 273, 279 Furer 140 G Gal 165 Gernentz 37 Giles 128, 227, 253 Golinski 1, 44, 128, 130, 224, 228 Goodman 273 Gooskens 17, 128, 227, 251, 260–2, 287, 292, 303 Gorter 251 Grainger 272, 273 Green 278 Grin 166 Groseva 310 Grosjean 77, 164, 165 Grünbaum 219, 227 Gumperz 173, 181, 192 H Haegeman 274 Hägi 141 Hall 94 Halliday 13, 199 Hammarberg 309 Hansen 105–6, 109, 128 Harley 277 Haugen 26, 105, 106, 109, 112, 219–21, 224, 226, 228, 250 Ház 1, 285, 292, 303–4 Heeringa 251, 257, 288, 291, 294 Heilenman 275 Heiniger 169 Heller 180 Herrlitz 9, 76, 77 Hickerton 250 Hill 174 Hinskens 96 Hoffmann 92 Hohenstein 197 Holmes 180

324 Receptive Multilingualism Hopper 202, 203 House 4, 6, 13, 19, 75–7, 181, 197, 202–4, 207–11, 279 Hufeisen 3, 5, 16, 18, 78, 110, 130, 266, 307, 310–17 Hüllen 198 Humphreys 268 Hunston 197 Hyland 197 I Iwasaki 196 J Jahr 31, 37 Jensen 106, 250 Jescheniak 270 Jessner 310, 313, 315 Jiang 270 Johnson-Laird 277 Jonkman 251 Jörgensen 128, 222 K Kamwangamalu 2 Karker 3 Kärkkäinen 197 Kärrlander 128, 222, 245 Kasper 165, 197, 207 Kellerman 78 Kennison 273, 279 Kilborn 275 Kim 269 Kintsch 277 Klein 5, 43, 106, 110, 130, 227, 266, 302, 317 Klin 276 Kloss 3, 26, 35, 106, 109, 110 Knapp 183 Koda 275 Kolde 141, 142, 160, 161 Konieczny 274 Koole 4, 9, 77, 81–3, 88 Kristensen 222 Kroschewski 300 Krumm 307 Kutas 272 L Lademacher 76 Lemhöfer 273 LePage 28 Levelt 309 Levy 278 Li 275 Linthout 76 Liu 275 Lleó 19 Longtin 271

Loos 9, 77, 181, 190 Lüdi 4, 12, 13, 77, 128, 138, 154– 6, 164–70, 173, 180, 219 Lundin 222 Lundin Åkesson 222, 223, 226, 236, 241, 245–6 Lutjeharms 5, 14–18, 78, 110, 130, 268–75, 286, 287, 293, 296, 300, 303–4 Lyons 196 M McCann 5, 266 McDonald 275 McNamara 174 McQueen 270 MacWhinney 274, 275 Malt 278 Maria Theresia 59 Marslen-Wilson 271 Marx 3, 5, 16, 18, 78, 130, 266, 314–17 Matthey 145 Matthiessen 199 Maturana 10, 107, 108 Mauranen 197 Maurud 15, 111, 112, 128, 221, 223, 226, 227, 232, 233, 237, 240–2, 250 Mavrodieva 316 Meierkord 183 Meisel 19, 173 Meißner 279, 280, 310–12 Melberg 3 Melinger 271 Menke 28 Mertens 316 Metzger 203 Meunier 271 Meuter 137 Milroy 77, 90, 165 Möller 16, 18, 75, 128, 261 Mondada 166, 170–2 Morton 270 Moscoso 271 Moscovitch 278 Mulac 202 Müller 76, 179, 187 Müller-Lancé 269 Murray 143, 163 Muysken 77, 90, 165 Myers-Scotton 180 N Nelde 76 Nesse 31, 37 Nettle 38 Neuner 314–16 Newerkla 55, 58, 59, 69 Nichols 38 Noyau 164

