Ethnic styles of speaking in European metropolitan areas 9789027282538, 9027282536

Table of contents :
1. Introduction (by Kern, Friederike)
2. The emergence and diffusion of Multicultural English (by Fox, Sue)
3. Kiezdeutsch as a multiethnolect (by Freywald, Ulrike)
4. Prosodic style-shifting in preadolescent peer-group interactions in a working-class suburb of Paris (by Fagyal, Zsuzsanna)
5. Emerging Moroccan and Turkish varieties of Dutch: Ethnolects or ethnic styles? (by Hinskens, Frans L.)
6. Prosody and unit-construction in an ethnic style: The case of Turkish German and its use and function in conversation (by Selting, Margret)
7. Rhythm in Turkish German talk-in-interaction (by Kern, Friederike)
8. Constructions with Turkish sey and its German equivalent dings in Turkish-German conversations: sey and dings in Turkish-German (by Simsek, Yazgul)
9. Joint attention and cooperation in the Swedish of adolescents in multilingual settings: The use of san 'such' and sahar 'like' (by Ekberg, Lena)
10. Linguistic variation and linguistic virtuosity of young "Ghetto"-migrants in Mannheim, Germany (by Keim, Inken)
11. Late modern youth style in interaction (by Madsen, Lian Malai)
12. Developing multiethnic youth language in Helsinki (by Lehtonen, Heini)
13. Index

Citation preview

Ethnic Styles of Speaking in European Metropolitan Areas

Studies in Language Variation The series aims to include empirical studies of linguistic variation as well as its description, explanation and interpretation in structural, social and cognitive terms. The series will cover any relevant subdiscipline: sociolinguistics, contact linguistics, dialectology, historical linguistics, anthropology/anthropological linguistics. The emphasis will be on linguistic aspects and on the interaction between linguistic and extralinguistic aspects — not on extralinguistic aspects (including language ideology, policy etc.) as such. For an overview of all books published in this series, please see http://benjamins.com/catalog/silv

Editors Frans Hinskens

Paul Kerswill

Jannis K. Androutsopoulos

Peter Gilles

K. K. Luke

Arto Anttila

Barbara Horvath

Rajend Mesthrie

Gaetano Berruto

Brian Joseph

Pieter Muysken

Paul Boersma

Johannes Kabatek

Marc van Oostendorp

Juhani Klemola

Sali Tagliamonte

Miklós Kontra

Johan Taeldeman

Bernard Laks

Øystein Vangsnes

Peter Auer

Universität Freiburg

Meertens Instituut & Vrije Universiteit, Amsterdam

Lancaster University

Editorial Board University of Hamburg Stanford University L’Università di Torino University of Amsterdam

Jenny Cheshire

University of London

University of Luxembourg University of Sydney The Ohio State University Eberhard Karls Universität Tübingen

Gerard Docherty

University of Tampere

Penny Eckert

University of Szeged

William Foley

CNRS-Université Paris X Nanterre

Newcastle University Stanford University University of Sydney

Maria-Rosa Lloret

Universitat de Barcelona

The University of Hong Kong University of Cape Town Radboud University Nijmegen Meertens Institute & Leiden University University of Toronto University of Gent University of Tromsø

Juan Villena Ponsoda

Universidad de Málaga

Volume 8 Ethnic Styles of Speaking in European Metropolitan Areas Edited by Friederike Kern and Margret Selting

Ethnic Styles of Speaking in European Metropolitan Areas Edited by

Friederike Kern Bielefeld University

Margret Selting University of Potsdam

John Benjamins Publishing Company Amsterdam / Philadelphia

8

TM

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

Library of Congress Cataloging-in-Publication Data Ethnic styles of speaking in European metropolitan areas / edited by Friederike Kern, Margret Selting. p. cm. (Studies in Language Variation, issn 1872-9592 ; v. 8) Includes bibliographical references and index. 1. Language and languages--Variation. 2. Intercultural communication--Europe. 3. Ethnic groups--Europe. 4. Europe--Languages. I. Kern, Friederike. II. Selting, Margret. P120.V37E87 2011 410.9’94--dc23 2011036371 isbn 978 90 272 3488 9 (Hb ; alk. paper) isbn 978 90 272 8253 8 (Eb)

© 2011 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Table of contents

Introduction Friederike Kern

1

The emergence and diffusion of Multicultural English Sue Fox, Arfaan Khan and Eivind Torgersen

19

Kiezdeutsch as a multiethnolect Ulrike Freywald, Katharina Mayr, Tiner Özçelik and Heike Wiese

45

Prosodic style-shifting in preadolescent peer-group interactions  in a working-class suburb of Paris Zsuzsanna Fagyal and Christopher M. Stewart

75

Emerging Moroccan and Turkish varieties of Dutch:  Ethnolects or ethnic styles?  Frans Hinskens

101

Prosody and unit-construction in an ethnic style:  The case of Turkish German and its use and function in conversation Margret Selting

131

Rhythm in Turkish German talk-in-interaction Friederike Kern

161

Constructions with Turkish şey and its German equivalent dings  in Turkish-German conversations: şey and dings in Turkish-German Yazgül Şimşek

191

Joint attention and cooperation in the Swedish of adolescents  in multilingual settings: The use of sån ‘such’ and såhär ‘like’ Lena Ekberg

217

vi

Ethnic Styles of Speaking in European Metropolitan Areas

Linguistic variation and linguistic virtuosity of young  “Ghetto”-migrants in Mannheim, Germany Inken Keim and Ralf Knöbl

239

Late modern youth style in interaction Lian Malai Madsen

265

Developing multiethnic youth language in Helsinki Heini Lehtonen

291

Index

319

Introduction Friederike Kern

University of Bielefeld

1. From learner varieties to ethnic styles: Research on linguistic variation in a changing Europe In recent years, ethnic styles of young people with a migrant background have become an important research object in sociolinguistics. It has become clear that similar ways of speaking have been developed by young people in many cities all over Europe. While linguists at first primarily took notice of the deployment and exploitation of these phenomena in the media, now more and more research groups are describing the use of these new ways of speaking in everyday talk in their social settings. Work on ethnic varieties of speech has been prospering in many European countries. This work is continued in the present volume, with the specific aim of bringing together scholars from projects on ethnic varieties of speaking that have developed in cities throughout Western Europe. All papers of the present volume focus on the description of the linguistic characteristics that constitute the nonstandard structures of ethnic ways of speaking. The book presents information about the recent development of young migrants’ language that displays specific lexical, syntactic, phonetic and prosodic patterns that have not been observed in standard or regionalized spoken varieties of the national languages. For this purpose, it brings together various research designs which explore the phenomenon from different research perspectives. More specifically, the volume concentrates on three approaches to linguistic variation: correlational methodology of sociolinguistic research, conversation analytical and interactional-linguistic methodology, and an ethnographic perspective on language use and the constitution of social identities and social relations. The aim of volume is to explore the scope of these different methodologies and to provide a basis for the discussion and evaluation of the theories of language variation associated with them. In addition, the volume takes a European perspective on the phenomenon in order to gain insight into the similarities and differences between ethnic ways of



Friederike Kern

speaking and to learn about their various functions in discourse. The collection intends to contribute to the debate about general processes of the development of non-standard, ethnic ways of speaking in situations of language contact in the first decade of the 21st century in Western European cities, and their possible impact on long-term linguistic change. This introduction is organized as follows: first, a general background to the topic of the volume is given, with a brief outline of different linguistic traditions of investigating migrant speech in post-industrial European societies. The next section starts with some remarks on the notion of ethnicity in the context of language change and variation. This is followed by a discussion of a range of terms that are currently used to denominate the research object in question (i.e. “ethnic ways of speaking”) because they – often implicitly – reflect different ways of conceptualizing migrant speech as descriptive objects, based on diverging linguistic theories and methodologies. Finally, a brief preview of the contributions is given, with an outline of a number of the ways they relate to each other. 1.1

Background

Due to a long history of job migration in Western Europe, aspects of language variation in migrant communities have been subject to linguistic studies since the 1970s. Most of these analyzed second language acquisition and/or communication by or with immigrant (guest) workers in a traditional sociolinguistic research paradigm (cf. e.g. Keim 1978; Klein & Dittmar 1979). Data were collected in sociolinguistic interviews and/or experimental settings. Studies focused on the structural linguistic properties of the immigrants’ learner varieties and correlated them to external categories such as the speakers’ age, origin, length of stay in the resident country etc. In this view, linguistic variation was understood as the product of a set of independent extra-linguistic variables, determining its specific forms and structure as dependent variables. Later, this theoretical perspective on the link between linguistic variation and social structure was reversed, and the attention was instead directed at the effects of language use on social structure. Studies in interpretive sociolinguistics, such as those that Gumperz (1982) and Blom & Gumperz (1972) engaged in, showed that . Other countries, or continents, such as the USA, Canada, Australia, New Zealand, or Africa are excluded from the research overview that is aimed for in this volume on the grounds of their different histories and politics on immigration. For a long time, Western European countries have not perceived themselves as countries of immigration, and job migration was often regarded as their main form of immigration. This has resulted in specific language policies and language ideologies that still form part of the discourse in which the language use of migrants is described and evaluated.



Introduction

linguistic patterns are used by speakers in order to create social structures between participants and change aspects of conversational meaning during the course of a verbal exchange. The notion of contextualization is of particular importance in this research context, bringing attention to the observation that linguistic cues are deployed by participants as resources to create situational and interactional meaning (cf. Gumperz 1982). With contextualization cues, such as prosody, codeswitching, speech styles etc., speakers make interpretive frames available and thus signal how their utterances are to be understood by the recipient. As a consequence, context is no longer considered to be a stable, extra-linguistic unit but a flexible and ever-changing concept, wherein participants activate certain aspects in order to establish and convey meaning in conversation. Gumperz’ work inspired subsequent studies on language variation in social interaction, such as work on the negotiation of social identity in conjunction with distinct ethnic varieties (cf. e.g. Sebba 1993; Hinnenkamp 1998) or on the functions of code-switching in bilingual talk (Backus 1993, 1996; Hinnenkamp 2005). Compared to earlier sociolinguistic work on the relation between language variation and ethnic background, methods of data collection and analysis in this new framework were very different. Studies were now based on large collections of authentic verbal interactions such as mundane face-to-face and telephone conversations that were combined with ethnographic data from interviews and field observations. These data give access to the speech communities’ and speakers’ underlying norms and values relevant to their language use. This line of research (cf. e.g. Jørgensen 2003 for a collection of studies in this area and Auer 2007 for a collection of papers on linguistic heterogeneity and social identity) is very much alive today, and the contributions in the present volume connect to it in many ways, theoretically and methodologically. With the emergence of new ethnic ways of speaking among the second and third generation young migrants in many urban centres in Europe, a fresh research interest has developed in the new millennium which transcends the traditional sociolinguistic view on migrants’ language which originated in the 1970s. Studies on the new ethnic ways of speaking share the view that the recently developed linguistic resources differ from those described for migrants’ varieties in the ’70s. Above all, they cannot be conceptualized as ‘learner varieties’ because the young speakers today, unlike their parents, also show linguistic proficiency in various regionalized varieties of the respective standard languages; they form only part of the speakers’ repertoires. The ethnic ways of speaking have thus less . Cf. e.g. Gumperz (1964) for an elaboration on the term “repertoire”. He argues that speakers have a verbal repertoire at their disposal they can select from in accordance to the situative and communicative needs at hand.





Friederike Kern

to do with imperfect acquisition but with the ways in which speakers actively construct and perform ethnic identity (cf. Hoffman & Walker 2010) and other social meanings. In the light of this insight, the sociolinguistic description of ethnic language variation has to be based on new theoretical and methodological considerations. In order to gain a better understanding of the diverging approaches that have been employed so far, a discussion of the terminology used for the denotation of the research object that reflects different research traditions will be provided in the next section.

2. Ethnic ways of speaking: Different terminology, different conceptualizations 2.1

Ethnicity and language variation

Ethnicity can be understood as one of the key concepts of anthropological, sociological and sociolinguistic research (Deumert 2004). In the context of language variation and change, two approaches to ethnicity can be distinguished (cf. Hoffman & Walker 2010). In the “objective” or “etic” approach ethnicity is defined by a common language, race or lineal descent. In this view, ethnicity is a predefined category and is thus comparable to other sociological categories such as sex, age, class etc. Ethnicity, in this view, induces an identity that is ascribed to people rather than achieved by them (cf. Labov 2001: 245). In the “subjective” or “emic” approach, ethnicity is not viewed as an immutable category with a fixed set of properties but as an interpretive members’ category that is based on social constructions of perceived cultural and historical similarities. “Acts of identity” (Le Page & Tabouret-Keller 1985) by members of an ethnic group are expressions of such social constructions; at the same time they show that ethnicity is not always a central aspect of people’s identity that may be evoked according to situated social needs. This line of thought is also found in Barth’s considerations (1969) on the formation of ethnic groups. Barth points out that it is often the boundaries to other groups that are more important for the formation of an ethnic group rather than “the stuff it encloses” (Barth 1969: 15). Thus, practices of distinction (“Distinktionspraktiken”, cf. Wimmer 1996: 412) become constitutive for the creation of an ethnic group, even more so, if there are relatively little cultural differences from the majority group. Such a focus on boundaries has proved especially useful for the analysis of modern ethnic groups in Western immigrant societies (cf. Deumert 2004).



Introduction

The influence of ethnicity on linguistic chance is still a controversial issue. Labov (2001) provides a number of examples in which language use differs across ethnic groups but is apparently not based on substratum influences. Other studies offer different explanations, implying that ethnicity does in fact has long-term effects on language variation and change (see, e.g., McCafferty 2001 and Hoffman & Walker 2010). 2.2

Linguistic classifications of ethnic ways of speaking

Several classifications have been used for the description of the new ways of ethnic speaking, with the notion of ‘ethnolect’ and the notion of ‘ethnic style’ as the most common ones. Both terms refer to the same surface phenomenon (i.e. ethnic ways of speaking) but imply different research traditions and designs. Some researchers classify their object of study as ethnic varieties, thus adhering to more traditional sociolinguistic terminology and methodology (cf., among others, Kotsinas 1998; Bodén 2005 for Swedish ethnic ways of speaking; Nortier 2001 and Dorleijn & Nortier 2005 on “straattaal” (“street language”) in the Netherlands and Androutsopoulos 2001 and Auer 2003 for German ethnic varieties). Those studies focus on the description of linguistic properties as much as on the complex sociolinguistic processes the varieties’ use is embedded in. Other research groups speak of ethnic styles, thus focussing the speakers’ use of these ways of speaking as a resource in interaction that is deployed to negotiate in-situ social practices and meaning (cf. Kern & Şimşek 2006; Kern & Selting 2006 and Selting 2007). Others began categorizing their objects of study as pan-ethnic or multi-ethnic styles when it became clear that the ways of speaking created by adolescents in European cities cannot be ascribed to influences from a single substratum language and, moreover, that they share various structural features (cf. e.g. Quist 2000, 2005 on Danish multi-ethnolects, Svendsen & Røyneland 2008 on Norwegian multiethnolectal speech and Togersen, Kerswill & Fox 2006 and this volume on Multicultural London English). Even though the distinction between these different notions is not clear-cut, in what follows an attempt is made to describe some of the diverging underlying theoretical concepts and the analytical foci that derive from it. 2.2.1 Ethnolects The term ‘ethnolect’ is used mostly in variation theory that is based on correlational sociolinguistics. Ethnolects are understood as varieties of a spoken non-standard; they are clearly non-native and, in addition, associated with specific ethnic groups (cf. Auer 2003), or are even seen as their linguistic markers (cf. Clyne 2000). Clyne





Friederike Kern

offers a conceptual distinction between ethnolects and multi-ethnolects (see below). Ethnolects, as opposed to multi-ethnolects, are characterized by substratum influences from a single language system and can thus be understood as emerging from contact situations with two specific languages (cf. Clyne 2000). Ethnolectal descriptions of ethnic ways of speaking highlight features of linguistic structure, and focus on their functions as markers of identity and social relations. The linguistic description is combined with detailed accounts of patterns of use within modern urban societies. This particular focus helps to arrive at comprehensive descriptions of the non-standard variants that are part of and indeed constitute ethnic ways of speaking. In this spirit, many linguistic studies have addressed the task of exploring the specific linguistic features of European ethnolects and the complex sociolinguistic processes connected with their use. An insight into the development and use of German ethnolects is provided by Auer (2003). He makes a distinction between primary, secondary and tertiary ethnolects. The primary ethnolect is conceptualized as a linguistic variety developed by adolescents of non-German origin. The secondary ethnolect is a stylized version of the primary ethnolect that has been transformed by media genres such as comedy or films. If the secondary ethnolect is used by German adolescents with no further knowledge of the primary ethnolect, Auer calls it tertiary. This classification helps to understand the ways in which ethnolects used by different groups of speakers mix and intertwine. However, it is, albeit implicitly, based on the same presumption as many other studies on ethnolects: i.e. that ethnolects constitute linguistic entities with unique structural properties and clear-cut boundaries. It is this view on ethnolects that can be challenged by other approaches to linguistic heterogeneity. Drawing on new insights into the relationship between linguistic and social structure, Eckert (2008) points out that the term ‘ethnolect’ reflects a view on language as a fixed rather than a fluid entity. Eckert argues that ethnolects are not discrete entities that signal ethnic identity but emerge out of shared social practices. Investigating the use of ethnic linguistic markers of White Anglo English and Chicano English by adolescents in California high schools, she shows how single markers of ethnic speech can be employed to indicate other social categories such as “coolness” that, at the same time, reproduce the peer-based social order at school. Eckert’s research thus corroborates the hypothesis that the relation between ethnic markers of speech and ethnic social categories is not as straightforward as is suggested by other studies on the topic. Instead, linguistic markers that were originally ethnic can gain new social meaning in the context of peer

. See also the critical discussion of the term ‘ethnolect’ in Jaspers (2008).



