Can Language be Managed in International Business?: Insights into Language Choice from a Case Study of Danish and Austrian Multinational Corporations (MNCs)

Bellak, Nina Doctoral Thesis Can Language be Managed in International Business?: Insights into Language Choice from a Case Study of Danish and Austrian Multinational Corporations (MNCs) PhD Series, No. 23.2014 Provided in Cooperation with: Copenhagen Business School (CBS) Suggested Citation: Bellak, Nina (2014) : Can Language be Managed in International Business?: Insights into Language Choice from a Case Study of Danish and Austrian Multinational Corporations (MNCs), PhD Series, No. 23.2014, ISBN 9788793155473, Copenhagen Business School (CBS), Frederiksberg, https://hdl.handle.net/10398/8973 This Version is available at: https://hdl.handle.net/10419/208899 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/ Nina Bellak LIMAC PhD School Department of International Business Communication PhD Series 23-2014 PhD Series 23-2014 Can Language be Managed in International Business? copenhagen business school handelshøjskolen solbjerg plads 3 dk-2000 frederiksberg danmark www.cbs.dk ISSN 0906-6934 Print ISBN: 978-87-93155-46-6 Online ISBN: 978-87-93155-47-3 Can Language be Managed in International Business? Insights into Language Choice from a Case Study of Danish and Austrian Multinational Corporations (MNCs) CAN LANGUAGE BE MANAGED IN INTERNATIONAL BUSINESS? Insights into Language Choice from a Case Study of Danish and Austrian Multinational Corporations (MNCs) Nina Bellak PhD thesis Supervisor: Associate Prof. Maribel Blasco LIMAC PhD School Department of International Business Communication COPENHAGEN BUSINESS SCHOOL Nina Bellak Can Language be Managed in International Business? Insights into Language Choice from a Case Study of Danish and Austrian Multinational Corporations (MNCs) 1st edition 2014 PhD Series 23.2014 © The Author ISSN 0906-6934 Print ISBN: 978-87-93155-46-6 Online ISBN:978-87-93155-47-3 LIMAC PhD School is a cross-disciplinary PhD School connected to research communities within the areas of Languages, Law, Informatics, Operations Management, Accounting, Communication and Cultural Studies. All rights reserved. No parts of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage or retrieval system, without permission in writing from the publisher. To my family and dear friends Meiner Familie und treuen Freundinnen I Acknowledgements First of all, I extend my sincere thanks to the various supervisors that have accompanied and supported me in this unique experience of doing a PhD (Associate Prof. Lisbeth VerstraeteHansen, Associate Prof. Mette Skovgård Andersen and Associate Prof. Inger Mees). More importantly, I wish to express my deep gratitude to Associate Prof. Maribel Blasco who has eventually taken over and made it possible for me to complete this PhD. Her supervision, based on detailed reading of my work in progress, not to mention messy drafts, her sharp thinking, constructive and comprehensive feedback even including linguistic editing, fruitful discussions in lengthy meetings, keeping me on track and providing guidance and direction, has been brilliant and impeccable. I could not have done it without her support. THANK YOU, Maribel! In 2010 I got the great opportunity to visit Prof. Anne-Wil Harzing at the University of Melbourne in Australia who helped me establish a network within the research community, provided me with valuable comments on my work and with the technical tools to start my data analysis. I am deeply thankful for all her support. Of course, I would like to thank all my colleagues in particular the ‘junior’ peers from the Department of International Business Communication at Copenhagen Business School who have created a lovely not only work but also social environment of which it has been a pleasure to be part. In this context, special thanks go to our former colleague Assistant Prof. Kristian T. H. Jensen who translated the abstract into Danish. Mange tak! I also thank colleagues outside the department for their advice they gave me in the early stages of the project. The selected Danish and Austrian case companies have enabled me to undertake this research; I am especially grateful for decision makers’ willingness to participate in the study and to provide access, the individuals’ openness, not least their time devoted to my fieldwork and most importantly their thoughts and experiences they shared with me. Also, I extend my sincere thanks to the assessment committee consisting of the internal chair, Associate Prof. Ole Helmersen, and the two external examiners Associate Prof. Karen M. Lauridsen (Aarhus University) and Prof. Rebecca Piekkari (Aalto University). They have not only evaluated this thesis but also provided me with constructive feedback and invaluable input on my work. I am very thankful for all the time and effort they have put into assessing my PhD and writing a comprehensive assessment report. Finally, I extend my heartfelt thanks to my beloved family, in particular my parents, the good old soul mates in Austria1 and close friends in Denmark2, who have always been there for me when I needed them and encouraged me to complete this project. / Nicht zuletzt gebührt außerordentlicher Dank meiner lieben Familie, insbesondere meinen Eltern, guten Freundinnen und Freunden in Österreich3 und Dänemark4, die immer für mich da waren, wenn ich sie gebraucht habe und in schweren Stunden aufmunternde Worte fanden, aber mir vor allem die nötige Energie und Motivation gaben, um dieses Forschungsprojekt zu Ende zu bringen. Ihr seid die Besten [end ei laf ju]! J Without all these people, their invaluable support and inspiration of various kinds, it would have been impossible to reach the end of this both challenging and rewarding ‘PhD journey’. Thank you all! Tusind tak! Dankeschön! 1 Special thanks to the ‘Chicken-Forum’ and SuperSy! 2 Special thanks to my ‘Øsi-Mädels’! 3 Besonderer Dank geht an das ‚Chicken-Forum‘ und SuperSy! 4 Besonderer Dank geht an meine ‚Øsi-Mädels‘! II Preface When reading the title of the present PhD thesis, one may wonder about the motivation for selecting Danish and Austrian companies. One of the reasons is personal; it all started when I came to Denmark in 2009 to start this PhD project centering around language in international business (IB). I was introduced to a new culture and society quite open towards the English language and also its use. I had not experienced anything like that before, neither in my home country Austria nor Belgium where I had lived for a year. When arriving in Denmark, my second language was French, not English, which meant I did not feel comfortable using English. However, I usually chose English (as a lingua franca) when approaching Danes and I was impressed by their vocabulary, oral communication skills and especially by their code-switching abilities, which I could not keep up with. They might have addressed me in Danish, realized I was a foreigner who did not speak Danish, and would immediately switch to English without any major problems. In Denmark the public exposure to the English language is pronounced as on national television, radio and other media. Apart from children’s programs being dubbed, everything is broadcast in the original language version supplemented with Danish subtitles, which facilitates foreign language learning. According to an EU survey entitled Europeans and Their Languages (Special Eurobarometer 243), Denmark is also the country with the highest population rate (46 %) selfevaluating its English skills as “very good” (European Commission 2006: 14). Almost half (44 %) of those speaking it as a foreign language (86 %) also use it “often” (European Commission 2006: 17). These facts are indicators for an overall Anglophile Danish society which is perceived controversially, thus the subject of lively debate (e.g. Phillipson 2001, Preisler 2003, Lund 2003). My personal impression gained was that this ‘focus’ on the English language and a ‘general’ selfconfident attitude towards English use had an impact on the perceptions and values of languages other than English within Danish society. Being immersed in Danish culture and Copenhagen Business School (CBS), I also gained insights into tertiary education. Most Danish business schools have recently been confronted with a og flytning af personale). De fleste valg bliver i virkeligheden truffet både på det individuelle og på koncernniveau, og de er svære at adskille fra hinanden. Koncernniveauet er opsplittet i individuelle ledere som selvstændigt foretager sprogvalg, der overhovedet ikke er blevet harmoniseret. Et ekstra niveau udgøres af ydre kræfter (fx myndigheder, love) der pålægger multinationale selskaber brugen af flere sprog. Endelig varierer de sproglige valg på tværs af de multinationale selskabers organisatoriske enheder, deres interne og eksterne kommunikation og kommunikationssituationer. Vi kan konkludere at sprogvalg er et socialt, komplekst, kontekstafhængigt og flersproget fænomen, som er svært at kontrollere eller regulere. Sammenfattende viser min forskning at sprogplanlægning i internationale forretningsmæssige sammenhænge i multinationale selskaber næppe kan være centraliseret eller monolingval. Når de er påvirket af ydre kræfter, bliver sprogplanlægning uden for deres kontrol. Dette tyder på at sprogplanlægning skal være lokaliseret, multilingval og modtagelig for den sociale kontekst. Til syvende og sidst kan man spørge om det overhovedet er nødvendigt at foretage sprogplanlægning eller om det ville være bedre at overlade sprogvalg til den enkelte. Denne viden kan bidrage til både forskning og praksis i erhvervslivet. Contents Acknowledgements ................................................................................................................. I! Preface ................................................................................................................................. II! Summary ............................................................................................................................. III! Resumé ............................................................................................................................... IV! 1!Introduction .................................................................................................................... 1! 1.1!Problem Definition: Language Choice in International Business – Can Language be Managed? ........................................................................................................................................ 1! 1.2!Research Question (RQ) ....................................................................................................... 3! 1.3!Overall Aim and Contributions ............................................................................................ 3! 1.4!Structure of the Thesis .......................................................................................................... 4! 2!Positioning the Study in the Existing Literature ............................................................... 5! 2.1!Overview of a Cross-disciplinary Field .................................................................................. 5! 2.1.1!Language (choice) within business research – language as a ‘barrier’ ........................... 5! 2.1.2!Language (choice) within language research – language as a ‘resource’ ........................ 7! 2.2!Empirical Findings of Prior Research on Language Choice in International Business ..... 10! 2.2.1!Language choice in MNCs ........................................................................................... 10! 2.2.2!Language choice for external communication purposes: the ‘forgotten’ dimension in MNC contexts .......................................................................................................................... 17! 2.2.3!Positioning the study at the cross-road of two non-Anglophone nations ................... 20! 2.2.3.1!Language choice in Danish businesses ............................................................................... 21! 2.2.3.2!Language choice in Austrian businesses ............................................................................ 23! 2.2.4!(Empirical) knowledge gaps summarized ..................................................................... 25! 2.3!Theoretical Discussions in a Cross-disciplinary Field ......................................................... 27! 2.3.1!Language choice made at different levels: language policy/strategy, planning/management, practice – useful distinctions or conceptual confusion? .................. 27! 2.3.2!Language choice in MNCs: a sociolinguistic approach ............................................... 37! 2.3.2.1!MNCs as multilingual speech communities and language functions ............................... 40! 2.3.2.1.1!HQ language, local and customer languages .............................................................. 41! 2.3.2.1.2!Lingua francas and corporate languages ..................................................................... 42! 2.3.2.2!Language choice classifications (from a seller’s perspective) ............................................. 45! 2.3.2.3!Different types of language choices .................................................................................... 46! 2.3.2.3.1!Passive multilingualism ............................................................................................... 46! 2.3.2.3.2!Code-switching/-mixing .............................................................................................. 47! 2.3.2.3.3!Language choices involving translation and interpretation ....................................... 47! 2.3.2.3.4!Language choices for educational purposes (language training and acquisition, language buddies) ........................................................................................................................... 51! 2.3.2.3.5!Language choices for HRM purposes (selective recruitment, staff relocation) .......... 52! 2.3.3!Language choice embedded in social context: a working framework ......................... 53! 2.3.3.1!Language proficiency and norms of a speech community ................................................. 54! 2.3.3.2!Domains of language use: participants, their roles and relationships, setting, topic within the employment domain .................................................................................................................... 56! 2.3.3.3!Politeness strategies (and social relationships) ................................................................... 58! 2.3.3.4!Attitudes to language (vs. beliefs and ideology) ................................................................. 60! 2.3.3.5!Personal motivation (integrative vs. instrumental/extrinsic vs. intrinsic) ......................... 64! 2.3.3.6!Internal and external forces ............................................................................................... 66! 2.3.4!(Theoretical) knowledge gaps summarized .................................................................. 68! 2.4!Summary ............................................................................................................................. 69! 3!Research Design and Methods ........................................................................................ 71! 3.1!A Qualitative Research Approach ...................................................................................... 71! 3.2!Philosophical Underpinning .............................................................................................. 72! 3.3!A Qualitative Case Study .................................................................................................... 77! 3.3.1!A collective case study of four MNCs headquartered in Denmark and Austria ........ 79! 3.3.1.1!Selecting multinational corporations (MNCs) .................................................................. 80! 3.3.1.2!Selecting Danish and Austrian MNCs ............................................................................... 81! 3.3.1.3!Selecting two case companies per country ......................................................................... 84! 3.3.1.4!Initial problems encountered ............................................................................................. 86! 3.4!Data Collection ................................................................................................................... 87! 3.4.1!Empirical materials: research interviews (RIs), corporate documents and websites ... 87! 3.4.2!Collecting interview data: a first overview ................................................................... 88! 3.4.2.1!Interview respondents across three levels of employment ................................................. 90! 3.4.2.1.1!Further problems encountered ................................................................................... 91! 3.4.2.2!Entering the field and conducting research interviews (RIs) ............................................. 91! 3.4.2.3!The interview protocol (see Appendix) .............................................................................. 94! 3.4.3!Treating alternative data sources: corporate documents and websites ........................ 96! 3.5!Data Analysis ....................................................................................................................... 97! 3.5.1!Data preparation: transcribing research interviews ..................................................... 98! 3.5.2!Coding phase I: exploring and reducing data digitally (using NVivo) ........................ 99! 3.5.2.1!Developing a coding tree following an abductive logic ................................................... 100! 3.5.3!Pattern coding phase I: revisiting coded data manually to further reduce data and look for relationships ............................................................................................................. 103! 3.5.4!Pattern coding phase II: understanding and interpreting relationships using withincase displays ............................................................................................................................ 104! 3.5.5!Crafting case study reports in an iterative fashion: ‘telling the story…’ .................... 107! 3.5.5.1!Drawing and verifying conclusions .................................................................................. 108! 3.5.5.1.1!Treating alternative data sources (data triangulation) .............................................. 109! 3.5.6!Finalizing reports and synthesizing findings across cases .......................................... 110! 3.6!Summary ............................................................................................................................ 112! 4!Language Choice in a Multinational Corporation (MNC): A Case Study (I) of DK_1 .... 113! 4.1!About DK_1: Facts and Figures ........................................................................................ 113! 4.2!Language Choices on DK_1’s Website ............................................................................. 113! 4.3!DK_1 as a Multilingual Speech Community: Language Functions and Language Choice Norms ......................................................................................................................................... 115! 4.3.1!A language guide (on English usage) or language policy? .......................................... 117! 4.4!Language Choice at DK_1 Informed by Social Context .................................................. 119! 4.4.1!Social-linguistic dimension: participants’ language proficiency informs language choices .................................................................................................................................... 119! 4.4.1.1!Proficiency informs the choice of English ....................................................................... 120! 4.4.1.2!Proficiency informs the choice of the HQ language ........................................................ 122! 4.4.1.3!Proficiency informs the choice of local languages (and transition to English) ............... 123! 4.4.1.4!Proficiency informs the choice of (translation into) local languages (double standard) ........ .......................................................................................................................................... 124! 4.4.1.5!Proficiency informs the choices of English and local/customer languages (involving interpretation) .................................................................................................................................. 125! 4.4.1.6!Proficiency informs code-switching/-mixing .................................................................... 125! 4.4.1.7!Proficiency informs passive multilingualism .................................................................... 127! 4.4.1.8!Proficiency informs the choice of the language of instruction for corporate training .... 128! 4.4.1.9!Proficiency informs the choice of English for educational purposes (language training) ...... .......................................................................................................................................... 129! 4.4.2!Social-relational dimension: participants’ role relationships and politeness strategies inform language choices ......................................................................................................... 129! 4.4.2.1!Buyer-seller relationship and politeness strategies inform the choice of (translation into) customer languages .......................................................................................................................... 130! 4.4.2.2!Buyer-seller relationship and politeness inform the choice of (interpretation into) customer languages .......................................................................................................................... 135! 4.4.3!Social-psychological dimension: participants’ attitudes to languages inform language choices .................................................................................................................................... 136! 4.4.3.1!Ethnocentric attitude informs (imposes) the choice of Danish (internal forces) ............ 136! 4.4.3.2!Positive attitude to the HQ language informs the choice of Danish (via help of expatriate) .......................................................................................................................................... 138! 4.4.3.3!Geocentric attitude informs the choice of English for educational purposes (language training and acquisition) .................................................................................................................. 139! 4.4.3.4!Varying attitudes inform language choices for HRM purposes (selective recruitment and staff relocation) ................................................................................................................................. 141! 4.4.4!Social-psychological dimension: participants’ motivations inform language choices (including language acquisition) ............................................................................................ 143! 4.4.4.1!Integrative, instrumental/extrinsic and intrinsic motivations inform the choice (and acquisition) of HQ, local and customer languages .......................................................................... 143! 4.4.4.2!Intrinsic motivation informs the choice of local languages ............................................. 144! 4.4.5!Social-regulatory dimension: external forces impose languages ................................ 144! 4.4.5.1!National (home-/host-country) laws enforce the choice of (translation into) customer languages .......................................................................................................................................... 145! 4.4.5.2!National (host-country) authorities enforce the choice of (translation into) local languages .......................................................................................................................................... 146! 4.4.5.3!National (host-country) education systems (acquisition planning) influence selective recruitment ....................................................................................................................................... 146! 4.5!Conclusions Drawn from Case Study I: Language Choice at DK_1 Embedded in Social Context ....................................................................................................................................... 146! 5!Language Choice in a Multinational Corporation (MNC): A Case Study (II) of DK_2 ... 153! 5.1!About DK_2: Facts and Figures ........................................................................................ 153! 5.2!Language Choices on DK_2’s Website ............................................................................. 153! 5.3!DK_2 as a Multilingual Speech Community: Language Functions and Language Choice Norms ......................................................................................................................................... 155! 5.3.1!A language policy statement and its understandings ................................................. 158! 5.4!Language Choice at DK_2 Informed by Social Context .................................................. 163! 5.4.1!Social-linguistic dimension: participants’ language proficiency informs language choices .................................................................................................................................... 163! 5.4.1.1!Proficiency informs the choice of English ....................................................................... 163! 5.4.1.2!Proficiency informs the choice of local/customer languages ........................................... 166! 5.4.1.3!Proficiency informs the choice of the HQ language (involving translation) ................... 167! 5.4.1.4!Proficiency informs the choices of English and local languages (involving translation) ....... .......................................................................................................................................... 167! 5.4.1.5!Proficiency informs the choices of English and local languages (involving interpretation) .. .......................................................................................................................................... 168! 5.4.1.6!Proficiency informs code-mixing ...................................................................................... 169! 5.4.1.7!Proficiency informs passive multilingualism .................................................................... 169! 5.4.2!Social-relational dimension: participants’ role relationships and politeness strategies inform language choices ......................................................................................................... 171! 5.4.2.1!Buyer-seller relationship and politeness strategies inform the choice of (translation into) customer languages (translation quality imposed by external forces) ............................................. 171! 5.4.2.2!Buyer-seller relationship and polycentric attitude inform the choice of customer language for educational purposes (language training) .................................................................................. 174! 5.4.3!Social-psychological dimension: participants’ attitudes to languages inform language choices .................................................................................................................................... 175! 5.4.3.1!Ethnocentric attitude informs (imposes) the choice of Danish (including language acquisition) (internal forces) ............................................................................................................ 175! 5.4.3.2!CEOs’ ethnocentric and geocentric attitudes inform(ed) the choices of Danish and English (internal forces) ................................................................................................................... 177! 5.4.3.3!Geocentric attitude informs the choice of English (divergence) combined with ethnocentrism .................................................................................................................................. 179! 5.4.3.4!Geocentric attitude informs the choice of English for educational purposes (language training and acquisition) .................................................................................................................. 180! 5.4.3.5!Varying attitudes inform language choices for HRM purposes (selective recruitment and staff relocation) (double standard) ................................................................................................... 181! 5.4.4!Social-psychological dimension: participants’ intrinsic motivation informs the choice of local languages ................................................................................................................... 183! 5.4.5!Social-regulatory dimension: external forces impose languages ................................ 184! 5.4.5.1!International scientific community enforces the choice of English within Research & Development .................................................................................................................................... 184! 5.4.5.2!Supranational institutions and legal regulations enforce language choices .................... 185! 5.4.5.3!National (host-country) authorities and customers enforce the choice of local/customer languages (involving translation) ..................................................................................................... 186! 5.4.5.4!National (host-country) education systems (acquisition planning) influence selective recruitment ....................................................................................................................................... 187! 5.5!Conclusions Drawn from Case Study II: Language Choice at DK_2 Embedded in Social Context ....................................................................................................................................... 188! 6!Language Choice in a Multinational Corporation (MNC): A Case Study (III) of AT_1 ... 195! 6.1!About AT_1: Facts and Figures ........................................................................................ 195! 6.2!Language Choices on AT_1’s Website ............................................................................. 195! 6.3!AT_1 as a Multilingual Speech Community: Language Functions and Language Choice Norms ......................................................................................................................................... 197! 6.4!Language Choice at AT_1 Informed by Social Context .................................................. 201! 6.4.1!Social-linguistic dimension: participants’ language proficiency informs language choices .................................................................................................................................... 201! 6.4.1.1!Language proficiency informs the choice of English ....................................................... 202! 6.4.1.2!Language proficiency informs the choices of English and local languages (involving translation) ....................................................................................................................................... 204! 6.4.1.3!Language proficiency informs the choices of local languages and English (involving interpretation) .................................................................................................................................. 205! 6.4.1.4!Language proficiency informs code-switching/-mixing .................................................... 207! 6.4.1.5!Language proficiency informs passive multilingualism ................................................... 208! 6.4.2!Social-relational dimension: participant roles, role relationships, politeness strategies and power relations inform language choices ........................................................................ 209! 6.4.2.1!Buyer-seller relationship and politeness strategies inform the choice of customer languages for HRM purposes (selective recruitment) ...................................................................... 209! 6.4.2.2!Politeness strategies and participant role (relative power) inform the choice of the HQ language (double standard) .............................................................................................................. 210! 6.4.2.3!Politeness strategies and social relationships inform the choice of the local language for educational purposes (language acquisition) ................................................................................... 210! 6.4.3!Social-psychological dimension: participants’ attitudes to languages inform language choices .................................................................................................................................... 212! 6.4.3.1!Ethnocentric attitude informs the choice of German and interferes with standardization (double standard) ............................................................................................................................. 212! 6.4.3.2!Ethnocentric attitude informs the choice of German for educational purposes (internal forces) .......................................................................................................................................... 213! 6.4.3.3!Local ethnocentric attitude informs the choice of the local language (involving interpretation/translation) ............................................................................................................... 214! 6.4.3.4!Varying attitudes inform language choices for HRM purposes (selective recruitment and staff relocation) ................................................................................................................................. 215! 6.4.4!Social-psychological dimension: participants’ motivations inform language choices (including language acquisition) ............................................................................................ 218! 6.4.4.1!Integrative and instrumental/extrinsic motivations inform language choices (including acquisition) ....................................................................................................................................... 219! 6.4.4.2!Instrumental/extrinsic and intrinsic motivations inform the choice of a local language ..... .......................................................................................................................................... 219! 6.4.5!Social-regulatory dimension: external forces impose languages ................................ 220! 6.4.5.1!International stock market enforces the choice of English within Investor Relations (IR) .......................................................................................................................................... 221! 6.4.5.2!National (host-country) work regulations enforce the choice of (translation into) local languages .......................................................................................................................................... 222! 6.5!Conclusions Drawn from Case Study III: Language Choice at AT_1 Embedded in Social Context ....................................................................................................................................... 223! 7!Language Choice in a Multinational Corporation (MNC): A Case Study (IV) of AT_2 ... 229! 7.1!About AT_2: Facts and Figures ........................................................................................ 229! 7.2!Language Choices on AT_2’s Website ............................................................................. 229! 7.3!AT_2 as a Multilingual Speech Community: Language Functions and Language Choice Norms ......................................................................................................................................... 231! 7.4!Language Choice at AT_2 Informed by Social Context .................................................. 233! 7.4.1!Social-linguistic dimension: participants’ language proficiency informs language choices .................................................................................................................................... 233! 7.4.1.1!Proficiency informs language choices for foreign market expansion .............................. 233! 7.4.1.2!Proficiency informs the choice of English ....................................................................... 234! 7.4.1.3!Proficiency informs language choices (involving interpretation) .................................... 237! 7.4.1.4!Proficiency informs language choices (involving translation) .......................................... 239! 7.4.1.5!Proficiency (and channel) inform(s) code-switching (English, German) ......................... 240! 7.4.2!Social-relational dimension: participants’ role relationships and politeness strategies inform language choices ......................................................................................................... 