Ó Ó Laoire 309 Oakhill 277 Ogris 62 Östman 203 P Palmer 196 Paradis 267, 268, 273 Pekarek 166, 173 Petersen 182–6 Peterson 271 Pfaff 180 Pierce 250 Porquier 164 Posner 308 Prescher 76 Py 145, 154, 156, 164–8 Pynte 275 R Raven 76 Rawson 267, 277 Rayner 269 Rehbein 10, 75–7, 80–2, 94, 95, 98, 131, 181–3, 186, 192, 199, 279 Reuter 219, 227 Ribbert 4, 9, 16, 97 Ridell 128, 223 Ritchie 2 Roelands 75, 78–80, 88 Romaine 2 Rothweiler 19 Rumelhart 268 Rydén 316 S Sanford 270 Schade 287, 296 Scharloth 141 Scheerer-Neumann 287 Schegloff 164 Scheibman 196, 197, 205 Schiffrin 207 Schriefers 270, 274 Schröder 14, 76, 128 Schutz 164 Selting 155 Senger 311 Shannon 107 Shapiro 274 Shetter 94 Smith 27, 33, 196, 197, 227 Snedeker 274 Stankulowa 316 Stegmann 5, 43, 110, 227, 266, 317 Stotz 137 Stourzh 53, 57

Name index 325

Stoye 317 Stubbe 180 Suchsland 274 Swain 165 T Tabouret-Keller 28 Taft 270 Teich 197 Teleman 1, 16, 130 ten Thije 4, 9, 10, 16, 77, 78, 81– 3, 88, 92, 96–8 Thomas 197, 272 Thompson 197, 202 Tiisala 39 Traugott 197, 202, 203 Trautmann 96 Traxler 267, 275 Trudgill 34, 35, 253 Trueswell 274 Tsui 203

Turner 250 U Ungeheuer 107 V Van Bezooijen 17, 128, 227, 250, 260–2, 287, 292 Van den Berg 250 Van Dijk 277 Vetter 8, 77, 131 W Wagner 180–6 Wallnig 53, 54 Watts 143, 163 Weaver 107 Wenzel 292, 303–4 Werlen 6, 11, 12, 77, 128, 130,

143, 145, 151, 155, 175, 219 Westheide 9, 76, 77, 96 Wetzel-Kranz 167 Williams 309 Wolff 250, 260 Wright 38–9, 43–4 X Ximenes 3 Z Zahn 76 Zeelenberg 280 Zeevaert 1–6, 10, 13–4, 19, 43, 52, 74–8, 106, 112, 117, 118, 128–30, 181, 182, 192, 218, 219, 228, 231, 246, 249, 250, 303, 308 Zwitserlood 272 Zybatow 110

Index of subjects A Abstand languages 106 see also Ausbau languages action space 80, 81 active competence 105 adult L2-language learning 42 articles in post-position 40 asymmetric communication 30 attentional processing 5, 268 attitudes 17, 96, 145, 201, 234, 244, 250–4 Ausbau languages 26, 35, 36, 106 see also Abstand languages automatic processing 267–9 B Baltic 33–8, 224 Biel/Bienne 11, 141–9, 153–6 bilingualism 25–7, 37, 144–5, 160, 166, 180, 272 bilingual cantons 138, 141 bilingual city 11, 138–46 bilingual district 138–43 bilingual mode 164–5 bridge language 302, 310–13 business communication 37, 179 business negotiation 181–3, 191 see also merchants C closely related languages 27–31, 96, 105–10, 218–9, 261, 279 cloze test 260, 301 code switching 87–90, 165, 186– 8, 278 see also dialect mixing cognates 16–18, 88, 255–61, 273, 279, 285–8, 291–6, 299–303 cognate words 256, 279, 302 misleading cognates 296, 300 non-cognates 255–61, 286 collocations 198, 201 communication asymmetric communica tion 30 business communication 37, 179 crucial communication 107, 118 ELF communication 195, 198, 210

exolingual communica tion 152, 164 face-to-face trading communication 27 intercommunity communica tion 163–6, 173 inter-Scandinavian communication 52, 222 interscandinavian semicommunication 103 multilingual communica tion 1–4, 49–53, 61–6, 74– 80, 106, 118–9, 126–7, 148, 173, 180–1, 191 semi-communication 1,10, 26, 76, 105–9, 111, 219–20, 227 see also conversation communicative competence 110, 166, 173, 308 active competence 105 multilingual competence 52, 67, 165, 173 passive competence 4, 105 communicative styles 197–8 Competition Model 274–5 complement clause constructions 196, 202–5 comprehension 15–18, 179, 191, 233–45, 252–6, 259–62, 265–72, 275–9, 286–7, 316, 318 see also intercomprehension contextualisation cues 181 conversation 120, 145–52, 169, 186–91, 195–9, 209 correspondence rules 292–8, 311–2 creoloidisation 35 creoloids 34 crucial communication 107, 118 cues 181, 191, 268, 274–5, 302 Cultural Apparatus 75, 94 D DaFnE 308–10, 315–6 Danish 3–11, 34–6, 111–16, 119–23, 217–27, 231–9, 242–4 default language 143, 154, 167 default model 11, 143 definite article 40