Introduction

interaction. Eckert’s research shows that notions of ethnicity should be used carefully when describing linguistic variation in multi-lingual settings. Madsen (this volume) elaborates on this point, rejecting any ethnic label to avoid premature ethnicizing of patterns of linguistic heterogeneity. 2.2.2 Multi-ethnolects In the light of the fact that the term ‘ethnolect’ is more and more criticized, researchers have proposed to used the term ‘multi-ethnolect’ instead. In contrary to the notion of ethnolect, it is employed to describe ethnic ways of speaking that have been developed on the basis of heterogeneous linguistic and ethnic backgrounds. As mentioned above, Clyne proposed the distinction between ethnolects and multi-ethnolect. In contrast to ethnolects, the linguistic structures of multi-ethnolects cannot be understood as characterized by substratum influences of a single language. Rather, they emerge from influences from several linguistic systems (Clyne 2000). Some structures may derive from processes of second language acquisition. This latter aspect, however, is not followed up in any of the contributions of this volume. It would be a worthwhile topic of future research. The term ‘multi-ethnolect’ thus adheres to the observation that ethnic ways of speaking are not always associated with any particular ethnic group but are used by speakers with various ethnic backgrounds. Indeed, neither the speakers of a multi-ethnolectal (or multi-cultural – see Fox et al. this volume) variety nor this variety’s linguistic features are homogeneous. They reflect the multilingual background of ethnically mixed communities that can be found in many urban areas of European cities. The notion of multi-ethnolects is thus considered a better way to capture the heterogeneity and fluidity of modern urban life in Europe. As Fox, Khan & Torgersen (this volume) point out, multi-ethnolects are used actively to constitute new communities and shared social identity within these communities. They can be regarded as an indication of social change in urban areas where people meet and interact in a linguistically heterogeneous environment (see also Quist 2005). 2.2.3 Ethnic styles With the concept of ‘ethnic styles’, a more practice-oriented approach is targeted. As a consequence, the main attention is shifted to the social meaning of language variation in talk-in-interaction. Structural analysis is supplemented with the analysis of interactional and sociostylistic functions. The focus is now on investigating how speakers use ethnic styles as meaningful resources to achieve particular goals . For an extensive discussion of the ideological implications of different ethnic labels used in linguistic studies see Androutsopoulos (2007).





Friederike Kern

in talk-in-interaction, such as the constitution and organization of practices and actions in conversation; and how speakers employ ethnic styles as important resources for the establishment and maintenance of social identities and relations. The concept of style has a long tradition in interpretive sociolinguistic approaches. It goes back to Hymes (1974) ‘ways of speaking’ which are linked explicitly to the notion of style. In Hymes’ view, speech communities comprise sets of styles, which are used by speakers for the creation of social meaning. However, significant speech styles should not be conceptualized simply as correlations of linguistic features and extra-linguistic, contextual parameters. Hymes points out that styles are always interpreted by speakers as socially meaningful. They can be used outside their “normal” contexts; in this case, speech styles are recognized and interpreted by other speakers in other than the initial defining contexts. With this approach, Hymes shifts the focus from a formal linguistic description and a one-dimensional linkage between style and context to a principally functional description of linguistic structures, and he adverts to a more complex connection between styles and contexts. Hinskens (this volume) also points out that styles are not determined by extra-linguistic context but they are “continuously being constructed as socially and interactively significant products”. From this perspective, styles are actively deployed by speakers to construct facets of social reality, such as social relations between participants and social identities (cf. Coupland 2007; Eckert 2001; Selting 2007 and 2009). Accordingly, the analysis of styles is “the analysis of how style resources are put to work creatively” (Coupland 2007: 3). Styles are flexible linguistic resources that speakers construct from their linguistic repertoire. In contrast to linguistic varieties, styles are conceptualized not with clear-cut, but rather with fuzzy boundaries. They are composed of a set of more or less prototypical co-occurring linguistic features on different levels of linguistic description: phonetic-phonological, morpho-syntactic, lexico-semantic, prosodic and pragmatic but also gestural, kinetic and proxemic, to name just a few (Selting 2007 and 2009). Style is thus “a holistic and a multilevel phenomenon” (Auer 2007a: 11) that needs to be described on various linguistic levels. Selting (2007) further calls attention to the fact that styles are not only used as social styles, e.g. for indexing membership of a certain social group, but also for the organization of conversational tasks and activities in talk-in-interaction, such as the construction of narrative and assessment sequences (cf. also Selting 2001: 5). In this volume, style analysis will cover both aspects of stylistic functions: several papers concentrate on ethnic styles as resources to demonstrate affiliation with in-group members or co-present friends. Other papers investigate how speakers deploy ethnic styles to bring social categories into the discourse. In addition, the papers in Part II show how sets of linguistic features that are marked



Introduction

as “ethnic” are used by participants to fulfill different organizational and activityrelated tasks in talk-in-interaction. Thus, the notion of style as an analytical category for the description of the respective ethnic way of speaking is important in many papers of this volume. This choice reflects the insight that ethnic ways of speaking do not constitute discrete linguistic entities but are rather fuzzy phenomena with some core linguistic features and more variable features at the boundaries. More importantly, the notion of style opens up the possibility to focus on the functions of such features to establish and convey social and interactive meaning in conversations. It is this focus that is pursued by many of the papers collected in this volume. However, it is important to point out that both the -lect and -style approaches focus on different aspects of linguistic heterogeneity; not so much on the level of linguistic description, as the papers in the volume show, but more so on the level of functional explanations. Approaches with the -lect or -multi-lect-perspectives concentrate on macro-social processes of use and linguistic change, whereas the -style-approaches center on micro-level interactional analysis. Both approaches, as the collection in this volume show, complement one another to help us gain more insight into the patterns of ethnically based language variation and change. 2.2.4 Pan-ethnic styles Parallel to the notion of multi-ethnolect, the notion of pan-ethnicity also emerged in the context of analyzing ethnic ways of speaking. Erickson & Shultz (1982: 30) define pan-ethnicity as a “similarity across ethnic categories that differ in terms of national origin and race but are similar in other respects”. According to De Klerk (1996: 9), typical pan-ethnic terms are “colored, black, white and Indian” that are used as collective terms for the description and classification of a wide range of different ethnic groups on the basis of similar physical characteristics or a common language or religion. One the one hand, the use of such pan-ethnic terms reflects the insight that ethnic groups are difficult to define and language boundaries between them are often fluid. On the other hand, it puts diverging ethnic groups into a single, all-embracing category and thus neglects important ethnic and linguistic differences. In this respect, the notion of ‘pan-ethnicity’ is similar to the notion of ‘multiethnolect’. The difference is, however, that ‘pan-ethnicity’ emphasizes similarities between diverging ethnic groups by using collective labels for them. In contrast, the notion of multi-ethnicity is based on the idea that people from different ethnic . E.g. Stroud (2004) labels the famous Rinkeby Swedish as a “pan-immigrant contact variety”, implying that it is not used by one particular ethnic group but by several migrant groups who live in the same district in Northern Stockholm.



10

Friederike Kern

backgrounds who mix in specific social environments such as suburban neighborhoods form a self-identified and self-sustained multi-ethnic speech community with discrete linguistic features. However, despite this being an interesting theoretical issue, a further discussion is abandoned here because the notion of pan-ethnicity is not used as a categorial description in this volume.

3. Outlook on the following chapters The contributions to this volume explore the link between ethnic ways of speaking and their social meaning in more detail. They investigate ethnic linguistic variables in multilingual contexts on several levels of linguistic description (Part I), the use of such variables as stylistic devices for the organization of conversational tasks at hand (Part II), and how ethnic styles are used for the constitution and signaling of shared identity (Part III). Even though the contributions differ in their methodologies and their conceptualizations of ethnic styles of speaking as multi-ethnolects or ethnic styles, they have several aspects in common. First, all contributions systematically address the question of which linguistic variables become part of what is recognized as ethnic ways of speaking in contrast to non-standard and/or regional varieties. Whereas a number of papers focus on phonetic features (Fox, Khan & Torgersen, Hinskens, Keim & Knöbl), others investigate prosodic features such as intonation contours (Fagyal, Selting, Şimşek), rhythmic structures (Kern) or the constitution of prosodic units (Selting). Ethnic ways of speaking are also found to be created on the level of lexical choice (Ekbert, Şimşek), or syntactic structure (Selting, Wiese et al.). Many contributions make an attempt to integrate different levels of linguistic description into a comprehensive account of ethnic ways of speaking (Keim & Knöbl, Selting, Kern, Şimşek, Lehtonen, Madsen). Nevertheless, the collection of papers also shows clearly that phonetic and prosodic features seem to be particularly relevant for the constitution and recognition of ethnic styles all over Europe (Fox, Khan & Torgersen, Hinskens, Fagyal, Selting, Kern, Şimşek and Keim & Knöbl). Secondly, ethnographic information plays an important role in most papers. This reflects the understanding of language variation as being fundamentally context-bound. Indeed, the social meaning of linguistic resources cannot be satisfyingly interpreted without drawing on ethnographic and contextual knowledge. In the case of the papers at hand, this knowledge comprises information about the existence and composition of peer-groups and their members‘ interaction (Fox, Khan & Torgersen, Lehtonen, Madsen), changing speech situations from more to less formal (Hinskens, Fagyal & Stewart), and, on the microlevel of interaction,



Introduction

the sequential development of talk-in-interaction and the activities for which linguistic resources are employed (Selting, Kern, Şimşek, Lehtonen, Madsen). Thirdly, most papers of the volume take into account that mechanisms of linguistic change and the constitution of social meaning take place in spoken dialogue. Accordingly, much of the data consist of naturally occurring talk-in­interaction such as telephone conversations (Selting, Şimşek, Kern) or peer group interactions (Ekberg, Lehtonen, Keim & Knöbl, Madsen). Some authors (Fox et al., Hinskens, Lehtonen) use sociolinguistic interviews as their databases; others employ picture description tasks that are combined with occurrences of natural talk stemming from the same speech situation (Fagyal & Stewart). Even though the contributions all deal with similar questions – i.e. which linguistic features constitute ethnic ways of speaking and what is their social meaning – they do so in different methodological and theoretical frameworks, with different analytical foci. The papers by Fox, Torgersen & Khan, Wiese et al., Fagyal and Hinskens focus on the quantitative analysis of single linguistic variables that are clearly marked as ethnic and that indicate linguistic change taking place in multilingual environments. Those papers combine quantitative evidence with more interpretive notions, such as social networks (Fox, Torgersen & Khan) or style (Fagyal,  Hinskens), in order to shed light on how social context leads to linguistic change and which interactive setting provides for the use of ethnically marked variables in speech. These papers are assembled Part I of the volume. The contributions in Parts II and III choose a more practice-oriented approach. They use interpretive, qualitative methodology, partially combined with ethnographic fieldwork. Part II contains a collection of papers that focus on the use of ethnic variation in speech for the organization of practices and actions in talk-in-interaction. The papers by Selting, Kern and Şimşek originate from the same research project in which the use of ethnic styles in natural interactions were investigated on the basis of telephone conversations between adolescents with Turkish German ethnic background. Employing methods from conversation analysis (cf. Sacks, Schegloff & Jefferson 1974) and interactional linguistics (cf. Couper-Kuhlen & Selting 2001), the authors explore the ways the speakers make use of ethnic styles in order to organize narratives (Kern, Selting), information structure (Ekberg Selting, Şimşek), and accounts and the introduction of new topics in talk-in-interaction (Kern). It is shown that ethnic styles are composed of clusters of co-occurring linguistic resources from, e.g., prosody, syntax and lexic. These resources are combined with one another in what emerges as typical patterns of ethnic styles. The integrated description of these resources and their use in their sequential context leads to a comprehensive account of the forms and functions of the observed ethnic styles in social interaction. Both Kern and

11

12

Friederike Kern

Şimşek additionally discuss the status of specific linguistic variants as substratum influences on the German-based ethnic styles. Ekberg focuses on the use of lexical items that gain new meaning in the context of multilingual settings. In Part III, the papers use ethnographic methods of data collection in combination with conversation analytical (Keim & Knöbl, Lehtonen) and/or interpretive sociolinguistic (Madsen) methods of data analysis. These papers explore the complex ways in which linguistic markers of ethnicity are employed for communicative goals in talk-in-interaction, such as the constitution of social identity (Keim & Knöbl, Madsen) and the signaling of affiliation with peer groups (Lehtonen). In summary, the contributions to this volume show that the constitution of ethnic ways of speaking takes place on all levels of linguistic description, and they explore their context-bound social and interactional meaning. Altogether, the papers cover a wide range of ethnic ways of speaking in different European urban communities. The following countries are represented: England (Fox et al.), the Netherlands (Hinskens), France (Fagyal & Stewart), Germany (Wiese et al., Selting, Kern, Şimşek, Keim & Knöbl), Sweden (Ekberg), Finland (Lehtonen) and Denmark (Madsen). The collection thus provides a thorough insight into the new ethnic ways of speaking developing in urban areas all over Europe. 3.1

Overview of the contributions

3.1.1 Part I. Linguistic markers of multi-ethnic speech: Sociolinguistic approaches to language variation and change Fox, Khan & Torgersen focus on aspects of linguistic change that take place in several English multilingual urban communities. They conceive of multicultural English as constituting a part of new urban varieties, although there are regional and in-group differences. The authors opt for the term “multicultural” to describe the new urban varieties because these variations do not seem to be associated with a particular ethnic group and are in fact also adopted by monolingual English speakers. In a quantitative study, phonetic details of three vowels in urban areas of two large British cities, London and Birmingham, are investigated and contrasted with traditional variants. Additional information on the current demography of the areas and an outline of their history of migration is provided to facilitate the understanding of linguistic change. It is, however, the design of social networks in particular that seems to play an important role for the development of new, ethnic variants of vowels: multi-ethnic social networks show inter-ethnic linguistic diffusion, whereas mono-ethnic networks tend to maintain the traditional local forms. In summary, the authors show that social network analysis helps to



Introduction

comprehend the underlying mechanisms of language change that is currently happening in many British cities. Wiese et al. provide a perception study of Kiezdeutsch, which they describe as a multi-ethnolectal variety of German used in multi-ethnic neighbourhoods for informal, everyday communication. After summarizing previous findings on linguistic patterns of Kiezdeutsch, focusing on syntax and information structure, they present results from a recognition experiment which tested whether Kiezdeutsch is recognized as a discrete variety and whether it is associated with specific speech communities. Furthermore, the authors tried to gain insight into attitudes towards Kiezdeutsch on the basis of the participants’ comments on the verbal stimuli. The quantitative analysis shows that Kiezdeutsch is indeed recognized as a distinct variety by the participants and is moreover associated with multi-ethnic neighbourhoods but not with any specific ethnic background. An additional qualitative analysis of the participants’ comments yielded that negative attitudes towards Kiezdeutsch, namely that it is regarded as ‘bad’ or ‘wrong’ German, are maintained by participants from both mono-ethnic and multi-ethnic communities. Drawing on the concept of style as audience design (cf. Bell 2001 and Coupland 2007), Fagyal and Stewart argue that linguistic outcomes of language contact provide the speakers with additional stylistic resources to convey social meaning by signaling affiliation with or distance from their respective audience. In a semi-experimental interview setting, the authors asked their adolescent informants to produce single target words. In the context of the interview setting, the informants produced the target words with pragmatically neutral rising or falling intonation contours. In side-comments to peers present during the interview, however, the speakers employed a characteristic falling-rising intonation contour that is associated with multi-ethnic language use. This observation strongly suggests that prosodic resources are used as stylistic resources to create social relations between interlocutors in conversational interaction. In the case at hand, the informants used falling-rising intonation to signal affiliation with their peers, in contrast to the neutral contours that they used in interaction with the field experimenter. Hinskens delivers a quantitative study of a single variable of Dutch, the phoneme /z/ and its ethnic variant /z:/, the latter being clearly recognized as ethnic and non-standard. Additionally he studies the use of the same variable in conversations by two Moroccan Arabs from a qualitative perspective. As a result of both quantitative and interpretive analyses of the same variable, Hinskens arrives at the conclusion that this variable carries social meaning insofar as it signals affiliation with interlocutors on the grounds of a shared social identity.