241! 7.4.2.1!Buyer-seller relationship and politeness strategies inform the choice of (translation into) customer languages .......................................................................................................................... 242! 7.4.2.2!Buyer-seller relationship and politeness strategies inform the choice (and acquisition) of the customer language ...................................................................................................................... 243! 7.4.3!Social-psychological dimension: participants’ attitudes to languages inform language choices .................................................................................................................................... 244! 7.4.3.1!Varying attitudes inform language choices for HRM purposes (selective recruitment and staff relocation) ................................................................................................................................. 244! 7.4.3.2!Geocentric attitude informs (imposes) the choice of English (internal forces) ............... 250! 7.4.3.3!Negative attitude to English and participant role inform the choice of the local language (involving interpretation) ................................................................................................................. 251! 7.4.4!Social-psychological dimension: participants’ instrumental/extrinsic motivation informs the acquisition and passive use of local languages ................................................... 252! 7.4.5!Social-regulatory dimension: external forces impose languages ................................ 253! 7.4.5.1!(Supra)national funding enforces language choices (involving translation) .................... 254! 7.4.5.2!National (home/host-country) authorities enforce the choices of HQ/local languages ....... .......................................................................................................................................... 255! 7.4.5.3!National (host-country) regulations enforce the choice of local languages (involving further training) ................................................................................................................................ 256! 7.4.5.4!Local suppliers enforce the choice of the local language (involving translation) ............ 257! 7.4.5.5!National (host-country) education systems (acquisition planning) influence selective recruitment ....................................................................................................................................... 258! 7.5!Conclusions Drawn from Case Study IV: Language Choice at AT_2 Embedded in Social Context ....................................................................................................................................... 258! 8!Synthesis and Discussion of Findings Across Cases ........................................................ 265! 8.1!Language Choice in four MNCs with Particularities and Commonalities ...................... 265! 8.2!Four MNCs as Multilingual Speech Communities: Language Functions and Language Choice Norms ............................................................................................................................ 269! 8.2.1!Corporate languages and their understandings ......................................................... 270! 8.2.2!Language policy, language guide and their understandings ...................................... 272! 8.3!Language Choice in four MNCs Informed by Social Context ......................................... 274! 8.3.1!Social-linguistic dimension: participants’ language proficiency informs language choices .................................................................................................................................... 274! 8.3.2!Social-relational dimension: participant roles, role relationships, politeness strategies inform language choices ......................................................................................................... 278! 8.3.3!Social-psychological dimension: participants’ attitudes to languages and motivations inform language choices ......................................................................................................... 282! 8.3.4!Social-regulatory dimension: external forces impose languages ................................ 288! 8.4!Summary ............................................................................................................................ 293! 4 | Page the notion of language management in international business contexts at all. In closing this knowledge gap, this study can contribute (empirically and theoretically) to both business and language research communities in an interdisciplinary way. Furthermore, the findings can be primarily applied by MNCs including their language managers or policy makers but might also be insightful to IB practitioners in general. Ultimately, my research hopefully raises language awareness and helps improve international business communication. 1.4 Structure of the Thesis The remainder of the present PhD thesis will be structured as follows; Chapter 2 presents the current state of the art and positions the study both empirically and theoretically within the broader fields of international business and language studies. Empirical and theoretical knowledge gaps are highlighted in order to contextualize and problematize the research question. In Chapter 3 a detailed account of the qualitative research approach and case study methodology is provided. The case study findings based on two Danish and two Austrian MNCs are presented in Chapters 4-7. Chapter 8 synthesizes the results across cases and discusses them in the light of prior research. The final Chapter (9) summarizes the conclusions that can be drawn, i.e. the study’s empirical, theoretical and practical contributions as well as its limitations and avenues for future research. This is followed by the list of references and the Appendix (Interview Protocol). 5 | Page 2 POSITIONING THE STUDY IN THE EXISTING LITERATURE 2.1 Overview of a Cross-disciplinary Field The purpose of this chapter is to present the state of the art and thereby position the study in the existing literature, that is, within business and language research. More specifically, I will look into language choice investigated in international business and management research, organization studies and business communication (with an overall focus on multinational corporations), to which I will refer to as business studies. Also, I will review literature that focuses on language choice in corporate settings studied from a (socio)linguistic perspective, which I will subsume under the umbrella of language studies. Sociolinguistic theory attempts among others to explain how a language choice is made in its (social) context and thus useful for answering the study’s research question (RQ), which centers on language choice in MNC contexts and languages’ manageability (see Chapter 1, Section 1.2). Within these two major fields, empirical research gaps will be highlighted. Furthermore, I will provide first insights into the rationale behind selecting Danish and Austrian businesses for this study, review seminal pieces focusing on Denmark and Austria, and position my research within these national contexts. This will be followed by a more detailed account of the theoretical debates and conceptual challenges as to language choice both within and across business and language disciplines; these debates will be taken as a point of departure and shape the overall sociolinguistic theoretical framework underpinning this research. Overall, both empirical and theoretical knowledge gaps, being the essence of this chapter, aim to contextualize the RQ and put the motivation for this study into perspective. 2.1.1 Language (choice) within business research – language as a ‘barrier’ The cross-disciplinary field of language choice is characterized by various understandings of language and language choice in corporations. First, within the business community, the notion of languages (multilingualism) often has negative connotations. Although occasionally referred to as a facilitator, language(s) tend(s) to be associated with a barrier to e.g. communication in general, market entry, subsidiary control or local operations and overall coordination (Brannen 2004, Luo 6 | Page & Shenkar 2006, Harzing & Feely 2008, Harzing et al. 2011). Language may also be conceptualized as a source of power which potentially has a negative impact on organizational structure and group dynamics or interpersonal relationships (e.g. Marschan et al. 1997, MarschanPiekkari et al. 1999b, Vaara et al. 2005, Brannen 2004). Multilingualism, in particular, is often linked to expenses (e.g. high transaction costs) or business loss, and perceived as an obstacle to achieving business goals, knowledge sharing and efficiency (e.g. Luo & Shenkar 2006, Lauring & Selmer 2012, Neeley 2012, 2013, Neeley et al. 2012). This also shapes the community’s overall understanding of or approach to language choice which often centers around overcoming language diversity. Thus, a commonly studied but also controversial phenomenon is the choice of a lingua franca – often English – in an attempt to facilitate communication. It may be perceived as a profitable choice to ease ‘the language barrier’ (Feely & Harzing 2003) or solution to the problem of multilingualism, and is commonly referred to as ‘standardization’ (Marschan-Piekkari et al. 1999a, 1999b, see also Vandermeeren 1998). As Piekkari & Tietze (2011) most recently summarize: The term ‘language standardization’ (Marschan-Piekkari, Welch and Welch, 1999, p. 379) refers to efforts by top management to instill a common corporate language and harmonize internal and external communications through general rules and policies. It is driven by the assumption that ‘one language fits all’ communication needs. Advocates of this concept share the strong belief that language and language diversity can be managed, i.e. by a centralized single-language policy suggesting lingua franca use within and outside the MNC network. In broader terms, “language design is the product of deliberate language choice” on the part of corporate management (Luo & Shenkar 2006: 321). This standardizing approach to language has been subject to criticism since it is perceived as being too simplistic. In more detail, Piekkari & Tietze (2011) argue that “language use is very much context-dependent and therefore difficult, if not impossible to dictate by general policies” and thus try to “set the agenda for language-sensitive research in international business and management” (p. 267). This has led to an intra-disciplinary “debate on standardization versus contextualization” (ibid.). In this vein, advocates of ‘language contextualization’ have taken up a more ‘language-sensitive’ stance which is open to languages other than English. They object to 7 | Page ‘English only’ and question its effectiveness or applicability to contexts of various natures (e.g. global, national, local, cultural, functional/occupational, situational) (e.g. Fredriksson et al. 2006, Steyaert et al. 2011, Barner-Rasmussen & Aarnio 2011). In other words, the debate within the business community centers on the question whether English can fulfill all communication purposes, whether language can be regulated or managed in such a centralized, standardizing, topdown manner (e.g. Piekkari & Tietze 2011, Janssens et al. 2004, see also Dhir & Gòkè-Paríolá 2002). This discussion and research problem is basically the foundation for my RQ; in order to contribute to this discussion and solve some of the intra-disciplinary issues raised, the thesis focuses on the notion of language choice in context seen from a sociolinguistic perspective. I posit that such an approach can shed light on the circumstances under which English and languages other than English might be chosen and the rationales behind. More specifically, sociolinguistic theory helps identify the social context that defines agents’ language choices and thus helps us better understand how language choices are made in everyday workplace situations. These contextual insights will hopefully enable us to revisit the usefulness of centralized policies (i.e. promulgated by headquarters) in general and single-language approaches in particular, and consequently theorize in a more informed fashion about the manageability of languages in MNCs (see RQ). This in-depth knowledge is helpful to (theoretically and practically) embrace language choice in its integrity, refine language policy and management for business purposes from a new – sociolinguistic – angle, to ultimately improve international business communication. 2.1.2 Language (choice) within language research – language as a ‘resource’ We also find various insights about the role of language and language choice in organizations within the language sciences. More specifically, within sociolinguistics, the notion of language(s) in societal or organizational contexts differs from the one predominant in business research. Language or multilingualism is usually associated with being a resource (rather than a barrier), also in workplace settings (Li 2007, Roberts 2007, Meyer & Apfelbaum 2010). A number of language scholars have investigated the relationship between linguistic resources or language choice and business performance, also referred to as the ‘economics of language’ (Grin 1996, Grin 2007, Coulmas 1992, Dhir 2005, García & Otheguy 1994). For instance, choosing a customer’s or 8 | Page business partner’s first language (L1) has been found to represent a competitive edge that can positively affect profitability and success (Vandermeeren 1998, Grin et al. 2009, Mettewie & Van Mensel 2009, Mettewie et al. 2006, Vandermeeren 1999). Consequently, it is often argued that the English language chosen as a lingua franca (standardization) cannot fulfill all communicative purposes in international business per se (Davignon 2008, Adiyaman 2011), and can in fact lead to the loss of business opportunities (Vandermeeren 1998). Similar insights can be gained from European projects like FLAIR (Hagen 1993), Elucidate (Hagen 1999), ELISE (Hagen 2001) or ELAN6 (Hagen 2006). The ELAN report, for instance, highlights that some of the companies sampled lose trade opportunities due to insufficient language resources.7 Also, the PIMLICO8 report illustrates the positive effects of a company’s (multilingual) language management strategy on export performance and turnover (Hagen 2011). More recently, the DYLAN project explores the dynamics of multilingualism among others in businesses and stresses that actors’ multilingual practices are not necessarily solutions or answers to problems but rather assets for corporations (Berthoud et al. 2013). This means language can be a resource or can offer a competitive edge and is not necessarily the source of a problem. In a similar vein, Spolsky (2009) wonders how it is possible that global businesses are that slow in developing multilingual policies given that international business is not a new development but has existed for centuries. Most interestingly, the author raises some questions that are quite similar to the one elaborated in the previous section. The author writes an entire book about language management where he finally argues that language management requires a detailed understanding of multilingualism and social structure, as well as of multidimensional social and demographic space. […] Because so much of language management produces questionable results, apparently supporting monolingual hegemony and discouraging pluralism and multilingualism, is this not an area (like religious belief) better left to individual free choice? […] We are left then with two basic questions; can language be managed? And if it can, should it be managed? (Spolsky 2009: 260f) 6 Effects on the European Economy of Shortages of Foreign Language Skills in Enterprise 7 Although often cited, the study has also been subject to some criticism such as not being representative of various countries (Grin 2008). 8 Promoting, Implementing, Mapping Language and Intercultural Communication Strategies 9 | Page In this sense, Spolsky’s (2009) views very much align with the ‘language-sensitive’ business scholars’ claims, as detailed above. Together, they support and justify the study’s RQ. A gap specific to language research, more precisely to the area of language policy and planning (LPP) or language policy and management is the scarce investigation of corporate contexts. LPP has traditionally dealt with national contexts such as planning/managing language for entire, usually multilingual, nations (i.e. macro contexts). Micro contexts such as businesses have yet seldom been studied possibly because they are not recognized as “an ‘authentic’ research genre” within the research tradition of LPP (focusing on the national level) (Baldauf 2006: 159). Baldauf (2006) also mentions the greater difficulties in gaining access to corporate than to public or governmental research sites. Similarly, Roberts (2007) notices that the field of language policy and planning is dominated by studies conducted in educational, community or family rather than workplace settings, and also mentions the enhanced difficulty of getting access to business environments. My overall impression of the language field is also that if businesses are investigated, it is mostly smalland medium-sized enterprises (SMEs) (e.g. Hagen 2006, 2011, Incelli 2008, Vandermeeren 1998, Grin et al. 2009, Mettewie & Van Mensel 2009). Furthermore, many of these studies adopt a quantitative methodology, which means they provide more general answers to which languages are chosen (and maybe their effect on business performance), rather than insights into situational or contextual language choice. In this sense, the field of language studies could benefit from qualitative insights into the relatively new micro dimension of business in general and MNC contexts in particular. This also helps to see whether it is feasible to transfer theoretical concepts such as policy or planning/management, initially generated in nation-state contexts, to corporate environments or whether they may have to be refined in one way or another. Overall, the business and language communities have different understandings of language, languages (multilingualism) and language choice, creating this kind of barrier-resource dichotomy. Yet, I believe they fruitfully complement each other, which is why I will try to reconcile the literatures in the framework of this research. In this vein, the present thesis attempts to contribute to both business and language research but in different ways. The field of business studies can benefit theoretically in that an in-depth study taking on a sociolinguistic approach can shed light on the 10 | Page phenomenon language choice in international business in a new – context-sensitive – way. Sociolinguistic theory can help decipher how language choice operates in social context. This helps reassess the usefulness of centralized and English-only policies and the degree to which languages can – if at all possible – be managed. The language community on the other hand can benefit empirically and methodologically from my study in that it provides new insights into one of the still under-researched micro-level areas, i.e. businesses, and into MNC rather than SME contexts. Also, I adopt a qualitative case study design to complement the mostly quantitative language studies and deepen our understanding of contextual language choice. In conclusion, my study engages in both business and language discussions by studying language choice in selected MNCs and thereby addressing the question of whether language can be managed. In the following, I will discuss seminal pieces of research in both fields and focus on their empirical findings. 2.2 Empirical Findings of Prior Research on Language Choice in International Business In an era of globalization, the choice of English is a common phenomenon; one can observe a hegemony of the English language – also referred to as ‘Englishization’ – in many different areas of modern society (e.g. academia, education, supranational institutions, tourism, arts, film and music industry) due to historical (i.e. colonial), political, cultural, social, economic and demographic or military forces with Anglo-Saxon roots (Phillipson 1992, 2003, 2008, Crystal 1997, Maurais 2003, Tietze 2004, Dor 2004, Graddol 2006, Spolsky 2009, Neeley 2012). The area of international business practice and research is no exception, as I shall discuss below. 2.2.1 Language choice in MNCs Many business studies focus on the choice of English as a corporate language in various contexts. In the following, I will discuss seminal pieces of research with a focus on the choice of English in such contexts. The extensive body of literature is twofold; first, a number of international business and management studies shed light on the choice of a company or corporate language in global 11 | Page operators, often studied from a (theoretically) human resource or management perspective (Marschan et al. 1997, Marschan-Piekkari et al. 1999a, 1999b, Piekkari et al. 2005, Vaara et al. 2005); second, there is the (international) business communication stream – with a focus on English chosen as a (business) lingua franca (B)ELF) – where partly applied linguistics, (inter)cultural and discursive frames are adopted (Charles & Marschan-Piekkari 2002, Charles 2007, van den Born & Peltokorpi 2010, Fredriksson et al. 2006, Tange & Lauring 2009, PalmerSilveira et al. 2006, Vollstedt 2002, Haegeman 2002, Louhiala-Salminen et al. 2005, Planken 2005, Nickerson 2005, Bargiela-Chiappini et al. 2007). Although business communication is the overarching term or label of the second stream, and one might think of various languages, there is an almost exclusive interest in English, which also explains the comparatively scarce scholarly attention languages other than English suffer from. Thus, it remains to include languages other than English into current research to unravel the circumstances under which these are chosen and for what reason in order to get a holistic picture of international business communication. In other words, an open-ended approach to language choice is needed. As briefly mentioned in the thesis’ introduction, English is commonly chosen as a corporate language by companies located in Anglophone countries (e.g. Crick 1999, Clarke 2000). SanAntonio (1987) ethnographically studies language choice and related problems from an ethnic identity perspective in a Japanese subsidiary of an American company that chooses English as the company language and prescribes the use of English in all local units. However, in order to compete with the Japanese business network and to attract highly-skilled personnel, the subsidiary hires people without English competences. The American expatriate managers try to enforce the English-only policy by e.g. forbidding the translation of documents into Japanese. The policy becomes a power and control mechanism which negatively affects intergroup relations. Only English-skilled Japanese staff has access to scarce information and better career opportunities (ibid.). This study illustrates that a centralized policy suggesting English only may be difficult to implement at the subsidiary level, especially in environments where the local language and ethnic identity are strongly linked, and English is not widely spoken.9 9 Interestingly, this article is published in a language journal but regularly cited in the business community. 12 | Page Another seminal article in the business literature is Marschan et al.’s (1997) case study of a Finlandbased MNC (Kone Elevators). Adopting a human resource management (HRM) perspective, they focus on the implications of the choice of English as a corporate language for local units and internal communication. The analysis (based on interview data from 25 local units) suggests that the use of English only facilitates communication on the upper management level but not necessarily on middle management and operating levels because of lacking English skills among staff members. The authors thus discuss the dialectical nature of language as a facilitator versus a barrier. Varying or lacking language proficiency can also become a source of power. Languageskilled personnel can control and impede communication by filtering or distorting information. The authors suggest the integration of language into business strategy, the development of language policies and the appointment of a language officer in order to reduce communication problems and improve internal communication. In a follow-up study, Charles & MarschanPiekkari (2002) specifically discuss lingua franca communication problems among the subsidiaries of Kone Elevators. The single case study based on interview data reveals language problems due to decreasing English skills lower down the hierarchy. This leads to the centralization of power on the one hand and the exclusion of those with low or lacking English skills on the other hand. However, the units offer comprehensive training in a variety of languages with different learning objectives based on local needs (ibid.). This study exemplifies that approaches to language training can differ across locations (organizational units) and from the centralized decision promoting a single corporate language. Fredriksson et al. (2006) conduct an interview-based (36 interviews) case study of a German MNC (Siemens AG), including units in Finland and Germany, and investigate the role or rather perception of English as a common corporate language. Although Siemens operates in almost 200 countries, the language of the home country (i.e. German) still has a dominant role. There is no common agreement on whether English or German, or both are the official corporate languages. Also, different languages (including languages other than English) are used in different contexts and required for e.g. top management positions. Based on their findings, the authors question the concept of a common corporate language in that English is neither commonly chosen nor perceived as the common language (ibid.). Their study shows that languages may be perceived differently across locations (organizational units) and staff members (levels of employment).! This 13 | Page means language choices made at the corporate level (such as choosing a corporate language) are subject to individual interpretations and perceptions, which opens up the discussion on language beliefs, ideology and attitudes which I address in further detail later in this chapter. Overall, these studies have provided insights into the choice of English at the corporate level in various business contexts and the implications of such a top-down decision for local units and employees. Various discrepancies between an English-only policy and business practice indicate that languages other than English seem to play a role but we do not know exactly which one. I adopt a more bottom-up, open-ended approach along the lines of Steyaert et al. (2011) who suggest that English (as a corporate language for internal communication purposes) “is not the endpoint of a language policy but forms one of the possible anchor points around which to deal with multilingual complexity” (p. 271). In this sense, the present study aims to complement the body focusing on English and the effects of adopting a corporate language by including languages other than English and exploring the broader concept of language choice, which I will further discuss below. One could argue that the choice of English as a corporate language also represents a form of language policy and/or language management decision. That is to say, both language and business scholars have studied language choice (at both employer/corporate and employee/individual levels) under the umbrella of language policy/management mostly understood as solutions to language and communication problems. (The individual concepts and their similarities and differences in meaning will be explained in detail in Section 2.3.) Within the language community, Kingsley (2010) addresses language choice at different levels, i.e. policy/management and practice levels, and the influence of beliefs in the multilingual workplace. Adopting language policy and management theories (Spolsky 2004, Shohamy 2006, Jernudd & Neustupný 1987), interview data with managers in ten Luxembourg banks shed light on top-down language policy, while questionnaires and focus-group discussions at the employee level from three case banks capture language practices and language beliefs. The comprehensive analysis shows that even in banks, where English has been explicitly managed, i.e. chosen and implemented as a working language, alternative ‘mechanisms’ or choices for different purposes are subsequently pursued (i.e. different language courses, recruitment of language-skilled candidates). This creates a pluralistic top-down implicit policy which also fosters multilingual use practices among employees. Staff both highly 20 | Page 2.2.3 Positioning the study at the cross-road of two non-Anglophone nations The thesis engages in the vivid debate, outlined above, on standardization versus contextual language choices and multilingual language policy frames (contextualization) and explores nonAnglophone MNC contexts contrasting English-speaking and U.S. perspectives (e.g. Neeley 2012). Several calls for comparative research can be found in the business literature. Fredriksson et al. (2006) state that “Anglo-Saxon companies are likely to suffer from certain ‘blindness’ to the language problem due to the current dominance of English as the lingua franca” (p. 411) and suggest that “[f]uture research should therefore compare firms from different language backgrounds in order to assess the possible effects of home-country language and ‘language blindness’” (p. 420). A potential ‘blindness’ of Anglophone businesses has also been stressed by other scholars (e.g. Thomas 2008, Graddol 2006, Crick 1999, Clarke 2000, Hagen 2005). Similarly, Welch et al. (2001) would welcome an in-depth exploration of “language issues in firms from different countries” (p. 206). In other words, “there is a need to extend the analysis into a wider range of company languages, not just those using English, and of parent-country languages” (Marschan-Piekkari et al. 1999b: 437). The controversial discussion in the extant literature in general and these calls in particular have motivated me to study MNCs headquartered in nonEnglish-speaking countries (with different parent-country (HQ) and corporate languages) and their responses to the question of language choice. In search of comparable research contexts, I came across an interesting report entitled Languages mean Business – Companies work better with languages published by the European Commission (Davignon 2008). The report claims that linguistic diversity represents a competitive advantage for businesses and “a lingua franca can never be enough to satisfy every communication need” (Davignon 2008: 3), along the lines of most language studies. In order to support this claim, the author not only refers to the often cited large-scale ELAN study (Hagen 2006), but also to research that has been conducted at the national levels. In this context, two countries, namely Denmark and Austria, are used as a positive and a negative case example respectively, as the following quotations show: The Confederation of Danish Industries carried out a survey among their member companies in 2007 which showed that other languages than English are used by more than a third of all companies and that 4 companies out of 10 have experienced more or 21 | Page less serious communication problems with trade partners in other countries as a result of linguistic deficits. (Davignon 2008: 11) This quotation lets one assume that Danish companies seem to be open towards the choice of languages other than English, and that their business partners lack the language skills needed for doing business globally. The Austrian market, on the other hand, is described in a more negative way: The Austrian Institut für Bildungsforschung der Wirtschaft published a report in 2006 projecting future problems in Austria, which is one of the few countries where the language skills among the general population have actually gone down according to the Eurobarometer. In addition to English, the companies will need Italian and the languages of the trading partners in Central and Eastern Europe, in particular Czech, Slovak and Hungarian. (Davignon 2008: 10) On the basis of the quotation above, one could speculate that the Austrian companies are linguistically not equipped for international trade as the general population speaks fewer foreign languages. Since the EU report (Davignon 2008) does not contain any references, we do not know exactly to which Danish survey and Austrian report it is referred. This has aroused my interest in looking into these two national research communities in more detail. In the following, seminal studies concerned with language choice and related problems from Danish and Austrian perspectives will be reviewed. It will turn out the picture drawn at the national levels is slightly different from Davignon’s (2008) postulates. 2.2.3.1 Language choice in Danish businesses In a Danish context we find a large-scale survey based on 267 Danish businesses, all members of the Confederation of Danish Industry (Dansk Industri). It sheds light on foreign language (FL) needs from the firms’ perspective, attitudes towards foreign language use and language choices in interaction with foreign trading partners. 60 % of the businesses in the sample conduct business in English whereas 8.7 % of the SMEs and 4.9 % of the larger companies use the trade partners’ first languages. Language barriers in international business are encountered with China, France, 22 | Page Germany and Russia. From a Danish perspective though, the implications and problems such as the loss of business opportunities are mostly attributed to the poor English skills of the foreign counterparts. In other words, Danish companies largely do not believe they need to choose any languages other than English (Verstraete-Hansen 2008, see also Verstraete-Hansen 2009). Similar insights can be gained from Millar & Jensen’s (2009) (mixed-method) study on language choice and management in five Danish multinationals, with an interest in the role of English in business and society. The findings reveal a prevalence of English as a lingua franca and corporate language in the Danish business world, also for external communication purposes. The authors conclude that “it is the customer that is seen as having the problem, not the company” (ibid. p. 99), if s/he does not speak English. In other words, perceptions of English predominant in Danish businesses and society color “perceived needs in relation to other foreign languages” (ibid. p. 102). Further qualitative research focuses on the choice of English as a corporate language within Danish MNCs. Lønsmann’s (2011) ethnographic study of a Danish pharmaceutical company sheds light on perceptions of English and Danish (language ideologies) within the company. It turns out that both English and Danish are chosen whereas Danish is considered important since the company is located in Denmark (“one-nation-one-language ideology”, p. 225). This also means that international staff members are expected to learn Danish to facilitate social integration. In this vein, Tange & Lauring (2009) qualitatively investigate the effects of English chosen as a corporate language on social interaction, knowledge sharing and organizational culture in 14 international Danish organizations. They identify a discrepancy between the Danes’ open-minded or positive attitude towards the adoption of a corporate language and their actual language practices; Danish is often used for informal communication, so-called language clustering, which largely prevents international staff from socializing with Danes. These two studies are interesting insofar as they not only highlight the role of English but also the relevance of Danish internally, specifically in the home-country environment (headquarters). More specifically, they stress the social implications of international staff members, i.e. their social integration or exclusion within a setting. By contrast, another Danish study shows that HQ-subsidiary interaction may also include local languages (e.g. French) if English lingua franca communication is not an option because local staff do not speak fluent English (Andersen & Rasmussen 2004). 23 | Page Based on these insights, one could get the impression that Danish companies tend to be ‘Englishfriendly’ when it comes to international business encounters and do not overtly consider languages other than English, unless they ‘have to’, which is more or less the opposite of what is postulated in the EU report (Davignon 2008). Also, within this body of Danish studies, language policy and management is often associated with or even limited to the adoption of English as a corporate language and its implications (see also Sørensen 2005). As a consequence, a scarcity of research into alternative language choices (including languages other than English or Danish, language training, translation/interpretation, etc.) is noticeable. This is a gap the present thesis wishes to address by studying language choice in its broader, multifaceted sense in Danish MNC contexts. 