dialectal continuum 37, 218 dialect mixing 32 diamorphs 42 diasystem 29, 32–5, 38 diffused language 28 diglossic situation 160 discourse knowledge 77, 186, 190–1 discourse organization 203 discursive interculture 77 divergeolect 36 E educational system 140, 161–3 EEC 33, 232 EES 232 ELF communication 195, 198, 210 ELF communities 210 English as a foreign language 143, 195 English as lingua franca (ELF) 2, 6, 180 English L1 198–200 EU 1–3, 6, 49–53, 61, 65–6, 232, 307–8 EuroCom 269–73, 310, 316–18 EuroCom-method 265, 269– 73 EuroComGerm 308, 316–17 exolingual communication 152, 164 experiential domains 200, 209 F face-to-face interaction 30–1, 182 face-to-face trading communication 27 false friends 288, 299–300 Faroese 41, 106, 217–19 feedback interview 90 first person pronoun 196–9 focused language 28 foreign language 3–6, 76 142–5, 174, 231, 265–72, 285–6, 307–10, 313–15 English as a foreign language 143, 195 Fribourg/Freiburg 142–7 functional diglossia 31 functional pragmatic approach 81

Subject index 327

G gambit 207 genetic relationship 34, 36, 41 German 32–42, 50–64,73–83, 87–97, 110, 137–55, 159–64, 167–74, 183–90, 218–9, 271–5, 278, 285–8, 291–303, 314–19 Gesamtsprachenkonzept 162 globalization 137 global languages 33, 307 grammaticalization 198, 202, 204 H Habsburg Empire 50–6, 62–7 Hanseatic chancellery 32 historical language 106 Holistic language-policy plan 162 homoglossic 164 I illocution 83 see also speech action immigrants 235–8, 241–2 institutional constellation 77 institutional discourse 81, 92, 94 institutional keywords 88–91 institutional knowledge 81, 88–92, 180 insular Scandinavian languages 217–8 intelligibility 106–7, 110, 195, 217, 250–61 mutual intelligibility 106–7, 195, 250, 255–6 inter-Scandinavian communica tion 52, 222 interactional meaning 207 interactive planning 81–3, 87 intercommunity communica tion 163–6, 173 intercomprehension 27, 110 163–5, 232, 286–7, 292, 303, 316–7 interdialect 34–5 interdialectal differences 218 interlanguage 34–5, 311–2 interlingual strategies 316 interscandinavian communica tion 11, 15 interscandinavian semi communication 103 J jargon 34

K koinéisation 35 koinés 34 L language language areas 138–9 language awareness 313–5 language choice 12, 77,146–9, 167–73, 273, 309 language policy 4–12,50–6, 61–6, 80–1,140, 154–5, 165, 218 language tests 236, 240, 260 see also Abstand languages; Ausbau languages; closely related languages; bridge language; default language; diffused language; focused language; local language; minority language; national language; neighbouring languages; supranational languages Latin 7, 26, 32–4, 38–9 60, 104, 144, 166, 185, 218, 293, 301 learner varieties 195, 204–9 learning processes 308 let it pass-strategy 118, 219 Levenshtein distance 255–9, 291, 298–9 lingua franca 6–14, 26–33, 51– 2, 67, 76, 127, 142–3, 154–5, 164, 183, 223 lingua franca communica tion 14, 127 see also English as lingua franca (ELF) linguistic accommodation 7, 119 linguistic distances 250–1, 255, 258 linguistic diversity 49–50, 126, 172 linguistic minorities 159 listening tasks 316 local language 59, 139–43, 161, 163 Low German 32–41, 105, 295 M mainland Scandinavia 34–6 mainland Scandinavian languages 29, 34–5, 217– 19, 227 main language 139, 143–4, 191 majority 141, 154, 203, 294 medial 160 medial diglossia 141–2 medieval London 38