13

14

Friederike Kern

3.1.2 Part II. Ethnic styles in an interactional perspective: The use of linguistic resources in talk-in-interaction Selting’s investigation of the ethnic style of Turkish German is based on telephone conversations between adolescents with a Turkish ethnic background. Focusing on the composition of prosodic units, she shows how prosody is used as an important resource for the constitution of Turkish German. In contrast to more standard and regional varieties of spoken colloquial German, the speakers of Turkish German construct short prosodic units by pre- and post-positioning syntactic units before and after their host sentences as well as by juxtapositioning short clauses and phrases. As a consequence, information units are split up into several short prosodic units, which is unfamiliar in more standard or regionalized varieties of spoken German. Selting points out that speakers do not use these features throughout the conversations but only in particular sequences. A functional analysis demonstrates how speakers deploy those short prosodic units in narratives as a discourse-focusing device. Kern’s paper, based on the same corpus as Selting’s, describes the forms and functions of rhythm in utterances with recognizable Turkish German ethnic style. Two types of rhythmic principles are identified, namely a pattern based on accent isochrony and a pattern based on metrical stress distribution. It is demonstrated how rhythmic patterns locally govern prosodic and syntactic structures which may differ from those common in colloquial spoken German as a result. Sometimes, the primary accent of a turn constructional unit is shifted for rhythmic reasons. This practice is unknown to colloquial spoken German rules of accentuation. In other cases, syntactic structures are changed in accordance with a rhythmic pattern, also a practice distinctive to Turkish German. It is suggested that speakers make use of phonetic and phonological features from both Turkish and German in order to establish and maintain rhythmic patterns. Finally, the functions of rhythm in Turkish German as a resource for the organization of sequences and story-telling and for the construction of accounts are discussed. Şimşek explores the use of the Turkish dummy word şey (“thing”) and its German lexical equivalent ding as a resource in Turkish German. Drawing on a large corpus of data from both telephone conversations between Turkish German and German adolescents and narrative interviews with Turkish speakers, she compares the forms and functions of şey/ding in German, Turkish and TurkishGerman talk-in-interaction. Even though şey and ding are embedded in similar syntactic constructions, there are remarkable prosodic differences: in Turkish, şey never carries the main phrasal accent and is usually part of a larger prosodic unit; in Turkish German, on the contrary, both ding and şey may constitute separate prosodic units and carry the main accent. The careful structural analysis is



Introduction

complemented with a functional analysis demonstrating how speakers make use of these constructions as a focusing strategy. Ekberg investigates the two comparative-demonstrative expressions sån (‘such’) and såhär (‘like’) as pragmatic devices in multilingual talk in different areas of Sweden (Malmö and Stockholm). For this purpose, she uses a mix of data, such as spontaneous and semi-directed group interviews, class-presentations and self-recordings of peer interaction with monolingual and multilingual peer groups. Even though the use of sån and såhär is not clearly associated with a particular ethnic way of speaking, Ekberg observes that the lexical items are employed with expanding meanings in multilingual settings. Apart from added structural functions, both lexical items adopt an additional pragmatic function as a focusing device: they direct the recipients’ attention to a linguistic expression in order to establish joint attention. 3.1.3 Part III. Ethnic styles in an ethnographic perspective: The description of repertoires and the constitution of identities Keim & Knöbl provide a comprehensive account of the environment and the linguistic repertoire of a 17-year-old of Turkish origin, living in Mannheim, Germany. On the basis of two extensive interviews – an ethnographic one with a member of the research group and one with a TV reporter that took place two years later – they describe the young migrant’s environment and educational development and show how the speaker uses his wide linguistic repertoire and communicative competence when performing in accordance with the different demands of the speech situations at hand. Especially in the more formal TV interview, the young man shows high linguistic proficiency when switching expertly between different linguistic forms to carry out various discoursive and social tasks. Lehtonen also adopts a practice-oriented, style-based perspective on language variation. She demonstrates how lexical items from different languages and ethnic varieties such as huono sumoi (‘bad Finnish’) can be used as stylistic resources to signal affiliation with a multi-ethnic peer group. However, representations of ethnicity in talk-in-interaction are intertwined with subcultural youth styles. Lehtonen points out that language variation in multi-ethnic youth groups is relatively new in Finland and has not yet received the stage where it can be recognized as a distinct Finnish-based multi-ethnic variety. In conclusion, Lehtonen argues that linguistic analyses should focus on describing the kinds of stylistic resources multilingual environments offer to the participants. She claims that the social meaning of these resources is negotiated in situ rather than fixed. Madsen questions the foregrounding of ethnicity as an explanatory category in research on linguistic behaviour of young (multilingual) speakers in interactions.

15

16

Friederike Kern

After an extensive discussion about the terms ‘ethnolect’ and ‘multi-ethnolect’ she concludes that the notion of style, referring to a larger system of linguistic resources with fuzzy boundaries, is a more useful concept for the description of linguistic variation because it avoids a priori assumptions about the link between linguistic practices and ethnicity. Madsen then offers analyses of sequences of interactions between young males in a Kopenhagen TaekWanDo club that goes beyond the boundaries of ethnicity. She shows that, rather than its link with ethnic categories, it is the streetwiseness and toughness associated with what she calls the late modern youth style. Speakers draw on this as a vehicle to negotiate social status and emphasize mutual affiliation. What she calls the “polyphony of semiotic resources” available to young people in multilingual environments emphasizes the complex links between language and social meaning and, at the same time, points to the fluidity of linguistic boundaries.

References Androutsopoulos, Jannis. 2001.“From the streets to the screens and back again: On the mediated diffusion of ethnological patterns in contemporary German.” LAUD Linguistic Agency, Series A, No. 522. Essen: Universität Essen. Androutsopoulos, Jannis. 2007. “Ethnolekte in der Mediengesellschaft. Stilisierung und Sprachideologie in Performance, Fiktion und Metasprachdiskurs”. In Standard, Variation und Sprachwandel in germanischen Sprachen, Christian Fandrych and Reinier Salverda (eds), 113–155. Tübingen: Narr. Auer, Peter. 2003. “Türkenslang: Ein jugendsprachlicher Ethnolekt des Deutschen und seine Transformationen.” In Spracherwerb und Le­bensalter, Annelies Häcki Buhofer (eds), 255– 264. Tübingen/Basel: Francke. Auer, Peter. (ed.). 2007. Style and social identities. Alternative approaches to linguistic heterogenity. Berlin etc: de Gruyter. Auer, Peter. 2007a. “Introduction”. In Style and social identities. Alternative approaches to linguistic heterogenity, Peter Auer (ed.), 1–21. Berlin etc: de Gruyter. Backus, Ad. 1993. Patterns of Language Mixing. A Study in Turkish-Dutch Bilingualism. Wiesbaden: Otto Harrassowitz. Backus, Ad. 1996. Two in One. Bilingual Speech of Turkish Immigrants in the Netherlands. Diss. Tilburg University. Barth, F. (ed.). 1969. Ethnic Groups and Boundaries. London: George Allan & Unwin. Bell, Alan. 2001. “Back in style: Reworking audience design.” In Style and Sociolinguistic Variation, Penelope Eckert and John Rickford (eds), 139–169. Cambridge/New York: Cambridge University Press. Blom, John & John Gumperz. 1972. “Social Meaning in Linguistic Structures: Code-Switching in Norway.” In Directions in sociolinguistics. The ethnography of communication, John Gumperz and Dell Hymes (eds), 407–443. Oxford/New York: Basil Blackwell.



Introduction

Bodén, Petra. 2005. “The sound of ‘Swedish on Multilingual Grounds’”. Proceedings, Fonetik 2005: 37–40. (http://www.ling.gu.se/konferenser/fonetik2005/papers/Proc_fonetik_2005. pdf, 15.3.2010). Clyne, Michael. 2000. “Lingua Franca and ethnolects in Europe and beyond.” Sociolinguistica 14: 83–89. Couper-Kuhlen, E. & Margret Selting. 2001. “Introducing Interactional Linguistics”. In Studies in Interactional Linguistics, Margret Selting and Elizabeth Couper-Kuhlen (eds), 1–22. Amsterdam/Philadelphia: Benjamins. Coupland, N. 2007. Style. Language Variation and Identity. Cambridge/New York: Cambridge University Press. De Klerk, Vivian. 1996. “Introduction”. In Focus on South Africa, Vivian De Klerk (ed.), 7–17. Amsterdam: Benjamins. Deumert, Ana. 2004. “Ethnicity”. In Sociolinguistics. An International Handbook of the Science of Language and Society, Ulrich Ammon et al. (eds), 355–360. Berlin etc.: de Gruyter. Dorleijn, Margreet & Jacomine Nortier. 2005. “Is Moroccan-Dutch developing into a general ethnolect?” Presentation, given on the International Conference on Language Variation in Europe (ICLaVE) 3, June 23.–25. Amsterdam. Eckert, Penelope. 2001. “Style and social meaning”. In Style and Sociolinguistic Variation, Penelope Eckert and John Rickford (eds.), 119–126. Cambridge/New York: Cambridge University Press. Eckert, Penelope. 2008. Where do ethnolects stop? International Journal of Bilingualism 12: 25–41. Erickson, Frederick & Jeffrey Shultz. 1982. The counselor as gatekeeper: Social interaction in interviews. New York: Academic Press. Gumperz, John. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Hinnenkamp, Volker. 1998. “Mehrsprachigkeit in Deutschland und deutsche Mehrsprachigkeit. Szenarien einer migrationsbedingten Nischenkultur der Mehrsprachigkeit.” In Das 20. Jahrhundert. Sprachgeschichte – Zeitgeschichte, Heidrun Kämper and Hartmut Schmidt (eds.), 137–162. Berlin/New York: de Gruyter. Hinnenkamp, Volker. 2005. “ ‘Zwei zu bir miydi?’– Mischsprachliche Varietäten von Migrantenjugendlichen im Hybriditätsdiskurs”. In Sprachgrenzen überspringen. Sprachliche Hybridität und polykulturelles Selbstverständnis, Volker Hinnenkamp and Katharina Meng (eds.), 51–103. Tübingen: Narr. Hymes, Dell. 19892. “Ways of Speaking”. In Explorations in the Ethnography of Speaking, Richard Bauman and Joel Sherzer (eds.), 433–452. Cambridge: Cambridge University Press. Jaspers, Jürgen. 2008. “Problematizing ethnolects: Naming linguistic practices in an Antwerp secondary school”. International Journal of Bilingualism 12(1–2): 85–104. Jørgensen, J. Normann. (ed.). 2003. “Bilingualism and Social Relations: Turkish Speakers in North-west Europe”. Journal of Multilingual and Multicultural Development 24(1–2) (Special Issue). Keim, Inken. 1978. Gastarbeiterdeutsch. Untersuchungen zum sprachlichen Verhalten türkischer Gastarbeiter. Tübingen: Narr. Kern, Friederike & Margret Selting. 2006. “Einheitenkonstruktion im Türkendeutschen: Grammatische und prosodische Aspekte.” Zeitschrift für Sprachwissenschaft 25(2): 239–272.

17

18

Friederike Kern

Kern, Friederike & Yazgül Şimşek. 2006. “Türkendeutsch: Aspekte von Einheitenbil­dung und Rezeptionsverhalten.” In Mehrsprachige Individuen – vielsprachige Gesellschaften, Dieter Wolff (ed.), 101–119. Frankfurt/Main: Lang. Klein, Wolfgang & Dittmar Norbert. 1979. Developing Grammars. The Acquisition of German Syntax by Foreign Workers. Berlin: Springer-Verlag. Kotsinas, Ulla-Britt. 1998. “Language contact in Rinkeby, an immigrant suburb”. In Jugendsprache – langue des jeunes – youth language. Linguistische und soziologische Perspektiven, Jannis Androutsopoulos and Arno Scholz (eds.), 125–148. Frank­furt/Main: Lang. Nortier, Jacomine. 2001. “ ‘Fawaka, what’s up?’ Language use among adolescents in monoethnic and ethnically mixed groups”. In Meetings at the crossroads, Anne Hvenekilde and Jacomine Nortier (eds.), 61–72. Oslo: Novus Forlag. Nortier, Jacomine & Margreet Dorleijm. 2008. “A Moroccan accent in Dutch: A sociocultural style restricted to the Moroccan community?” International Journal of Bilingualism 2008/12: 125–142. Quist, Pia. 2000. New Copenhagen ‘Multi Ethnolect’. About language use among adolescents in linguistic and culturally heterogenous settings. Danske Talesprog, Vol. I. Institut for Dansk Dialektforsnik. Kopenhagen: C.A. Reitzels Forlag, 143–212. (English summary at URL: http://www.piaquist.dk/Summary%20of%20Danish%20Multiethnolect.pdf, 15.3.2010). Quist, Pia. 2005. “New speech varieties among immigrant south in Copenhagen – A case study.” In Sprachgrenzen überspringen. Sprachliche Hybridität und polykulturelles Selbstverständnis, Volker Hinnenkamp and Katharina Meng (eds.), 145–161. Tübingen: Narr. Sacks, Harvey, Emanuel Schegloff & Gail Jefferson. 1974. “A simplest systematics for the organisation of turn-taking for conversation”. Language 50: 696–735. Sebba, Mark. 1993. London Jamaican. Language systems in interaction. London: Longman. Selting, Margret. 2007. “Interactional Stylistics and Style as a Contextualization Cue”. In Rhetorik und Stilistik. Ein internationales Handbuch, Ulla Fix and Andreas Gardt (eds.), 1038–1053. Berlin/New York: de Gruyter. Selting, Margret. 2009. “Communicative Style”. In The Pragmatics of Interaction. Sigurd D’hondt, Jan-Ola Östman and Jef Verschueren (eds.), 20–39. Amsterdam: Benjamins. Svendsen, Bente & Unn Røyneland. 2008. “Multiethnolectal facts and functions in Oslo, Norway.” International Journal of Bilingualism 2008/12: 63–83. Torgersen, Eivind, Paul Kerswill & Sue Fox. 2006. In Ethnicity as a source of changes in the London vowel system. Language Variation – European perspectives, Frans Hinskens (ed.), 249–263. Amsterdam: Benjamins. Wimmer, A. 1996. “Kultur. Zur Reformulierung eines sozialanthropologischen Grundbegriffs”. Kölner Zeitschrift für Soziologie und Sozialwissen­schaften 48(3): 401–425.

The emergence and diffusion of Multicultural English Sue Fox,* Arfaan Khan* and Eivind Torgersen**

*Queen Mary University of London / **Sør-Trøndelag University College

We draw on the observation that many features of new urban language varieties in multilingual contexts do not seem to be associated with any one particular ethnic group. Rather, many young speakers draw on and share at least a portion of the wide linguistic resources reflecting the ethnically mixed nature and multilingual backgrounds of their communities. We discuss new varieties of this kind in Britain’s two largest cities, London and Birmingham. Our results support the existence of a Multicultural English, with social and regional variation. An important factor appears to be the amount of social interaction that takes place between individual speakers and we demonstrate that multi-ethnic friendship networks play a key role in its emergence and diffusion.