2.2.3.2 Language choice in Austrian businesses In the Austrian context, we find a number of studies, mostly quantitative ones, that account for language choices with regard to both internal and external communications. For example, Archan & Dornmayr’s (2006) study is a country-wide survey based on 2,017 businesses of varying sizes. It focuses on FL needs, language skills of newly recruited staff and language training offered by companies. The main business lingua franca relied on is English (except for interactions with Hungary) and 10 % of the sample have adopted English as a corporate language. Between 17 % and 31 % of the businesses interacting with Latin America, Spain, Italy and France rely on the buyers’ L1s. German may be used in Eastern European and former Yugoslavian markets (ibid.).12 Linguistic adaptation to Romance-language speaking partners is also found in a case study of three Austrian middle-sized exporters (Bäck 2004). The choice of German in Eastern European collaboration is similarly highlighted in a regional quantitative study of Lower-Austrian SMEs bordering the Czech Republic and Slovakia (Schweiger 2008). Another survey gives an overview of the role of foreign languages in 40 Austrian MNCs. The majority of the companies investigated – with an average number of approx. 6,500 staff – have German as a corporate language (52 %). The rest (48 %) have either one (English), two (German 12 See also Ammon (2001) on the role of German as a lingua franca in Europe and Busch (2010) on language change and policy in former Yugoslavia. 24 | Page and English, English and Italian) or three (German, English and Italian or French) corporate languages. 60 % of the firms believe in the ‘very important’ role of foreign languages, in particular those companies with English as a corporate language or multiple corporate languages. In addition to adopting corporate languages, a high number of the firms either support externally organized language classes (79.5 %) or in-house courses (74.4 %), in addition to stays abroad for language training purposes, selective recruitment and participation in international conferences. Professional translators and interpreters are relied on for promotional material, contracts and technical texts, occasionally also for negotiations. Besides English, also Slavic languages, Hungarian, Chinese, Romance languages, etc. are reported to having gained importance since the mid-1990s. In an era of globalization and Eastern European market expansion (fostered by the EU enlargement in 2004), the choice of English is perceived as not being enough (Weber 2008). Overall, it seems that Austrian businesses choose a mix of languages, including the use of English, of customers’ or business partners’ languages and ‘their own’ language (i.e. German). They tend to adopt different corporate languages – which makes them interesting to study (e.g. MarschanPiekkari et al. 1999b) – engage in a number of alternatives to English and may regard foreign languages as important for global business operations. These impressions shed a different light on the ‘Austrian business situation’ than Davignon (2008) does, in particular in contrast to the Danish contexts. Moreover, with regard to research on Austrian firms and methodologies applied, Lavric (e.g. 2008) has repeatedly stressed the lack of qualitative in-depth insights into language choice in business in comparison to the larger body of quantitative studies (e.g. Kubista-Nugent 1996, Rheindt 1997, Seeböck 1999, Daublebsky 2000). This is also reflected in my review. A valuable qualitative contribution is based on SMEs (Bäck 2004). Thus, the research community focusing on Austrian businesses can benefit from the present project in that it is of an exploratory and qualitative nature and includes MNC (instead of SME) contexts. In conclusion, it is argued that a comparison of Danish and Austrian multinationals having different national language backgrounds, i.e. Danish and German, can hopefully yield novel insights into the area of language choice in international business by potentially offering 25 | Page sophisticated or diversified answers to the question of language choice and alternatives to the choice of English and single-language policies. 2.2.4 (Empirical) knowledge gaps summarized The previous sections of this chapter have given an overview of seminal empirical studies shaping a cross-disciplinary field of inquiry and identified several research gaps. First, a number of studies rooted in the wider business community examine the notion of English as a lingua franca or corporate language and are interested in the implications of such a choice for internal communication. This means an extensive body of literature has provided valuable insights into the role of English and related problems in various organizational contexts (Marschan et al. 1997, Marschan-Piekkari et al. 1999a, 1999b, Charles & Marschan-Piekkari 2002, Louhiala-Salminen et al. 2005, Piekkari et al. 2005, Vaara et al. 2005, Fredriksson et al. 2006, Tange & Lauring 2009). These studies or rather the problems and discrepancies identified let one wonder whether English can be enough and what functions languages other than English assume in MNC contexts. In addition, they question the usefulness of centralized choices and the degree to which languages can be managed at all. Second, both business and language literatures have shed light on alternative choices (including both employer/corporate and employee/individual levels) for e.g. internal/external communication purposes, educational or HRM purposes (adoption of several corporate languages, language training and acquisition, reliance on or recruitment of language-skilled individuals, translation/interpretation services, code-switching/-mixing, etc.). The crossdisciplinary field lacks in-depth knowledge about the context in which these various choices are embedded to learn about the manageability of language. Third, with regard to an MNC’s communication, it has been noticed that there is often a strong focus on internal communication. While interunit communication is possibly dominated by English, the external dimension, i.e. host-country interaction with customers, requires the choice of customer languages. Thus, subsidiaries assume the role of mediators between HQ and local markets. In this context, some authors stress that single-language policies do not fit local settings 26 | Page and their encounters with customers. Overall, language choices for external communication purposes need to be empirically investigated, if we strive for a more holistic picture of language choice in MNCs and languages’ manageability. By integrating both internal and external communication dimensions in the present study, I hope to contribute new insights to the field. Fourth, with regard to methodology, several business scholars suggest a comparison of nonAnglophone research sites (Fredriksson et al. 2006, Marschan-Piekkari et al. 1999b, Welch et al. 2001). This may help gain new perspectives on language choices, alternatives to the choice of English and counterbalance Anglophone perspectives (e.g. Neeley 2012). Danish and Austrian businesses are mentioned in an EU report representing ‘good’ and ‘bad’ examples of languageoriented business activities. A closer look into other sources including Danish and Austrian firms has drawn a controversial picture, which renders these national contexts scientifically interesting to compare. Prior studies have shown that the Danish MNCs investigated prioritize English and Danish whereas the Austrian companies tend to rely on German, English and other languages. A cross-national comparative study aims to provide multifaceted insights into language choice in international business, possibly contrasting English-only perspectives. (Further reasons for the adoption of a cross-border design including these headquarters locations will be discussed in Chapter 3, Section 3.3.1.2.) Another methodological observation concerns prior research’s focus on either one or more headquarters or the subsidiary level only. Some studies include both levels, supposedly for (unit) triangulation purposes, but do not explicitly differentiate between language choices made at the various organizational levels (HQ, divisions, subsidiaries) in the discussion of their findings. Since many of the extant findings do yield differences across the locations (settings) resulting in discrepancies of various kinds, the distinction seems crucial and will be considered in the present study, in order to advance the area of language choice in the context of MNCs. In conclusion, we have gained an overview of how the notions of language and language choice have so far been empirically investigated in different organizational contexts. The theoretical discussions within the field are dealt with in the next sections of this chapter. 27 | Page 2.3 Theoretical Discussions in a Cross-disciplinary Field In the following three sections, I will discuss the research phenomenon language choice theoretically and elaborate on the theoretical frameworks that underlie this research. First, the review of a number of studies and their empirical findings, as presented in the previous sections, has highlighted that language choice in international business is not a clear-cut phenomenon. Authors within and across business and language communities seemingly have difficulties conceptualizing language matters in organizational contexts, which has stirred some conceptual confusion, which I will attempt to elucidate by adopting the overarching concept of language choice. Second, I will conceptualize language choice in IB, and MNCs particularly, from a sociolinguistic perspective. This includes a detailed description of an MNC’s idiosyncrasies from a sociolinguistic angle and the various language choices that are made within the MNC network. The third and last part of this chapter focuses on a number of primarily sociolinguistic theories, combined with concepts from linguistic anthropology, applied linguistics and social psychology, that provide possible factors that influence language choices made in their social context. The chapter finishes off with a summary of the theoretical knowledge gaps. 2.3.1 Language choice made at different levels: language policy/strategy, planning/management, practice – useful distinctions or conceptual confusion? If one reviews the phenomenon language choice in an interdisciplinary fashion, one discovers that multiple conceptualizations of language choice can be found in both language and business communities, as also reflected in the previous sections of this chapter. In fact, we are dealing with a myriad of concepts and terms that have both similar and different meanings within and across disciplines, e.g. (explicit, implicit) language policy, (simple, organized) language management (solutions), language barrier solutions, (discursive) practices and language use can be found (e.g. Kingsley 2010, Nekvapil & Nekula 2006, Spolsky 2009, Feely & Harzing 2003, Harzing et al. 2011, Steyaert et al. 2011). In addition, the definitions not only partly overlap but may also suffer from vagueness or implicitness. 28 | Page One theoretical problem within the extant literature is that authors have tried to conceptualize language choice at different levels, i.e. at the employer/corporate and the employee/individual levels, which creates a distinction between the policy/management and practice dimensions. Yet, this distinction does not seem totally clear-cut. Prior research, as reviewed earlier, has shown that the dimensions do overlap in that policies or management may have an implicit character and thus resemble practices rather than top-down policies or management activities. Also, policy may not be approved by all staff members and thus be disregarded, etc. In this sense, the conceptual distinction (between policy/management and practice) may not be useful in a business context after all. This has led me to adopt a more open, inductive approach to the subject matter and start out with the overall concept of language choice. This will allow me to see at which level a choice is made, by whom, in which communicative situation and for which reason. In this way it will be possible to see whether policy/management can de facto be distinguished from practices, that is whether language can be managed at all within an international business context (see RQ – Chapter 1, Section 1.2). Terms and concepts like policy or management have their roots within sociolinguistics, as only briefly touched upon in the introduction of this chapter. More specifically, the concepts originally embrace language choice at an institutional level, where authoritative bodies such as governments choose one or more languages and promote or declare them as e.g. ‘national’, ‘minority’ or ‘regional’ languages, in an attempt to regulate or change language use among participants (practices) within a (multilingual) speech community such as a (multilingual) nation (Fishman 1987, Ising 1987, Berthoud & Lüdi 2011, Coupland 2010b, De Swaan 2010, Ricento 2013). These overt or deliberate interventions on language use have been studied in the research tradition of language planning and language change (e.g. Haugen 1987, Cooper 1989, Jernudd & Das Gupta 1971, Rubin & Jernudd 1971), and language policy (e.g. Spolsky 2012a, Ricento 2006, Shohamy 2006, Phillipson 1992, Skutnabb-Kangas & Phillipson 1995), all contributing to varying degrees to a field which is today mostly known as language policy and planning (LPP) (Ricento 2006, 2013, Baldauf 2006, Jernudd & Nekvapil 2012) or language planning and policy, depending on where the focus lies (Deumert 2009, Spolsky & Lambert 2006),13 and increasingly 13 Language policy and planning (LPP) has emerged as an integrated field in an era of decolonization and nation building in the 1960s, when newly independent African and Asian countries needed to choose or rather agree on a 29 | Page also language policy and management (Spolsky 2009). The broader field is characterized by various concepts. Spolsky (2012b, 2009, 2004) summarizes that over time authors have defined and used similar concepts (e.g. language policy, language plan, language planning, language intervention, (organized versus simple) language management, language engineering, language treatment) in different ways or interchangeably. One could add language regimes (Coulmas 2005a) or the regulation of linguistic behavior (Seargeant 2009) to the list. There are also scholars who have used and still prefer the general term language choice, also for discussing it at a broader societal scale, instead of policy, planning or management (Hunt 1966, Romaine 2000). The broader field is a branch of sociolinguistics, possibly studied under the umbrella of (societal) multilingualism and language contact (Wodak et al. 2011), applied sociolinguistics (Mesthrie 2011, Coulmas 1997, Fishman 1971), applied sociology of language (Fishman 1972b) or understanding and intervening (Wardhaugh 2006). Since they are somewhat similar and interrelated concepts, it is difficult to draw a precise definitional line between policy and planning. While the definition was originally based on Kaplan & Baldauf’s (1997) work, which drew on ideas from Rubin & Jernudd (1971), Baldauf (2006: 19f) revises previous definitions as follows: Traditionally, language planning has been seen as the deliberate, future-oriented systematic-change of language code, use and/or speaking, most visibly undertaken by government, in some community of speakers. Language planning is directed by, or leads to, the promulgation of a language policy(s) – by government or some other authoritative body or person. Language policies are bodies of ideas, laws, regulations, rules and practices intended to achieve some planned language change (Kaplan & Baldauf, 1997: 3). Language policy may be realised in very formal (overt) language planning documents and pronouncements (e.g. constitutions, legislation, policy statements, educational directives) which can be either symbolic or substantive in form, in informal statements of intent (i.e. in the discourse of language, politics and society), or may be left unstated (covert). While the distinction between language policy (the plan) and language planning (plan common ‘national’ language (Ricento 2013, Deumert 2009, for a historical overview see Dhir & Gòkè-Paríolá 2002, Jernudd & Nekvapil 2012). This was considered necessary in order to “improve efficiency, develop nationalistic attitudes, and promote economic development” in times of modernization (Ricento 2013: 525). On the one hand, it seems LPP has become an integrated field of study due to the interrelationship of the concepts policy and planning, which I will define later in this section. On the other hand, originated as two separate disciplines both language policy and language planning still have their own journals (e.g. Language Policy, Current Issues in Language Planning) with different agendas, that is, policy and planning are not necessarily studied in consolidation. There does not seem to be a consensus on what the field is called. What seems to be commonly agreed upon is that the broader field of LPP lacks an overarching theory due to the complexity of matters involving language in society (Ricento 2006). 36 | Page documented in explicit documents or statements but need to be investigated in their implicit nature. This aligns with the implicit nature of policy, as discussed earlier (e.g. Baldauf 2006). The business scholars Feely & Harzing (2003) conceptualize language management as solutions or options undertaken by the employer at the corporate level which aim at overcoming the language barrier (as between headquarters and their national subsidiaries) (see also Harzing & Feely 2008). MNCs can then formulate a language strategy based on the range of “options for managing language problems” (Feely & Harzing 2003: 42). Among those are e.g. the use of a lingua franca, functional multilingualism, translation and interpretation services. The impact or intervention on employees’ language use is not included. As also mentioned in the literature review, later on Harzing et al. (2011) omit the term management but adhere to language barrier solutions. (As mentioned above, not all matters are probably about a language problem or barrier and a fitting solution; in a multilingual setting like the MNC, linguistic borders need indeed to be crossed but they are not necessarily associated with a barrier. The question might just be which language to choose in which context.) Among language scholars, one also finds the notion of ““solutions” for overcoming potential problems” (Lüdi et al. 2010: 212), such as the use of a lingua franca, receptive multilingualism, or translation and interpretation, but they are categorized as ideologies and practices of communication (ibid., cf. Nekvapil and Nekula 2006). These solutions may also be grouped as foreign language use in businesses: asymmetrical use (buyer’s or seller’s language), symmetrical use (e.g. lingua franca, polyglot dialogue) and translation/interpretation (Vandermeeren 1998). This extensive overview illustrates the complexity of the subject matter. Given its cross-disciplinary nature, one would expect inconsistencies across research disciplines. Though the terminological and conceptual variety or multiplicity also exists within the communities, i.e. language and business research, which renders the clarification of ideas even harder. The terms and concepts have different meanings for different authors, their definitions overlap with one another, and no terminological consensus has yet been found in the research communities. More specifically, regardless of the research tradition they come from, both language and business scholars seem to struggle with distinguishing between policy (statements), strategy, planning, management, solutions, interventions and partly also practices and ideologies. To reach some conceptual clarity, I will use the inclusive term and concept language choice which does not imply per se any form of 37 | Page agency, hierarchal dimension or intervention in the form of policy or management but covers both the policy and practice levels. To provide an example; translation and interpretation services are often grouped as (organized) language management (e.g. Nekvapil & Nekula 2006, Hagen 2006, Feely & Harzing 2003) or some formalized solution to language problems (e.g. Harzing et al. 2011). But what if a subsidiary once chooses Chinese as the target language of a business letter and sends it to a local translation agency, or if the individual employee translates/interprets for her/his boss on an occasional basis? They are all translation/interpretation activities but they are not necessarily formalized in any way. Another problem concerns the dual purpose of most language management definitions; i.e. management intends to solve language or communication problems (e.g. by translating information) and to change language users’ practices (individual choices). One might ask in which way the reliance on external translation agencies, even if this might be a formalized solution, de facto influences or intervenes with employees’ skills and choices or practices. The fact that some documents are translated into different languages (as a result of language choices) may not interfere with staff’s language choices at all. So, it might be more useful to pay special attention to the context and specify who chooses which language in which setting for which purpose in order to advance the field in micro contexts such as business environments, and to reassess the role or usefulness of policies and management, i.e. the manageability of languages, in workplaces (i.e. MNCs) and thereby possibly help improve international business communication (see also RQ). Besides, critical voices asking whether language can and should be managed can be heard in both language (Rubin & Jernudd 1971, Deumert 2009, Spolsky 2009) and business research communities (Piekkari & Tietze 2011). In order to study language choice in context, more precisely, how language choices are made in MNCs, this research adopts a sociolinguistic framework, which will be discussed below. ! 2.3.2 Language choice in MNCs: a sociolinguistic approach My decision to adopt a sociolinguistic approach to this study has been inspired by some business scholars. For instance, Brannen et al. (2012) argue that the business community needs “interdisciplinary insights gained from a plethora of fields such as [...] linguistics, [...] in order to generate genuinely innovative frames of reference for understanding the role of language in 38 | Page international business”. Similarly, Harzing & Feely (2008) invite sociolinguists to contribute to explaining how language (and language barriers) impact on international business and management. Most of the business studies investigating language choice and related matters rely on frameworks rooted in communication, culture, management and organization, postcolonialism, (social) psychology, strategy, to mention a few, occasionally (socio)linguistics (e.g. Harzing & Feely 2008, Selmer 2006, Brannen 2004). By adopting a sociolinguistic perspective in the present study, it is hoped to shed light on language choice in MNCs in a new way, generating insights of use to the broader business studies community and advancing the field of international business in an unconventional way, i.e. by stressing the social dimension and nature of language choice. More specifically, we may get a more sophisticated picture, that is a context-specific, in-depth understanding of how the firm’s participants or agents choose languages in a social context. In other words, once we know more about the contextual dimensions that influence language choice, we will hopefully learn whether language can be managed by the employer. In this sense, sociolinguistic principles are believed to answer the study’s RQ (see Chapter 1, Section 1.2). Broadly speaking, the field of sociolinguistics is concerned with the study of language in society (Romaine 2000, Holmes 2008). It has also been described as the study of speakers’ choices (Coulmas 2005b), the study of language in its social context (Spolsky 2011) or the science of language and society (Ammon et al. 1987, see also Peng 1982). 18 When talking about language, sociolinguists usually distinguish between a language (e.g. English, Chinese, Spanish) and its varieties of national, regional, social/ethnic, specialized, etc. nature within a single language (e.g. British English, Boston English), together making up a code (Fishman 1968, 1972b). A code can be described as “a culturally defined, rule-governed system of shared arbitrary symbols that is used to transmit meaning” (Cooley 1983: 242). Cooley’s (1983) broad definition includes both oral and written language (i.e. media) as well as paralinguistics (e.g. 18 Looking at the history of sociolinguistics, important names that have substantially shaped and contributed to the field with their areas of expertise are Charles Ferguson (sociolinguistics), Joshua Fishman (the sociology of language), William Labov (language variation and change), Basil Bernstein (codes and social class), Dell Hymes (the ethnography of communication/speaking) and John Gumperz (interactional sociolinguistics) (see Wodak et al. 2011 for an overview), but also Einar Haugen (language planning) and Uriel Weinreich (language contact) (see Spolsky 2011). Spolsky (2011) explains that the field is sometimes called sociolinguistics (Fishman 1971) and sometimes the sociology of language (Fishman 1968, 1972b), terms which might be used interchangeably, as Fishman seems to do himself (cf. Fishman 1971, 1972b). At times they are demarcated in that the sociology of language is associated with a stronger focus on sociological aspects (Grimshaw 1987). 39 | Page pitch), nonverbal cues (e.g. body language) and discourse encompassing among others the choice of language – which is this study’s main interest. A code also comprises language style (e.g. level of formality in language usage) and language register often related to occupational areas (e.g. legal, technical, medical fields or vocabularies) (Scotton 1983, Holmes 2008, Romaine 2000). When talking about society, the concept may be associated “with dyadic encounters, small group interaction, large group functioning, the articulation of social classes and sectors, contacts and contrasts between entire nations, etc.” (Fishman 1968: 5). Based on the above, sociolinguistics can be described as “the study of the characteristics of language varieties, the characteristics of their functions, and the characteristics of their speakers as these three constantly interact, change, and change one another within a speech community” (Fishman 1971: 4). It also “seeks to discover the societal rules or norms that explain and constrain language behavior and the behavior toward language in speech communities” (Fishman 1971: 3). This also implies that even at a societal level languages are not necessarily overtly planned/managed, and that explicit language policies may not always exist. Yet, language users are acquainted with the community’s behavior norms, including language choice. Put simply, sociolinguistics aims to explain why individuals speak in different ways in different social contexts, identify the social functions of language and discover how language is used to convey social meaning (Holmes 2008). The overall goal of sociolinguistics is “to move towards a theory which provides a motivated account of the way language is used in a community, and of the choices people make when they use language” (ibid. p. 13). This makes it eminently suited to studying language choice in a social context such as the workplace in general and an MNC network in particular. In other words, sociolinguistic theory helps get a preliminary understanding of language choice in MNCs (see RQ – Chapter 1, Section 1.2). Generally speaking, sociolinguists try to answer the question ‘who communicates in what language (or variety) to whom, where, about what, and for what reason?’ within a speech community (Fishman 1965, 1972b, Holmes 2008). That is to say, participants (language users) – sometimes divided into senders/addressers and receivers/addressees/recipients of information – choose a language (or languages), including a medium (written, read or spoken/oral) and a channel (e.g. email, letter, telephone) in a 40 | Page communicative situation or event (Hymes 1968, Fishman 1971, 1972b),19 within a setting of interaction (location), in order to communicate about a subject matter (topic), with a certain function and purpose (Hymes 1968, 1972, 1974, 1987, Holmes 2008).20 Applying a sociolinguistic framework to the present study enables to describe and distinguish between different participants (e.g. employer, employee, customer, etc.), languages with different functions (e.g. lingua franca), communicative events/situations (e.g. conference, meeting), the setting of the workplace including different physical locations (e.g. MNC’s home country) and different purposes (e.g. internal versus external communication), all forming a multilingual speech community, i.e. the MNC, which I will further detail below. 2.3.2.1 MNCs as multilingual speech communities and language functions From a sociolinguistic perspective, the multinational corporation (MNC) can be conceptualized as a multilingual speech community. A speech community, as also mentioned earlier, can be defined as a group of people that “share at least a single speech variety” (Fishman 1972b: 22), that communicate regularly and frequently, and share a set of rules and conventions (norms) about language usage (Gumperz 2009). (I will elaborate on community norms later in this chapter.) It 19 I do not exclusively nor entirely rely on Hymes’ (1968) model of the ethnography of speaking, rather take the elements relevant for this study, and combine them with other key terms commonly used within sociolinguistics (see e.g. Fishman 1972b on the use of medium or media). 20 The terms function and purpose are hard to distinguish, often used interchangeably, and can also have different meanings within sociolinguistics; for instance, the function of speech (i.e. utterances) may be of e.g. expressive/affective, directive or referential/informative nature (Hymes 1968, Holmes 2008). The purpose or function of a communicative event or situation (e.g. meeting, job interview) is the reason(s) for the interaction such as organizing a business meeting, applying for a job, attracting customers, etc. In this sense, purpose/function may also overlap with or is closely related to the notion of topic (Holmes 2008). By contrast, the function of an entire language or variety within a linguistic system can also ‘stand for’ the language of one nation, of law/court, education or trade, etc. In this sense, languages serve national, international, official, educational or commercial functions; adding to the confusion, these types of functions are often also associated with languages’ statuses or roles, the latter being less formalized. Terms like function, status or role are used in various ways and difficult to distinguish; note, Esperanto serves the function of an auxiliary language, English has the status of a world language, or German has the role of a lingua franca in Eastern Europe (Ammon 1987). Assigning national or other official status to a language may also involve some form of national language policy and planning in a speech community, also known as status planning. I will primarily adhere to the term function capturing at best the sometimes implicit or not well-defined role of languages within the MNC and employment domain. The literature, for instance, has shown that the meaning or interpretation of a corporate language is ambiguous in that it does not seem to be perceived or associated with an official status (as the status of a national language). Thus, the notion of function seems more appropriate for describing the nature of e.g. a lingua franca or corporate language in an MNC context. Furthermore, in this study the purpose of a language choice grasps what the language (with a certain function) is chosen for; e.g. English in its function as a corporate language is chosen for internal communication purposes. 41 | Page has traditionally been associated with a nation or region where the social rules or conventions imply the use of the national or regional language within the community (Gumperz 2009, Gumperz 1968). But the concept may be applied to any other unit or entity of language users such as a workplace, company or other organizational context (Spolsky 2009). As opposed to a language community (e.g. English, German speakers), a speech community is not necessarily monolingual but may be diglossic (bilingual), heteroor polyglossic (multilingual), or characterized by a number of language varieties (Fishman 1972b, Romaine 2000, Ricento 2013). Similarly, in the field of international business and management, the MNC is often described as a ‘multilingual’ organization (Luo & Shenkar 2006, Fredriksson et al. 2006, Steyaert et al. 2011, see also Thomas 2008). This is because the MNC network consists of geographically dispersed organizations including a parent company or headquarters (HQ) and the national subsidiaries (e.g. Ghoshal & Bartlett 1990), which are all embedded in different linguistic environments (Thomas 2008).21 2.3.2.1.1 HQ language, local and customer languages These geographic and linguistic idiosyncrasies of an MNC also provide languages with different functions. The headquarters is located in a home country and dominated by its national language, which I will label ‘HQ language’ (cf. Thomas 2008, Fredriksson et al. 2006, Luo & Shenkar 2006, Harzing & Pudelko 2013). The subsidiaries’ locations are usually referred to as ‘host countries’ (e.g. Ghoshal & Bartlett 1990). The languages primarily chosen at subsidiaries assume the 21 In the field of international business, MNCs’ environments usually represent industries. A global environment or industry, for instance, requires a high degree of global integration (i.e. centralized, globally scaled operations) and little local responsiveness (i.e. decentralized, national operations aligned with local needs). A multinational/multidomestic environment promotes a strong focus on local responsiveness and a weak on global integration. The international environment in which both dimensions (integration and responsiveness) are weak fosters dissemination of parent-company (HQ) knowledge. There is also the transnational environment requiring an integrated strategy; the transnational approach tries to simultaneously unite global efficiency, flexibility in responding to domestic differences and needs as well as worldwide learning or knowledge exploitation (Bartlett & Ghoshal 1998, Ghoshal & Nohria 1993, Harzing 2000a, Harzing 2000b). The different environments may foster different structures (i.e. organizational designs) and corporate strategies impacting on interunit communication and language choices in one way or another (Luo & Shenkar 2006). Research has shown that it could also go the opposite way, that communication flows and language affect an MNC’s organizational structure. That is, a ‘shadow’ structure based on language develops which is different from the firm’s formal one. This means units are officially grouped together geographically (e.g. Southern Europe including Belgium, France, Italy, Luxembourg, Spain) but on a more informal basis they are clustered according to languages or language groups with varying language distance from the headquarters (Marschan-Piekkari et al. 1999b). 42 | Page functions of ‘local languages’ (cf. Luo & Shenkar 2006, Thomas 2008, Harzing & Pudelko 2013). Outside of or external to the MNC one finds independent buyers, clients, customers, business partners and agents or joint ventures in home-country and host-country markets that all ‘speak’ different languages whose function will be referred to as ‘customer languages’. (In this study’s case, HQ, local and customer languages are primarily languages other than English except for Anglophone units or subsidiaries and buyers/customers, of course.) 