mental lexicon 16, 270–3, 279 merchants 37, 41, 42 Middle Ages 28–37, 40, 42 minority 49, 141 minority language 11, 28, 106, 41 see also linguistic minorities miscommunication 13, 95, 203 see also trouble sources misleading cognates 296, 300 models of interlingual communication 138 monolingual 4, 12, 33, 57–9, 108–9, 111–19, 141–5, 160– 9, 173 morphemes 38, 42, 269–72, 279, 298–300 multilingualism 1–7, 29–33, 36– 7, 52–8, 63–7, 137–42, 164, 173, 179–81, 308–109 multilingual communica tion 1–4, 49–53, 61–6, 74– 80, 106, 118–9, 126–7, 148, 173, 180–1, 191 multilingual competence 52, 67, 165, 173 multilingual discourses 104, 108–9 receptive multilingualism 25– 34, 37–43, 52–3, 61, 73–81, 104–6, 111–14, 125, 174, 181, 249, 308, 319 unspecified multilingualism 52, 58, 67 see also bilingualism mutual intelligibility 106–7, 195, 250, 255–6 mutual language understanding 220–3 mutual understanding 3–7, 36, 88, 91, 103–7, 110–12, 120, 217, 220–3, 307 N nation-states 50 national cohesion 137 nationalism 27–9, 56, 67 national language 27–9, 33, 50, 137–41, 160–3, 166, 180 neighbouring languages 26, 28, 162, 232–8, 242–4 non-cognates 255–61, 286 Northern Europe 29, 33, 36 Norwegian 26, 34, 41, 105, 111– 2, 119–22, 186–90, 217–22, 231–9, 242–4, 260, 261 O official language 2, 51, 61–2, 138–41, 144, 154, 250 opaque 26, 179, 187

328 Receptive Multilingualism P participant framework 207 passive competence 4, 105 phonological recoding 269, 273 pluri-lingual discourse 308 plurilingual interactions 171 plurilingual techniques 171 polyadic conversations 171 polyglot dialogue 52, 66, 109, 308, 319 possessive pronoun 41 pragmatic knowledge 191 prior knowledge 276–7 process types 13, 199 prosodic patterns 190 psychotypology 78 R reading process 265–70 receptive multilingualism 25– 34, 37–43, 52–3, 61, 73–81, 104–6, 111–14, 125, 174, 181, 249, 308, 319 see also polyglot dialogue; semi-communication; pluri-lingual discourse receptive skills 109, 315–6 reporting 81–3, 87 S Scandinavian languages 7–9, 34–7, 40–2, 217–19, 231–7,

240, 243, 260, 279 Scandinavian linguistic fellowship 231, 233, 245 Schengen 232 semantic processing 265, 276–8 semi-communication 1,10, 26, 76, 105–9, 111, 219–20, 227 see also receptive multilingual ism; polyglot dialogue sensitization 316 similarity 105, 257, 271–2, 291– 8 sound correspondences 266, 279, 286–92 speech action 81–4, 180, speech community 103– 4, 112, 126 spontaneous grammar 311–2 stance 196–200, 207–10 standardisation 27, 32, 33 strategy 13, 31, 118, 219, 268– 71, 311- 2, 315–18 let it pass-strategy 118, 219 subjectivity 195–200, 209–10 supranational languages 33 SVO-patterns 42 Swedish 34–6, 111–16, 119–27, 217–27, 231–44, 307–8 Swiss model 142, 153, 161–9 Swiss multilingualism 141 Switzerland 78, 105- 6, 137–43, 159–63, 218 syntactic 198–9, 266–70, 273–6, 317–8

T team discussions 81–3 territoriality principle 140–2, 160 trading contacts 37 transfer 5, 104, 107–8, 270, 275– 6, 280, 310–12, 317 translation equivalent 270–2, 299 translinguistic wording 165 trouble sources 11, 112, 117–20 typological distance 75, 279–80 typological proximity 75, 267 U unilateral convergeolect 36 unspecified multilingualism 52, 58, 67 V V1 patterns 8, 42 varieties 27–9, 106, 173, 210, 250, 291 learner varieties 195, 204–9 W word order 34, 38–9, 42, 255, 274–6, 279, 280 word recognition 266–73, 287, 302

In the series Hamburg Studies on Multilingualism the following titles have been published thus far or are scheduled for publication: 6 5 4 3 2 1

Thije, Jan D. ten and Ludger Zeevaert (eds.): Receptive Multilingualism. Linguistic analyses, language policies and didactic concepts. 2007. x, 328 pp. Rehbein, Jochen, Christiane Hohenstein and Lukas Pietsch (eds.): Connectivity in Grammar and Discourse. x, 447 pp. + index. Expected May 2007 Lleó, Conxita (ed.): Interfaces in Multilingualism. Acquisition and representation. 2006. xiv, 284 pp. House, Juliane and Jochen Rehbein (eds.): Multilingual Communication. 2004. viii, 359 pp. Braunmüller, Kurt and Gisella Ferraresi (eds.): Aspects of Multilingualism in European Language History. 2003. viii, 291 pp. Müller, Natascha (ed.): (In)vulnerable Domains in Multilingualism. 2003. xiv, 374 pp.