1. The emergence and diffusion of Multicultural English in London and Birmingham In this chapter we draw on the observation that many features of new urban language varieties in multilingual contexts do not seem to be associated with any one particular ethnic group. Rather, many young speakers draw on and share at least a portion of the wide linguistic resources reflecting the ethnically mixed nature and multilingual backgrounds of their communities. These developments have been noted across Europe in urban areas with significant minority ethnic populations and have often been assigned a name to indicate that its structure (phonological or grammatical) differs from the local variety traditionally associated with the area: Kiezdeutsch in Kreuzberg, Neukölln and Wedding in Berlin, Germany (Paul, Wittenberg and Wiese 2008; Wiese 2009), Rinkeby Swedish in Stockholm, Sweden (Kotsinas 1988; Bijvoet 2003; Stroud 2004), Copenhagen Multiethnolect in Denmark (Quist 2000, 2008). Here, we document new varieties of this kind in Britain’s two largest cities, London and Birmingham. We refer to these emerging varieties as Multicultural English to indicate that they are varieties adopted

20 Sue Fox, Arfaan Khan and Eivind Torgersen

by speakers from different ethnic backgrounds, including monolingual English speakers, and that they can be traced to influences from a multilingual environment. We argue that unlike the adoption of out-group language used for stylistic purposes described in earlier work (Hewitt 1986; Rampton 1995), Multicultural English, while drawing on different ethnic varieties, has become the unselfconscious vernacular of some young people regardless of their ethnic origin. The results presented here come from three studies of adolescent speech, two in London and one in Birmingham. As well as being Britain’s two largest cities, both locations have been popular ‘points of arrival’ (Bermant 1975) for large numbers of immigrants from minority ethnic groups. They are both urban areas of considerable social and ethnic diversity, although this is not to suggest that the demography for each location is the same. The London studies were carried out in two different boroughs. The first study was conducted in Tower Hamlets, home to a 65,000-strong Bangladeshi community, the largest group outside Bangladesh and who represent over 33% of the Tower Hamlets population. The second study was conducted in the adjoining borough of Hackney where the Bangladeshi group represents less than 3% of the population but where there are large Black African and Black Caribbean communities. Although the largest minority ethnic group in Birmingham is the Pakistani community, the percentage of Black Caribbean residents is above the national average too. There are also differences in the settlement patterns of immigrant groups in these locations. In Tower Hamlets and Birmingham, the Bangladeshi and Pakistani communities respectively, have tended to settle in areas which were vacated by the indigenous white population as it moved further out of the city. In Birmingham in particular this has led to residential segregation of different ethnic groups and therefore has implications for the level of interaction that takes places between the different communities. In Hackney, though, the different ethnic groups are much more evenly spread throughout the borough. The phonological features we discuss are the price, face and goat vowels (Wells 1982). For each study, a brief overview of the literature is given followed by a description of the study and the results. As well as the traditional regional varieties of London and Birmingham English, distinctively ethnic varieties of English have also been observed in these cities (Sebba 1993; Thorne 2003). Such varieties are usually a mixture of local and minority ethnic (e.g. Jamaican Creole and Indian English) linguistic features which, in the case of the features discussed here, may include the substitution of the monophthongs [oː] and [eː] for goat and face respectively, while [aɪ] is reported for the price vowel. We therefore compare our results both to the traditional variants and to the reported minority ethnic varieties to determine whether inter-ethnic diffusion is taking place. For each study, we also provide an outline of the history of migration into the areas



The emergence and diffusion of Multicultural English

and their current demography, as we consider this social data to be important both in terms of methodological issues and in the interpretation of the results. Despite the demographic differences of the locations, our results support the existence of a Multicultural English, with social and regional variation. The varieties used in each location appear to be replacing some aspects of the traditional local urban vernaculars, while drawing on input from them for a number of their characteristics. Furthermore, an important factor in this process of change appears to be the amount of social interaction that takes place between individual speakers. We demonstrate here that multi-ethnic friendship networks play a key role in the emergence and diffusion of Multicultural English. 1.1

The London vowel system

Wells (1982) notes that one of the most striking characteristics of Cockney London English is that it exhibits a set of phonetic changes referred to as Diphthong Shift, as shown below. In effect, the overall pattern for the onsets of the front-closing diphthongs in London is a counter-clockwise shift from RP forms, without implying that Cockney had precisely the RP forms in the past. RP iː eɪ aɪ ɔɪ Popular London ɪ̜i ʌɪ ɑɪ ɔ̣ɪ Cockney әi aɪ ɑɪ oɪ

(from Wells 1982: 308)

It is assumed that the Diphthong Shift originated in London but it is also found among local accents of much of the south of England and, of particular interest here, also in the local accents of the Midlands, including Birmingham (Wells 1982: 256). Labov’s recording of a London woman in the 1960s accords with this diphthong-shifted pattern and illustrates that the onset of the price vowel is back and raised (Labov 1994: 169). Sivertsen (1960: 64), in her study of Cockney phonology, also reports that the initial element of the price vowel has a more back quality than that found in RP and data from the Survey of English Dialects (Orton and Tilling 1971) provide further evidence that the phonetic quality of the first element of the price vowel is characteristically fully back in London. For the face vowel, Labov’s 1960s recording of a London man shows a very open onset and accords with Labov’s claim that this vowel occupies the lowest position in the system for many speakers in London (Labov 1994: 209). Further evidence for this is provided by data from The Survey of English Dialects (Orton and Tilling 1971) which show that the starting point for the London face vowel is lower than the standard RP accent. Sivertsen (1960: 56) also reports that Cockney differs from RP in having a much more open initial element but states that there

21

22

Sue Fox, Arfaan Khan and Eivind Torgersen

is considerable variation, with the more open realisations being considered by the speakers themselves as ‘more characteristically Cockney’ than the closer ones. The third vowel which we consider is the goat vowel, which has traditionally had a lowered central onset in London. Sivertsen (1960: 88) describes it as [œʊ], which is more front than RP, and also ‘wider’. The starting point is between front and central and open-mid, which Wells (1982: 308–309) states is between [æ̈] and [ɐ]. The end point may be rounded or unrounded. Hurford (1967: 462), in a study of a Bethnal Green family, found diphthongal realisations starting with [ɐ] but also [æ̈]. End points were around [ә] and [ʊ], but for one female speaker (born 1929) they were fronted to [ʉ] or even [ʏ]. This may be the first time that goatfronting is reported (see below). Beaken (1971: 242–243) also found fronted offsets among young girls in Old Ford (Bethnal Green). 1.2

The Tower Hamlets study

The modern day borough of Tower Hamlets is situated directly to the east of the City of London. We refer to this area as the ‘traditional’ East End of London (see Fox 2007 for further discussion). Since around the 1950s, this area has undergone one of the most rapid social and economic transformations that has taken place anywhere in Britain. The reasons for this transformation have been threefold. Firstly, following redevelopment of the area after World War II, vast numbers of the population were moved out of the East End of London through a process of decentralisation, leaving a mainly ageing population (Forman 1989: 14). Secondly, the London Docks, which made up the largest port in the world, closed between 1967 and 1981 causing thousands of job losses and rapid decline of the docklands area. The area was subsequently redeveloped and is now home to the Canary Wharf Development, described as ‘a city within a city’ (Bird 1993: 120) but which has failed to attract the original residents back to the area. Thirdly, and importantly for this study, there has been a huge influx, mainly since the 1970s, of immigrants from Bangladesh. The area now contains the largest group of Bangladeshis in the UK and the largest group of Bangladeshis outside of Bangladesh. This community constitutes over 33% of the total population of Tower Hamlets and is still growing. The community tends to be concentrated in the western neighbourhoods, where the Bangladeshis sometimes represent between 75% and 90% of the population. In contrast to the ageing population identified nationally, the population of Tower Hamlets has got younger and in the under-19 age group, the Bangladeshis are the numerically dominant ethnic group. According to DfES (Department for Children, Schools and Families) returns on ethnicity, in January 2007, 63% of all Tower Hamlets primary school-age children were of Bangladeshi origin.



The emergence and diffusion of Multicultural English

Tower Hamlets then, once the home of white working-class families who spoke the Cockney dialect and were themselves known as Cockneys, is now an area of social, economic and ethnic diversity and this study attempts to investigate the linguistic consequences of these changes. The study hypothesises that interethnic linguistic diffusion may be taking place and looks for features of ethnically marked variants in the spontaneous conversations of adolescents. 1.2.1 The Tower Hamlets sample The sample for the Tower Hamlets study is made up of the following ethnic and gender groups as shown in Table 1 below: Table 1.  Adolescent participants in the Tower Hamlets study

Male Female Total

White British

Bangladeshi

Mixed race White British/Black Caribbean

Total

11  9 20

17

2

17

2

30  9 39

The project is an ethnographic study carried out in 2000–2001, over a period of nine months, of adolescents at a youth centre in Tower Hamlets and makes use of participant observation to examine friendship networks as well as sociolinguistic/ethnographic interviews for data elicitation. The interviews were mainly with pairs or small groups of friends. The young people who attended the youth centre were drawn from the local area and came from various ethnic backgrounds, with the exception that it was not attended by any Bangladeshi girls. For cultural reasons, Bangladeshi adolescent boys and girls are not encouraged to mix socially in this context. The girls who attended the youth centre and whose language use was analysed were all of white British origin. The data therefore cannot provide us with any information on language use by Bangladeshi girls. 1.2.2 The price vowel Over 5,000 tokens of the price vowel were subject to auditory analysis by the researcher. To control for accuracy, every 30th token was rechecked and random samples were checked by an independent analyst. The tokens were sorted according to the onset of the starting point of the diphthong and a total of six variants were identified: [æ] [aɪ] [ɐɪ] [ɑ̘ɪ] [ɑɪ] [ɑː]. Each variant was noted within its context so that possible constraints or conditioning factors could be investigated. These variants demonstrate that at least three of the six identified are not normally associated with the traditional Cockney dialect. It is true that the central

23

Sue Fox, Arfaan Khan and Eivind Torgersen

to fully back onsets [ɑ̘ɪ] [ɑɪ] and [ɑː], as reported in the literature for this area, continue to exist but the results indicate that the newer variants of [æ] [aɪ] and [ɐɪ] are more commonly being used among the adolescents. The new variants are strikingly different to the vowel systems of Labov’s London speakers recorded in the late 1960s and those described by Sivertsen (1960). The variant described as [ɐɪ] is a narrow diphthong with a starting point which is slightly raised and not fully open. The second element of the diphthong is often weakened, producing a diphthong of the type [ɐɛ]. The variant [aɪ] has an open, unrounded front onset, similar to RP, although again it tends to be a narrower diphthong in this data set, with the second element closer to [ɛ]. Both the [ɐɪ] and [aɪ] variants can at times be near monophthongal, with the second element of the diphthong being much reduced. These types of realisations bring the latter two variants closer towards the final variant [æ], a monophthongal realisation similar to the RP trap vowel. Figure 1 shows the overall distribution in this corpus of the price tokens between the six different variants, according to ethnicity and gender. Figure 1 illustrates that the variants [æ] and [aɪ] are predominantly used by the Bangladeshi boys (11% and 53% respectively), with some use by the mixed race (14% total) and white British boys (12% total) but only minimal use among the white British girls (1%). On the other hand, the white British girls are the only speakers who use the traditional monophthongal [ɑː] and then only 1% overall. Use of PRICE variable according to ethnicity and gender 90 80 Percentage of use

24

70 60 50 40 30 20 10 0 [æ]

[aɪ]

[ɐɪ]

[ɑ̘ɪ]

[ɑɪ]

[ɑː]

PRICE variants Bangladeshi boys Mixed origin boys white/Afro-Caribbean

White British boys White British girls

Figure 1.  Use of PRICE vowel among East London adolescents according to ethnicity and gender



The emergence and diffusion of Multicultural English

However, their preference is to use the traditional London [ɑɪ] variant (78%). While the Bangladeshi boys use more of the front variants [æ] and [aɪ] and the white British boys use more of the back variant [ɑɪ], it is in the use of the central variants, particularly [ɐɪ] that we start to see similar frequencies in the boys’ groups. The white British boys use this variant 34% of the time, the Bangladeshi boys use it 26% of the time and the boys of mixed origin use it 23% of the time. This would seem to indicate that levelling may be taking place, with the more extreme variants at opposite ends of the continuum in the process of being levelled away and a new multicultural variant emerging. Indeed, even the girls’ data indicate that they use this form 14% of the time, moving in the same direction, albeit at a slower rate. We return to possible reasons why this is happening after the following section. 1.2.3 The face vowel Over 3,000 tokens of the face vowel were subject to auditory analysis (see Section 1.2.2 for controls). The tokens were sorted according to the onset of the starting point of the diphthong and a total of five variants were identified: [e̝ɪ] [eɪ] [ɛɪ] [æɪ] [aɪ]. Once more, each variant was noted within its context so that possible constraints or conditioning factors could be investigated. Of the five variants identified, only two are associated with previous descriptions of Cockney in the literature, namely [æɪ] and [aɪ]. These continue to exist among the adolescents but the results indicate that the newer pronunciations [e̝ɪ] [eɪ] [ɛɪ] are more commonly being used. [ɛɪ] has also been identified as a London variant but it is often used as a variant comparable to [ʌɪ] to describe a more open and central starting point (Wells 1982; Hudson and Holloway 1977) and to describe a more ‘popular London’ accent (Wells 1982). However, the variant we describe as [ɛɪ] has a front onset slightly lower than [e] but higher than [æ]. It is also a narrow diphthong, at times almost monophthongal, one of the most striking features of the newer pronunciations. The variant [eɪ] has an onset similar to RP but again tends to be a narrower diphthong in this data set, at times almost monophthongal. The variant [e̝ɪ] has a higher onset than RP and is again near monophthongal, at times sounding closer to the RP kit vowel. Figure 2 shows the overall distribution in this corpus of the face tokens between the five different variants, according to ethnicity and gender. Figure 2 illustrates that it is predominantly the Bangladeshi boys who use the front raised variants [e̝ɪ] and [eɪ], suggesting that they have arisen among this group. However, they are also used by the mixed race (see Table 1) and white British boys to some extent (9% and 23% respectively). While the Bangladeshi boys use more of the front raised monophthongal-type variants and the white

25

26 Sue Fox, Arfaan Khan and Eivind Torgersen

Use of FACE variable according to ethnicity and gender

70

% of use

60 50 40 30 20 10 0

[е ɪ] ▴

[е еɪ]

[ɛɪ]

[æɪ]

[aɪ]

FACE variants Bangladeshi boys Mixed race White/Black Afro-Carib boys

White British boys White British girls

Figure 2.  Use of FACE vowel among East London adolescents according to ethnicity  and gender

British/mixed race boys make greater use of the more open diphthongs, it can be seen that the boys’ groups are all frequently using [ɛɪ], bringing their speech much closer to each other than to the girls’ speech. The frequencies for this variant are 55% for the white British boys, 29% for the mixed race boys and 26% for the Bangladeshi boys. Again, this would seem to indicate that there may be a process of levelling taking place, with the extreme variants at either end of the continuum being levelled away and a new multicultural variant emerging. The girls, however, do not show signs of taking part in the adoption of the new variants, with only 4% of their usage attributed to [ɛɪ], although it could mean that the boys are initiating the lead with the girls following at a slower pace. At present though, the girls use the traditional London pronunciations [aɪ] and [æɪ] 96% of the time. 1.2.4 Friendship networks The quantitative results have shown that for the vowels analysed, the Bangladeshi boys do not, in general, use pronunciations that are associated with the reported local dialect of the area. In as much that the Bangladeshi community is now the numerically dominant group, particularly in the adolescent age group, we can say that the English spoken by the majority of people in this particular area of London is no longer (at least for the vowels analysed) the variety traditionally considered to be Cockney. Furthermore, the variants associated with the Bangladeshi boys are being adopted by the non-Bangladeshi boys in this study but not generally by the white British girls. However, these results in themselves do not give an explanation of why or how these changes have come about and we turn



The emergence and diffusion of Multicultural English

7

5

6

1

4

Girls

8 11

12

14 17

18

9

19

15

21

22

Girls

Older White British boys

16

29

Younger white/mixed race boys

39

30

33 36

13

2

3

32 31

Younger Bangladeshi boys

35

23 26 25

34

20 40

27 24 28

38

37 Older Bangladeshi boys

Figure 3.  Representation of youth members’ friendship groups

now to the role of friendship groups, which appears to be an important factor in the process of spreading innovative forms. As discussed previously, the data collected for this project were the result of several months’ ethnographic participant observation and recording in a youth club and consequently, ongoing close friendships within the club could be observed. A ‘sociogram’ was developed in order to investigate the relationship between language use and friendship group membership. The information concerning the friendship groups was compiled on the basis of the individuals’ own representations of their friendship groups and also on close observation of the social interaction between the youth club members. The relationships are represented in the sociogram in Figure 3. The youth club members largely divided themselves into six main groups, represented by the large circles in Figure 3. The numbered smaller circles inside the large circles represent individual speakers. The lines joining the large circles indicate that there was some interaction between the members of each of the groups. The girls have been divided into two groups. The first group, speakers 5, 6 and 7, were new young members of the youth club and their interaction with other club members tended to be with the younger group of white boys, although they did know the other girls in the youth centre. The second group of girls consisted of