2.3.2.1.2 Lingua francas and corporate languages Apart from HQ and local/customer languages, it is necessary to elaborate on alternative functions such as a lingua franca which is discussed in both language and business communities (e.g. Vandermeeren 1998, Feely & Harzing 2003). Rooted in applied linguistics, the widely known and broadly used notion of a lingua franca – originally a pidgin for commercial purposes22 – describes “all instances of using a language different from the speakers’ mother tongues for specific purposes” (Knapp & Meierkord 2002a: 9). In other words, “when we find a language which is commonly used by people whose native languages are different, we describe it as a lingua franca […] used to communicate across linguistic barriers” (Samarin 1968: 661). In international business this lingua franca is very often the English language (Coulmas 1992, Thomas 2008, Fredriksson et al. 2006, Ammon 1991, St John 1996), possibly representing the only ‘shared’ language among interactants (Feely & Harzing 2003, Nekvapil & Nekula 2006).23 Usually ‘English as a lingua franca (ELF)’ (e.g. Seidlhofer 2001) ‘belongs’ to nonnative speakers (NNSs) of English.24 The concept applied to business or cross-border mergers may also be referred to as ‘Business English Lingua Franca (BELF)’ and BELF users as “communicators in their own right – not ‘non-native speakers’ or ‘learners’” (Louhiala-Salminen et al. 2005: 404). These 22 Samarin (1968) offers a detailed overview of the existing kinds of lingua francas and their different purposes; they may serve as trade, contact, international languages, etc. 23 Apart from lingua franca, English or its function could also be associated with the terms ‘global’ or ‘world’ language and ‘Englishes’ (accounting for the multiplicity of English varieties) (Graddol 2006, Phillipson 2001, Crystal 1997). 24 ELF is not only an applied language management concept but a whole research discipline in itself with its own journal (Knapp & Meierkord 2002, Seidlhofer 2002, Mauranen 2005, Mauranen & Ranta 2009, House 2003, House 2007). In addition, it is often the research focus in the (international) business communication literature and related areas (Vollstedt 2002, Haegeman 2002, Louhiala-Salminen & Charles 2006, Nickerson 2005, Planken 2005, BargielaChiappini et al. 2007, Palmer-Silveira et al. 2006). 43 | Page classifications do not seem theoretically useful for the present study. Regardless of whom English ‘belongs’ to, business communication in English is very likely to be heterogeneous, occurring across different groups of speakers. This means we can hardly study isolated (B)ELF communication since also English native speakers (NSs) might have to communicate with NNSs of English within and across the multilingual MNC borders. For the purpose of this research, ELF simply describes the phenomenon where a shared language is used or needed; this comprises not only NNSs or English learners but also NSs in business contexts with different levels of English skills. The varying fluency of speakers presumably impacts on communication, in that it facilitates or impedes mutual understanding (Phillipson 2008, Steyaert et al. 2011). One exception applies; if it concerns NSs only, I will adhere to the label ‘English as a native language (ENL)’ (Seidlhofer 2004).25 Apart from a lingua franca, a language can also assume the function of a corporate language. As the literature review has already illustrated, English (or any other lingua franca) may be chosen as the ‘(shared) company language’ (Marschan et al. 1997, Marschan-Piekkari et al. 1999b) or ‘(common) corporate language’ (Marschan-Piekkari et al. 1999a, Thomas 2008, Feely & Harzing 2003, Harzing et al. 2011), which could be defined as follows: A common corporate language (CCL) is the language of global operations. This is the language in which official information is transferred between subsidiaries and their parent organisations. Due to the spread of English as a global language of communication, CCLs are frequently English. (Thomas 2008: 309) The choice of a common corporate language also aims to “harmonize internal and external communications through general rules and policies” (Piekkari-Tietze 2011: 267), also referred to as ‘(language) standardization’ (Piekkari-Tietze 2011, Marschan et al. 1997, Marschan-Piekkari et 25 Furthermore, it has to be mentioned that ELF is often considered a ‘neutral’, culture-free language detached from Anglophone culture and ELF users’ own (i.e. L1) cultural norms and understandings (Crystal 1997, Krumm 2003). Louhiala-Salminen et al. (2005) conduct research in two Swedish-Finnish mergers and identify different, culturebound discursive practices regarding the level of politeness, directness and the amount of chitchat in their ELF interactions. In the present study, it is therefore assumed that in practice users do have different cultural backgrounds that affect ELF communication in one way or another (Henderson 2005, Meierkord 2002, Phillipson 2008), but I do not investigate this in a systematic manner. 44 | Page al. 1999a, 1999b).26 Furthermore, a corporate language intends to facilitate corporate interunit knowledge transfer, formal reporting and corporate documentation or to harmonize information systems (Vollstedt 2002, Feely & Harzing 2003, Marschan-Piekkari et al. 1999a, BarnerRasmussen & Björkman 2007). The notion of a common corporate language has also been challenged by prior research (Fredriksson et al. 2006, Millar & Jensen 2009). Staff members may not perceive it as common, i.e. do not commonly share or approve this idea of having one common language. Similarly, Sørensen’s (2005) survey results among 70 corporations based in Denmark show that having English as an (official) corporate language “is no indicator of the scope nor the extent of its English usage” (p. 67). Also conceptually, he critically notices that “there is no collective perception of the concept, its scope, and how various aspects of CL [corporate language] interact” (p. 68). I will follow Sørensen’s (2005) terminology and use the term ‘corporate language’ which is not limited to the English language but can be applied to any language. Besides, there might be more than a single language, possibly two or even three languages assuming the function of corporate languages within an MNC (Weber 2008, Fredriksson et al. 2006). Based on the above, it seems the different functions and choices are conceptually both similar and different; whether English being assigned the function of a corporate language is a form of policy and way of formally standardizing or harmonizing language choice (for internal and/or external purposes), or whether a corporate language merely describes the function it assumes within an MNC and the way it is used, is unclear and needs to be confronted with empirical evidence. Overall, MNCs as multilingual speech communities are characterized by multiple languages assuming different functions, internal and external communication dimensions, as illustrated below (see Figure 2-1). The nature of language choices in different communicative situations will be detailed in the following sections. 26 Note, in the LPP community the term is used in its broader understanding of regulating or standardizing language usage and structure by means of corpus planning (e.g. Ricento 2013), which is not necessarily limited to the English language if one considers multilingual speech communities. 45 | Page Figure 2-1: MNCs as multilingual speech communities 2.3.2.2 Language choice classifications (from a seller’s perspective) Aligned to the above, Vandermeeren (1998) classifies language choice in seller-buyer interaction which is relevant to this study and can also be applied to MNCs. If a company chooses the first languages (L1s) of its customers – or customer languages as I call them – the language choice can be classified as ‘linguistic adaptation’ resulting in an asymmetrical communication. This kind of language choice has also been described as ‘language accommodation’ (Giles & Powesland 1975, cited in Holmes 2008, see also Callahan 2006).27 If the business chooses its ‘own’ language (HQ 27 Similarly, Giles & Powesland’s (1975, cited in Holmes 2008) speech accommodation theory (SAT) encompasses different speech strategies from the sender’s perspective, namely speech divergence or convergence. Divergence describes the phenomenon where a speaker wants to mark social distance, i.e. distance herself/himself (from the interlocutor) by deliberately choosing a language other than the addressee masters. This language behavior can be driven by e.g. political motives, negative attitudes towards the other person or group of speakers, etc. In a workplace environment, this could mean that a corporation intentionally chooses English or any other language although it is well known that (the) counterpart(s) or addressees do not speak this particular language. By contrast, speech vs. External Communication MNC Network External interactants in home market (customer language) HQ in home country (HQ language) External interactants in host market 2 (customer language 2) MNCs as Multilingual Speech Communities Internal (Interunit) S 2 in host country 2 (local lang. 2) Subsidiary 1 in host country 1 (local language 1) Language choice Lang. choice Lang. choice Lang. choice Lingua franca / corporate language 52 | Page 2.3.2.3.5 Language choices for HRM purposes (selective recruitment, staff relocation) Selective recruitment Language(s) may also be chosen for HRM purposes, namely recruitment. Different languages may be chosen as relevant for different positions. In other words, it may be decided that for certain tasks or responsibilities specific language competences are required and job applicants’ linguistic repertoires need to account for that. For instance, native speakers of a particular language or candidates with various foreign language skills may be hired, which is referred to as selective recruitment (Hagen 2006, Feely & Winslow 2005, Feely & Harzing 2003, see also Reeves & Wright 1996). In other words, languages or language proficiency can become a hiring criterion for a job at hand. In this sense, the recruiting executive or employer chooses one or more languages for a given position or job profile. In-/expatriation of staff (language nodes) Another language choice with an HRM purpose is the strategic positioning of personnel within the MNC network via staff transfers; language-skilled staff members may be inpatriated or expatriated with the purpose of facilitating (interunit) communication (Feely & Harzing 2003, Thomas 2008, Hagen 2006, Feely & Winslow 2005). Inpatriates are staff coming from local units to the headquarters, whereas expatriates are personnel leaving the home country and being assigned to subsidiaries in host countries. Employees being neither homenor host-country nationals will be referred to as third-country nationals or international staff that can be located at both HQ and subsidiaries (Harzing 2000b, Tayeb 2000). Marschan-Piekkari et al. (1999a) discover that an expatriate might be fluent in the HQ language and thus be in charge of the communication with HQ during the assignment. Prior to or during the stay abroad s/he may also acquire the local language if it is deemed necessary for the local position. After repatriation – redeployment at HQ – these local language skills may prove useful for post-assignment HQsubsidiary communication. They become “language nodes” (ibid. p. 377), which also puts them in an influential and powerful position, going beyond their initial responsibilities. Based on their case-study insights, Marschan-Piekkari et al. (1999) conclude that language plays an important role in the area of IHRM which is not always recognized by corporate management. Consequently, the 53 | Page authors argue that language should have a more strategic position in the MNC by incorporating language proficiency into HR policies, more specifically staff recruitment, selection, training or transfers, etc. (These bior multilingual employees are also described as “bridge individuals” (Harzing et al. 2011: 281), “key personnel” or “critical gatekeepers” filtering information where language skills actually represent a source of power (Piekkari et al. 2005: 340f).) When comparing conceptual with empirical insights, the question arises whether in-, exor repatriates are deliberately positioned as ‘language nodes’ due to their language skills or whether they simply become these mediators during their assignments out of contextual or circumstantial necessity. The business literature suggests that they become language nodes and concludes that they could be positioned more strategically (e.g. Marschan-Piekkari et al. 1999a, Feely & Harzing 2003). Also, Hagen’s (2006) language research shows that corporations do in-/expatriate staff but the underlying rationale does not seem to be related to language matters. In other words, the empirical insights reflect more the linguistic implications or ‘side-effects’ of in-/expatriation rather than a strategic positioning of language-skilled staff. We need further empirical findings to better align them with conceptual ideas. In sum, HRM is an area that may include language choices with various purposes which need to be looked into in a more contextual fashion. The language choices of various kinds made for different purposes, as described above, are of a rather descriptive and applied or practice-oriented nature; that is to say, they give an overview of the language choices that are possibly made in an MNC but they do not tell us anything about the underlying rationales for a given choice. In the following, I will therefore elaborate on a number of frameworks that attempt to grasp the interaction between language choice and social context. 2.3.3 Language choice embedded in social context: a working framework As mentioned earlier in this chapter, sociolinguistic theory helps build a working framework that captures how language choice operates in social context. I will exemplify how I apply the frames to the MNC context. 54 | Page 2.3.3.1 Language proficiency and norms of a speech community In the present study, participants’ language proficiency describes language users’ native or first language (L1) and possibly nonnative or foreign language (FL) skills (i.e. L2, L3, L4, etc.) forming or adding up to the linguistic repertoire of a speech community (Gumperz 1968, 2009, see Llurda 2000 on proficiency, competence and ability). For my purpose, the notion of proficiency merely describes the reported, perceived and socially constructed skills, variation or perceived limitations but does not encompass any form of objective evaluation or formal assessment. The way language proficiency operates in the speech community can be captured by Scotton’s (1983) work. She develops a theory of markedness and code choice that accounts for community members’ “variation in linguistic code choice”, i.e. why people use different codes in interaction. The model relies on the premise that participants in conversation interpret all code choices in terms of a natural theory of markedness. That is, as part of their communicative competence (Hymes 1972), speakers recognize choices as either unmarked or marked in reference to the norms of their speech community. Community norms designate specific linguistic choices as the unmarked realization of a specific set of rights and obligations holding between a speaker and addressee. It is this association between code choices and rights-and-obligations sets which enables us to calculate a conversational implicature from a code choice. (Scotton 1983: 115) In order to understand the essence of the theory, I need to further elaborate on the key terms ‘communicative competence’ and ‘norms’, as already touched upon earlier. What Hymes (1972, 1987) describes as ‘communicative’ or ‘sociolinguistic competence’ is the knowledge language speakers have about how to interact and use language appropriately in context. This also includes being familiar with and able to follow the rules or behavior norms of a community. Rules or norms regarding language and social behavior are recurring key terms when defining and describing the field of sociolinguistics (Cooley 1983, Fishman 1971, see also Ammon et al. 1987, Spolsky 2009). For instance, Gloy (1987) sees ‘language norms’ as a subcategory of ‘social norms’ and defines them as “expectations and/or explicit manifestations of modal circumstances which (ought to) regulate the formation, functions, usage and evaluation of linguistic units of varying 55 | Page complexity” [my translation].30 Communicative competence and conventional rules or norms are mostly acquired unconsciously or implicitly by community members as they grow up and are socialized within the speech community (Spolsky 2004, Fishman 1972b). Fishman (1972b) argues that norms and behaviors of a speech community are shaped by participants’ roles and relationships. Role relationships “are implicitly recognized and accepted sets of mutual rights and obligations between members of the same sociocultural system” (p. 37). This includes choosing the appropriate code in a given social context or specific situation (Fishman 1972a, Spolsky 2004). In this sense, it could be argued that the buyer-seller relationship, as mentioned earlier and referred to in the business literature, implies the commonly recognized customers’ right to information in their L1 and the seller’s obligation to comply with that e.g. by translating product information into customer languages. Role relationships will be further discussed below. Only if these kinds of social conventions, norms or “practices” are articulated or “spelled out” by an authoritative body, we can talk about language management (Spolsky 2004: 10). This suggests, as mentioned earlier, that members of a community do not necessarily need to be told which language to choose (under which circumstances) by top-down policies or management. In other words, language choices may be the result of social implicatures rather than policy and management. If we link this theoretical discussion to our literature review, one might wonder if businesses actually manage languages or if we observe more of a natural theory of markedness, as Scotton (1983) suggests, applied to a multilingual speech community; more specifically, her model suggests that language users negotiate and interpret code choices in interaction as ‘unmarked’ (or ‘marked’) with respect to the norms within a speech community. These norms define unmarked language choices by the specific rights and obligations that speaker and addressee hold. That is, a speaker knows when to employ which language/code, e.g. language X with addressee (1) in a specific situation and language Y when talking to addressee (2) in another context. Applied to the workplace domain, HQ staff may choose the HQ language in the home country while subsidiaries choose the local languages in host-country settings. Both choices may represent unmarked languages since they might be interpreted as the ‘natural’ thing to do, as appropriate choices or the recognized norm and conventions within the MNC. A violation of the norms or deviance 30 “Erwartungen und/oder explizite Setzungen modaler Sachverhalte, die ihrem Inhalt zufolge die Bildung, Verwendungsabsicht, Anwendung und Evaluation sprachlicher Einheiten der verschiedensten Komplexitätsgrade regulieren (sollen)” (Gloy 1987: 121). 56 | Page from the conventionalized unmarked choice is manifested in the use of a ‘marked’ language (e.g. language X instead of Y). Participants of multilingual speech communities normally speak one (e.g. the dominant language of a speech community), two or more languages and may master them to varying degrees (i.e. L1, L2, L3, etc.). In this sense, language users’ linguistic abilities limit communication or code choice to some shared competences which may interfere with the social order or norms and conventions (Scotton 1983, see also Ammon et al. 1987, Holmes 2008). Scotton (1983: 125) labels this the “virtuosity maxim”, which means that a marked (i.e. unconventional) language choice is the result of varying or lacking language fluency in the unmarked (i.e. conventional) language. Put simply, in incidences where varying language proficiency ‘gets in the way’, it may supersede social norms and dimensions (Scotton 1983, see also Hunt 1966). Applied to this study, one might imagine a business environment whose norms suggest the use of Danish (i.e. unmarked language) because the company is located in Denmark, for instance. If the interacting participants do not have Danish in their linguistic repertoires, they may choose a language other than Danish diverging from the norm (i.e. marked choice). In the business literature (Vaara et al. 2005, Louhiala-Salminen et al. 2005), this kind of contextual language choice determined by the linguistic repertoires and language skills of the participants involved in a communicative situation is associated with a ‘pragmatic’ view on language choice. Choosing an alternative to Danish is pragmatic in the sense that it is the only possible option to ensure some form of communication. 2.3.3.2 Domains of language use: participants, their roles and relationships, setting, topic within the employment domain Fishman (1972a, see also Mioni 1987) divides multilingual speech communities into domains of language use to conceptualize language behavior; among those is the employment domain characterized by three social(-relational) factors (influencing language choice): participants (e.g. employer/seller, employees, buyers/customers), the setting (workplace with a physical location) and specific topics typical of the domain (e.g. professional or business matters). What distinguishes participants from individuals is that they assume certain roles. The typical roles or labels for participants in the employment domain are employers, employees, managers, workers, 57 | Page customers, clients, etc. In an MNC context, one needs to add (regional) headquarters and subsidiaries. Participants can assume more than one role within the domain; consider the MNC network; HQ may choose a language in its role as a seller for external communication purposes or as an employer for communicating internally with national subsidiaries. The different participant roles may result in different language choices. Thus, participants also have different role relationships, e.g. HQ-subsidiary, subsidiary-subsidiary, employer-employee (internal) or buyerseller (external) relationships. Participant roles and relationships might have an impact on language choice. The setting or location has physical characteristics (e.g. size, design of factories, office space), social meaning and interpretation. In the case of an MNC, we are faced with several types of settings or locations (corporate, regional, divisional headquarters environments, foreign subsidiaries in host countries) with varying executive power, autonomy and responsibilities. In this sense, domains reconcile people and physical realities (places). As such, the setting or location may affect language choices. The third characteristic of a domain is topics that are conceived appropriate to talk about or discuss in the domain. Certain topics may be linked to specific languages and therefore influence choices. The changing of topics, e.g. from business to social matters, in interaction may explain switching between languages (code-switching) (Spolsky 2009). It has to be noted that the employment domain exists next to other domains such as family, friendship, religion and education which are embedded in or together form the multilingual speech community. The underlying idea of Fishman’s concept of domains is that one specific language (variety) is associated with a given domain which renders a speech community or whole society bi-or multilingual. For instance, participants choose language X at home (family domain) but language Y at work (employment domain). This concept is too narrow for our purposes because I conceptualize the MNC not only as an employment domain as a part of the larger global society but also as a multilingual speech community in itself, given its participants all speak different languages by definition. As Holmes (2008: 23) argues, domains of language use help “draw a very simple model summarising the norms of language use for the community”. Nonetheless, the social dimensions of the employment domain (participants and their inherent relationships, setting and topic) are indeed useful for the present study. 58 | Page 2.3.3.3 Politeness strategies (and social relationships) In a business context (employment domain) a language may also be chosen for reasons of politeness. That is, participants may intend to establish a polite image with the counterpart and therefore use the latter’s first or preferred language (accommodation). The concept might be applied to interpersonal interaction or communication with e.g. customers in general. Selecting the counterpart’s language is perceived as a politeness act whereas using one’s own language or a lingua franca might be regarded as less polite or even rude (Bäck 2004). In Holmes’ (2008: 281) general terms, politeness involves taking account of the feelings of others. A polite person makes others feel comfortable. Being linguistically polite involves speaking to people appropriately in the light of their relationship to [the speaker]. Inappropriate linguistic choices may be considered rude. […] Making decisions about what is or is not considered polite in any community therefore involves assessing social relationships along the dimensions of social distance or solidarity, and relative power or status. This also ties in with the sociolinguistic norms or rules as discussed above in the context of a speech community or employment domain; speakers or participants assume different roles, such as seller, employer, boss, employee or customer, etc. which also says something about their inherent relationships, social distance or closeness and relative power or status per se; these interact with politeness in different ways. In other words, whether a language choice is considered (in)appropriate or (im)polite in a given context is influenced by the roles and social relationships of the interactants. More specifically, the basic assumption of Brown & Levinson’s (1987) politeness theory is that every adult member of a society has ‘face’ which describes their public self-image they wish to claim for themselves. There are two types of face; (1) negative face refers to “freedom of action and freedom from imposition” (ibid. p. 61), while (2) positive face describes the self-image or ‘personality’ any member claims and wants to be appreciated in interaction. Face can be maintained (polite behavior) but also threatened or even lost (impolite behavior). There are (negative and positive) politeness strategies to maintain some other person’s face in different ways. A negative politeness strategy is characterized by respectful, specific and focused behavior (on part of the speaker) which does not impede the addressee’s freedom of action or attention (negative 59 | Page face). In a buyer-seller relationship, negative politeness could be the reason for choosing a business partner’s first or preferred language (accommodation) instead of a lingua franca (standardization). A positive politeness strategy is addressed to and satisfies the addressee’s positive face. It usually occurs between socially close interactants that approve their personalities and share knowledge. By contrast, face may also be threatened (face-threatening act) or lost which describes the concept of being embarrassed or humiliated. In this context, the relative power of the speaker over the addressee, the social distance between the interactants and the degree of imposition involved when performing face-threatening acts are important factors. If we link this to language choice at workplaces and their organizational hierarchies, it could be argued that executives have more power than their employees and thus more freedom of action including choosing languages with different implications for others. They may choose or rather ‘impose’ a language on their staff and thereby possibly perform a face-threatening act to varying degrees; if inferiors are not so fluent in this language or do not master it at all, they might feel uncomfortable or embarrassed.31 Partly similar to the concept of domain, politeness theory, too, takes into consideration participant relationships, some form of (a)symmetry of power or status and social distance/proximity (Brown & Levinson 1987). As Scotton (1983) summarizes Brown & Levinson’s (1987) work, politeness strategies attempt to preserve interactants’ face and their social relationships. 31 Yule (1996: 65) describes a positive politeness strategy as “solidarity strategy” which marks closeness between speaker and addressee through e.g. the use of a shared dialect or inclusive pronouns and terms, and the exchange of personal information. This is similar to Holmes’ (2008) solidarity/social distance scale which is concerned with participant relationships. The better participants know each other the higher the solidarity index and the more socially intimate they are. This kind of relationship may foster an informal speech style, positive politeness strategies, vernacular languages or vernacular forms within a code, the L (low) variety of a language or a particular code. This may also be an expression of shared values and attitudes, of a shared identity with the same ethnical background. Thus, the language shared by an ethnic group is also described as “the language of solidarity” (Holmes 2008: 440). Applied to this study, this could mean that expatriates might want to show solidarity and mark social closeness by using the local language (of the setting/host-country) instead of a lingua franca. By contrast, if the participants are socially distant the solidarity level is also low. Conversations between strangers are most likely dominated by a formal speech style, negative politeness strategies, the H (high) variety of a language or lingua francas and official languages (Holmes 2008). This could explain why HQ-subsidiary communication is in English, and business-to-customer interaction or buyer-seller relationships are dominated by customer languages (Feely & Harzing 2003). 60 | Page 2.3.3.4 Attitudes to language (vs. beliefs and ideology) From a language policy perspective, language beliefs or ideology, i.e. “the beliefs about language and language use” (Spolsky 2004: 5), represent a crucial and influential element with regard to language choice within a speech community, as explained below: The members of a speech community share also a general set of beliefs about appropriate language practices, sometimes forming a consensual ideology, assigning values and prestige to various aspects of the language varieties used in it. These beliefs both derive from and influence practices. They can be a basis for language management, or a management policy can be intended to confirm or modify them. Language ideology or beliefs designate a speech community’s consensus on what value to apply to each of the language variables or named language varieties that make up its repertoire. (Spolsky 2004: 14, Spolsky 2005: 2153f) Language ideology or beliefs assume a significant role in Spolsky’s (2004) language policy model since they have the potential to account for language choices at both the practice and policy/management levels. It is unclear whether beliefs and ideologies are used interchangeably. Regardless, both concepts seem to center on assigning values (and prestige) to languages. The author argues that in most nations there are a number of ideologies coexisting but usually there is one dominant one. This also implies that values are not assigned equally to languages but variance can be observed in most speech communities or domains. Overall, Spolsky’s (2004) elaboration on beliefs and ideology is not very comprehensive which is why I need to look into disciplines other than LPP, i.e. linguistic anthropology and the broader framework of sociolinguistics. Language ideology or linguistic ideology as an entire field of inquiry is a branch of linguistic anthropology. Language ideology can be defined as “the way in which we think about language” (Seargeant 2009: 346/348), more specifically as “entrenched beliefs about the nature, function, and symbolic value of language” (ibid. p. 346). This also helps understand why Spolsky at times uses beliefs and ideology interchangeably. Another, well-cited, definition is Silverstein’s (1979: 193, cited in Seargeant 2009) proposition, that describes linguistic ideologies as “any sets of beliefs about language articulated by users as a rationalisation or justification of perceived language structure or use”. This means that language users are aware of language, its structure and use, and 61 | Page able to express that. More specifically, ideology seems to be understood as the articulated explanation for language use. Other authors acknowledge the implicit nature of language ideology. For instance, Kroskrity (2004: 498) defines language ideologies as “beliefs, or feelings, about languages as used in their social worlds”. He understands beliefs as “local understandings” of an explicit or tacit nature, and feelings as a “relatively automatic aesthetic response” (p. 512). In this sense, these thoughts about or conceptions of language may be explicitly articulated or implicit in communication, and can be seen as more or less successful “attempts to rationalize language usage; such rationalizations are typically multiple, context-bound, and necessarily constructed from the sociocultural experience of the speaker” (p. 496). I would argue that if they are multiple and subjective, this also means that they are not necessarily consensual within a speech community. That is to say, Spolsky’s (2004) and Kroskrity’s (2004) views on ideology are oppositional. Woolard (1998) gives an overview of four major strands of language ideology research where ideology may be seen as: (1) “ideational or conceptual, referring to mental phenomena; ideology has to do with consciousness, subjective representations, beliefs, ideas” (p. 5), (2) “derived from, rooted in, reflective of, or responsive to the experience or interests of a particular social position, even though ideology so often (in some views, always) represents itself as universally true” (p. 6), (3) “ideas, discourse, or signifying practices in the service of the struggle to acquire or maintain power” of a social, political or economic nature, and “the tool, property, or practice of dominant social groups” (that attempt to disguise, legitimize or distort the unequal distribution of power), and (4) subject to “distortion, illusion, error, mystification, or rationalization” which “can derive from the defense of interest and power” or “limitations on human perception and cognition” (p. 7). The author also argues that there is a clear divide between the first two approaches which are more neutral and descriptive – often also covered under the label of worldviews or beliefs – whereas the third and fourth notions are certainly critical and more negative (Woolard 1998). In sociolinguistics, language beliefs or ideologies, in particular the elements of assigning values and prestige to languages or subjective representations are often conceptualized as language attitudes grounded in Labov’s (1966, cited in Mesthrie 2009) seminal work. He defines them as hearers’ immediate evaluative response to or reaction toward different accents, providing varieties 68 | Page here, comprises the potentially limited effectiveness of centralized management; this problem of effectiveness may also apply within the MNC such as discrepancies between the headquarters’ decisions and local choices at the subsidiary level, as the literature review has also shown. 2.3.4 (Theoretical) knowledge gaps summarized The previous sections have intended to give an overview of the study’s conceptual and theoretical frameworks. The cross-disciplinary nature of the subject matter has had theoretical implications. There is a myriad of conflicting, similar and different concepts of language choice making it almost impossible to work across but also within disciplines, and consequently hard to compare works of different authors or even the same author(s). For instance, the ideas of language policy and management within sociolinguistics (cf. Baldauf 2006, Spolsky 2004, 2009, Nekvapil & Nekula 2006), within international business and management (Feely & Harzing 2003, Harzing et al. 2011) or across the two are far from united. However, I have tried to work in an interdisciplinary fashion and highlighted the similarities/differences and the conceptual confusion the broader field suffers from. In order to reach some clarity or rather theoretical simplicity to begin with, I have adopted the concept of language choice (instead of policy/management, practice, use, etc., cf. Baldauf 2006, Spolsky 2004, 2009, Berthoud & Lüdi 2011, Lauridsen 2008, Lüdi et al. 2010, Nekvapil & Nekula 2006, Hagen 2006, 2011 Feely & Winslow 2005) and thereby abandoned the idea of policy/management and practice levels per se. Furthermore, I have combined insights from primarily sociolinguistics (e.g. Fishman 1968, 1971, 1972a, 1972b, Hymes 1968, 1972, 1974, Scotton 1983, Brown & Levinson 1987, Deprez & Persoons 1987, Vandermeeren 1998, Holmes 2008, Spolsky 2009), and related fields such as applied linguistics (lingua franca communication, language acquisition) (e.g. Knapp & Meierkord 2002, Seidlhofer 2001, 2004, Lambert et al. 1968, Noels 2001 – partly drawing on social psychology), and Translation Studies (e.g. Reiß & Vermeer 1991, Holz-Mänttäri 1984, SnellHornby et al. 1999, see also Steyaert & Janssens 1997, Janssens et al. 2004), as well as international business and management studies (e.g. Feely & Harzing 2003, Harzing & Feely 2008, Harzing et al. 2011, Marschan-Piekkari et al. 1999a, Steyaert et al. 2011), including corporate communication (Dhir & Gòkè-Paríolá 2002, Fredriksson et al. 2006). 