Receptive Multilingualism: Linguistic analyses, language policies and didactic concepts (Hamburg Studies on Multilingualism)

Language Contact and Contact Languages (Hamburg Studies on Multilingualism (Hsm))

Multilingual Communication (Hamburg Studies in Multilingualism)

Aspects of Multilingualism in European Language History (Hamburg Studies on Multilingualism)

Multilingualism

Connectivity in Grammar and Discourse (Hamburg Studies on Multilingualism)

Multilingualism & nation Building

Multilingualism In India

The Exploration of Multilingualism: Development of research on L3, multilingualism and multiple language acquisition (Aila Applied Linguistics Series)

Emotions and Multilingualism (Studies in Emotion and Social Interaction)

Emotions and Multilingualism (Studies in Emotion and Social Interaction)

Multilingualism, Cultural Identity, and Education in Morocco

Handbook of Multilingualism and Multilingual Communication

Living Languages: Multilingualism across the Lifespan

Milton's Languages: The Impact of Multilingualism on Style

Beyond Bilingualism: Multilingualism and Multilingual Education (Multilingual Matters)

Linguistic landscapes in the Netherlands: a study of multilingualism in Amsterdam and Friesland

Blackwell Guide to Research Methods in Bilingualism and Multilingualism

Medievalism, Multilingualism, and Chaucer (The New Middle Ages)

Multilingualism and Nation Building (Multilingual Matters, No 91)

Towards global multilingualism: European models and Asian realities

Linguistic Perspectives on Language and Education

Linguistic Perspectives on Language and Education

The Human Rights of Linguistic Minorities and Language Policies

Language and Linguistic Nonviolence

Multilingualism in Mathematics Classrooms: Global Perspectives (Bilingual Education and Bilingualism)

Beyond Bilingualism: Multilingualism and Multilingual Education (Multilingual Matters)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Multilingualism in European Bilingual Contexts: Language Use And Attitudes (Multilingual Matters)

Literary Rhetoric. Concepts-Structures-Analyses

Receptive Multilingualism: Linguistic analyses, language policies and didactic concepts (Hamburg Studies on Multilingualism)

Language Contact and Contact Languages (Hamburg Studies on Multilingualism (Hsm))

Multilingual Communication (Hamburg Studies in Multilingualism)

Aspects of Multilingualism in European Language History (Hamburg Studies on Multilingualism)

Multilingualism

Connectivity in Grammar and Discourse (Hamburg Studies on Multilingualism)

Multilingualism & nation Building

Multilingualism In India

The Exploration of Multilingualism: Development of research on L3, multilingualism and multiple language acquisition (Aila Applied Linguistics Series)

Emotions and Multilingualism (Studies in Emotion and Social Interaction)

Emotions and Multilingualism (Studies in Emotion and Social Interaction)

Multilingualism, Cultural Identity, and Education in Morocco

Handbook of Multilingualism and Multilingual Communication

Living Languages: Multilingualism across the Lifespan

Milton's Languages: The Impact of Multilingualism on Style

Beyond Bilingualism: Multilingualism and Multilingual Education (Multilingual Matters)

Linguistic landscapes in the Netherlands: a study of multilingualism in Amsterdam and Friesland

Blackwell Guide to Research Methods in Bilingualism and Multilingualism

Medievalism, Multilingualism, and Chaucer (The New Middle Ages)

Multilingualism and Nation Building (Multilingual Matters, No 91)

Towards global multilingualism: European models and Asian realities

Linguistic Perspectives on Language and Education

Linguistic Perspectives on Language and Education

The Human Rights of Linguistic Minorities and Language Policies

Language and Linguistic Nonviolence

Multilingualism in Mathematics Classrooms: Global Perspectives (Bilingual Education and Bilingualism)

Beyond Bilingualism: Multilingualism and Multilingual Education (Multilingual Matters)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Linguistic Diversity And Language Theories (Studies in Language Companion Series)

Multilingualism in European Bilingual Contexts: Language Use And Attitudes (Multilingual Matters)

Literary Rhetoric. Concepts-Structures-Analyses

Recommend Documents