27

28

Sue Fox, Arfaan Khan and Eivind Torgersen

all the other girls who attended the club, not all of whom are represented in the study. These girls had all known each other from childhood and many had family associations. Within this group there were smaller sub-groups. There were four main groups of boys. The first group consisted of younger boys, mainly of white or mixed race origin. Sometimes they hung out with the older boys and there were some kinship ties and family affiliations between the two age groups but most of the time, these younger boys formed their own group, sometimes staying within the club but often hanging around on the streets outside. This group also interacted freely with the two groups of girls. The second group also consisted of younger boys, but mainly of Bangladeshi origin. Again, there were some kinship ties between the younger and older Bangladeshi boys but the younger boys tended to stay within the club premises, playing pool and table tennis. The boys from the two younger groups interacted freely when taking part in the youth club activities but outside the club, the groups were clearly demarcated. The younger Bangladeshi boys had very little contact with the girls. Among the older boys there was far more inter-ethnic mixing, both within the club and on the streets outside. Here the boundaries between groups became blurred and the groups were based more on the speakers’ own representations of their friendship groups than on any particular observed behaviour patterns. The first of these groups consisted of an older set, made up mainly of boys of white, black and mixed race origin. Although they always came into the youth club to play a game of pool or play music, most of this group’s time was spent outside on the street, smoking, riding mopeds and generally hanging out. The second group of older boys consisted mainly of boys of Bangladeshi origin. Some tended to stay within the club and would help out with the younger members but others spent much of their time taking part in the street activities described above. Only a small number of boys from this group seemed to have any interaction with the girls from the youth club. The results of the quantitative analysis showed that the girls in this study do not adopt the innovative variants of the price and face vowels to any great extent. It can be noted, though, that the sociogram points to the fact that they have no direct close connections to the Bangladeshis, who use the innovative variants the most. The distribution patterns are therefore likely due to the levels of interaction between individual speakers and their friendship groups rather than being specifically characteristic of particular gender or ethnic groups. In contrast to the girls, there is much more inter-ethnic mixing between the boys, particularly between the older groups of boys who adopt the same social practices both in the youth centre and outside on the streets. The results of the quantitative analysis showed that the boys are using a mixture of variants, from those favoured by the Bangladeshi boys to those associated with the traditional dialect and that common



The emergence and diffusion of Multicultural English

7

5

6

1

4

Girls

2

3

8 11 12

14

18

17

13

19

15

21

22

Girls

9 Older White British boys

16

29

Younger white/mixed race boys

33

30

36

32 31

Younger Bangladeshi boys

39

34

23 35

25 26

20 40

27 24 28

Key

38 37

1–15% 16–30% 31–45% 46–60% Over 60%

Older Bangladeshi boys

Figure 4.  Distribution of the PRICE variant [ɐɪ] among different friendship groups

ground is being sought somewhere in the middle of the continuum, as seemingly part of an accommodation process. To illustrate the role that friendship networks may play in the adoption and diffusion of innovations, the results for the price variant [ɐɪ] are plotted directly onto the sociogram, as shown in Figure 4. Recall that this variant is used by the white and mixed race boys 34% and 23% of the time respectively and by the Bangladeshi boys 26% of the time. Note however that the results demonstrate very little use of this variant by the younger Bangladeshi boys (who predominantly use the more front variants). Again, the sociogram points to the fact that this group has a low level of interaction with other groups. Among the older boys however and, to a certain degree, the second group of girls, there is much more variability. Again, these results seem to suggest therefore that the distribution patterns are linked to the degree of interaction between individual speakers and that friendship networks appear to be playing a key role in the emergence and diffusion of this multicultural variety of London English. 1.3

The Linguistic Innovators study – London

The Linguistic Innovators study (Kerswill, Cheshire, Fox and Torgersen 2008) compares two areas of London – the boroughs of Hackney in inner London and

29

30

Sue Fox, Arfaan Khan and Eivind Torgersen

Hackney ethnic group percentages Other

2 1.17

Chinese Black other

2.39 11.98

Black African

10.29

Black Caribbean Asian other

0.82

Asian Bangladeshi

2.94

Asian Pakistani

1.07

Asian Indian

3.76

Mixed race other

1.11

Mixed race white Asian

0.78

Mixed race white black African

0.79

Mixed race white black Caribbean

1.52

White other

12.26

White Irish

3.02

White British

44.12 0

5

10

15

20

25

30

35

40

45

50

Figure 5.  Population by ethnic group in Hackney. Data source: Census 2001

Havering in outer London. The two areas are, demographically, quite different. Havering is predominantly white – 95.2% according to the 2001 Census figures. In general, the adolescents from Havering who took part in the study did not have multi-ethnic friendship networks, unlike those from Hackney. In the following discussion we focus on the results from Hackney. Hackney is the inner-London borough directly adjoining, and to the north of, Tower Hamlets. Like Tower Hamlets, it is an area of ethnic and social diversity but the historical migration patterns differ to some extent. The Bangladeshi community so dominant in Tower Hamlets constitutes just 2.94% of the Hackney population, while the main minority ethnic groups are Black African (11.98%) and Black Caribbean (10.29%). Figure 5 shows the population of Hackney by ethnic group. The sample for the Hackney Linguistic Innovators study is made up as shown in Table 2. Data collection was carried out over a period of approximately ten months with adolescents at a college in Hackney. Participant observation and sociolinguistic/ethnographic interviews were used for data elicitation and fieldwork was conducted by the same researcher who carried out the Tower Hamlets study (see



The emergence and diffusion of Multicultural English

Table 2.  Adolescent participants in the Hackney sample Male Female Total

Anglo

Non-Anglo

Total

12 10 22

15 12 27

27 22 49

previous section). We interviewed ‘Anglo’ speakers (whose families had relatively local roots) and ‘non-Anglos’ (who were children/grandchildren of immigrants). In general, the participants were asked to define their own ethnicity. The interviews were mainly with pairs or small groups of friends. In addition, we recorded conversations with 8 working-class Anglo adults aged 65–80 (4 women, 4 men), to act as a reference point for comparison with the speech of the adolescents. For this study, the vowels were subject to acoustic analysis. While this differs to the other two studies, the outcome of both acoustic and auditory analysis is the plotting of vowels in relation to each other in the vowel space. By using phonetic transcription the results can be directly compared. First formant (F1), representing vowel height, and second formant (F2), representing front/backness, readings were taken at two measurement points, onset and offset of the diphthongs. All readings were normalised (using the Lobanov formula) together with monophthong data for each speaker. Groups of speakers were compared using multivariate ANOVA tests. For complete discussion of the diphthong data see Kerswill, Torgersen and Fox (2008). 1.3.1 Results Figure 6 shows the main differences between old and young speakers for face, price and goat. The figures have been calculated by averaging the normalised diphthong onset and offsets across speakers and then across speaker groups. The results show that face is becoming less open. While the elderly speakers have an onset in the area of [æ], the young speakers have an [ɛ] or [e] onset. The onset for price has moved to a central position and is lowered from [ɑ̝] to [ɐ̞]. The onset for goat is raised and backed and the offset is slightly fronted. The quality for all three diphthongs is less wide for the young than old speakers and demonstrates a movement away from the diphthong-shifted qualities earlier reported in London (see Section 1.1). These average results, however, hide variation within the group of young speakers, where we found effects of ethnicity. price shows a significant effect of ethnicity: non-Anglo speakers have a more fronted (p < 0.05) and more open (p < 0.05) onset than the Anglo speakers. We find the same effect for the onset of goat. The non-Anglo speakers have a more raised onset for goat compared to the Anglo speakers (p < 0.05), while the Anglo

31

–1.5

Sue Fox, Arfaan Khan and Eivind Torgersen

GOOSE

–1.0

KIT KIT

GOOSE

FOOT

FOOT

–0.5

CHOICE CHOICE

0.5

F*1

0.0

FACE DRESS DRESS

1.0

MOUTH FACE TRAP

LOT GOAT LOT GOAT PRICE START START START PRICE

1.5

STRUT TRAP MOUTH

2.0

32

2

1

0 –1 F*2 Lobanov normalised means

–2

Figure 6.  Average FACE, PRICE and GOAT for young (filled circle) and old speakers (cross) in Hackney

speakers have a more fronted offset than the non-Anglo speakers (p < 0.05). Girls have a more fronted offset for goat than boys, but the difference is not significant (p = 0.08). For face, there is a statistically significant gender difference, with boys having a less open onset than girls (p < 0.001). These results are also consistent with the pattern found among the Tower Hamlets adolescents (see previous section). The non-Anglo speakers have a less open (p < 0.001) and more front (p  236 237 -> 238 -> 239 -> 240 241 (b) -> 242 -> 243 -> 244 245 246 247 248 249 (c) -> 250 -> 251

Gül: guck mal früh MORgens, look early in the morning NEE; no nach der ARbeit, after work isch geh so DINGS;= I go like thingy =zu a äh’ zum AUto; weißt du, to uh to the car you know (1.0) daNACH, after that v: vor meinem FENSter, in front of my window is so BRIEF; there is letter (1.2) isch GUCK so, I take a look isch dachte erstmal so STRAFzettel; at first I thought like ticket Zel: [((lacht)) ((laughs)) Gül: [isch GUCK so, I take a look da ist lie’ also BRIEF; there is well letter von TAner; by ((name))

Throughout this extract, we can see short units being produced in different kinds of constructions. (a) Lines 236–240: pre- and post-positioned constituents + sentences with V2 word order, etc.

In line 236, Gül starts with the discourse marker guck mal (‘look’) and the time adverbial früh MORgens (‘early in the morning’), ostensibly the beginning of a larger unit-to-come. This stretch of talk shows a primary accent in the first syllable of the word MORgens and ends with rising-to-mid final pitch, which could at



Prosody and unit-construction in an ethnic style 137

this point still be prosodically integrated in a complete unit-in-production. What follows in lines 237 and 238, however, is an exposed self-initiated self-repair: the repair marker NEE; (‘no’) starts with a drop in pitch and ends with mid-falling pitch. nach der ARbeit, (‘after work’) the repair of früh MORgens, begins with a slight upstep in pitch and ends with pitch rising-to-mid, thus realizing a repetition of the pitch chosen for the repairable. The beginnings of new prosodic units for the repair here retrospectively allow the inference that the unit begun in line 236 has either been abandoned or is intended to stand as a prosodic unit of its own. What is most interesting here, however, is that the repair itself, nach der ARbeit,, in line 238, is not integrated in the unit and sentence-to-follow either, but again ends with pitch rising-to-mid, thus repeating the potentially unit-final pitch of line 236. Furthermore, the next line 239 shows the beginning of a sentence with verbsecond word order, starting with a drop in pitch. As the comparison in (1’) (see below) shows, in colloquial and vernacular Berlin German (collG), sentences starting with an adverbial such as nach der ARbeit would require the verb in second position, yielding nach der ARbeit geh ich so etc. in a single prosodic unit. Yet the syntax and prosody exhibited in lines 238 and 239, namely nach der ARbeit, isch geh so DINGS; (‘after work’ and ‘I go like thingy’) in separate prosodic units, is a construction typical of the Turkish German style of speaking. This suggests that indeed line 238 was again intended to be phrased as a prosodic unit of its own. Next in our extract, in line 240, the item DINGS (‘thingy’) from line 239 is, with latching, but also after some hesitation signals, replaced by zum AUto; (‘to the car’), again in its own prosodic unit and ending with the same falling-to-mid pitch that was used for the repairable DINGS;. The tag question weißt du, (‘you know’) is added to this unit. In colloquial or regional vernacular German, the information phrased by Gül here in three separate prosodic units would rather be phrased in one single syntactic and prosodic unit, as the comparison in (1’) demonstrates: (1’) Lines 238ff. in Turkish German and a possible colloquial German version Turkish German: nach der ARbeit, isch geh so DINGS; … zum AUto; after work I go like thingy to the car adverbial subj V part ‘thingy’ PP Poss. collG version: nach der Arbeit geh ich so zum AUto after work go I like to the car adverbial V subj part PP

So far, thus, we have seen a first example of the following phenomenon: Information that in colloquial German would rather be phrased in one single syntactic

138 Margret Selting

and prosodic unit, in Turkish German is split up into several separate prosodic units. In this example, as in many other cases, the information is structured as follows: we can distinguish a pre-positioned adverbial or prepositional phrase (nach der ARbeit,) and a post-positioned prepositional phrase (zum AUto;) which are placed around something like a kernel sentence isch geh so DINGS;. Yet, for the units dealt with so far, you could still argue that the short prosodic units are a result of the repair processes becoming necessary here. They would then just be epiphenomena of the incremental nature of formulating and repairing one’s talk in interaction. That this is not the case is more clearly revealed in the following lines of the extract. (b) Lines 242–244: single sentence split up into three separate prosodic units

In lines 242 through 244, Gül produces an utterance that is syntactically constructed as a pre-positioned item danach (‘after that’), followed by a sentence with verb-second word order. As shown in (1’’), a possible version in colloquial German (collG) would be danach is vor meinem fenster so n brief,: (1’’) Lines 242ff. in Turkish German and a possible colloquial German version Turkish German: daNACH, v: vor meinem FENSter, is so BRIEF; after that in front of my window is such letter adverbial PP V subject Poss. collG version: da is vor meinem fenster so n brief; after that is in front of my window such a letter adverbial V PP subject

In colloquial German, as we saw before, in normal verb-second sentences, only one constituent may precede the finite verb. The pre-positioning of an item as the first constituent of a sentence with verb-second word order thus demands the finite verb to appear in second position; with all other constituents following after the finite verb. In Turkish German, however, the pre-positioning of an item does not affect the following sentence in this way: In the example in lines 242ff. we again see, syntactically, a normal sentence with verb-second word order following the pre-positioned adverbial. Incidentally, here, the sentence with verb-second word order has another topicalized adverbial as its first constituent before the finite verb in second position.

. As for colloquial German danach in this context seems odd, I replace it with the more adequate da.



Prosody and unit-construction in an ethnic style 139

Figure 1

In colloquial German, the entire sentence would be packaged in one single prosodic unit. In the Turkish German example, both the pre-positioned items daNACH, as well as v: vor meinem FENSter, as well as the verb phrase is so BRIEF; are phrased in prosodic units of their own and prosodically separated from the next constituents. The pitch trajectory of lines 242f. is depicted in Figure 1. daNACH exhibits a primary accent, its pitch is rising like at the end of a prosodic unit, and the next constituent, v: vor meinem FENSter, is set off with a drop in pitch in comparison to the pitch at the end of daNACH. And again, v: vor meinem FENSter, exhibits another primary accent, its pitch is again rising like at the end of a prosodic unit, and the next constituent, is=so BRIEF;, is again set off with a drop in pitch in comparison to the pitch at the end of v: vor meinem FENS­ ter,. All these prosodic structures contribute to make both pre-positioned items daNACH and vor meinem FENSter interpretable as prosodic units of their own, separated from the following prosodic unit. Nevertheless, as neither daNACH nor vor meinem FENSter are items here that can stand alone, they syntactically project continuation. In the remainder of Extract (1), after some utterances which I need not comment on here, we can see a similar thing again: (c) Lines 250–251: sentence + post-positioned constituent

In line 250, Gül produces the sentence da ist lie’ also BRIEF; (‘there is well letter’), with an internal repair and a primary accent, and adds the information that this letter was from her admirer, von TAner; (‘by Taner’), in a post-positioned prepositional phrase in its own prosodic unit with its own primary accent. As the information about the sender of the letter, von Taner, is obligatory here, it could be expected to be packaged into the same prosodic unit with the prior information. Here, however, it is set off into its own prosodic unit. In the analysis so far, I have identified the same kinds of phonetic and prosodic features used for packaging prosodic units as are also used in colloquial German. In Turkish German, however, they seem to be deployed for the packaging of different syntactic structures, i.e. they seem to be associated with different domains. In order to make this difference clearer, we need to briefly look at some of the previous results of research on unit construction in colloquial German more generally.