69 | Page A synergy of several fields with a primary focus on sociolinguistics has allowed me to conceptualize the MNC as a multilingual speech community and employment domain. Then, I have tried to conceptualize language choice in an open-ended and simplistic fashion, resulting in several language choices (for various purposes). Sociolinguistic frames attempt to better understand language choice embedded in its social context. In other words, I have developed a theoretical and analytical framework that attempts to give a preliminary idea of language choice in MNCs and the influence of social context on choices. Also, these frameworks will guide the analysis in an openended and exploratory fashion. 2.4 Summary This chapter has provided an overview of the current state of the art and positioned this research in the existing literature, at the intersection of business and language research, with a focus on Danish and Austrian multinationals. Empirical, theoretical and partly methodological gaps have been illuminated. This has helped to further contextualize, problematize and justify the research question, and to put the study’s contribution into perspective. In sum, the present thesis addresses contextual language choice in an exploratory and multidimensional way in multinational corporations (MNCs) from a primarily sociolinguistic perspective with the overall aim of shedding light on the manageability of language in such international business contexts (see also Chapter 1, Sections 1.1-1.3). 71 | Page 3 RESEARCH DESIGN AND METHODS This chapter aims to exemplify the various methodological choices made in the course of this study; first, it will shed light on the study’s overall research approach – being of a qualitative nature – and the case study research design chosen. The latter encompasses the selection of case companies and the way (i.e. method) and by what means (i.e. data sources) these cases were studied. Second, the chapter will provide a detailed account of the data analysis adopting an interpretive perspective. 3.1 A Qualitative Research Approach Following my overall research aim (see Chapter 1, Section 1.3), this work is a purely qualitative piece of research in order to gain in-depth insights into language choice in international business. Denzin & Ryan (2007: 580) provide a comprehensive description of what a qualitative inquiry implies: Qualitative research is multi-method in focus, involving an interpretive, naturalistic approach to its subject matter. This means that qualitative researchers study things in their natural settings, attempting to make sense of or interpret these things in terms of the meanings people bring to them. Qualitative research involves the studied use and collection of a variety of empirical materials—case study, personal experience, introspection, life story, interview, and observational, historical, interactional and visual texts—which describe routine and problematic moments and meanings in individuals’ lives. By employing a broad spectrum of intertwined interpretive activities and practices (methods), qualitative researchers aim to grasp as many “subtle variations in ongoing human experience” (Denzin & Lincoln 2000b: 19) and thereby “get a better understanding of the subject matter at hand” (ibid. p. 3f). Furthermore, “[t]hey seek answers to questions that stress how social experience 72 | Page is created and given meaning” (ibid. p. 8).34 Given the nature of the study’s objective, that is language choice as a social phenomenon in MNC contexts (see Chapter 1, Section 1.2), a qualitative inquiry or design was considered most appropriate (e.g. Denzin & Lincoln 2000a, 2011, Lincoln & Guba 1985, Lamnek 2005, Marschan-Piekkari & Welch 2004). 3.2 Philosophical Underpinning The way research is conducted very much depends on one’s approach to philosophy of science, which is about the ways researchers perceive the world or reality (ontology), they believe the latter can be captured and studied in order to reach acceptable scientific knowledge and position themselves in relation to the known (epistemology), and the different ways of generating that knowledge (methodology). Qualitative researchers are philosophers (with different worldviews) in that they are guided by ontological, epistemological and methodological principles forming the paradigm in which they navigate (Denzin & Lincoln 2000b). Paradigms can be defined as “sets of basic beliefs” (Guba & Lincoln 1994: 108). Denzin & Lincoln (2000b) argue that the underlying philosophy renders all qualitative research interpretive in nature. The qualitative research tradition has over time been shaped by four general or major interpretive paradigms (positivist & postpositivist, constructivist-interpretive, critical, and feminist-poststructural) (Denzin & Lincoln 2000b). Similarly, Saunders et al. (2012) distinguish between four philosophies of science: positivism, realism, interpretivism and pragmatism. At a more general level, it could be argued that paradigmatic controversies often center around positivist and nonpositivist worldviews. If we take a look at several handbooks of qualitative research the overall contemporary trend seems to favor or imply nonpositivist philosophies (e.g. social constructivism), as can be illustrated by generalizing statements as the following: “Qualitative researchers stress the socially constructed nature of reality, the intimate relationship between the researcher and what is studied, and the situational constraints that shape inquiry” (Denzin & Lincoln 2000b: 8). Miller & Glassner (2004) critically argue that “[r]esearch cannot provide the mirror reflection of the social world that 34 These characteristics stand in sharp contrast to quantitative research, where the focus is on studying the objective physical reality, controlling causes and effects or the causal relationships between variables, measuring and quantifying phenomena, testing hypotheses, generalizing findings to a population and formulating general laws that are able to predict future outcomes (Flick 2009: 13, see also Denzin & Lincoln 2000b). 73 | Page positivists strive for, but it may provide access to the meanings people attribute to their experiences and social worlds” (p. 126). In other words, qualitative research is “grounded in a philosophical position which is broadly ‘interpretivist’ in the sense that it is concerned with how the social world is interpreted, understood, experienced, produced or constituted” (Mason 2002: 3). Similarly, I will follow a nonpositivist or interpretivist, more precisely the constructivistinterpretive path (Denzin & Lincoln 2000b). The (social) constructivist paradigm fundamentally shaped by Berger & Luckmann (1966) is widely adopted in qualitative (social science) research (Guba & Lincoln 1994, Lincoln & Guba 2000, Schwandt 2000, Alvesson & Sköldberg 2009). Schwandt (2007: 38) defines constructivism in its “ordinary sense”, as follows: [C]onstructivism means that human beings do not find or discover knowledge so much as construct or make it. We invent concepts, models, and schemes to make sense of experience, and we continually test and modify these constructions in the light of new experience. Furthermore, there is an inevitable historical and sociocultural dimension to this construction. We do not construct our interpretations in isolation but, rather, against a backdrop of shared understandings, practices, language, and so forth. In this vein, it is worthwhile mentioning the role of language within the social constructivist paradigm (for an extensive overview see Irwin 2011). Schwandt (2000) explains that the philosophy of social constructionism35 endorses “an expressivist-constructivist theory of language, in which, broadly conceived, language is understood as a range of activities in which we express and realize a certain way of being in the world”. This means humans “are self-interpreting beings” and “language constitutes this being” (p. 198). In sociolinguistic terms, “[l]anguage constructs social reality” (Holmes 2008: 436). From an ontological point of view, this means that within the social constructivist paradigm reality is constructed by mankind and thus ‘relativist’. These multiple realities can be described as constructions that are mentally, socially, experientially and individually informed. Such an 35 Constructionism, as Schwandt (2000) uses the term, may be used instead of constructivism. Often used interchangeably, Crotty (1998), though, distinguishes between the individual that makes sense of the world (constructivism) and society as a collective group or culture that generates meaning and shapes individuals in their way of thinking (constructionism). 74 | Page ontology makes its clear delineation from epistemology disappear; the latter understands knowledge generation as a ‘transactional’ and ‘subjectivist’ process. That is to say research products are conjointly and interactively created by the inquirer and their study objects or better described as their subjects (Guba & Lincoln 1994). This process of co-constructing meaning and knowledge (through e.g. interviews) makes it “impossible and undesirable for researchers to be distant and objective” (Hatch 2002: 15). As opposed to quantitative researchers who try to detach themselves from the research activity and thereby reach objectivity, qualitative scholars rather immerse into the field and deliberately choose to be present and involved in their research (Patton 2002). Thus, researchers’ impressions, reflections, observations, feelings, etc. become part of the research process. In qualitative research the subjectivity of both the inquirer and the study participants are acknowledged. Consequently, most qualitative researchers do not claim to be objective or value-free, rather reflexive, which means that they critically reflect on their selves, their relationships with respondents and how these might impact on the research activity (Flick 2009, Lincoln & Guba 2000). They may also strive to be sensitive, which stands for the “ability to respond to the subtle nuances of, and cues to, meanings in data” (Strauss & Corbin 1998: 35). In this sense, the qualitative researcher is often associated with the image of an (interpretive) ‘bricoleur’ who by drawing on multiple sources and analytical steps attempts to make sense of the data and build social worlds (Denzin & Lincoln 2000a, 2011). These ontological and epistemological considerations have methodological implications, namely that the inquirer-participant interaction is necessary to decipher and interpret the different constructions through a hermeneutical and dialectical (contrastive) approach, eventually leading to a consensus construction (Guba & Lincoln 1994). Hatch (2002) tries to translate Guba & Lincoln’s (1994) ideas as follows: Researchers spend extended periods of time interviewing participants and observing them in their natural settings in an effort to reconstruct the constructions participants use to make sense of their worlds. Hermeneutic principles are used to guide researchers’ interpretive coconstruction of participant perspectives. (Hatch 2002: 15) As already mentioned, the social constructivist paradigm belongs to the interpretive traditions or paradigms where the researcher’s aim is to achieve understanding (Verstehen) through (subjective) 75 | Page interpretation. Put simply, analysts interpret their data in search of meaning or to make sense of them. Hermeneutics has been described as “the theory of interpretation” (Smith et al. 2009: 21). Hermeneutic principles, as mentioned in the quotation above, thus describe the analytical process of sense-making, which is not a linear process, as it is often the case in quantitative positivist paradigms, but a circular one (Flick 2009). It is circular in that “[t]he analyst works back and forth between the data or story (the evidence) and his or her own perspective and understandings to make sense of the evidence” (Patton 2002: 477f). Also, the activity is characterized by going back and forth between smaller and larger data entities, between ‘the part’ as well as ‘the whole’ in order to reach authentic conclusions. This is the reason this analytical method is usually referred to as the hermeneutic circle (e.g. Schwandt 2007, Patton 2002, Smith et al. 2009). Strictly speaking, the analyst “is engaged in a double hermeneutic because the researcher is trying to make sense of the participant trying to make sense of what is happening to them” (Smith et al. 2009: 3). This means the analyst’s sense making is only a second-order or second-degree construction while the participants’ accounts are of a first-order nature (Smith et al. 2009, Schütz 1962, cited in Flick 2009). Hermeneutics as a methodology has its origin in interpreting texts (e.g. biblical and religious texts). Kvale & Brinkmann (2009) argue that “[f]rom hermeneutics, qualitative researchers can learn to analyze their interviews as texts and look beyond the here and now in the interview situation” (p. 51). The authors, by referring to Schwandt’s (2000) work, state that the meaning of interviews (as texts) or other documents emerges from a cultural, historical, political and literary context as well as background information and other surrounding meanings, perceptions or values (Kvale & Brinkmann 2009). Overall, an analytical approach that involves subjective judgment and (contrastive) interpretation of the reflexive inquirer leads to subjective rather than objective or absolute findings (Hatch 2002, Flick 2009, Denzin & Lincoln 2000b). The latter are often represented in the form of case studies or extensive narratives that describe the interpretations constructed as part of the research process. Accounts include enough contextual detail and sufficient representation of the voices of the participants that readers can place themselves in the shoes of the participants at some level and judge the quality of the findings based on criteria other than those used in positivist and postpositivist paradigms. (Hatch 2002: 15f) 76 | Page In other words, evaluation criteria such as validity and reliability become trustworthiness and authenticity in qualitative (naturalistic) inquiry (Guba & Lincoln 1994, see also Lincoln & Guba 2000). It is important to remember that qualitative research attempts to draw a “holistic picture” of phenomena and settings, and to “understand the fundamental nature of a particular set of activities and people in a specific context” (Patton 2002: 480). As mentioned in the very first section (3.1) of this chapter, qualitative research focuses on understanding subject matters in greater detail (rather than explaining them). By contrast, quantitative research typically has explanatory purposes, i.e. testing hypotheses or cause-effect relationships in order to arrive at generalizable laws and theories that are ideally able to predict future events. The demarcation line between descriptive and exploratory versus causal and explanatory approaches is not totally clear-cut. In other words, explaining is not limited to quantitative research, but has a different meaning in qualitative studies. Schwandt (2007: 105) argues: In commonsense usage, to explain something is to make it intelligible or understandable. However, usually, when we ask for an explanation, we are asking for an account of why something happened. Qualitative inquiry is often concerned not simply with what human actions mean but why they occur. In search of the ‘why’ and explanations, the qualitative researcher is dependent on their subjects (respondents) who provide their individual explanations for human actions and decision making. To the extent that you are describing the causal linkages suggested by and believed in by those you’ve interviewed, you haven’t crossed the line from description into causal interpretation. And, indeed, much qualitative inquiry stops with the presentation of case data and cross-case descriptive comparisons aimed at enhancing understanding rather than explaining “why”. (Patton 2002: 478) This means I am interested in reported, subjective ‘explanations’ for human decision making and experience, i.e. the underlying rationale(s) why a particular language is chosen in a given context. This is a distinct epistemological differentiation between my qualitative approach to ‘explanation’ and the one in the quantitative tradition where explanation is commonly understood as deciphering cause and effect relationships. In this study, explanations can be described as personal 77 | Page accounts or rationales that are multifaceted and context-dependent, hence valuable as they help deepen our “understanding of human experience” (Stake 1995: 38). In conclusion, a qualitative research approach positioned in the constructivist-interpretive or social constructivist paradigm which draws on relativist ontologies, subjective epistemologies and naturalistic methodologies (Denzin & Lincoln 2000b) appeared appropriate for the purpose of this study, namely gaining a deep understanding, ‘inside’ knowledge (Dyer & Wilkins 1991) on the manifestation of language choice in international business contexts (i.e. MNCs) including participants’ underlying rationales. 3.3 A Qualitative Case Study As briefly mentioned, case studies are often the research product created or strategy of inquiry applied within the qualitative, interpretive tradition in general and the social constructivist paradigm in particular (Stake 2000).36 In alignment with my philosophy of science, a case study design seemed useful. Broadly speaking, within the qualitative tradition, the case study is understood as a strategy of inquiry that “relies on interviewing, observation, and document analysis” (Denzin & Lincoln 2000a: 371). From a sociological perspective, Orum et al. (1991: 2) define a case study “as an in-depth, multifaceted investigation, using qualitative research methods, of a single social phenomenon. The study is conducted in great detail and often relies on the use of several data sources”. From an organizational perspective, Hartley (1994: 212) argues that “[a] case study allows for a processual, contextual and generally longitudinal analysis of the various actions and meanings which take place and which are constructed within organizations”. Within the field of business and management research it has been defined as “a research strategy that examines, through the use of a variety of data sources, a phenomenon in its naturalistic context, with the purpose of ‘confronting’ theory with the empirical world” (Piekkari et al. 2009: 569). The relationship between evidence and theory in qualitative research in general and in case study research in particular is viewed controversially across disciplines and paradigms, depending on philosophical groundings (see Piekkari & Welch 2011a for an overview of the case study approach 36 Stake (2000) does not limit the case study to a qualitative methodology per se, though focuses on the latter. 84 | Page case study of this type has been conducted in an MNC context. I will now turn to the selection of the specific companies. 3.3.1.3 Selecting two case companies per country In case study research, the sample size is usually far too small to pursue a random sampling strategy (aiming at representativeness and generalizability); since these were not my goals, cases were selected in a purposeful (rather than random) manner. In a purposeful and balanced design, which takes into account both balance and variety in case characteristics, the cases selected ought to give the opportunity to learn about the phenomenon in varied contexts. Usually, case researchers work with typologies or attributes in order to select their cases (Stake 2000, see also Ghauri 2004, Welch et al. 2011). In organizational or MNC contexts one might think of e.g. industry, company size, (HQ and subsidiary) locations, etc. Considering my research goal and the purpose of finding a balance between particularization (thick description) and comparison, I chose a small number of cases for this collective study, that is two cases per country. To observe how language choice operates and functions across different contexts, I focused on the attributes HQ location (home country), industry and company size. Accordingly, one larger manufacturer and one smaller MNC operating in an alternative industry were purposefully selected for each country, as indicated by Table 3-2 below (Stake 2000, Yin 2009). Size Denmark (HQ) Austria (HQ) Larger 1 Manufacturer 1 Manufacturer 2 larger manufacturers Smaller 1 MNC from alternative industry 1 MNC from alternative industry 2 smaller MNCs from alternative industries 2 Danish MNCs 2 Austrian MNCs 4 MNC cases in total Table 3-2: Selecting two case companies per country 85 | Page The following table (3-3) summarizes in more detail the characteristics of relevance for choosing the study’s MNCs (HQ location, industry, company size measured by number of staff and subsidiaries, and turnover per anno or business year). For reasons of confidentiality the company names have been disguised, the numbers are only approximate values and the industries are not discussed in detail. In summary, I was striving for both balance and variety in the case study design which is by no means representative of MNCs as a whole, of national contexts or any basis for statistical generalization. It is important to remember that “[e]ven for collective case studies, selection by sampling of attributes should not be the highest priority. Balance and variety are important; opportunity to learn is of primary importance” (Stake 2000: 447). 4 Cases Industry Size (no. of staff) Size (no. of subsidiaries) Turnover (in billions of EUR) DK_1 Manufacturing 18,000 80 ≈ 3 (in 2012) DK_2 Pharmaceutical 6,000 57 ≈ 2 (in 2013) AT_1 Manufacturing 46,400 500 ≈ 11.5 (in 2012/13) AT_2 Construction 7,400 26 ≈ 1.5 (in 2012/13) Table 3-3: Characteristics of the four case companies Apart from HQ location, industry, size and turnover as selection criteria, each company had specific characteristics that rendered it theoretically interesting and relevant, and justified its selection. Based on desk-research and initial contacts with DK_1 and DK_2, it became apparent that they both had chosen English as their corporate language and formulated a language policy draft37 and a language guide. Also, the key contact person of DK_1 was a senior translator who provided first insights into language choice and related translation processes that seemed relevant. These characteristics rendered both DK_1 and DK_2 viable Danish case companies since diversified language choices could be identified and promising further insights were hoped for. The selection of AT_1 was theoretically based on two factors; first, it is so large that a high degree 37 Language policy statements were not a selection criterion because I was interested in language choices made in context and the manageability of languages in MNCs. Policies were no exclusion criterion either because they were believed to deepen our understanding of language choice in its integrity. 86 | Page of linguistic heterogeneity informing various language choices was assumed. Also, the pilot interview conducted with an executive revealed AT_1’ structure and organization into divisions and variation in language choice. HQ seemed to have two main corporate languages whereas one division (an Austrian-Swedish merger) had chosen English as a corporate or divisional language. This also tied in well with Marschan-Piekkari et al.’s (1999b) call for case companies with more than a single corporate language. Second, the corporate website’s appearance came across as totally bilingual and very professional from a Translation Studies’ perspective. These were the main criteria for including AT_1 in the study. The fourth company case (AT_2) was theoretically chosen due to its strong focus on German on the one hand and its primary business activities in Eastern Europe on the other hand, whose national educational systems and foreign language learning (macro language policy and planning) seem to differ from Central and Western Europe (Ammon 2001). In this sense, I assumed to find innovate approaches to language choice and thus included AT_2 into the study. Overall, one could say that there was no such thing as a specific attribute that was a universal selection criterion. This is because I was interested in learning about diversified and multifaceted choices. In other words, the case companies were all both similar and different to some extent, balance and variety being of importance (Stake 2000, Ghauri 2004). In selecting the different cases, I also attempted to strike a balance between the research foci of the study and pragmatic considerations (Ghauri 2004, Patton 2002, Stake 2000), such as temporal and financial resources, geographical distance and possibilities of access. 3.3.1.4 Initial problems encountered Generally speaking, it was difficult to find and gain access to case companies in both countries. The companies were approached via email and asked if they were willing to participate in the study. The response rate was rather low. Initially I had a more ethnographic approach to data collection in mind, i.e. immersing in the field by spending two or three days at HQ, doing participant observation (e.g. in meetings), identifying and interviewing participants (e.g. executives, employees, international staff) I consider scientifically relevant or interesting. This was believed to be a fruitful approach to getting an idea of the company in its context, i.e. its 87 | Page structure, employees, working atmosphere, languages used, etc. The first MNC (DK_2) I had contacted turned this idea down owing to “lacking resources”. Instead the key person, having a rough idea of what the project was about, suggested I could come and conduct a few interviews with people he considered interesting, including himself. On the basis of this response, some adaptations to the planned fieldwork and selection strategy were necessary; when contacting further case companies I kept the description and title of the project short and catchy (1-2 sentences), suggested a limited number of participants needed (5-10) who I could conduct individual interviews with (30 minutes each). (I will discuss the research interviews in the following sections.) These modifications to data collection methods were fruitful, and were a pragmatic balance between research goals and companies’ willingness to participate. Overall, the possibilities of access depended very much on the openness of an individual contact person with an interest in research projects (DK_1) and cooperativeness of people either my former supervisor (DK_2) or I myself (AT_1, AT_2) knew personally. This was crucial for awakening the companies’ interest, building confidence and obtaining further access. 3.4 Data Collection 3.4.1 Empirical materials: research interviews (RIs), corporate documents and websites Generally speaking, once a case is chosen there are further decisions to make as to which places, persons or events to study or observe (Stake 2000). Apart from selecting company cases, I had to decide how to study them, i.e. which type of evidence to collect when, where and from whom, also referred to as searching for “cases within the case” (Stake 2000: 447) or “casing” (Ragin 1992: 217). Research interviews (RIs) (Kvale & Brinkmann 2009, King 1994, Miller & Glassner 2004) represented the primary data source as they were considered most suitable for investigating human behavior, interpretation and understanding, i.e. participants’ perceptions, experiences, beliefs, problems and most importantly rationales and motivations behind language choices. In more philosophical terms, an open-ended interview (as text) can be regarded as “a window into human experience” (Ryan & Bernard 2000: 769) or “the opportunity for an authentic gaze into the soul of another” (Silverman 2000: 822f). Furthermore, through nonpositivistic, i.e. interactive, interviews “intersubjective depth” and “deep mutual understanding” can be achieved, 88 | Page eventually generating “knowledge of social worlds” (Miller & Glassner 2004: 126f). Despite their importance, RIs were not the only empirical material. In the interpretive tradition alternative sources to interview data usually play an important role (Piekkari et al. 2009). Following this convention, both corporate documents and websites were included in the study. The inclusion of different types of sources also aimed to triangulate data and to study the company cases in their entirety (Patton 2002). In other words, the use of a variety of data sources (triangulation) helps ensure an in-depth understanding of the research matter under investigation, and thereby attempts to strengthen a study (Patton 2002, Denzin & Lincoln 1994, Flick 2009). “Triangulation has been generally considered a process of using multiple perceptions to clarify meaning, verifying the repeatability of an observation or interpretation” (Stake 2000: 443, 1998: 97). However, ontologically speaking, it is important to remember that since there is no ‘objective truth’, triangulation is not a device or tool of, but an alternative to validation (Flick 2009). This means we are actually interested in the multiple realities (ontology), i.e. different perceptions of a single phenomenon which may be dialectical or controversial but that does not render them less valid, especially because the ‘particular’ cases are often the most interesting and insightful ones (Ragin 1992). 3.4.2 Collecting interview data: a first overview In total, I collected 34 semi-structured qualitative research interviews from two Danish and two Austrian MNCs (see Figure 3-4 below) at two points in time (December 2009-January 2010 and June-September 2010). It was decided to collect interview data at both headquarters and subsidiaries. This ensured “unit triangulation” (Marschan-Piekkari et al. 2004: 254), a common strategy in IB case study research. This is because the type of knowledge, information richness and degree of detail may differ across levels. For instance, the more decentralized translation and language training are administered, the less information can be gathered at the HQ level. So in order to get a holistic picture of all aspects of language choice in international business contexts, I needed to consider local units, too. Choosing local units (see Figure 3-4 below) was once again done in a purposeful manner, involving a snowball or chain sampling strategy leading to “information-rich cases” (Patton 2002: 242). This 89 | Page means that first or key informants were encouraged to suggest potentially interesting locations or respondents based on the study’s research foci. Austria, for instance, represented not only the home country (for AT_1 and AT_2) but also the host country of subsidiaries in all four cases; the key person of DK_1 suggested Austria and Germany as local options with potential interview candidates. Similarly, in the case of DK_2, one influential HQ executive had a good connection to the Austrian unit, had lived in Austria and was willing to pave the way. For both Austrian HQs, the subsidiaries selected were also located in Austria because they were very specialized in a field and operated globally and independently from the headquarters. They could therefore provide interesting insights, although the HQ and local languages were both German. It is worth mentioning that the Austrian company cases offered a great variety of language choices and were quite open about their ‘problems’. Thus there was less need to investigate in more depth at other local levels since data saturation was achieved more quickly. Obviously, there was also a personal and pragmatic dimension involved; it was easier for me – being Austrian – to obtain access to Austrian and German units. Australia as a subsidiary of DK_1, the only Anglophone environment, was included as a local unit because access could easily be gained – I spent a research semester in Australia – and HQ staff facilitated access to local interviewees (snowball principle). The following figure (3-4) summarizes the distribution of research interviews (RIs) conducted in Denmark and Austria, across two organizational units, i.e. HQs and subsidiaries in Australia, Austria and Germany. 90 | Page Figure 3-4: Interview data collected from Danish and Austrian HQs and subsidiaries (in Australia, Austria, Germany) 3.4.2.1 Interview respondents across three levels of employment A purposeful snowball sampling strategy was also applied to selecting interview respondents. These were employees (e.g. language consultant, assistant/administrative staff, translator, copywriter, lawyer), middle managers (project leader, HR, product, account or marketing managers with varying supervisory functions, heads of smaller departments) and top managers with different responsibilities (managing directors, (regional) directors or executives and heads of larger departments or whole divisions). (Former) in-/expatriates or repatriates and third-country nationals represented an interesting group regardless of location and responsibilities because of their linguistic repertoires, perspectives on different environments and diverse experiences. They 34 Research Interviews (RIs) AUSTRIA: 14 RIs DENMARK: 20 RIs AT_1: 8 Manufacturer DK_1: 8 Manufacturer AT_2: 6 Construction DK_2: 12 Pharmaceutical HQ: 3 S 1: 3 Australia HQ: 4 HQ: 3 HQ: 6 S 2: 1 Austria S 3: 1 Germany S 1: 6 Austria S 1: 5 Austria S 1: 2 Austria 91 | Page were also ‘door-openers’ to accessing additional sites (local units and staff). The following table (3-5) shows the distribution of research interviewees across these three levels of employment. 3 Levels DK_1 DK_2 AT_1 AT_2 Employees 4 3 3 0 Middle Managers 2 5 3 3 Top Managers 2 4 2 3 In total 8 12 8 6 Table 3-5: Distribution of interview respondents across three levels of employment 3.4.2.1.1 Further problems encountered A constructivist-interpretive research design including a purposeful sampling strategy is more flexible than others, and allows unique cases and insights to emerge (Denzin & Lincoln 2000a, Stake 2000, Ragin 1992, Patton 2002). This flexible approach also had its drawbacks such as limitations of access which explain the uneven distribution of participants per case study. More specifically, at the Austrian subsidiary of DK_2, the local contact person was particularly helpful and provided unlimited access by introducing me to almost all her colleagues, which doubled the number of participants in one day. At AT_2 it was the opposite; the key person at HQ considered six informants in total sufficient and basically denied further access (to the employee level). This means access is often dependent on influential staff and can therefore also differ across organizational units and employment levels. Yet, this did not seriously compromise data collection since data saturation (including diversified insights into language choices and ‘particular’ cases) could still be achieved within AT_2. 3.4.2.2 Entering the field and conducting research interviews (RIs) The final participants and interviewees were all contacted via email, briefed on the project, and practicalities were sorted out, such as date, time and location of the interview. The RIs were conducted face-to-face in the respondent’s usual environment (e.g. office, meeting room) with a 92 | Page few exceptions. In six cases I could not visit the site and phone interviews were opted for. Except for one focus group interview (with three participants and suggested by one of them), I only conducted individual (one-to-one) interviews, considered most appropriate to gain as many diversified and personalized insights as possible.38 All interviews were digitally audio-recorded (with the respondents’ permission) and lasted between 30 and 45 minutes on average (min. 12 and max. 85 minutes).39 The interview language was either German or English, depending on the interactants’ preferences and language proficiency. I was not fluent enough to conduct an interview in Danish, so English was chosen instead. The same was applied to all the other cases where the first language of the interviewee (i.e. Croatian, Dutch, Estonian or Swedish) could not be used. It was decided not to rely on an interpreter since the interviewees in question had agreed on doing the interview in English. Conducting a lingua franca interview in interor cross-cultural settings might lead to reduced data richness due to difficulties expressing ourselves in a foreign language (e.g. MarschanPiekkari & Reis 2004, Welch & Piekkari 2006). I was aware of these potential problems and thus sensitive to possible misunderstandings. The interviews were conducted in a semi-structured fashion: I had an interview guide with an outline of the research phenomena of interest operationalized through several questions, which will be discussed in detail in the following section (3.4.2.3, see also Appendix). I tried to keep the questions rather short with a simple syntax and colloquial terms. Since interviewing is also a form of oral speech production, a simple language makes it easier for the respondent to follow. Semistructured interviewing also allowed the inclusion of probing questions (i.e. asking for further 38 Focus group interviews are more suitable for exploring new phenomena, where the researcher presents a topic and only facilitates rather than leads the conversation, opening up for a broader, unstructured discussion (Kvale & Brinkmann 2009). Since I had a semi-structured research agenda with specific research interests and related questions, focus group interviews were disregarded in the first place. But the key person (from DK_1) and respondent suggested the focus group interview as an opportunity for him and his two colleagues to get an idea of my project and consequently reflect together on other possible interview candidates (snowball principle). The interview was not as rich in data as the individual ones because not everyone answered all questions. Besides, the turn taking between informants and group dynamics may have influenced responses and viewpoints. 39 The shortest interview was technically embedded in another interview. This was when a repatriate unexpectedly interrupted an interview. In this case, the actual respondent asked his colleague next door to join the interview and to share his experience as a former expatriate with us when it came to the topic of expatriation and the use of local languages. After 12 minutes the repatriate left again and the interview was continued with the actual respondent. Although it was not an entire interview but input on only one area (i.e. expatriation), it was counted as an interview representing a new, ‘particular’ case due to its special insights (Ragin 1992). 93 | Page description, explanation or examples) and interpreting questions (i.e. repeating and paraphrasing information for clarification purposes), if needed. In addition, structuring questions were used in a meta-communicative way to introduce new topics and summarize previously discussed ones (Kvale & Brinkmann 2009). The questions were open-ended, i.e. the informants needed to express and formulate replies themselves and thereby recreate their subjective understandings of the themes of interest. Such an approach let the interviewee and me (interviewer) build up an interactive setting which led in various directions, depending on the respondent’s position (management vs. employee level), language proficiency, language choices, worldview, values and beliefs, working area, experience, interest in the topic, and also willingness to provide information and their time available. I tried to adhere to the interview script but if informants’ answers opened up new directions, follow-up and specifying questions were added. Consequently, the order of the questions asked was not identical. This means the research interviews were embedded in their interpersonal and situational contexts, providing me with a diversified and multifaceted data set. Such a flexible and exploratory data collection method allowed the integration of additional themes with the emergence of new data, as suggested by Miles & Huberman (1994) or Ghauri (2004). In other words, qualitative research and semi-structured interviews are not neat and clear-cut but allow the researcher to explore more deeply by asking additional questions regarding a specific theme, omit certain questions in another context, change the order of the questions, etc., all depending on the flow of interaction and discussion, which is unique to every interview (King 1994, Saunders et al. 2012). Further characteristics of the research interviews were that they contained narrative, factual and conceptual elements (Kvale & Brinkmann 2009). The RIs were largely narrative, because they had ‘story-telling’ and retrospective components when expressing experiences. Talking about a particular topic or concept might evoke associations concerning an event in the past, illustrating a point of view or justifying an argument. The interviews also provided factual information, e.g. facts about the company’s history, organizational structure, numbers and figures, which could partly be cross-checked and confirmed by studying the websites and by multiplying informants (i.e. triangulation, data saturation). To some extent the RIs were also conceptual, as they enabled me 100 | Page (Coffey & Atkinson 1996: 110). More specifically, I did a microanalysis which is a “detailed lineby-line analysis” that helps “generate initial categories (with their properties and dimensions)” (Strauss & Corbin 1998: 57). This line-by-line analysis can also be applied to single words, sentences, or paragraphs. In my case, the data chunks were whole paragraphs rather than single sentences in order to describe and analyze the evidence in its context. I adopted an open-coding strategy (Glaser & Strauss 1967), read the first three interviews on hard copy, highlighted themes of interest and made marginal notes including tentative labels. After this initial reading and preliminary coding, all the texts (of the first data set) were imported to NVivo, which is a software facilitating qualitative analysis of a huge amount of data (e.g. Bazeley 2007). The codes, called nodes in NVivo, represented common themes which were grounded in data, i.e. emerged inductively, but shaped by my conceptual understandings and theoretical framework (abductive logic), in accordance with my qualitative (constructivist-interpretive) design. In this sense, the codes (nodes) were my own creations. 3.5.2.1 Developing a coding tree following an abductive logic I continued my detailed line-by-line analysis digitally with the first set of interviews, then with the second data set and gradually developed a (hierarchical) coding tree in NVivo based on tree nodes, nodes and sub nodes (see Figure 3-6 below). The nodes represent the themes that emerged from the data, also tied to the interview protocol, i.e. the questions asked to operationalize research phenomena of interest (see Section 3.4.2.3, Appendix (Interview Protocol) and RQ – Chapter 1, Section 1.2). Coding phase I was characterized by a detailed descriptive analysis (Miles & Huberman 1994, Stake 1995). This means that exploring the data set dominated the first analytical steps. Coding in NVivo (and using its features such as memos) provided a diversified and detailed account of the data and facilitated the analytical process. It allowed me to store the same data chunk in multiple categories (nodes). The coding tree followed more or less the structure of the interview protocol. Therefore, 101 | Page the first tree node stored background information on the respondents, for instance.42 Some nodes do not have any sub nodes, but were part of the tree. For instance, none of the companies had a language department, yet I coded all the answers provided in the node. Furthermore, the complexity of the phenomenon language choice led to many similar nodes containing similar data. The HQ language was reported to be used in different contexts creating several codes ‘at HQ’ (representing the location) or ‘interunit communication’ (the communication purpose). Or, the tree node ‘external communication’ needed to be subdivided into ‘foreign market expansion’ which was in one Austrian case dominated by the ‘HQ language’ while for other external communication purposes ‘customer languages’, ‘English’ and ‘HQ language’ needed to be created. Data stored in ‘customer languages’ may contain information also stored under the tree node ‘translation’ and its sub-nodes. This means, initial coding was not only about data simplification (reduction) but also about “complication”, that is, “coding usually is a mixture of data reduction and data complication. Coding generally is used to break up and segment the data into simpler, general categories and is used to expand and tease out the data, in order to formulate new questions and levels of interpretation” (Coffey & Atkinson 1996: 29f). NVivo’s features such as annotations and memos were useful for ‘thinking about’ the data during the digital coding process. Remarks were added in the form of NVivo annotations; this could be something striking about the data. Later on memos were used to write longer descriptions or summarizing statements, and to clarify meanings. Memos can be defined as a “researcher’s record of analysis, thoughts, interpretations, questions, and directions for further data collection” (Strauss & Corbin 1998: 110). Both annotations and memos were particularly useful for understanding and interpreting interviewees’ perceptions and for reflecting on possible linkages between phenomena at an early stage of analysis. After this first digital coding phase, it was crucial not to lose track and ‘drown in data’ but regain an idea of ‘the whole’ empirical picture as the hermeneutic circle suggests, which led to pattern coding phase I. 42 The characteristics of each respondent (Respondent’s profile), in alignment with the interview’s introductory questions (#1, 2), only represented complementary knowledge, which was coded in order to be able to retrieve it more easily later in the analytical process, notably to integrate the information needed in the case reports (i.e. the respondents’ positions, responsibilities or work areas). Their language proficiency (i.e. L1 and FLs) – see Question 3 of Appendix (Interview protocol) – was not coded but registered in the case book, another function in NVivo, where different kinds of attributes (i.e. location of the employee and highest degree of education) were created to get a better description of the interviewees, their environment and background. 102 | Page Tree nodes Main nodes Nodes Sub-nodes Respondent’s profile Position Responsibilities Educational background Organizational structure Centralized (strong HQ control) Autonomous subsidiaries Internal communication At HQ (location) HQ language At subsidiaries Local languages Interunit communication English Problems HQ language Local languages External Communication Foreign market expansion HQ language Customer languages English HQ language Associations with language policy Set of rules English as a corporate language Language training and acquisition Emails/meetings in English In-/expatriation management Corporate language English (group language) English (common language) German Language department Translation In-house Nonprofessional Machine translation Problems & advantages 103 | Page Professional Terminology management (term base, translation memory systems) Outsourced/prof. Text type, genre, register Interpretation In-house/nonprof. Problems Professional Communicative event Language training and acquisition English HQ language Local language Customer language Paid by company On site During working hours HRM Selective recruitment Expertise more important than language Staff relocation Language nodes Local language Code-switching/- mixing Written medium Oral medium Positive associations Trigger channel Passive multilingualism Written medium Oral medium Figure 3-6: A (hierarchical) coding tree 3.5.3 Pattern coding phase I: revisiting coded data manually to further reduce data and look for relationships As a next step, I reread the entire data set coded to further reduce, complicate and eventually better understand the data. Though, rather than using NVivo again, it was performed manually in this second round. In the following steps, it was attempted to regain an overview by merging these small, diversified entities with partly overlapping information, resulting from coding phase I, into 104 | Page bigger units again. This was done by rereading coded data on hard copy according to themes (nodes), first case study by case study, then according to themes across cases. Since the same data chunks existed in more than one category (node), I read the same information repeatedly, adding value to the process of understanding and learning about the phenomena. This also helped revisit and refine nodes, i.e. ‘sensitize’ the concepts (Coffey & Atkinson 1996). Existing and newly emerging nodes were written in the form of marginal remarks on the hard copies and key words representative for the themes were highlighted in the texts. Through rereading and thereby reanalyzing an entire case study in this fashion, I was able to first regain a good overview of the four individual cases and second to see patterns, reoccurring themes across cases. In other words, this kind of systematic data revision also assisted the process of seeking linkages and relationships in the data set, also referred to as pattern coding; this is usually a step following initial descriptive coding because interconnections tend to be implicit and require analytical interpretation through repeated reading of data and revision of findings (Miles & Huberman 1994). For instance, from rereading and comparing respondents’ rationalizations vis-àvis the choice of the ‘HQ language’ ‘at HQ’ and of ‘local languages’ ‘at subsidiaries’ the notion of ‘community norms’ emerged. Furthermore, when going through the hard copies and reinterpreting data in their context, I identified rationales underlying language choice. So, I could for instance start to see a link between the theme ‘English’ and ‘proficiency’, or ‘language acquisition’ and ‘customer language’ and ‘motivation’, or between ‘HRM’ and ‘attitudes to language’, or between ‘translation’ and ‘buyer-seller relationship’ or ‘external forces’. The next step was then data display. 3.5.4 Pattern coding phase II: understanding and interpreting relationships using within-case displays Data display can be described as a visual representation of the data via concepts, models, key quotations or tables and matrices filled with summary statements (Ryan & Bernard 2000). Pattern coding (phase I), as just explained, was the basis for displaying relationships in the form of matrices. For each company one conceptual matrix was drawn on a whiteboard and summary statements were entered (Miles & Huberman 1994) (see Table 3-7 below). The matrix focused on 105 | Page the main themes of interest, i.e. the social dimensions and rationales informing language choice in the MNCs. More specifically, the most relevant nodes from the coding tree in the very left column and rows were placed side-by-side with five columns defining the Social Context. The summarizing statements entered formed the basis for the next step in the analysis (i.e. writing case reports – see Sections 3.5.5 and 3.5.6). Note, the matrix below does not display the entire data set but visualizes some insights across cases. It only aims to demonstrate how I proceeded in the analytical process. Therefore, the table is to be understood as a simplified illustration of the actual four matrices (i.e. one per company case). The social context language choice was embedded in could be divided into the following categories: (1) Participants’ language proficiency, (2) Participant roles, roles relationships & politeness, (3) Participants’ attitudes, (4) Participants’ motivations, and (5) External forces. They informed different language choices including translation and interpretation, choices for HRM purposes (selective recruitment, staff relocation), choices for educational purposes, codeswitching/-mixing, passive multilingualism etc. including various languages. The choices could be made at both the individual and corporate levels. This transition from the descriptive coding phase to this pattern coding was crucial in that it refracted the idea of levels or vertical (top-down) formalized language management such as translation and interpretation undertaken by the employer or corporate management. By ‘looking behind’ the themes of interest such as translation, that is uncovering the decision maker’s underlying rationales for a choice involving translation in a given communicative situation, we could learn that translation could be initiated by a single employee or the corporate level (which also needed to be divided into HQ, regional and local management levels). This multilayered notion of social context and multidimensional language choices (across levels of employment and organizational units) uncovered the complex and multilingual nature of language choice in the cases. Ultimately, this helped answer the research question and learn about the limited manageability of language in MNC contexts. These categories were later aggregated and further abstracted into (1) social-linguistic, (2) socialrelational, (3 & 4) social-psychological and (5) social-regulatory dimensions to reach a higher level of abstraction (Miles & Huberman 1994). 106 | Page Table 3-7: Relationships between social context and language choice SOCIAL CONTEXT LANGUAGE CHOICE Participants’ language proficiency Participant roles, role relationships, politeness Participants’ attitudes Participants’ motivations External forces Choice of English Participants have varying proficiency à English is chosen as lingua franca à pragmatic choice if a shared language is needed Positive value assigned to English (geocentric attitude) informs the choice of English for HRM purposes (selective recruitment, expatriation), language training including operative level International scientific community, international stock market, internat. funding of construction projects impose English Choice of HQ language Participants have varying proficiency à Danish/German is chosen in (regional) meetings à internationals (NNSs) accommodate NSs Executives (participant role) are addressed in German owing to politeness (high status and power) Ethnocentric attitude informs the choice (imposition) of Danish/German (can include acquisition) à internationals feel they need to use Danish/German (internal forces) Home-country authorities (tax office) impose HQ language, including translation (English à German) Choice of local languages Translation into local languages for staff with limited English proficiency (HQ level) Home-country executive is asked to learn the local language of the newly acquired site to show goodwill (politeness) à positively affects social relationships Decision makers’ polycentric attitude informs the choice of local languages for HRM purposes (expatriation) Linguistic adaptation to colleagues (local language users) ‘for fun’ Host-country authorities impose local languages, including translation and translation quality (English/HQ language à local languages) Choice of customer languages Some participants would also choose English with customers but they do not speak English à their L1 (customer lang) is chosen (involving interpretation) Customers are addressed in their L1 due to buyerseller relationship and politeness (‘English is not enough’) à involves (non)professional translation Buyer-seller relationship and polycentric attitude let executive think about offering customer language training to staff Employee acquires customer language for instrumental/ extrinsic motivation à identifies a utilitarian value Customers communicate ‘back’ (product complaint) to subsidiary which has to respond Codeswitching/- mixing (different languages) Email chains switch between English and Danish/German Switch to L1 for clarification purpose Participants borrow words from other language à positive impact on social relationships Passive multilingualism (different languages) Differences in language fluency (L1 vs. L2) Mutual intelligibility of L1s 107 | Page This process of rereading and cognitively absorbing data, paraphrasing them in my own words and entering the summarizing statements into the matrices helped capture the ‘holistic picture’ of the data. In this analytical phase I also tried to ‘put myself into the respondent’s shoes’ and to remember the interview situation, the overall impression gained and the respondent’s attitude to my project or how s/he expressed thoughts and articulated meanings. This attempted to reach more authentic interpretations of the data. After this visualization of the empirical data (i.e. within-case displays), I started writing them up in the form of four individual case reports (see Chapters 4-7). 3.5.5 Crafting case study reports in an iterative fashion: ‘telling the story…’ What is commonly referred to as ‘writing up’ sounds easier and more mundane than it actually is. It is more like an art or craft of storytelling (e.g. Stake 1995) or scientific narration that requires representational, e.g. stylistic, argumentative, rhetorical, textual and literary, considerations that shape the resulting case study report. The case researcher needs to provide enough contextual description in order to convey the scientific message in a meaningful way. It is the analyst who makes these strategic analytical choices that inform the story s/he wants to tell. The mode of writing and representation are not arbitrarily chosen but scientifically informed or rigorous and yet subjective. This means that analysts decide what experiences of their participants to reveal and what information to omit depending on research foci and interests; there are different ‘units of narrative’ (e.g. individual social actor versus a social group or institution) that entail different ‘levels of generality’ (from specific, context-sensitive to more general statements about social actors, behaviors, experiences or acts and events). In other words, the focus of the narrative, be it an individual or act, functions as “the main carrier of the argument”, hence influences the analytic level of generality (Coffey & Atkinson 1996: 112-115). Flick (2009) stresses that subjects transform everyday experience, events or activities embedded in a natural and social environment into knowledge, which needs then to be interpreted and understood by the researcher (through the ascription of meaning) and eventually translated into readable texts which in their sum construct the world. (This creates a triangular relationship between experience, interpretation and construction.) Scientific reports as representational accounts of participants’ multiple realities do 108 | Page not speak for themselves, rather are subject to interpretation too; they need to be interpreted by their readers. This means both authors of social science texts and their readership are all involved in and contribute to the social construction of reality based on texts. “Reading and understanding texts become active processes of producing reality, which involve not only the author of (in our case social science) texts, but also those for whom they are written and who read them” (Flick 2009: 79). As Stake (1995: 126) formulates it: “The researcher should try to anticipate what vicarious experiences will do for the reader, should try to organize the manuscript so that naturalistic generalization is facilitated. By providing information easily assimilated with the readers’ existing knowledge, the writer helps readers construct the meanings of the case.” In sum, any qualitative data set has the potential of telling multiple tales constructed by the (case) researcher and their understanding of the social world. In the present study, the crafting of the written case reports was done in a cyclical and iterative fashion. First, drafts were written based on the coding phases, in particular coding phase II, pattern coding and the matrix formation (data display), as explained above. The four reports drafted were initially rather chaotic and unstructured. Several revisions and editorial work were necessary during which the reports’ order and structure changed multiple times, information was cut out or revised and new data added. Both ‘individual’ and more ‘general’ units of narrative or statements can be found since I studied language choice at all levels. Representative quotations were carefully selected, and if necessary translated, indicated as “[my translation]”, and included in the reports since they served as exemplars of themes and concepts (Ryan & Bernard 2000, Miles & Huberman 1994), of ‘particular’ cases (Stake 2000) and of the respondents’ subjective understandings. Quotations also served the purpose of making the case stories or accounts more lively, partly personalized and authentic. In other words, this was an attempt to raise the findings’ quality as to credibility (trustworthiness and authenticity) (Guba & Lincoln 1994, Hatch 2002). 3.5.5.1 Drawing and verifying conclusions The findings were cross-checked through revisiting coded and non-coded (‘raw’) data (interview transcripts) to make sure I had interpreted them as authentically as possible. Absorbing and reinterpreting coded and raw data with fresh eyes and comparing them with the interpreted, 109 | Page displayed data in the reports helped finalize them. In this process, when making the final adaptations, I made sure to always indicate where the evidence came from. I had realized that data were sometimes presented as if they had been universal perceptions within a case study, which they were not, and so the source(s), i.e. who reported what, were made explicit. The phase of conclusion drawing and verification was also characterized by looking for negative evidence (Miles & Huberman 1994). However, not all of the authors’ suggestions for verifying results could be included. Neither the entire reports nor individual quotations were cross-checked with respondents which would have increased the trustworthiness of the findings. Due to the size of the sample and data set, and time constraints, I was unable to pursue and administer this kind of verification process which represents a methodological limitation. 3.5.5.1.1 Treating alternative data sources (data triangulation) In the phase of writing the reports, the alternative data sources were analyzed, not in NVivo but manually, in order to firstly triangulate sources and secondly to gain potentially new insights into language choice in MNCs. The documents were analyzed in an exploratory fashion by looking for themes of interest, as generated in the coding phases. Again, through semi-inductive manual coding on hard copy and an interpretive analysis, I aimed to understand the contents of the documents. The latter were summarized and integrated into the case reports. DK_1’s language guide (see Chapter 4, Section 4.3.1 of Case Study I) revealed information about translation solutions and quality, terminology management (corporate term base) and language ambassadors (similar to the notion of language buddies). Relevant information was extracted and quoted in the case report for the purpose of displaying data and raising authenticity. DK_2’s language policy statement (see Chapter 5, Section 5.3.1 of Case Study II) provided insights into guidelines on language choice. It was vital to compare the draft and the final version of the language policy to check if any fundamental changes had been made. The final version did contain new themes such as ideas about the choice of English for HRM (selective recruitment) and educational purposes (English training). These themes were included and discussed in the findings (case reports). 116 | Page As the respondent (#1a) explains, English has been defined as the group or corporate language that assumes the function of an internal lingua franca. It is for instance chosen for interunit communication purposes (“when we communicate among the companies”) since the participants, i.e. different organizational units or their agents, all speak different languages. Other than that, participants affiliated with a setting (location) will choose the national language of the setting assuming the functions of the HQ language and local languages. The way the respondent (#1a) expresses his thoughts on language choices within the speech community (the repeated use of “of course”) also indicates that these choices seem somewhat logical or obvious to him. Similarly, respondent #1b stresses the function of local languages within their national environments: But we have employees from other countries who naturally may communicate internally in the language they have in, between the administrative people from Germany or Holland or Hungary and when they communicate just two or three persons, then they would communicate in their native language. (Respondent #1b_Assistant to Group President/Management_HQ_Danish) The choice of local (national) languages among local language speakers is considered a ‘natural’ phenomenon or language choice behavior. These choices are not described as some top-down policy or language management, rather as common-sense, i.e. unmarked, choices. In other words, both respondents’ (#1a, #1b) rationalizations of language choices (“of course”, “naturally”) also tell us something about the behavior norms of DK_1 as a multilingual speech community. That is, the communicative implicatures suggest that Danish (as the HQ language) and host-country languages (as local languages) all represent unmarked languages that are chosen among L1 users in their respective settings. English is the unmarked lingua franca (for interunit communication purposes) whose function is also described as ‘group language’, ‘official corporate language’, ‘company language’, ‘primary language’ or ‘our global language’. 117 | Page 4.3.1 A language guide (on English usage) or language policy? In relation to the function of English within the MNC network, we learn that the company has a “DK_1 language guide”. The ‘language group’ is a team of four employees,46 “who were given the task of working out a language guide for [DK_1]’s corporate language English” (p. 76 of the language guide). Although stated as if it had been a top-down approach and a decision had been made at the top management level, the focus group interview indicates otherwise; it was more the language group and colleagues with an interest in languages from different departments who thought that DK_1 should have a language guide. We could say the idea emerged at the department level and was subsequently proposed to the corporate management (bottom-up). The coordinator of this project was respondent #1a, a Danish in-house translator of English, affiliated with the Technical Marketing Department, who collaborated with three colleagues from Corporate Branding, Group Production and Corporate Communication, all having an interest in languages or even a degree in linguistics. Overall, it is defined as their “common reference book in questions relating to English” (p. 1). More specifically, it is a text production manual; it advises staff on how to structure and write a coherent text in English, which standard variety to use (i.e. British English), how to spell corporate names correctly, which style, terms, expressions and phrases to prefer and which one(s) to avoid in different business scenarios. It tries to standardize and improve corporate language usage. Furthermore, it is explained how cultural differences can affect one’s text. In addition, it offers language assistance to employees; copywriters, translators and secretaries, the language group, in-house “language ambassadors” can be contacted and asked for help. A language ambassador is an employee that volunteers to promote the language guide locally and may act as a language consultant within the local company or even offer language assistance (e.g. proofreading). More specifically, it says: “A big corporation like [DK_1] employs many different specialized people, such as copywriters, translators and secretaries. These colleagues write and translate internal and external texts every day and thus have great experience in the language field” (p. 77 of the language guide). This reflects that in-house translation is common practice characterized by both professional and nonprofessional solutions. Furthermore, the corporate 46 Three of those were the focus group interviewees (Respondents #1a, 1b, 1c). 118 | Page glossary (which will be further detailed in the context of terminology management – see Section 4.4.2.1), and other (online) dictionaries that can be consulted are listed. This language guide exists as a PDF file and hardcopy booklet in Danish and English. The English version sometimes gets requested by local units since they use it as learning material for local English training. Some informants associate this language guide with a sort of language policy, while others clearly delineate it from an explicit language policy, as follows: Yeah, we do. I’m hesitating whether… I am thinking, is it a policy or just guidelines? I can’t remember. […] The document is trying to put forward some good habits, some good views on what is good language. It is not a set of rules, that in this situation you do this, and in this situation you do that and you can look it up in this book, you can’t. Only a few pieces of good advice as to how you write or how you contact other people. But it’s more a question of trying to spread the view that good language is important for a global organization as DK_1. Because good language is also part of our image. So if we are not concerned about the language both when we send out some written material or when we communicate with people face-to-face then good language is also part of our image. And that’s what we try to do in the language guide, try to point out to our colleagues all over the world how important it is to be aware that language is part of our common image. (Respondent #1a_Senior Translator_HQ_Danish) The respondent (#1a) explains that the language guide is not a language policy, which he sees as a set of rules regulating situational language choice, but rather some guidelines on English usage and where to find language assistance within the company. Although the guide is primarily about English text production, it seems to have another purpose, that is, raising awareness as to the importance of languages including external communication and their potential positive or negative impact on the corporate image. In this sense, the language guide tries to draw employees’ attention to language and communication matters at a generic level. The notion of a language policy encompasses further interpretations. For instance, respondent #3 (Regional Business Director_S/AT_Danish) argues that there is no language policy, neither does he mention the language guide: “Not a policy but the only thing is, we have a company language and that’s English. It was actually Danish for a lot of years. We were expanding internationally but now it’s English.” He does not associate the choice of a company language with a language policy, rather describes the functions of languages that have changed over time related to international expansion. Others link the notion of a language policy to the level of English proficiency that is 119 | Page required for attending corporate training sessions or to in-/expatriation policies. The understandings of a language policy are multiple which means there does not seem to be an explicit or well-defined policy within the corporate context of DK_1 upon which respondents commonly agree. One could say that these first insights indicate that DK_1’s languages serve various functions that are manifested in behavior norms within the community; L1s are chosen within the units and English as a shared (corporate) language across units. There is no explicit language policy (statement) defining these functions and norms or guiding language choice per se. Yet, there is a language guide that attempts to standardize English usage and suggests language assistance (in the form of translators, language ambassadors or dictionaries, etc.). These preliminary insights indicate that language choices are embedded in social context, yet do not grasp the rationales underlying language choices, which will be the focus of the remainder of this case study. 4.4 Language Choice at DK_1 Informed by Social Context The following sections will detail how social context influences language choices, more specifically, the contextual dimensions that inform language choices at DK_1. These include English, the HQ language (Danish) and local/customer languages (see also Chapter 2, Section 2.3.3). 4.4.1 Social-linguistic dimension: participants’ language proficiency informs language choices The first contextual dimension that informs language choices is of a social-linguistic nature. Generally speaking, language proficiency stands for participants’ linguistic repertoires which include usually one first language (L1) and possibly foreign languages (FLs), i.e. L2, L3, L4, etc. More specifically, the notion of proficiency embodies both ends of the continuum; proficiency ranges from high to low (nonexistent) proficiency in languages. In this sense, varying language proficiency not only informs language choices but may also induce positive and negative associations with participants, thus language can represent both a resource and a barrier. Note, 120 | Page participants are to be understood in a broader sense, i.e. can represent the individual language speaker but also larger entities such as departments or organizational units (HQ, subsidiaries) that have different L1s (i.e. HQ versus local languages) and make different language choices (see also Chapter 2, Section 2.3.3.1). 