140 Margret Selting

4.2 Construction and delimitation of units in colloquial German  talk-in-interaction In order to describe the ways in which units are built and delineated through prosody in colloquial German talk-in-interaction, we need to distinguish between unmarked and marked cases. 4.2.1 Unmarked cases of unit construction in colloquial German The most important features can be summarized as shown in (a)–(c). For the purposes of the analysis at hand, (a) is most important here. (a) Co-occurrence of syntactic and prosodic units

In general, in unmarked cases, prosodically packaged units co-occur with syntactically (and semantically and pragmatically) possible units and thus constitute TCUs in their sequential context (cf. Selting 1995, 2000). Sometimes, there are items that syntactically could be part of either a previous unit, or a unit of its own, or part of the following syntactic unit. In such cases, the prosodic packaging signals the item’s status as intended by the speaker. (b) Beginnings more clearly signalled than endings

In general, the beginning of (new) units is signalled more clearly than the end of units. This makes sense – even though recipients must attend to possible completions – because in talk-in-interaction the possible completion of a unit may nevertheless have to be adapted to a recipient’s lack of response or other little troubles. For this reason, possible unit completions need to be flexible and adaptable to the exigencies of the interaction, by way of, for instance, extending it until another possible unit completion is reached that then the recipient might respond to in the designed way. In view of this necessary flexibility of unit-completion, it is in many cases only the clear beginning of a new unit that retrospectively allows the analysis that the prior unit is now treated as being completed. – This means that for analysing unit construction it might be easier to look at the beginning of new than at the ending of prior units. (c) Phonetic features at the beginning, throughout and at the end of units

As I have shown elsewhere, in German talk-in-interaction syntactically possible units may be packaged and delineated through the following bundle of phonetic and prosodic features for the organisational practices indicated (Selting 1995: Chapter 2, 2000):



Prosody and unit-construction in an ethnic style 141

i. The beginning of new units: New units often, but not always, exhibit the following cues to realize a delimitation from the prior unit: a. a break in pitch, realized through, e.g., an upstep or drop in pitch in relation to the end of the prior unit, b. increased tempo at the beginning of the new unit, or at the end of the old and at the beginning of the new unit (cf. ‘rush through’; Schegloff 1982: 76). These cues may but need not co-occur with pauses or ‘gaps’ between units. ii. The ending of units: a. Units normally realize at least one primary, focal accent, mostly constituted as a pitch accent. The final accent of the unit can be shown to function as the beginning of the transition relevance place/space (TRP). Here, the prosodic contextualization of either turn-holding or turn-yielding starts and is continued up to the end of the unit (cf. Schegloff 1982, 1998; Selting 1995, 1996, 2000). b. In northern colloquial German, it is in particular the final pitch chosen in and after the final pitch accent of the unit that signals turn holding or turn yielding: Level or mid rising pitch are used to signal turn holding; all other pitch movements at the ends of units and turns can – unless the sequential context has projected continuation – be used for turn yielding. c. In co-occurrence with the final pitch movement, the final unaccented syllables of the unit normally also show reduced loudness, compared to the accented syllable. In addition, final unaccented syllables may be lengthened and/or there may be a pause or ‘gap’ before the beginning of the new unit. Alternatively, final syllables may be produced faster and combined with latching in order to ‘rush through’ into the next unit (cf. Schegloff 1982: 76). In addition, final syllables may but need not be produced with creaky voice. iii. Signalling the current state within and throughout the unit: Within and throughout the unit delineated through the above described features, pitch and loudness normally follow the trajectory projected from the start and the first accented syllables. In particular, after unit-internal hesitations, hitches and/or pauses, the pitch and loudness that was deployed before . This of course fits together nicely with the system of turn taking as described by Sacks, Schegloff and Jefferson (1974) (cf. also Selting 2000): Just as in their rule system turn holding is the more marked case, prosodically turn holding is also the more marked case which is realized with particular, restricted, contours.

142 Margret Selting

the ‘trouble’ is being taken up again and continued, thus signalling cohesion with the prior part of the unit. In general, i.e. in unmarked cases, there is a co-occurrence of possible syntactic and prosodic units in their sequential conversational context. This means that in unmarked cases TCUs are constructed in such a way that possible syntactic units like sentences, clauses, phrases or single-word items are being packaged as a unit via prosody, in particular through beginning and ending with the above mentioned prosodic cues co-occurring with possible syntactic beginnings and endings of units. 4.2.2 Marked cases of unit construction in colloquial German In marked cases, however, possible syntactic units can be separated into several prosodic units. There are at least two possible uses for this in colloquial German. Only the second one is plausible here. i. Speakers can deploy the separation of syntactic units into several short prosodic units as a cue to suggest the interpretation of their talk as ‘insecure’, ‘tentative’ or the like; in this case, speakers also use other cues of hesitation and insecurity with a higher than normal frequency, such as hesitation tokens, self repairs, pauses, reduced loudness, etc. (cf. Selting 1995: 77f.). ii. Speakers can deploy the separation of syntactic units into several short prosodic units as a cue to suggest the interpretation of their talk as ‘emphatic’, ‘highly relevant’ or the like. Each prosodic unit then is presented as an information unit with its own primary accent that suggests the interpretation of focus. The separation of, say, a single clause into three prosodic units will thus suggest the interpretation of three foci for the entire syntactic unit. It is evident that this strategy can be deployed as a highlighting strategy for focussing on each prosodically packaged unit (for an example of this usage in an argumentation cf. Selting 1995: 322, lines 699–704; cf. also Selting 1994). In colloquial German, however, the splitting-up of syntactic units into several separate prosodic units is a marked strategy, not used very frequently. In Turkish German, in contrast, it is used fairly frequently and seems to be much less marked. After this excursion, I will now continue with some more examples of short prosodic units in Turkish German. . This is similar to what Halliday (1967) described as ‘tonality’, or what Pheby (1980) described as “markierte Kongruenz” in which “der Satz fällt mit mehreren Tongruppen zusammen” (Pheby 1980: 862).



4.3

Prosody and unit-construction in an ethnic style 143

More data and more analyses

In Extract (2), lines 64–65 and 68–69, reported action sequences are presented in short prosodic units: (2) tkdtw_Vor_Ha/Ze 63 FR =ECHT, really -> 64 HA =isch hab GEStern gesEhn; I saw yesterday -> 65 sO ne ThÜte? (.) such a bag (drugs) 66 dis war bestimmt ZEHN gramm;= that were surely ten grams 67 =oder t’ sch’ zwanzisch GRAMM; (-) or twenty grams -> 68 der mann hat geGEben, (.) the man gave -> 69 70 ZE einfach SO: [ simply like that oh my god (oh boy) 71 HA [

In line 64, a predicate is formulated that lacks its object; this object is only alluded to in line 65, where Hacer tells her recipient the topic of her story in a grammatically elliptical form, a noun phrase only, namely that she watched drug dealers dealing with a huge bag full of drugs. The predicate formulated in line 64 has its own primary accent in the word GEStern (‘yesterday’) and final mid-falling pitch. The obligatory object is packaged in a separate post-positioned unit which is shown in line 65: it begins with an upstep in pitch in relation to the end of the prior unit, has its own primary accent in the word ThÜte (‘bag’) and ends with high-rising pitch. The pitch of these lines can be depicted as shown in Figure 2. In lines 68–69, the paying for the drugs by the buyer and the giving of the bag by the seller are again presented in short prosodic units with elliptical sentences

144 Margret Selting

Figure 2

that, compared to colloquial German, lack grammatical objects. I will not detail on these lines. Extract (3) shows a later passage from the same conversation as Extract (2), again with short prosodic units: (3) tkdtwVor_Ha/Ze 97 HA 98 99 100 -> 101 -> 102 103 FR -> 104 HA 105 FR -> 106 HA

dann STEHN die da, (--) then they stand there und da REden n bisschen auf äh so and talk there for a bit uh like zwEi minuten drEi minuten, two minutes three minutes (-) und dann ge’ (.) nimmt der von der TAsche raus? and then he takes out of his pocket [diese] TÜte, this bag (drugs) [hm? ] [dann GIBT] der? then he gives [hm- hm- ] und dann gi’ zahlt der die GELD. and then he pays the money

In lines 101f., where Hacer describes the drug dealer‘s action of taking the bag with the drugs from his pocket, the obligatory constituent diese TÜte (‘this bag’) is formulated as a prosodically separate post-positioning in a separate prosodic unit after the prior part of the sentence und dann ge’ (.) nimmt der von der TAsche raus? (‘and then he takes out of his pocket’), which by itself is not a complete sentence. Both prosodic units have their own primary accents, the post-positioned constituent diese TÜte is set off from the prior prosodic unit through a drop in pitch at its beginning. Both prosodic units have final rising pitch, but the unit in line 101 ends higher than the one in line 102. The pitch trajectory of lines 101–102 can be sketched as shown in Figure 3.



Prosody and unit-construction in an ethnic style 145

Figure 3

Furthermore, in the remainder of the extract, in line 104, Hacer describes the drug dealer’s action of giving the drugs to the buyer again in a short prosodic unit which packages a minimally short, elliptical syntactic construction. And again, in line 106, the handing over of the money by the drug buyer is presented in a relatively short unit which displays an internal self-repair. In Extracts (4)–(6), you can see some more examples which I will not deal with in detail here. I only give rough translations. (4) tkdtw07_Sem5: Z.181ff. (3:30) 181 SEM 182 -> 183 -> 184 -> 185 186 TAN 187 188 SEM 189 TAN 190 SEM 191 192 TAN 193 194 SEM -> 195

((schnieft)) ey;= =ich bin RUNter gegangen, I went down von EUCH, from you Eben, just now, hm=hm; hast dus geKRIEGT, did you get it was, what hast du es geKRIEGT, did you get it nee; no auto war ja nicht DA; car wasn’t there ach SO; oh hm=^hm; ich GUCK so, I look meine SCHWESter, my sister

146 Margret Selting

-> 196 -> 197 -> 198 -> 199 200 201 202 TAN 203 204 205 SEM 206 207 TAN 208 SEM 209 TAN 210 SRM 211 212 213 214 TAN 215 SEM

(---) meine cousi couSINS, (-) my cous cousin und meine Oma; and my grandma ist UNten; are down there (1.2) die woll[ten they wanted [unten auf der STRASse,= down in the street =oder b’(-) or a bei EUCH auf der strasse; at your place in the street bei EUCH auf der strasse; at your place in the street stetTINner strasse; stettin street hä, ja vorm: pizza MAX; (-) yes in front of pizza max

((lacht)) ((laughs)) .h ich so was SUCHTS was SUCHT ihr hier; me like what are what are you looking for here ich wollte niCOL bis zu die: I wanted nicol as far as the bis zu: ecke (-) as far as corner BRINgen; take beGLEIten;=weisste, accompany her you know

(5) tkm02_Erc1: Z.153ff. 153 ERC ey ALter;= hey old boy



154 155 156 157 158 NIM -> 159 ERC -> 160 -> 161 -> 162 -> 163 -> 164 -> 165 -> 166 167 NIM 168 ERC 169 170 171 172 173 -> 174

Prosody and unit-construction in an ethnic style 147

=auf ARbeit hatten wir heute wieder so ne (.) at work we again had such a DRECKSdiskussion; dirty discussion ich HASse so was; I hate it ja;=ich weiß AUCH nich wieso ey; yeah I don’t know either why WASne für eine; what kind of mann über DINGS so; man about things like über so AUSländer; about like foreigners über DEUTsche so:; about germans like und ähm (.) und alles MÖGliche;= and ähm and all kinds of things =weißt weißt du wir ESsen noch; you know you know we are eating still wir ESsen;= we are eating =wir haben doch MITtagspause, we have our lunch break immer eine STUNde lang, always an hour long hm=hm, und heute hatten wir wieder unseren OPI da,=alter, and today we had this granddad there again der dieses jahr AUFhört, who retires this year der in RENte geht, who goes into retirement .h und (.) KAUM is der seinen ersten tag wieder da, and barely is he there again on his first day er kommt wieder aus SCHWEden, he comes back from sweden und äh LAbert soand uh jabbering like fängt so dieses diskussionsthema AN so; starts like this topic of discussion like

148 Margret Selting

-> 175 -> 176 177 178 179 -> 180 -> 181 -> 182 183 NIM 184 ERC 185 -> 186 -> 187 -> 188 -> 189 -> 190 -> 191 -> 192 193

=über die ganzen ähm extreMISten so; about all these ähm extremists like islaMISten;= islamists =ich mein’ I mean kann ja natürlich SEIN; weißt du, can of course be you know ähm es gibt ja von: uhm there are of es gibt ja von JEder kultur also SCHLIMme;= there are of every culture well bad ones =so LEUte;= like people =und alles MÖGliche; and everything else ja auf jeden FALL; yeah of course und (.) es gIbt halt ähm: and there is uhm hier die denken AUCH so; here they also think like that ich weiss AUCH nicht;= I don’t know either =isLA:M;= islam =es gibt dann ähm norMAle leute die das ausüben so, there are then uhm normal people who practice that like isLAM(is) so, islam like und es gibt wieder diese extreMISten; and there are again these extremists die fundamentaLISten; the fundamentalists so alles MÖGliche; like everything else und(.)ich find das son bißchen(.)SCHEISse dass das so-= and I think this is a little shit that that like



194 195 196 197 198 199 NIM 200 ERC 201 NIM 202 ERC 203 204 205 206 207 NIM 208 ERC

Prosody and unit-construction in an ethnic style 149

=

das is voll KOMisch; it is really funny wir haben doch den ANderen lehrling; we have this other apprentice der kommt doch aus dem OSten; ne, he comes from the east you know (.) (h)ja; yeah (h)der aus HELlersdorf kommt; who comes from hellersdorf ja; yeah wenn DER da is, when he is in wird über OSsis hergezogen, the easterners are being pulled to pieces wenn ICK da bin, when I am in wird über äh DINGS hergezogen; thingies are being pulled to pieces über [ä:h about uh [AUSländer; foreigners ja:;=so irgendwie voll KOmisch ey; yes like somehow completely funny ey

(6) tkdtw04_Nev3b: Z.337ff. 337 NEV 338 DAN 339 NEV 340 341

ich muss meine STIMme aufnehmen; (-) I have to record my voice waRUM denn, why is that na das ist so=ne UNI; well there is such a uni(versity) die FORschen wie hier ähm (--) who does research about how here ähm tÜrkische LEUte, turkish people

150 Margret Selting

342 343 344 345 346 347 348 349 350 DAN 351 NEV 352 -> 353 -> 354 -> 355 -> 356 357 358 359 360 361 362 363

die hier AUFgewachsen sind; who have grown up here wie die REden; how they talk was für=n diaLEKT die haben; what kind of dialect they have (1.7) haben die geFRAGT;= they asked =willst du MITmachen,= do you want to cooperate =hab ich gesagt oKAY, I said okay (1.3) ach `SO:; oh hä also NICHT so aus SPASS, hä so not just for fun (3.0) .h mein BRUder;= my brother =der ist Oben;= he is upstairs =Immer noch;= still =SCHLÄFT;=weißt du, sleaping you know ich ver↑STEH ihn nicht; I don’t understand him GEStern;=monika; yesterday monika ich war vielleicht !EI!ne stunde am computer, I spent perhaps one hour at the computer hab für ihn so MITfahrzentrale gesucht; looked for like agency for arranging lifts for him er hat am MONtag DINGS; he has on monday things PRÜfung; (---) exams er hat GAR nix, he doesn’t have a thing



364 365

Prosody and unit-construction in an ethnic style 151

er verSTEHT nix, he doesn’t understand a thing und dann MEINT ich; and then i meant

These extra examples are intended to indicate how frequent and common the phenomenon is. They cannot be dealt with in detail here. The cues that are used to signal and delimitate the short prosodic units are the same as those described for the Examples (1) through (3) above. 4.4 Conclusion of the structural analysis What we have seen so far is the following: Information that in colloquial or regional German would be phrased in one single TCU with co-occurring syntactic and prosodic units, in Turkish German is often split up into several separate prosodic units. These separate prosodic units may be syntactically obligatory or optional constituents of the prior syntactic structure. They are set off from the prior prosodic unit via using the cues for the beginning of a new prosodic unit; they are given their own primary accents; and they have their own final pitch. They are thus phrased as separate, yet fully “well-formed” prosodic units, even though they are not complete TCUs. So, in Turkish German prosody often does not delimitate TCUs, but only smaller parts of TCUs; prosodic units are often associated with smaller domains than possible syntactic units in their sequential context. In other terminology: In Turkish German, the domains of prosodic phrasing may be – apart from possibly complete syntactic units – also smaller constituents of possibly complete syntactic units. Some of these are pre- and post-positioned constituents in relation to kernel sentences (cf. Kern & Selting 2006a, b), but in many cases normal sentences or clauses are separated into several prosodic units.

. This is the difference between the short units in Turkish German and what Günthner (2005) calls ‘Dichte Konstruktionen’: a ‘dense construction’ is equivalent to what has before been conceived of as an ‘ellipsis’, i.e. a short construction that nevertheless is a correct complete syntactic unit in its co- and context (cf. also Selting 1997). . These smaller prosodic units in Turkish German do not seem to correspond to, for instance, intermediate phrases that are being integrated into intonation phrases on a higher level of organization, through a shared declination line. They sound rather separate and delimited from each other, cf. again Figures 1–3.