4.4.1.1 Proficiency informs the choice of English Participants’ language proficiency informs the choice of English. Participants, organizational units (headquarters, regional headquarters and national subsidiaries), more precisely their agents, all speak different first languages. English is chosen in many different communicative events when a shared language is needed as in interunit communication. These events include both written and oral media (e.g. email correspondence, presentations, meetings, corporate training, etc.). DK_1’s participants perceive the choice of English in various ways and often associate it with some challenges since English fluency varies. The focus group interviewees (Respondents #1a, #1b, #1c), for instance, acknowledge that the choice of English (in its function as a group language) influences their daily work insofar as it is a foreign language for the majority of staff and everyone is a better communicator in her/his native language (L1). Respondent #6 (Product Manager_S/AU_English) is slightly more critical by arguing that “there is a natural understanding that most people can, you know, work with English, there [are] some that can’t, to be honest”. Proficiency not only informs language choice but sometimes also the choice of media and channels. Respondent #2 (Member of Regional Management_S/DE_German) once had a Czech employee (with Russian skills and marginal knowledge of English and German). They chose English as a lingua franca but stuck to a written medium since the Czech colleague preferred email correspondence, so that he could consult a dictionary. In this case, participants’ language proficiency not only informs the choice of English but also limits it to a specific medium (written) and channel (email) of communication. The choice of English as a shared or group language for interunit communication purposes also includes English native speakers (NSs) such as the Australian subsidiary, and can be associated with problems. Locally, Denmark is brought up and Danes’ English skills are commented on in 121 | Page two ways; on the one hand, they are described as being generally very proficient, on the other hand, “there’s always mistakes and then we call it ‘Danglish’” (Respondent #7_Account Manager_S/AU_English/Punjabi). In another respondent’s (#6) words: You do have some challenges with certain colleagues who struggle with their English. And then, you know, even on a day-to-day basis when you deal with the Danes, it tends to be, you know, you don’t always get your message across and I suppose that’s probably not just a language thing but also a cultural thing. You know, different cultures sometimes say things differently, you know. So, that’s interesting. (Respondent #6_Product Manager_S/AU_English) This respondent (#6) adds another dimension to the problem of English choice and proficiency, that is, perceived cultural differences; they may also affect mutual understanding and lead to misunderstandings. English may also be chosen in external encounters. At a product launch in Taiwan organized by the regional HQ in Singapore including various participants with different proficiency, English has been chosen as the shared language. Though, some problems linked to this choice are mentioned. An English speaker (Respondent #7_Account Manager_S/AU_English/Punjabi) noticed limited English skills and different ways of speaking among the Taiwanese and other Asian cultures. Some participants seemed more shy and reserved when talking. As a reaction to the nonnative speakers’ (NNSs’) behavior and perceived cultural differences, the respondent (#7) and his Australian colleagues adapted their English and spoke very slowly. Similar impressions are gained from interactions with Europe. Another English NS (Respondent #6_Product Manager_S/AU_English) attempts to choose a ‘simple’ register and style (e.g. by excluding slang terms or Australian expressions) in written communication, in case the recipient’s English skills are poor. In these examples, the respondents choose English as a native language (ENL) but simplify it in oral and written media for communicating with NNSs while the latter accommodate the English NSs. Apart from these countries mentioned above, variation in English fluency is also brought up in the context of Eastern Europe, Germany, Austria, Southern Europe, South America. Apart from countries, job functions and functional areas are mentioned, too. Engineers, for instance, are a group of staff that is reported to struggle with the English terminology of the technical register. Within the financial and legal registers, a financial expert in the Ukraine is 122 | Page named “who knows the law by heart but his English is a disaster [my translation]” (Respondent #2_Member of Regional Management_S/DE_German). Overall, English is commonly chosen in interactions uniting different participants with different first languages or varying language proficiency. English represents a pragmatic choice if a shared language is needed, but is also associated with some problems. These perceived problems relate to counterparts’ (NNSs’) limited or lacking English skills and diverse cultural backgrounds. 4.4.1.2 Proficiency informs the choice of the HQ language Participants’ proficiency also informs the choice of the HQ language Danish. Danish gets chosen in different communicative situations such as meetings if all participants speak Danish, as explained below: Sometimes there are meetings where everything is presented in Danish. […] There are hardly any foreigners in the corporate management, there are a few. I know that because I have lived there myself. […] and not every Dane, even at the top management level, although they all speak fluent English, but if one starts speaking near-native or good English, some words are misunderstood. And you want to avoid that. So, Danish is chosen instead. I don’t mind because I understand almost everything except for a few dialect expressions [my translation]. (Respondent #2_Member of the Regional Management_S/DE_German) Danish is chosen at a meeting if all attendees master Danish either as an L1 or FL, as the respondent (#2) for instance. In this sense, the NNSs linguistically adapt to the NSs of Danish. Additionally, the respondent (#2) assumes that not all Danish executives feel totally comfortable speaking English which is why Danish is preferred to English. These internal meetings might also be set up across organizational units, among expatriated Danes in a regional (management) position and Danish-skilled international managers. If Danes represent the majority and nonDanes are outnumbered (e.g. 15:3), even this kind of “international meetings” are held in Danish. The respondent (#2) adds that he also senses a bit of national pride that favors the choice of Danish in these contexts. In sum, language proficiency of the participants involved in the 123 | Page communicative situation, supported by the number of languages speakers (majority of L1 users), and possibly an ethnocentric attitude inform the choice of the HQ language Danish. Regarding the choice of Danish, we also learn that proficiency informs the choice of Danish but limits it to easier subject matters (topics). This can be seen with an English NS with Danish proficiency (Respondent #4_Copywriter & Business Development_HQ) who chooses Danish only for socializing (purposes) because his Danish is not good enough to discuss more complex matters such as professional or business-related topics. (He chooses English for presenting his work in meetings.) In this sense, the choice of Danish is limited to simple conversation topics (e.g. small talk). That is to say language proficiency not only informs language choice but also the choice of possible topics in a foreign language. 4.4.1.3 Proficiency informs the choice of local languages (and transition to English) The following quotation highlights that participants’ proficiency can also inform the choice of local languages in the first place which may shift to English later in the internal communication process: I think it all depends on which language is most convenient in the situation. If you are in a situation where those people speak the same local language, then, of course, they would use that local language instead of switching to a language which is foreign to all of them. Then they would prefer to speak in their own language. But that would also be in that situation. And when the situation develops and they have to carry it further up [from the local to the HQ level] then they would have to switch to English. (Respondent #1a_Senior Translator_HQ_Danish) Local participants involved in a given situation would choose the local language in the first place. (This also confirms behavior norms, the function of local languages and their unmarked nature – see Section 4.3.) This example refers to market research where the regional and local managements would explore potentially interesting adjacent markets and carry out studies or analyses in the local language. Once the information becomes relevant for and needs to be exchanged with other organizational units such as HQ, then the lingua franca English would be chosen (or maybe translated into) in order to ensure mutual understanding between participants. 124 | Page 4.4.1.4 Proficiency informs the choice of (translation into) local languages (double standard) If information that is distributed internally needs to reach staff members lower down the hierarchy, such as production workers, local languages are usually chosen over English. The rationale is that most production workers do not speak English and thus need to be addressed in their L1 (accommodation). In the case of written information the choice usually involves some form of translation from English into local languages. The international staff magazine, for instance, is a text that is outsourced to an external translation agency and professionally translated into the major local languages English, German, French, Hungarian, Chinese and Portuguese and proofread in house by NSs of these languages. The target languages are selected according to market size and number of staff, i.e. language speakers, based on an ‘investment-return calculation’, i.e. economic considerations. This gives the local languages different priorities and creates a hierarchy among local languages, more accurately double standards in that only some host countries are privileged to get access to certain information in their local language while others do not (based on the number of language users). One could even argue that this is some form of divergence (see also Chapter 2, Section 2.3.2.2) given that the receivers are known for their limited FL skills, ergo knowingly excluded from the information distributed via the staff magazine. Some other internal documents are only relevant for a specific region and therefore produced regionally, i.e. localized, in a nonprofessional manner. In the region Central and Eastern Europe, information for sales staff (e.g. sales newsletter) exists in Russian, German and English. Usually a bilingual staff member who is “good at” (Respondent #3_Regional Business Director_S/AT_Danish) Russian and English, for instance, writes and translates this text genre. A good command of the source and the target languages is considered sufficient to fulfill the translation task for this regional internal communication purpose. It could be summarized that these text genres addressing local staff are translated (both centrally and regionally) into certain major local languages because most of the intended receivers do not master English but speak the local language only. In this sense, the participants’ language proficiency, once again, explains this kind of language choice (linguistic adaptation/accommodation). 125 | Page 4.4.1.5 Proficiency informs the choices of English and local/customer languages (involving interpretation) English might be chosen for oral interaction among colleagues but combined with Russian based on the participants’ poor or lacking English skills; this involves some form of nonprofessional interpretation, as respondent #4 (Copywriter & Business Development_HQ_English) explains: Yes, I guess so because one of the colleagues couldn’t explain himself [in English], then one of the guys would interpret for me [from Russian into English]. So it would be intercollegially, you might say, that is two Russians and me in a conversation, and one Russian says something that I don’t understand, then they speak in Russian, and this colleague would translate it for me [into English]. But that’s not a formalized thing. This shows that if English cannot be chosen by all interactants, a bilingual person may mediate between the parties. It also highlights that these choices are not formalized in any way but rather made informally. Similarly, in oral interaction with customers the choice of their language (L1) combined with nonprofessional interpretation is common in certain regional communicative events. At the subsidiary level, it is reported that 90 % of their customers (also businesses) in Eastern Europe (Russia, Ukraine, Serbia) and Central Asia (Kazakhstan, Turkmenistan, Uzbekistan) do not speak English. If the Regional Business Director (Respondent #3_S/AT_Danish) meets with a customer (e.g. a mining company), a bilingual employee, usually the local sales manager, serves as a mediator between the two parties (e.g. English ßà Russian or Serbian). (Interestingly, the counterparts’ nonexistent English skills, i.e. their language proficiency, are the reported reason for this language choice and the customers’ accommodation rather than the customers’ role within the employment domain and linked asymmetrical buyer-seller relationship.) 4.4.1.6 Proficiency informs code-switching/-mixing Code-switching or code-mixing commonly occurs as the result of participants’ various language proficiency and may apply to both written and oral media. Internal email correspondence can be bior multilingual. Danes communicate with Australians via email and somewhere along the way 228 | Page English for educational purposes; the Department Head himself has studied the financial and stock exchange register and sent his assistant on short-term stays in the UK. Other forces are related to HRM; national work regulations or requirements, maybe of a legal nature, request the choice of and translation into the local language Russian when it comes to work contracts. Overall, these social forces render the MNC and its choices multilingual by definition. It could be concluded that AT_1 is a multilingual speech community that unites numerous settings (locations) all speaking different languages, English being the lingua franca or divisional corporate language (lingua franca). This case study (AT_1) has shed light on four contextual dimensions that inform language choice: the social-linguistic dimension language proficiency, the social-relational dimension based on participant roles, role relationships and politeness, the socialpsychological dimension including attitudes to languages (with imposing power) and motivations, and the social-regulatory dimension external forces (with imposing power). The contextual dimensions inform different language choices and also impose languages, including the HQ, local/customer languages and English, in different communicative situations, for internal and external communication purposes. The dimensions influence the individual staff member and/or the corporate management level which has to be refracted into individual decision makers who choose languages in a given communicative situation in their own right rather than in a consensual or unified manner. This suggests that social context is multilayered which renders language choice a complex and multilingual social phenomenon. The multilevel and multidimensional analysis of language choice lets us conclude that language can hardly be regulated or managed in a centralized manner in an MNC network such as AT_1. In other words, the findings indicate a limited manageability of language since language choice is contextually bound. These case study findings also undermine the usefulness of single-language policies but suggest multilingual policies and management activities, if at all necessary. 229 | Page 7 LANGUAGE CHOICE IN A MULTINATIONAL CORPORATION (MNC): A CASE STUDY (IV) OF AT_2 7.1 About AT_2: Facts and Figures AT_2 is a construction business headquartered in Austria. It employs around 7,400 employees and has subsidiaries in all Austrian provinces, 16 European countries and Australia. The main markets are however adjacent ones, such as Hungary, Germany or the Czech Republic. The MNC has expanded to Eastern Europe relatively recently whereas business relations with Germany have existed for decades, (probably influenced by sharing a language or belonging to the same language community). The company is structured into smaller units, specialized in different construction areas, which work independently but are supervised by corporate management. The area of track construction and maintenance is highly technical and innovative. It is directed by the division Rail (subsidiary), which itself operates more internationally and centrally. The firm’s clients are primarily from the public sector, e.g. national governments and local authorities or institutions, occasionally private businesses and international firms from the industry. The company’s contracted turnover in the fiscal year of 2012/13 was EUR 1,463.52 million/1.5 billion (≈ DKK 10.9 billion).51 (This case company (IV) is the second smallest MNC investigated in this collective case study.) 7.2 Language Choices on AT_2’s Website First insights into language choices can be gained from the company’s online representation. The corporate webpage (www.AT_2.com) is in German with a condensed English version which gives more or less the same information, e.g. extensive descriptions of AT_2’s expertise and services or news on current projects in Austria. In a language drop-down menu, the user can change the language and choose between Hungarian, Czech and Slovakian. The local language versions are structured a little differently and contain information that is relevant within the local context (e.g. 51 These data have mainly been collected from the company’s website. [retrieved February 19, 2014] 230 | Page on local projects, financial situation, visual material). In the investor relations section of the German version, financial documents can be downloaded (since 2007/08); there are biannual reports (spring/fall) and a final financial report. Only the spring report is translated into English, which might be hard to find for an English-speaking readership because it is uploaded on the German website. The foreign language versions do not contain any financial reports. The Hungarian website differs from the other sites not only in content but also design. It is not possible to navigate back to the German corporate site. The local contents are in Hungarian which are translated into German and English. A user can easily switch back and forth between these three language interfaces. It seems to be a rather comprehensive trilingual local website with its own design and functions. The only documents that can be downloaded are press-related documents in the original language they have been published in, regardless of language interface. This illustrates that language choices regarding the company’s online representation can be both localized as in the Hungarian subsidiary’s case and fairly centralized as to the remaining local websites where one de facto stays on the corporate website but can swap between language interfaces. The fact that the local subsidiary also has a German version suggests a stronger focus on the HQ language. AT_2’s online job advertisements for Austria and Germany are available in German. No foreign language skills are required or specified in the job descriptions. Neither can one find any postings written in English. Overall, these first impressions on language choice gained from AT_2’s website reflect variation in language choice across organizational units (settings). One could say the company presents itself as a multilingual speech community where participants choose different languages in different social contexts (e.g. home and host countries). The corporate website alone (which opens up in German and has a multilingual interface) leaves the impression of AT_2 being a German-speaking MNC with a polycentric outlook. In the following, I will look into the notion of AT_2 as a multilingual speech community, language functions and its language choice norms in more detail. 231 | Page 7.3 AT_2 as a Multilingual Speech Community: Language Functions and Language Choice Norms AT_2 can be described as a multilingual speech community where different languages serve different functions which also lead to certain language choice norms within the MNC network. Within the AT_2 network, German being the national language of the home country assumes the function of both the HQ and corporate languages which are chosen within the HQ environment and for internal communication across organizational units (settings). English serves the function of a lingua franca (also for internal communication purposes). Other languages have the roles of local languages which are primarily chosen within the settings of the national subsidiaries but also in interunit communication. All languages become customer languages once they are chosen for external communication purposes such as engaging with buyers and customers in local markets. In order to get a more detailed picture of AT_2 as a multilingual speech community, it is crucial to look into these functions and their meanings in more depth and language norms within the community. First of all, the question whether there is an official corporate language is interesting since it reveals the overall unmarked nature of German: Well, that is of course German. I mean we are an Austrian business if I may point that out. And many of us… I might add to the question of a corporate language, we are of course an Austrian corporation that has expanded only in the past years. Well, that hasn’t been the case for a very long time. Also, the structures are still in the process of being developed. There are also only a few in house that speak English at an acceptable level [my translation]. (Respondent #1_Managing Director International_HQ_German)52 The respondent (#1) seems almost offended and stresses that German is undoubtedly the corporate language. This also means it is often chosen for communicating across organizational units (setting). Yet, he associates foreign market expansion and internationalization with the choice of English, which is a relatively new development. Similarly, the Head of Corporate Controlling (Respondent #3_HQ_German) argues that “we see ourselves as a German-speaking international business relying on English, only if necessary [my translation]”. In this sense, German 52 Each interviewee is assigned a number (e.g. #1). Further characteristics indicated are the person’s job title or function (e.g. Managing Director International), location (HQ or subsidiary) and first language (e.g. German). In the following, the subsidiary will be abbreviated ‘S’ and its location indicated by the country code (i.e. ‘AT’ for Austria). 232 | Page is the unmarked language while English seems to be the marked language, yet unmarked business lingua franca if German is not an option. The unmarked nature of German and its function as both the HQ and corporate languages which are chosen for internal communication purposes are confirmed at the division by respondent #6 (Head of Department_S/AT_German): The main communication, meaning internally is of course German. […] [The corporate language] is German, yes. The corporation is… I’d say 60-70 % of the corporate staff is located in Austria or Germany and key players and also the shareholders all master the German language. Therefore, the communication in the corporation is in principle German [my translation]. Also, the fact that local subsidiaries speak local languages and choose them in daily business in an unmarked fashion is described as a well-known fact by respondent #4 (Project Manager_HQ_German): Then he, of course, employs local people who only speak Serbian or Greek, and then it has to continue in the local language. […] The daily business in the host countries is of course conducted in the national language, that’s obvious, otherwise it would get too complicated. But at some point it stops since a construction worker on site will neither speak German nor English, obviously [my translation]. The choice of local (national) languages among local language speakers e.g. at a construction site is considered an ‘obvious’ phenomenon or language choice behavior. These language choices, as described above, are not perceived as some top-down policy/management measures, rather as common-sense, i.e. unmarked, choices. In other words, the respondents’ (#1, #3, #6, #4) rationalizations of language choices (“of course”, “obviously”) also tell us something about the behavior norms of AT_2 as a multilingual speech community. The communicative implicatures suggest that German (as the HQ and corporate languages) and host-country languages (as local languages) all represent unmarked, yet contextually-bound languages that are chosen in their respective contexts. The function of English could be described as an auxiliary business lingua franca which is chosen if German is not possible. These preliminary insights indicate that language choices are embedded in social context but do not 233 | Page grasp the rationales underlying language choices, which will be the focus of the remainder of this case study. 7.4 Language Choice at AT_2 Informed by Social Context The following sections will detail how social context influences language choices, more specifically, the contextual dimensions that inform language choices at AT_2. These include English, the HQ language (German) and local/customer languages (see also Chapter 2, Section 2.3.3). 7.4.1 Social-linguistic dimension: participants’ language proficiency informs language choices The first contextual dimension that informs language choices is of a social-linguistic nature. Generally speaking, language proficiency stands for participants’ linguistic repertoires which include usually one first language (L1) and possibly foreign languages (FLs), i.e. L2, L3, L4, etc. More specifically, the notion of proficiency embodies both ends of the continuum; proficiency ranges from high to low (nonexistent) proficiency in languages. In this sense, varying language proficiency not only informs language choices but may also induce positive and negative associations with participants, thus can represent both a resource and a barrier. Note, participants are to be understood in a broader sense, i.e. can represent the individual language speaker but also larger entities such as departments or organizational units (HQ, subsidiaries) that have different L1s (i.e. HQ versus local languages) and make different language choices (see also Chapter 2, Section 2.3.3.1). 7.4.1.1 Proficiency informs language choices for foreign market expansion The company’s language proficiency informs possible language choices which also influences or de facto limits foreign market expansion. The industry in which this particular MNC operates is quite specialized, competitive and nationally bound through national funding and local governments or authorities being crucial interactants. It is therefore hard to compete with 234 | Page national, local competitors, which also prevents the company from entering e.g. French-speaking markets. Foreign market entry often requires the choice of the local language. This also explains that France is very dominant in the African French-speaking countries, and Spain in South America, as reported by respondent #1 (Managing Director International_HQ_German). Furthermore, he details: We don’t even enter these markets because we simply lack experts who could navigate and operate in the French-speaking or in the Spanish, Portuguese areas [my translation]. (Respondent #1_Managing Director International_HQ_German) This reflects the local linguistic requirements, i.e. the choice of the local languages which prohibits AT_2 from entering these ‘closed’ markets. The account also indicates that English does not have the function of a business lingua franca in these major-language markets within the construction industry. In this vein, we learn that German is chosen for different business transactions (including entering new markets). One way of expanding globally within the construction industry is ‘following’ local internationalizing clients abroad, as respondent #4 (Project Leader_HQ_German) explains. This means that AT_2 constructs a building in a foreign market, but for a German-speaking client. The same phenomenon has been experienced with a Belgian joint venture in Hungary where the Belgian construction company had primarily French-speaking clients who entered the Hungarian market. One could say that participants’ language proficiency informs the choice of German for external communication purposes which not exclusively but to some degree also defines AT_2’s foreign market expansion. (Note, AT_2 also engages in NonGerman-speaking markets and construction projects, as we will see in the following sections.) 7.4.1.2 Proficiency informs the choice of English Participants’ proficiency informs the choice of English for interunit written communication, as explained by respondent #1 (Managing Director International_HQ_German): 235 | Page Every Friday we get a commercial report by email, which is sent to me, to Mr. X and Mr. Y. This one contains a list with four bullet points summarizing the essential things that have happened this week. And here you can read about what he has to report to us. And this is between our subsidiary in Serbia, in Belgrade and between us at HQ in English. That’s how it works. Well, and I work a lot in English, I’ve also been travelling the entire last week, am engaged in three projects in Bosnia. Then I actually speak English from morning to night [my translation]. The respondent (#1) explains that his work and responsibilities include interunit supervision and collaboration. In this communicative event English is chosen as a lingua franca given that the participants involved speak different L1s but share English. Participants’ proficiency informs the choice of English for interunit oral communication. According to a Department Head (Respondent #5_S/AT_German), the division (subsidiary) hosts international annual meetings with local managements visiting from e.g. England, Hungary, Romania, where English is chosen as a lingua franca. Though, not all participants master the English language, and so the Polish assistant from England interprets for two Polish colleagues in a nonprofessional fashion. The choice of English is also associated with some problems among participants. For instance, the business navigates in a highly technical area dominated by engineers whose English skills are described as being inadequate in some contexts. This technical industry includes a complex terminology (register), which renders working in English as foreign language more difficult. Apart from this group of staff, several host countries are mentioned where local staff has lower English proficiency. Respondent #1 (Managing Director International_HQ_German) mentions the Hungarian and the Czech units that are very localized in that they engage in mostly nationallyfunded construction projects: “It’s rare to find someone with English skills there [my translation].” Lower English proficiency is also associated with the former Yugoslavian countries, Italy or Greece. Some respondents believe that the linguistic situation will change in the future and English will become more dominant as L2 and eventually replace German, a phenomenon that is partly already observed among the younger generation. Also, respondent’s #6 (Head of Department_S/AT_German) experience with Romania and the unit there is that English has gradually become more important and notably senior managers have 236 | Page only started speaking English in the last five years. He recalls that first contacts and the start-up process were realized with interpreters because at the time they had no language in common. In the context of Romania, respondent #5 (Head of Department_S/AT_German) shares his critical thoughts on the choice of English as a lingua franca in communicative events like meetings, as detailed below: It also happens that we, I am thinking of Romania, that we sit around a table and say: ‘Well, do we try it in English, do we do it in Romanian or German?’ What I have noticed is, especially if you don’t sit together with [English] native speakers, a meeting conducted in English is often a problem because both parties cannot fully express themselves or often one party at least cannot express what it wants to say. You also lose quality when translating, of course, but if two nonnative speakers confer, the ambiguity gets worse. That’s also something I’ve observed; you discuss things for minutes until you realize we do mean the same, just use different labels for the same thing. So often it’s me who says: ‘No, please, stick to Romanian, our colleague translates into German for me instead we try in English.’ Of course, there are many native speakers, especially in Eastern Europe, or people who speak really good English, then it’s no problem. But if a Romanian communicates with a Hungarian in English, forget it; the conversation is going nowhere [my translation]. (Respondent #5_Head of Department_S/AT_German) In the respondent’s (#5) point of view, the choice of English as a lingua franca entails that one or both parties cannot fully express themselves, which may lead to a loss of information, inaccurate communication and misunderstandings. In interaction with Romanian colleagues, he often prefers mediation to the choice of English, since it seems more convenient and efficient to him in certain situations, as expressed above. Contrary to most of the other respondents, he has also been in contact with English NSs and fluent L2 users of English in Eastern Europe. With regard to external interaction, we also learn that the choice of English for doing business or engaging in international construction projects is a relatively new phenomenon, as explained below: Often the language [English] is a deal breaker because there is no one in house or there used to be no one who would understand the contents of the documents [my translation]. (Respondent #2_Managing Director Commerce_HQ_German) 237 | Page In the past, AT_2 often had to resign from ‘English-speaking’ projects which is seen as a competitive disadvantage or even business loss (“deal breaker”). (We will learn more about these large-scale projects later in this case study.) 7.4.1.3 Proficiency informs language choices (involving interpretation) Participants’ proficiency informs language choices for internal communication purposes; oral encounters such as meetings are characterized by the attendees with varying proficiency, which makes language choices context-dependent in that they differ across communicative events, as elaborated below: No, there is no policy. Before every meeting, it is decided, depending on who is present, in which language the meeting is conducted. Well, everybody assumes that it is conducted in German and if there is someone attending who does not at all master the language and if no interpreter is around, then we decide to have the meeting in English. But this is decided on situational grounds, there is no policy or rule for that [my translation]. (Respondent #3_Head of Corporate Controlling_HQ_German) According to the respondent (#3), the unmarked choice for an internal meeting is German; attendees may choose German as an L1 or FL, language assistance (interpretation) may be relied on, or English is chosen as a lingua franca which seems to be the last resort. (The account also reflects the respondent’s (#3) attitudes to languages, to which I will return in Section 7.4.3.) Respondent #1 (Managing Director International_HQ_German) draws a similar, maybe less ethnocentric but more neutral picture with regard to contextual language choice: It also always depends on who is sitting around the table. There are people who only speak the national language. There are people who speak the national language and English. There are people who speak the national language and maybe only German [L2]. And then we navigate or prepare a meeting with interpreters, so that everybody can understand everything [my translation]. 244 | Page related to the fact that he has an assistant with very good German skills and relatively little contact with externals in comparison to the engineer who still constantly interacts with clients, suppliers, who also talks to people. The business administration is more of a desk job, I’d say. The technical executive is the one who is out in the field, who meanwhile speaks perfect Romanian, I was told [my translation]. (Respondent #5_Head of Department_S/AT_German) The respondent (#5) reports on the case of two Austrian expatriated managers who both speak the host-country language Romanian. He stresses that the external interaction with clients and suppliers in the local market requires the choice of Romanian (“constantly interacts with clients”, “talks to people”) compared to the local business administration that is dealt with ‘from the inside’ and facilitated by a German-skilled assistant. Overall, the buyer-seller relationship and the related politeness strategy or appropriate choice in customer interaction suggest the choice of the latter’s L1 (accommodation/linguistic adaptation). 