152 Margret Selting

Syntax

[possible syntactic unit, e.g. clause, in its sequential context ] | | Prosody 1 [ prosodic unit/phrase ] often in collG Prosody 2 [ prosodic unit/phrase 1] ([{]) [ prosodic unit/phrase n] often in TkG

Figure 4.  Co-occurrence of syntactic and prosodic units/phrases in TCUs in colloquial and Turkish German (‘|’ is intended to denote: ‘associated with’)

Future research will have to analyze in more detail what the other constituents are that can be separated in this way. The (tendency of) difference between colloquial German and Turkish German with respect to the co-occurrence of possible syntactic and prosodic units or domains can be represented as shown in Figure 4.

5. Functional analysis Why and for what ends does Turkish German deploy this kind of prosodic phrasing? What are the rationale and benefits of this kind of unit construction in this style? Most of the examples that I have in my collection occur in sequences of story telling. I will go back and look at the extracts in order to examine the functions of the described way of unit construction. In Extract (1), Gül tells her interlocutor about an unexpected letter that she found tucked to the windscreen of her car. Co-occurring with the short prosodic

. An anonymous reviewer asked about the reasons why Turkish German tends to show this prosodic pattern. Could there be some influence or interference from a Turkish substrate? I am very grateful to Yazgül Şimşek and Christoph Schroeder, specialists on Turkish at the University of Potsdam, for a discussion of this point. The tentative result is this: There is hardly any research available on the prosody and intonation of spoken Turkish which could help to answer this question. But yes, preliminary analyses let them think it is very plausible to hypothesize that the prosody of unit construction in Turkish German is influenced by the prosody of spoken Turkish. Spoken Turkish might well routinely separate complex syntactic structures into several prosodic units or chunks, especially in youth style. It might well be that prosodic units in unrehearsed spoken colloquial Turkish routinely have a different length from those in spoken colloquial German, perhaps due to the processing of complex agglutinating structures? This kind of unit construction in co-occurrence with Turkish intonation contours might result in a typical rhythmic structure for Turkish, which adolescents might then transfer to their style of Turkish German. This is an open field for future research. For existing further research on the prosody of Turkish German, in parts as compared to Turkish, see Kern (2008) and Şimşek (in press).



Prosody and unit-construction in an ethnic style 153

units, she deploys the attention getter guck mal (‘look’) in line 236 and the focussing and projection strategy ich GUCK so (‘I looked’) twice, in lines 246 and 249. That the letter was unexpected is made clear when Gül in line 247 says that at first she thought it was a ticket from the police. Producing her story in short prosodic units enables her to focus on each prosodically packaged unit separately, and that means: producing the information about the unexpected event in a slow and highlighted way. She creates suspension. Evidence for this is that her interlocutor Zelal seems to listen attentively and only responds with laughter after Gül mentioned the ticket from the police. In Extract (2), Hacer tells her interlocutors about a scandalous event that she observed: drug dealing. Her allusions at the quantity of drugs, sOne ThÜte (‘such a bag’) and bestimmt ZEHN gramm (‘surely ten grams’) etc., indicate that she indeed thinks that what she observed and now tells about is scandalous. Furthermore, the recipient’s response einfach SO: (‘simply like that oh my god (oh boy)’), with increased loudness, confirms that she, too, assesses the told-about information as scandalous. Here, again, the splitting up of information into short prosodic units enables Hacer to produce her story telling about a scandalous event in a slow and highlighted way. Splitting up sentences into short prosodic units seems to project continuation and suggest the interpretation of emphasis and emotive involvement. Evidence of this is that the recipients seem to listen and wait and then respond to the story, Zelal here responding to the expressed emphasis and assessment of the information as scandalous with her own concordant and emphatic response tokens in line 70. – A similar analysis also holds for Extract (3), except that here, still in the middle of the story, we only see more neutral continuers as response tokens by the recipient in lines 103 and 105. In Extract (4), we find a story being told about an unexpected event. In Extracts (5) and (6), we find stories that involve strong negative assessments – see the assessments drecksdiskussion (‘dirty discussion’) and ich hasse so was (‘I hate that’) in Extract (5) as well as er versteht nix (‘he doesn’t understand a thing’) etc. in Extract (6). In Extract (6), short prosodic units also seem to be used for a more special purpose, namely to create a rhythmically integrated sequence of units that highlights the information in another kind of way. (For the more detailed analysis of this example and on the construction of rhythm in Turkish German as an interactional resource see Kern 2008.) All in all, then, information that in other varieties of German would be packaged in one single prosodic unit is in Turkish German split up into several separate short prosodic units. This way of unit construction seems to be used as a resource to achieve the following functions in interaction:

154 Margret Selting

– Initial prosodic units that do not correspond to possibly complete syntactic constructions in themselves project continuation of the syntactic unit and thus the TCU under production. – As every short prosodic unit is given its own primary accent and thus signals focus, the splitting up of information achieves the signalling of multiple foci. Multiple foci, in turn, highlight the information thus presented and suggest the interpretation of emphasis. – If this resource is used in the complication part of story telling, the construction of short prosodic units seems to create suspense and thus project continuation of story telling till some later climax/high point or assessment or the like. As shown above, evidence in favour of this analysis can be gained from recipient responses. The result of my analysis thus is as follows: The function of splitting up TCUs into several separate prosodic units is similar in Turkish German to what has been summarized in Section 4.2.2 for colloquial German. What, however, is typical and different in Turkish German then? Four points can be made: 1. As said before, in the sequences with Turkish German as a style of speaking we find a lot of separation of TCUs into small prosodic units, a lot more than we normally find in colloquial German. At the time being, this is an impressionistic estimation, of course. The quantitative validation of this claim is still a desideratum. 2. Furthermore, in colloquial German, the splitting up of TCUs into smaller prosodic units is a marked device that may be used to signal emphasis; it is used rather rarely there. In Turkish German, however, besides for emphasis, the strategy is also deployed in segments of storytelling in which the signalling of emphasis is not plausible yet, for instance for the beginning of story telling (cf. Extracts (1): lines 236–240, (4): lines 183–185, (6): lines 353–356, or for parenthetical remarks like in (5): lines 163–166). This attests to the interpretation that in Turkish German the splitting up of TCUs into small prosodic units is deployed as a routine and unmarked device, not just for marked cases such as the signalling of emphasis in particular sequential or narrative practices or environments. 3. The device described here is similar to another one: In Turkish German, we often find asyndetic constructions, that means clauses juxtaposed without connecting conjunctions. They also often sound short, giving only the absolutely necessary information, with no redundancy. 4. However, we do not find many particles and other elaborate lexical means of expressing coherence and presuppositions in talk.



Prosody and unit-construction in an ethnic style 155

All this suggests that speakers of Turkish German, rather than using conjunctions and particles and other lexical means, revert to prosodic means of expressing coherence, emphasis and highlighting of information. This means that speakers of Turkish German leave out many verbal means of unit construction and unit combination and routinely substitute them with prosodic means. And because this is their predominant or routine resource in passages of Turkish German, this is precisely why they use the strategy so often (cf. also Kern 2008 on the use of prosody for the signalling of contrast in Turkish German).

6. Conclusions The phonetic and prosodic cues deployed for accomplishing unit-construction and delimitation in Turkish German do not seem to be very different from those used for other styles of German. What is different and typical of Turkish German, however, is its tendency to construct short units, and in particular to split up information that in other styles of German would be phrased in one single prosodic unit or TCU into several separate prosodic units. This means that after an initial constituent of a sentence that was constructed like a normal prosodic unit with a primary accent and a possible unit-final pitch trajectory, the speaker for the next constituent of the sentence produces a prosodic break and continues the syntactic unit like – albeit – the beginning of a new prosodic unit which again is produced with its own primary accent and a possible unit-final pitch trajectory, and so on. In this way, a single possible syntactic unit is split up into several separate prosodic units which each are formed like “well-formed” prosodic units in colloquial German. In another terminology again: While in colloquial German possibly complete syntactic units are normally packaged in and associated with one single prosodic unit – notwithstanding pre- and post-positionings to that unit – in Turkish German prosodic units are often associated with only parts of possibly complete syntactic units, such as dependent constituents that do not end in possible completion points of syntactic units in their sequential context. While in colloquial German, thus, prosody normally packages TCUs, in Turkish German it does not. In Turkish German, often only several separate prosodic units can be analyzed as constituting TCUs. Future research will need to describe the syntax of these smaller parts more closely. As I have shown, the construction of small prosodic units, and in particular the splitting up of possible single sentences, clauses or TCUs into several separate prosodic units in Turkish German is deployed as a routine resource in order to highlight information and signal emphasis. This way of constructing units in

156 Margret Selting

Turkish German was argued to compensate for the non-use of lexical means of expressing these interactional meanings and functions. While in other varieties of German, the construction of this kind of non-cooccurrence of syntactic and prosodic units is only rarely used in marked cases of expressed insecurity or emphasis, in Turkish German this is a frequently used resource for the accomplishment of highlighting and emphasis in normal everyday conversational interaction.

References Auer, Peter. 2003. “ ‘Türkenslang’: Ein jugendsprachlicher Ethnolekt des Deutschen und seine Transformationen.” In Spracherwerb und Lebensalter. Kolloquium anlässlich des 60. Geburtstags von Harald Burger, Annelies Häcki-Buhofer et al. (eds), 255–264. Basel/Tübingen: Francke. Günthner, Susanne. 2005. “Dichte Konstruktionen”. InLiSt (Interaction and Linguistic Structures) 43 (). Halliday, Michael A. K. 1967. Intonation and Grammar in British English. The Hague: Mouton. Keim, Inken. 2004. “Kommunikative Praktiken in türkischstämmigen Kinder- und Jugendgruppen in Mannheim”. Deutsche Sprache 32(2): 198–226. Kelly, John & Local, John. 1989. Doing Phonology. Observing, Recording, Interpreting. Manchester, New York: Manchester University Press. Kern, Friederike. 2008. Das Zusammenspiel von Prosodie und Grammatik in der Interaktion – am Beispiel des Türkendeutschen. Habilitationsschrift, Universität Potsdam. Kern, Friederike & Selting, Margret. 2006a. “Einheitenkonstruktion im Türkendeutschen: Grammatische und prosodische Aspekte”. Zeitschrift für Sprachwissenschaft 25: 239–272. Kern, Friederike & Selting, Margret. 2006b. “Konstruktionen mit Nachstellungen im Türkendeutschen”. In Grammatik und Interaktion, Arnulf Deppermann, Reinhard Fiehler & Thomas Spranz-Fogasy (eds), 319–347. Radolfzell: Verlag für Gesprächsforschung. Kern, Friederike & Şimşek, Yazgül. 2006. “Türkendeutsch: Aspekte von Einheitenbildung und Rezeptionsverhalten”. In Mehrsprachige Individuen – vielsprachige Gesellschaften, Dieter Wolff (ed.), 101–119. Frankfurt/M. usw.: Lang. Pheby, John. 1980. “Intonation”. In Grundzüge einer deutschen Grammatik. Von einem Auto­ renkollektiv unter der Leitung von K. E. Heidolph, W. Flämig und W. Motsch, 839–897. Berlin: Akademie-Verlag. Sacks, Harvey, Schegloff, Emanuel A. & Jefferson, Gail. 1974. “A simplest systematics for the organisation of turn-taking for conversation.” Language 50: 696–735. Schegloff, Emanuel A. 1982. “Discourse as an interactional achievement: Some uses of ‘uh huh’ and other things that come between sentences.” In Analyzing discourse: Text and talk, Deborah Tannen (ed.), 71–91. Georgetown University Round Table on Languages and Linguistics 1981. Washington, D.C.: Georgetown University Press, 1982. Schegloff, Emanuel A. 1988. “Discourse as an interactional achievement II: An exercise in conversation analysis.” In Linguistics in context: Connecting observation and understanding, Deborah Tannen (ed.), 135–158. Norwood, N.J.: Ablex. Schegloff, Emanuel A. 1996. “Turn organisation: One intersection of grammar and interaction.” In Interaction and grammar, Elinor Ochs, Emanuel A. Schegloff & Sandra A. Thompson (eds), 52–133. Cambridge: Cambridge University Press.



Prosody and unit-construction in an ethnic style 157

Schegloff, Emanuel A. 1998. “Reflections on Studying Prosody in Talk-in-Interaction.” Language and Speech 41(3–4): 235–263. Selting, Margret. 1994. “Emphatic speech style – with special focus on the prosodic signalling of heightened emotive involvement in conversation.” Journal of Pragmatics 22: 375–408. Selting, Margret. 1995. Prosodie im Gespräch. Aspekte einer interaktionalen Phonologie der Konversation. Tübingen: Niemeyer. Selting, Margret. 1996. “On the interplay of syntax and prosody in the constitution of turn-constructional units and turns in conversation”. Pragmatics 6(3): 357–388. Selting, Margret. 1997. “Sogenannte Ellipsen als interaktiv relevante Konstruktionen? Ein neuer Versuch über die Reichweite und Grenzen des Ellipsenbegriffs für die Analyse gesprochener Sprache in Interaktionen”. In Syntax des gesprochenen Deutsch, Peter Schlobinski (ed.), 117–156. Opladen: Westdeutscher Verlag. Selting, Margret. 2000. “The construction of units in conversational talk.” Language in Society 29 (2000): 477–517. Selting, Margret. 2001. “Fragments of units as deviant cases of unit-production in conversational talk.” In Studies in Interactional Linguistics, Margret Selting & Elizabeth CouperKuhlen (eds), 229–258. Amsterdam: Benjamins. Selting, Margret. 2005. “Syntax and prosody as methods for the construction and identification of turn-constructional units in conversation.” In Syntax and Lexis in Conversation. Studies on the use of linguistic resources in talk-in-interaction, Auli Hakulinen & Margret Selting (eds), 17–44. Amsterdam: Benjamins. Selting, Margret, Peter Auer, Birgit Barden, Jörg Bergmann, Elizabeth Couper-Kuhlen, Susanne Günthner, Uta Quasthoff, Christoph Meier, Peter Schlobinski & Susanne Uhmann. 1998. “Gesprächsanalytisches Transkriptionssystem (GAT).” Linguistische Berichte 173 (1998): 91–122. Şimşek, Yazgül. In press. Sequenzielle und prosodische Aspekte der Organisation der SprecherHörer-Interaktion im Türkendeutschen. Berlin: Waxmann.

Appendix Transcription symbols (according to GAT, Selting et al. 1998) Sequential structure [ ] overlap and simultaneous talk [ ] = latching Pauses (.) (-), (--), (---) (2.0) (2.85)

micropause brief, mid, longer pauses of ca. 0.25-0.75 secs.; until ca. 1 sec. estimated pause, more than ca. 1 sec. duration measured pause (notation with two digits after the dot)

158 Margret Selting

Other segmental conventions und=äh assimilations within units :, ::, ::: segmental lengthening, according to duration äh, öh, etc. hesitation signals, so-called ‘filled pauses’ ’ cut-off with glottal closure Laughter so(h)o haha hehe hihi ((lacht))

laugh particles within talk laugh syllables description of laughter

Accentuation akZENT ak!ZENT! akzEnt

strong, primary accent extra strong accent weaker, secondary accents

Pitch at the end of units ? rising to high , rising to mid - level ; falling to mid . falling to low Conspicuous pitch jumps ↑ to higher pitch ↓ to lower pitch Changed register



low register high register

Changes in loudness and speech rate =forte, loud =fortissimo, very loud

=piano, soft =pianissimo, very soft =allegro, fast =lento, slow =crescendo, continuously louder =diminuendo, continuously softer =accelerando, continuously faster =rallentando, continuously slower Breathing .h, .hh, .hhh h, hh, hhh



inbreath, according to duration outbreath, according to duration



Other conventions ((coughs)) ( ) (solche) al(s)o (solche/welche) ((…))

Prosody and unit-construction in an ethnic style 159

para- und extralinguistic activities and events concomitant para- und extralinguistic activities and event with notation of scope interpretative commentaries with scope unintelligible according to duration uncertain transcription uncertain sounds or syllables possible alternatives omissions in the transcript

Rhythm in Turkish German talk-in-interaction Friederike Kern

University of Potsdam

In the following contribution, I describe some characteristics of the ethnic style of speaking ‘Türkendeutsch’. My focus will be on the constitution of rhythm, and its forms and functions in talk-in-interaction. Two types of rhythmic principles will be identified, namely a pattern based on accent isochrony, and a pattern based on metrical stress distribution. The discussion of the data will demonstrate how rhythmic patterns locally govern prosodic and syntactic structures that may differ from colloquial spoken German as a result: sometimes, the primary accent of a turn constructional unit is shifted so it no longer conforms to colloquial spoken German rules of accentuation; in other cases, syntactic structures are changed in accordance with a rhythmic pattern. It is also suggested that speakers use phonetic and phonological features from Turkish and German to establish and maintain rhythmic patterns. Finally, the functions of rhythm in Turkish German as a resource for sequence organization, story-telling and the construction of accounts are discussed on the basis of the empirical material at hand.