7.4.3 Social-psychological dimension: participants’ attitudes to languages inform language choices Participants have different attitudes to languages in that they assign positive and negative values or attributes to languages. We learn about positive attitudes to the HQ language (ethnocentrism), to English (geocentrism) and to local/customer languages (polycentrism). These multiple attitudes vary across participants, i.e. do not form a consensual ideology, and may even stir controversies among executives (rival attitudes) and have imposing power (internal forces). So, the different attitudes to languages inform different language choices (see also Chapter 2, Section 2.3.3.4). 7.4.3.1 Varying attitudes inform language choices for HRM purposes (selective recruitment and staff relocation) In the context of attitude to languages, we gain first insights into the HQ’s ethnocentric orientation as to market entry. New business units are always established by expatriated Austrians or Germans with the support of local partners. After four or five years it may be taken over by 245 | Page local managers, once trust has been established. Ideal are also people that once migrated to Austria, started working for the company (at HQ) and at some point want to go back to their home country (e.g. one of the Serbian Managing Directors has lived in Austria for many years). However, the organizational structure is rather centralized since the local management remains a shared responsibility between the subsidiary and HQ, where the former is in charge of the daily business and the latter has a supervisory function. An example is mentioned where the number of expatriates and the degree of centralization have decreased over time. Within the division’s (subsidiary) specialized area of track construction and maintenance (division rail), a project in Romania was inaugurated with 80 Austrians, of whom only two managers are left, together with 120 to 250 Romanian employees. Sometimes only top management, sometimes the top and middle management levels contain of (N)NSs of German, the rest are locals (with possibly some German proficiency). Expatriates have a supervisory function or disseminate knowledge in the first place and are often HQ personnel. Furthermore, we learn that ethnocentrism informs the choice of German for selective recruitment, as explained at HQ: …since we actually have the philosophy, from the corporation’s perspective, some departments see that a little differently, but generally speaking we have the philosophy that we are a German-speaking firm and for a lot of our employees, also in foreign countries far away, German is a hiring criterion. […] Well, as I said before, we see ourselves as a German-speaking international business relying on English, only if necessary [my translation]. (Respondent #3_Head of Corporate Controlling_HQ_German) The respondent (#3) shares his views on the company’s philosophy and thereby his attitudes to German in comparison to English. More specifically, we learn how a positive attitude to German (ethnocentrism) informs the choice of German for HRM purposes, namely for local selective recruitment. In other words, German-skilled workforce is explicitly sought, so German can be chosen for internal communication purposes (“German-speaking firm”). It seems the ethnocentric attitude goes hand in hand with a comparatively negative attitude to English in that the company or rather the respondent (#3) – others do not share his attitude – prefers German and is not willing to speak English, unless he has to, which could also be described as internal forces, which will be discussed subsequently. At first, the executive’s (#3) account resembles the description of 246 | Page an implicit language policy or commonly shared ideology (“philosophy”) that fosters the choice of German at different levels. Though, it also reveals that there are internal discrepancies among executives (“some departments see that a little differently”), namely rival attitudes which affect language choice in different ways or allow alternative choices to German. More specifically, what the respondent (#3) means by that is the attitudes to German and English respectively vary within the company, in fact within HQ, and are not universally shared; some departments or their agents disagree with the respondent (#3) in that they have a less ethnocentric but more geocentric attitude, which will be further shown below. Aligned with an ethnocentric attitude, as illustrated above, selective recruitment is operationalized in a way that local staff (with the local language as their L1) is often German-skilled (L2) or multilingual (+ English), so they can mediate between HQ and the local site as well as external third parties (e.g. local clients, authorities, governments). In Slovenia, for instance, a Germanskilled Slovenian has been appointed as the local project leader who had lived in Austria for 20 years. Also in Serbia, the local key contact persons master German. At the Polish subsidiary, a local linguist of German has been employed. What facilitates selective recruitment, although not explicitly mentioned in this context, are participants’ language proficiency; AT_2’s experience with the markets in which it operates is that the latter tend to speak German as a foreign language, less commonly English, as mentioned earlier, or no second language at all (see also Section 7.4.5 on external forces). German is not equally important to all HR Managers. Participants’ geocentric attitudes also inform the choice of English for various positions (selective recruitment). For instance, respondent #4 (Project Leader_HQ_German) does not mind using English, which also means he hires skilled workforce without German but English proficiency only (in addition to the local language). Similarly, respondent #5 (Head of Department_S/AT_German) values multilingualism but prioritizes technical expertise over German proficiency; if a qualified engineer with a local/customer language as his L1 speaks English as an L2 or FL, he does not need to master German. Aligned with a geocentric attitude, insights gathered at HQ reveal that English proficiency will be a hiring criterion for the International Department’s future assistant, when the current one not speaking English will retire. (The department was established in 2005, is still 247 | Page growing and has a consulting function in the initial decision-making process as to which projects to make a bid for.) An interesting expatriation case or rather the post-expatriation reflections reported at HQ shed light on the role and influential power of attitudes on language choice including language acquisition for expatriates and selective recruitment at the local level. More specifically, two respondents (#2, #4) recall an Austrian being expatriated to the local site in Serbia and describe various language choices and problems from their individual perspectives which also reveal different attitudes to English and the local language Serbian (which account for two rival language choices for HRM purposes.) First, the Managing Director Commerce (Respondent #2_HQ_German) recalls the case and shares his understanding as follows: We had a site foreman in Serbia, for instance, who was expatriated to provide local support. When I visited, he told me he would take Serbian classes…I was like: ‘First, you’ll never learn it anyway, forget it!’ I said: ‘You better learn English because that’s what most people understand. 95 % of those on site understand English, but no one will ever understand your Serbian [my translation].’ (Respondent #2_Managing Director Commerce_HQ_German) Interestingly, the respondent (#2) is frequently involved in business collaboration with Serbia, where (nonprofessional) interpretation is usually relied on because the Serbian senior executives do not speak English, as he himself explains, and also mentioned earlier. He seems to neglect or at least is not aware of his own experience where he has struggled with Serbian business partners not mastering English. Yet, the respondent (#2) has a positive attitude to English (geocentric attitude) that influences his language choices or advice for that matter regarding language acquisition. Thus, he recommends the site foreman that he learns English instead of the local language. (The geocentric attitude also informs his individual language choices, as he explains.) This expatriation case is also recalled by a Project Leader at HQ (Respondent #4_German) who draws a completely different picture. He shares his perspective on the local situation in Serbia, its unmarked language choice, and provides an explanation for the nonexistent English skills at a Serbian construction site or in Belgrade in general: 248 | Page He has been at the construction site for half a year. At any event, also at on-site meetings, he sat next to him [local mediator] and asked him: ‘What are they talking about?’ That’s of course tedious. But obviously, the local client is actually Serbian and as such doesn’t necessarily have to speak English. He doesn’t need to. It’s not necessary. He is in fact constructor of the city of Belgrade and in this capacity carrying out the project. And therefore they, of course, speak Serbian at the table; seems just logical. It’s very, very difficult and you need to hire and train local people who can run the daily business on the construction sites in the local languages [my translation]. (Respondent #4_Project Leader_HQ_German) One could say that the respondent (#4) has a positive attitude to the local language (polycentric), believes that it might be overtly chosen at the local site (unmarked language) and finds it more profitable or efficient to train and hire local staff in host countries. The problems the expatriate was facing locally, namely that he felt excluded, could hardly communicate with anyone but constantly needed an interpreter, and the local language choices appear rather obvious to the respondent (#4) (“of course”, “obviously”, “logical”). These two quotations illustrate how executives’ subjective opinions and evaluations (attitudes) can express the opposite as to which language local participants primarily choose and consequently which language skills are needed for expatriates to be. These attitudes influence language choices for HRM and educational (training) purposes in different ways; respondent (#4) would hire local candidates with local language proficiency and train them the professional skills needed for running a local construction site, whereas respondent (#2) would expatriate a German-speaking skilled worker and advise him to acquire English to be able to communicate with local staff. Put simply, one says the expatriate would need to learn the local language because local staff does not speak English, while the other executive believes learning English would be most useful or efficient. This expatriation case also illustrates that attitudes to languages may not only differ across the organizational units of an MNC (HQ versus subsidiaries) but also within an organizational unit as the headquarters. More on polycentric attitudes and their influence on HRM can be gained at the division (subsidiary); a Department Head (Respondent #5_S/AT_German) has employed their former Hungarian teacher as his assistant and a trilingual (i.e. Romanian, German, English) secretary in Romania who both have been in charge of most of the internal correspondence. He also says that 249 | Page in order to facilitate foreign market expansion and entry, the division (subsidiary) may recruit (or headhunt) nationals from potentially interesting markets with local/customer language proficiency and the local cultural knowledge. For instance, a Polish national working for a German joint venture partner and competitor has been headhunted because the Polish market being quite ‘closed’ seemed profitable. This could also include staff relocation; a German speaker with Italian skills has been located in the Italian-speaking part of Switzerland and is in charge of Switzerland and Italy. A polycentric attitude can further encompass the choice of local languages for educational purposes (language acquisition for expatriates). The division (subsidiary) once had a large project in England where approximately 50 German speakers (upper management) and 140 Polish speakers (operative staff) were expatriated to the local unit in Reading, west of London. English language classes were locally, externally organized at the University of Reading and also partially funded by the English government, the rest being paid by the MNC. The management staff did the three levels of the Cambridge language test whereas the operative employees passed the first level, so that they could communicate at a basic level in the host country. These training and acquisition measures were initiated in order to ensure communication at and integration into the foreign local unit (setting). In general, these kinds of classes are paid for by the company and take place during working hours or after work, depending on the offer of the language school. (On that note, it is mentioned that the willingness of participating in evening or weekend classes is greater among the higher educated staff members.) Besides, offering local language learning to staff is also regarded as a form of language policy (Respondent #5_Head of Department_S/AT_German). In 2005, AT_2 bought a company in Hungary closely collaborating with the (division) subsidiary. Although the Hungarian Managing Director of the local unit had a decent command of the English language and passive understanding of German, the Austrian core team to be expatriated wished to learn Hungarian, driven by polycentric attitudes. They acquired some basic knowledge in order not to be totally dependent on the language mediator but to get an idea themselves of what was going on in daily business and internal meetings. This team consisted of a Department Head (Respondent #5_S/AT_German), who initiated the language learning, a construction manager and two operative staff members. They had an external teacher coming to the office and 250 | Page teaching them Hungarian twice a week, sometimes during working hours, over a period of six months. Although the Hungarian counterpart knew English and German, and mediation was provided at meetings, Hungarian skills were believed to make daily business for the expatriated Austrians easier. This is an example where polycentric attitudes and motivation overlap, so do the executive and individual levels. We could say that the executive (#5) endorses local language acquisition (choice of a local language for educational purposes) for both integrative and instrumental (extrinsic) motives. In addition, one might also argue that the respondent (#5) has an overall positive attitude to the local language and believes he and his team should all learn Hungarian. 7.4.3.2 Geocentric attitude informs (imposes) the choice of English (internal forces) Participants’ attitudes may not only vary but literally compete with each other affecting interpersonal encounters, i.e. informing language choices at meetings. The Head of Corporate Controlling’s (Respondent #3_HQ_German) positive attitude to German (ethnocentrism) and negative attitude to English conflict with some colleagues’ positive attitude to English (geocentrism); the respondent (#3) claims that some of his coworkers intentionally take a person without German skills to a meeting, so the meeting has to be conducted in English, and they can ‘show off’ their English skills. To the question whether there are things he would like to improve or change, he replies the following: I would promulgate a much stronger directive that we are German. Some are simply, some find it cool to speak English, although German were possible too. And that’s not necessary. Just that I can show off, that I speak good English. Unfortunately, it happens that meetings are held in English because someone likes that and he intentionally takes someone with him to the meeting who doesn’t speak German well enough, so that the meeting is conducted in English, and all the others know that he is a great English speaker. That’s a shame. I find that stupid. That’s something I would change [my translation]. (Respondent #3_Head of Corporate Controlling_HQ_German) This is an indication that there is a variation in English proficiency in that these colleagues speak better English than others (maybe including the respondent himself). It is not totally clear 251 | Page whether the respondent (#3) feels uncomfortable conducting a meeting in English due to his comparatively lower English proficiency, or simply does not like such a behavior since it makes other meeting attendees feel (linguistically) inferior and could be avoided. In this sense, language is also regarded as a powerful or even manipulative tool that attempts to patronize and mark power relations. It could be argued that his perceptions and evaluation of this kind of communicative event can be described as an overall negative attitude to English and its choice as a lingua franca at meetings. More specifically, the personal account shows how participants’ varying attitudes collide in that the colleagues’ geocentric attitude and related behavior (of choosing nonGerman-skilled personnel to accompany them) competes with or manipulates and in this example overrides the respondent’s (#3) ethnocentric attitude since the outcome is that English is chosen at the meeting. This language imposition is a manifestation of social internal forces. Overall, these insights are an indication that language choices are shaped by participants’ attitudes which seem to vary among them, as will be further illustrated in the following sections. 7.4.3.3 Negative attitude to English and participant role inform the choice of the local language (involving interpretation) Participants might have a negative attitude to English that informs the choice of their L1 instead, especially if they are in a higher position. It is reported that the usual meeting among Managing Directors from the division and Italy is not conducted in English. Respondent #5 (Head of Department_S/AT_German) assumes Italian executives’ negative attitude to English to be the main reason for the choice of their L1: In Italy you have, and that is comparable to Montenegro, the young generation that speaks English, I’d say people between 25 and 30, 35, that manage in English and grew up with it very naturally. Executive staff that is somewhat around 50 is rarely willing to speak English. I don’t know if it is really related to their capacity or simply their willingness […] But I’d say the classic meeting, [where] Managing Directors meet and talk, is almost always assisted by an interpreter. English is seldom an option [my translation]. Although the respondent (#5) does not know the exact reason for the language choice, he shares his impressions; first, he relates age to language proficiency in that older participants seem to have 252 | Page lower English skills (than the younger generations). At the same time, though, the respondent (#5) doubts that language proficiency is the actual or the sole reason for Italian executives avoiding the choice of English. He believes that they simply do not want to speak English (“willingness”), that is it has more to do with their negative attitude to English rather than with their actual skills or ability. Second, it seems that the participants’ level of employment or their seniority and related status give them flexibility and power in choosing their first language (Italian) over English. Eventually, the participants choose their L1, i.e. Italians choose Italian and the Austrians choose German, and some form of language assistance or mediation is organized. To this end, a Germanspeaker, who is located in the Italian-speaking part of Switzerland and also responsible for Italy, is in charge of most of the (nonprofessional) interpretation tasks concerning communication between the division and counterparts in Italy. In this example, participants’ negative attitudes to English, their participant roles with relative status and power (level of employment or seniority), possibly age, and less likely limited English proficiency account for language choice, more specifically inform the choice of the participants’ first language (local language) in communicative situations such as meetings. Consequently, some form of interpretation is necessary. 7.4.4 Social-psychological dimension: participants’ instrumental/extrinsic motivation informs the acquisition and passive use of local languages Participants have an instrumental/extrinsic motivation that informs their language choices (see also Chapter 2, Section 2.3.3.5). Respondent #5 (Head of Department_S/AT_German) has acquired a passive understanding of several languages on the job, in addition to English, Latin and French he had at school. These passive skills are applied to written and oral media, as he explains: I understand, seldom speak [them] though. I understand Italian, French, Spanish, meanwhile Romanian relatively well. Reading is no problem. […] to the extent that if somebody hands me over a note, I understand it, get an idea. If it goes into detail, I use a translator, of course, also at meetings, in Romanian, in Italian, of course, in French. Well, my French learned at school got a bit rusty, but we rarely enter the French-speaking area. […] That’s a bit rusty but the other languages do the trick actually. I still sit with interpreters in meetings but I understand a lot of what is being said already before the translation. And as I said, in writing, given that all 253 | Page these languages are Latin-based, there is no problem [my translation]. (Respondent #5_Head of Department_S/AT_German) The respondent (#5) chooses local languages in a passive fashion for utilitarian purposes within the employment domain. The passive proficiency is beneficial or useful in that he can understand written and oral information in local languages. In the case of Romanian and Italian, the choice includes acquisition, i.e. he has learned them on the job for the same reason. The frequent exposure to these languages as in meetings and the shared Latin roots seem to have facilitated the acquisition process. In sum, the respondent’s (#5) instrumental/extrinsic motivation informs the passive choice and acquisition of local languages. Respondent #6 (Head of Department_S/AT_German) has Russian skills which sometimes prove useful for the clarification of German translations of Slavic source texts. Understanding the original language version of a contract or other legal documents helps better interpret the translation. The closeness of Russian and the southeast European (i.e. Slavic) languages they deal with makes this language check possible. In this example too, certain utilitarian benefits of passive local/customer language skills have been identified. In other words, the respondent (#6) is instrumentally/extrinsically motivated to rely on his passive language skills. These two examples might leave us with the question if passive proficiency can be classified as a language choice or rather a response to a choice others have made. Yet, the motivation for the behavior seems to be of an instrumental/extrinsic nature, that is beneficial in the employment domain. 7.4.5 Social-regulatory dimension: external forces impose languages There are also social forces that are external to the MNC community and have imposing power; they define some of the company’s language choices or rather impose certain languages on AT_2. This also makes this dimension unique and distinct from the previous dimensions since it is not the MNC’s participants any longer that choose languages but external forces of different kinds which will be elaborated below (see also Chapter 2, Section 2.3.3.6). 260 | Page associated with some problems such as counterparts’ low or lacking English proficiency, including the management level and specific units (e.g. the Czech Republic, Hungary, Romania), and the difficulty of the technical register. At the interactional level, we have learned that the language skills possessed by interactants inform the choice of the HQ language, local languages or English, usually involving some form of mostly nonprofessional interpretation that is perceived as timeconsuming, low-quality and at times not objective. We have also learned that engineers may lack terms of the technical register in English or the local language and then they visually illustrate what they mean on a piece of paper. Moreover, varying language proficiency can inform the choice of more than one language in interaction such as code-switching; it occurs at meetings where Austrians switch from the meeting language English to their L1 for clarification purposes. In one-on-one phone calls between HQ and subsidiary staff, code-switching involves German as an L1 and L3 (also being the HQ/corporate language) and English as an L2 (lingua franca). The channel seems to trigger the switching, too. The findings have illustrated that various documents are written in the local language in the first place (unmarked language). If documentation needs to be shared with HQ, the information is translated into English (for a broader internal readership) or German (for individual HQ staff members). In this vein, problems but also advantages in relation to in-house nonprofessional translators have been shed light on. Sometimes local bilingual staff lacks the language skills to translate an acceptable target text. Yet, the advantages are that personnel can gradually familiarize themselves with the company-specific terminology and technical register relevant for the construction industry, and they can always crosscheck with an in-house engineer. Social-relational dimension: participants’ role relationships and politeness strategies The second contextual dimension is of a social-relational nature; participants’ role relationships (based on participant roles within the employment domain) and politeness inform different language choices. The asymmetrical buyer-seller relationship and politeness inform the choice of customer languages (linguistic adaptation/accommodation) involving acquisition and translation. Documents such as biannual reports that target external receivers are translated into customer languages, usually locally. Subsidiaries autonomously decide what promotional material gets 261 | Page translated or produced in customer languages. Interestingly, the buyer-seller relationship and politeness strategies can also translate into the need for being able to talk to local customers. These have led Austrian expatriate managers, in particular the technical executive, acquire the local/customer language Romanian on the job. They did not take any formal language training. In this sense, this social-relational dimension has informed the choice of a customer language for educational purposes, i.e. language acquisition, though in a learning-by-doing manner. Social-psychological dimension: participants’ attitudes to languages and motivations The fourth contextual dimension is also of a social-psychological nature; participants’ varying attitudes to languages inform different language choices. This case study has provided interesting details on the manifestation of attitudes since they are not only very explicit but also distinct, controversial, rival and far from consensual, which leads to a variety of language choices, in particular for HRM purposes (selective recruitment and staff relocation). Participants’ ethnocentric attitudes inform the choice of German as to local recruitment in that it renders German a common hiring criterion in host countries. Also, new business units are always started up by Austrians or Germans (which is not only related to languages but also to the centralized structure and matters of trust). Yet, many expatriates in management positions are German speakers informed by an ethnocentric attitude and HQ orientation. By contrast, other executives have a more geocentric attitude that informs the choice of English over German proficiency when they hire people. In this vein, an expatriation case in Serbia has revealed interesting insights; an executive with a geocentric attitude suggests the expatriate acquires English (which is less useful in Serbia though) while an executive with a polycentric attitude argues for the recruitment and training of locals (and against the expatriation of Austrians based on the assumption Serbians speak Serbian only). This illustrates how subjective attitudes are and how they influence decision making when it comes to choosing languages for HRM purposes. Complicating matters further, other managers have a more polycentric attitude, so they value and prioritize local/customer languages; hence, a divisional executive has hired assistants with Hungarian and Romanian skills for divisional (in Austria) and host-country positions, relocated staff with Italian skills (to the Italian-speaking part of Switzerland) or would headhunt candidates with local/customer languages 262 | Page from competitors in potentially interesting markets (e.g. Poland). This can also entail local language acquisition for expatriates, including the executive himself (i.e. English, Hungarian). This is an example where attitudes overlap with personal motivations. Furthermore, this case study has elaborated that attitudes not only differ or vary among participants but can also be explicitly rival at an interactional level. An executive with an ethnocentric attitude combined with a negative attitude to English prefers the choice of German in communicative events like meetings. He dislikes it when coworkers with a geocentric attitude impose English on a meeting by bringing a non-German-skilled person along. In this example, the geocentric attitude has imposing power and enforces the choice of English as a meeting language. Another manager believes that Italian senior executives have a negative attitude to English in that they are less willing than able to speak English and thus choose Italian instead facilitated by their status and relative power. Overall, attitudes towards languages differ among participants, can be rival and have imposing power. In sum, there is no universal or consensually shared attitude forming a dominant ideology but multiple attitudes with positive and negative attributes or values that account for different language choices including the HQ and local/customer languages and English. The social-psychological dimension also comprises participants’ motivations; they inform their language choices (including foreign language acquisition). A participant has through frequent exposure acquired a number of local languages on the job of which he passively chooses (uses) in particular Romanian and Italian at meetings. Another respondent passively chooses (relies on) his Russian skills to crosscheck translations from Slavic languages into German. The reasoning behind these choices is instrumental/extrinsic motivation, which means they have a utilitarian value for the participants in their workplaces. Social-regulatory dimension: external forces (with imposing power) The last contextual dimension is of a social-regulatory nature, external to the MNC community and has imposing power in that it defines some of the company’s language choices, or rather imposes certain languages on AT_2. An interesting finding in this case study is external (supra)national financial forces that dictate the project language dependent on where the funding 263 | Page comes from. Large-scale projects financed via supranational funds are conducted in English partly in combination with local languages if the host-country authorities involved request it. If projects are nationally funded, the first project language is the local language and documentation needs to be translated into German or English to include HQ. External (certified) translation services might be relied on. Also, national including both homeand host-country authorities such as tax offices or financial institutions speak the local language only, which is considered somewhat common-sense knowledge. One might even say it is described as an unmarked choice (which involves translation or interpretation). Similarly, national host-country requirements impose local languages, which informs the choice of local languages for educational purposes. In Poland, the subsidiary often had to provide certified translations into the local language. Thus, AT_2 invested in training a linguist as a professional translator. In the UK, a test had to be passed to get a local work permit which informed the choice of English for educational purposes (language acquisition). In this sense, one could say external forces also impose different training measures. Another external force in the AT_2 context are local suppliers that issue offers in the local language and the subsidiary has to deal with them, i.e. translate them into German for HQ (where problems with nonprofessional translations have been encountered). Finally, there is national (macro) language policy and planning, i.e. acquisition planning, which shapes national educational systems and the average level of foreign language proficiency among the population. This kind of external forces facilitates the choice of German in business encounters, but also for HRM purposes (selective recruitment) specifically in Eastern Europe. It could be concluded that AT_2 is a multilingual speech community that unites many settings (locations) all speaking different languages, German being the HQ/corporate language and English as a lingua franca. This case study (AT_2) has shed light on four contextual dimensions that inform language choice: the social-linguistic dimension language proficiency, the socialrelational dimension based on role relationships and politeness, the social-psychological dimension including attitudes to languages (with imposing power) and motivations, and the social-regulatory dimension external forces (with imposing power). The contextual dimensions inform different language choices and also impose languages, including the HQ, local/customer languages and English, in different communicative situations, for internal and external 264 | Page communication purposes. The dimensions influence the individual staff member and/or the corporate management level which has to be refracted into individual decision makers who choose languages in a given communicative situation in their own right rather than in a consensual or unified manner. This suggests that social context is multilayered, which renders language choice a complex and multilingual social phenomenon. The multilevel and multidimensional analysis of language choice lets us conclude that language can hardly be regulated or managed in a centralized manner in an MNC network such as AT_2. In other words, the findings indicate a limited manageability of language since language choice is contextually bound. These case study findings also undermine the usefulness of single-language policies but suggest multilingual policies and management activities, if at all necessary. 265 | Page 8 SYNTHESIS AND DISCUSSION OF FINDINGS ACROSS CASES In the previous four chapters I presented the individual case study reports of two Danish (DK_1, DK_2) and two Austrian (AT_1, AT_2) multinational corporations (MNCs). While the individual reports intended to provide in-depths insights into each case company, this chapter’s focus is to synthesize those by highlighting both similarities and differences and reach cross-case findings. I will discuss each result in the light of the extant literature. 8.1 Language Choice in four MNCs with Particularities and Commonalities To begin with, this collective case study has compared Danish and Austrian MNCs for two reasons. First, MNCs, in particular Anglophone businesses, have been ‘accused’ of being ‘blind’ towards language matters in business (Fredriksson et al. 2006). Inspired by calls in the current literature to counterbalance Anglophone insights (e.g. Fredriksson et al. 2006, see also Thomas 2008, Graddol 2006, Crick 1999, Clarke 2000, Hagen 2005), this study investigated nonAnglophone companies. Second, I also compared different home-country environment with different national languages (HQ languages) and different corporate languages, as suggested too (Marschan-Piekkari et al. 1999b, Welch et al. 2001) (see Chapters 2 and 3). (I will deal with corporate languages in Section 8.2.1 below.) From comparing home-country settings and their HQ languages we can learn that they choose them differently for internal and external communication purposes. In the Austrian companies German is chosen as a first language (L1) or foreign language (FL) for interunit communication purposes, and partly also in external business encounters (in combination with customer languages) (AT_2). Danish as an L1 or FL, on the other hand, is limited to internal communication. As far as external communication goes, both Danish MNCs do not choose Danish but partly English instead (in combination with customer languages). This seems to be related to the characteristics of the HQ languages being major (German) and small (Danish) languages in terms of number of language users. This means German is spoken as an FL outside of Austria or German-speaking areas such as Eastern Europe which facilitates the choice of [Document text truncated for crawler view.]