1. Introduction: Theories and methods As in many urban centers all over Europe, an ethnic style of German has emerged in Berlin (Germany) among the second generation of young Turkish migrants. Turkish German shows specific combinations of prosodic and syntactic patterns that are never found in colloquial spoken or regionalized varieties of German but occur in various ethnic styles described in other European cities. It thus may be may be part of a pan-ethnic style that has developed in Europe. Extensive research on Turkish German’s similarities with and differences to other European styles is still missing, however.

162 Friederike Kern

In the context of the research project this study is based on, the term “Turkish German” is used to refer to a particular ethnic style employed by young speakers of Turkish background. The term “style” is preferred over “variety” to accentuate a specific research perspective: the focus in the project is on the use of Turkish German as a meaningful resource to achieve particular organizational goals in talk-in-interaction. As a style of speaking (cf. Selting 1999), Turkish German exhibits specific prosodic and syntactic features that can be described systematically with reference to the context in which it is used. Among the structures studied are specific kinds of pre- and post-positioned constituents before and after their ‘host’ sentences, the separation of turn-constructional units into very short prosodic units, the deployment of both lexical stress as well as turn constructional unit accentuation as a resource for stylistic variation, and the forms and use of Turkish and German recipient tokens. Furthermore, the conceptualization of Turkish German as a style was derived from the following two observations: firstly, Turkish German is only a part of the speakers’ repertoires. Secondly, it is usually not deployed throughout the conversations but only in particular turns or turn constructional units (TCU). Both observations show that our speakers have the choice between several varieties, and their use of Turkish German is stylistically motivated: they employ it for the organization of interactionally relevant, conversational tasks. The understanding of Turkish German as an ethnic style focuses on its forms and functions in talk-in-interaction, and thus results in a specific theoretical and methodological approach to the phenomenon. As Turkish German occurs in spoken language, it should be analyzed in its natural context, i.e. in the interactions in which it is used. As a consequence, the study is based on natural data, such as telephone and face-to-face conversations.

2. Research on German ‘ethnolects’ Research on Turkish German and comparable ethnic styles has shown significant linguistic differences to colloquial spoken German styles of speaking, especially

. The research project “Prosody in Turkish German” is located at Potsdam University (Germany) under the supervision of Margret Selting. It is funded by the German Research Foundation (DFG). . Cf. Kern & Selting (2006a and b), Kern & Simsek (2006).



Rhythm in Turkish German talk-in-interaction 163

on the level of prosody. The prosodic structures that have so far been described as most typical are, among others, the following (cf. Androutsopoulos 2001; Auer 2003; Keim 2001, 2004): – phonetic features such as the shortening of vowels, the aspiration of plosives, the coronalization of fricatives (/Iç/ → /I∫/), and the simplification of consonant clusters ((/ts/ → /s/); – a stomping speech rhythm (“stampfendes Sprechen”, cf. Keim 2001), a hammering rhythm (“hämmernder” Rhythmus, cf. Keim 2004), rhythmic speaking as part of a specific “prosodic richness” (Füglein 2000); – a tendency towards “syllable-timed speaking” (cf. Auer 2003; Dirim & Auer 2004; Keim 2004) that may involve a non-reduction of normally reduced syllables. These studies suggest that rhythm and rhythmic speaking is one of the most striking features of Turkish German. In colloquial spoken German, on the contrary, rhythm is said to be of relatively little importance (cf. Schwitalla 2001). However, rhythm has been described as a major feature in other European ethnolects as well. Kotsinas (1998: 136) observes that “Rinkeby-Schwedish”, a variety of Swedish spoken by young people of Arabic and Turkish ethnic background, is prosodically “choppy” or “uneven”; it is not clear, however, what this impression is grounded on. Moreover, detailed investigations about the types and forms of rhythmic patterns in Turkish German are still missing and there is even less knowledge about the functions of rhythmic speaking. Functional descriptions usually associate the ethnolect with stereotypical formulations such as “sounding macho” (cf. Androutsopoulos 2001). However, none of the studies look at the use of rhythm as a resource in interaction. In this contribution, I will attempt to close part of this gap, and to study more closely the constitution and functions of rhythm in Turkish-German talk-in-interaction. For this purpose, I concentrate on rhythmic patterns that are audibly extremely salient for German speakers, firstly because they consist of rather short rhythmic units, resulting in what has been described as a ‘staccato’ rhythm, and secondly, because they show patterns of syntactic and prosodic features that are somewhat uncharacteristic of standard spoken German. After an introduction to the data (Section 3) and a short review of the key research on rhythm in language with reference to metrical differences between

. Along with Couper-Kuhlen & Selting (1996: 11), I will use the term ‘prosody’ to comprise musical attributes of speech such as melody, dynamics, rhythm, tempo and pause.

164 Friederike Kern

Turkish and German (Section 4), I will describe two types of rhythmic principles that are commonly found in Turkish German (Section 5.1). It will be argued that rhythmic patterns locally supersede prosodic and syntactic structures that differ from colloquial spoken German. It is suggested that speakers creatively use phonetic and phonological features from Turkish and German to establish and maintain a rhythmic pattern. In Section 5.2, I turn to the discussion of functions of rhythmic speaking in Turkish German talk-in-interaction. First, the use of rhythm as a rhetorical device in story-telling and assessments will be discussed (Section 5.2.2). After this, it will be shown how speakers use rhythm for sequence organization, specifically to constitute “new beginnings” (cf. Couper-Kuhlen 2004) of topics and/or actions (Section 5.2.3).

3. Data The investigation is based on data from telephone and face-to-face conversations between ten adolescent women and men of ethnic Turkish background, and their friends. The informants, aged 16–22, have attended secondary schools in Berlin and have lived there most of their lives. In total, the informants recorded approximately ten hours of naturally occurring talk. For comparison, telephone and face-to-face conversations of young German men and women were also recorded. The data were transcribed according to an improved CA-style transcription system (see Selting et al. 2009). The investigation of rhythm is mainly based on auditory analysis, with attentive listening, accompanied by tapping, as a main resource (cf. Couper-Kuhlen 1993); additionally, the program PRAAT was used to confirm the auditory analysis by measuring the spans between the onsets of vowels of the accented syllables that constitute rhythmic beats. However, because measurements are much more exact than human hearing, a deviation up to 30% was tolerated (cf. Couper-Kuhlen 1993 and Section 5.1 below for more details).

4. Rhythm in language According to Auer & Couper-Kuhlen (1994: 85), rhythm – preliminarily defined as the recurrence of events in time – plays an important role in our everyday . The program PRAAT was designed by Boersma & Weenink for the acoustic analysis of speech and can be found at URL: .



Rhythm in Turkish German talk-in-interaction 165

life, even though its importance for language and speech has long been ignored. As a result, only few studies have investigated its role for the organization of everyday talk. However, by tradition, rhythm is a topic in phonetics and phonology, and many studies of rhythm in language are based on the belief that all languages try for at least one prosodic feature to be isochronous, namely either the interval between two accented syllables, or the duration of the syllables. This assumption has become widely known as the “isochrony hypothesis”. 4.1

The isochrony hypothesis

Ever since Pike’s (1945) observations about rhythmic units in spoken language, and the uptake of these observations by Abercrombie (1967: 96), languages have been divided into stress-timed and syllable-timed languages, with isochrony as the basis for distinction: As far as is known, every language in the world is spoken with one kind of rhythm or with the other. In the one kind, known as a syllable-timed rhythm, the periodic recurrence of movement is supplied by the syllable-producing process […]. In the other kind, known as a stress-timed rhythm, the periodic recurrence is supplied by the stress-producing process: the stress pulses, and hence the stressed syllables, are isochronous.

Thus, in stress-timed languages (such as German), it is the intervals between two stressed syllables that are presumed to be of approximately the same length in time; in syllable-timed languages (such as Turkish), it is the syllables that are supposed to be of equal duration. However, experimental evidence has shown that the distinction between stress- and syllable-timing cannot be held on the basis of isochrony alone (cf. Dauer 1983; Cruttenden 1986). Instead, it is more plausible to assume a rhythmic typology of languages that takes the constellation of specific phonetic and phonological characteristics into account. Dauer (1983: 55) argues that a collection of these characteristics, rather than mere isochrony, is responsible for our perception of languages as being rhythmically different. According to Dauer, three aspects are most relevant for the perception of rhythmic differences between

. See, however, Erickson (1992), Couper-Kuhlen (1993), and Auer, Couper-Kuhlen & Müller (1999).

166 Friederike Kern

languages: syllable structure, the possibility to reduce non-stressed syllables, and  the grammatical significance of word and sentence stress. Auer & Uhmann (1988) also claim that a rhythm-based typology should classify languages according to specific constellations of phonetic and phonological features. These considerations also lead to the conclusion that the rhythmic differences between syllable- and stressed-timed languages should be conceptualized as scalar rather than dichotomous. Laver (1994: 527f.) proposes the terms “syllable-based rhythm” and “stress-based rhythm” to capture the concept of rhythm as an emergent feature of languages that depends on their respective metrical properties. The latter rely on features such as syllable structure, sonority qualities, and lexical stress, among others. Laver argues that the proposed terms should be preferred to the more limiting terms “syllable-timed” and “stress-timed” to describe the perceived rhythms of languages. I would like to follow Laver and describe Turkish and German as languages with a syllable-based rhythm and a stress-based rhythm respectively. 4.1.1 Stress-based and syllable-based rhythm in Turkish and German In Table 1 I list some typical prosodic characteristics of Turkish and German that may be responsible for their rhythmic differences. The most relevant features concern the complexity of syllables, the existence of reduced vowels in non-stressed syllables, and the relevance of word accents. German, with complex syllable structures, long and short vowels, a strong word stress, and the option to reduce syllables, shows many elements of an accent based type of rhythm: non-accented syllables often contain vowels that are changed in terms of quality and quantity; sometimes these vowels are entirely deleted. As a result, non-accented syllables in German are often strongly compressed. A phonetically strong word stress has important functions for the information structure of the turn constructional unit. On the contrary, Turkish displays many features typical of a syllable-based rhythm: its syllable structure is simple, there is vowel harmony and no distinction between long and short vowels. Reduced (centralized) vowels do not exist, and a weak phonetic word stress fulfills only limited functions for signaling the semantic focus.

. Recent studies have made a new attempt to find a phonetic basis for isochrony in different languages by measuring the lengths of consonant and vowel intervals and the proportion of vowels (cf. Ramus, Nespor & Mehler 1999; Galves, Garcia, Duarte & Galves 2002; Low, Grabe & Nolan 2000; Grabe & Low 2002). It remains unclear, however, how this corresponds to perceiced rhythm in language (cf. Kern 2008).



Rhythm in Turkish German talk-in-interaction 167

Table 1.  Phonetic-phonological differences between Turkish and German Syllable-based rhythm (Turkish)

Accent-based rhythm (German)

simple syllable structures (CV or CVC)

complex syllable structure, with consonant clusters

no distinction between long and short vowels no centralized (reduced) vowels vowel harmony no reduction of non-accented syllables

distinction between short and long vowels centralized (reduced) vowels possible no vowel harmony reduction of non-accented syllables

word stress phonetically weak and phonologi- word stress phonetically strong and phonocally fixed (usually the last syllable of a word) logically free grammatical functions possible but rare grammatical functions possible sentence stress – few pitch accents – less relevant for the signaling of semantic focus – relevant for the expression of contrast and/or emphasis

sentence stress – pitch accents – important for the signaling of semantic focus

As will be shown in Section 5.1 below, German and Turkish phonetic and phonological features are both used as resources for the construction of rhythmic patterns in Turkish German talk-in-interaction. 4.2

Rhythmic patterns as ‘gestalts’

Also drawing on the experimental evidence against the isochrony hypothesis, Couper-Kuhlen (1993) claims that it is necessary to shift the focus from the physical properties of an acoustic signal to the ways in which rhythmic patterns are perceived and interpreted in talk-in-interaction. Taking a ‘gestalt’ psychological approach, she argues that rhythmic patterns cannot be said to exist as physical properties of an acoustic signal but they are actively constructed in the mind of the human observer. The analytic distinction between accents on the one hand, and rhythmic beats on the other hand, is central to this concept of rhythm. Accents are usually composed by changes of pitch contour, loudness and/or length. These physical cues have to undergo a process of transformation through interpretation by the hearer before they can be perceived as part of a rhythmic pattern, or ‘gestalt’, even if they occur regularly in time. It is through such an interpretive process that accents may be perceived as rhythmic beats. A ‘gestalt’ psychological approach to rhythm also implies that some constellations of events can be perceived more easily as rhythmic patterns than others.

168 Friederike Kern

The following linguistic properties support the perception of rhythmic gestalts in speech; they are called the ‘euphonic principles’ (cf. Auer et al. 1999: 27): – A rhythmic pattern contains few or no off-beat stresses. – The duration of the rhythmic intervals does not exceed certain upper and lower time limits (1 second and 2/10th of a second respectively). – Beats occur with stresses, and stresses are realized by pitch obtrusion. – The number of off-beat syllables is fairly regular. – Prosodic rhythm is supported by lexico-syntactic properties such as repetition or parallelism. Euphonic principles play an important role for the perception of rhythmic patterns: the more of the above characteristics a rhythmic pattern reveals, the better it can be perceived as a rhythmic pattern. According to Couper-Kuhlen (1993), at least three strong elements are required to perceive a stretch of talk as rhythmic: three prominent stresses constitute the minimum amount of two regular intervals that may be perceived as rhythmic units. A silent beat, however, could follow after two clear beats (Auer et al. 1999). As Imrie (2008) points out, this observation confirms the intuitive assumption that a third beat can be meaningfully expected after two preceding beats. 4.3

Rhythmic patterns as contextualization cues

Several studies that have focused on rhythm in discourse-oriented linguistic research point out that rhythm has no meaning of its own but can be used as a contextualization device to accomplish specific tasks in conversation. Rhythm is used as a resource for the organization and negotiation of conversational meaning that can be reconstructed only in its sequential context. Accordingly, the functions of rhythmic patterns must be described in their local context. This approach to rhythm in language entails the view that rhythm is not omnipresent in talkin-interaction but is locally constructed at specific sequential positions only to complete particular tasks in talk-in-interaction. Thus, within a theory of contextualization, rhythmic patterns are believed to be locally produced, and to have conversational meaning that has to be described with reference to its conversational context. In accordance with this, I will look . Cf. Couper-Kuhlen (1993), Auer & Couper-Kuhlen (1994), Auer et al. (1999). Cf. also Erickson (1992) for an approach to rhythm that differs to some extent from the one outlined here. . Cf. Gumperz (1982) for more details on the theory of contextualization.



Rhythm in Turkish German talk-in-interaction 169

at rhythm as a resource for the organization of turn construction and turn and sequence transition below, and describe its conversational meaning in terms of its functions in talk-in-interaction.

5. Rhythm in Turkish German talk-in-interaction The following sections deal with the description of the forms and functions of rhythm in Turkish German. The first sections (5.1 and 5.2) deal with two types of rhythmic forms and their underlying principles and characteristics. The section closes with a discussion about the possible ways in which rhythm in Turkish German may be influenced by both Turkish and German linguistic features (Section 5.1.3). In the following section (5.2), the functions of rhythmic speaking as a resource for turn construction (5.2.1) and sequential organization (5.2.2) will be discussed. The chapter closes with a summary on functions of rhythm in Turkish German (5.2.3). 5.1

Forms of rhythmic patterns

Two types of rhythmic patterns that are based on different linguistic principles could be observed in Turkish German. One of them correlates well with previous descriptions of instances of rhythm-in-conversation: it is based on the principle of accent isochrony. Examples of this type feature a series of accents on syllables that are thus marked as prominent. The second type of rhythmic pattern is based on recurrent stress patterns. Two, sometimes three, isochronous utterances show identical distributions of accents and thus form a specific rhythmic pattern. This type of rhythmic structure has not been described so far. 5.1.1 First rhythmic principle: Accent isochrony Accent isochrony was found regularly in Turkish German talk-in-interaction. First, isochronous stretches of talk were identified by auditory analysis. Then, measurements were taken to locate the phonetic basis of isochrony. Following CouperKuhlen (1993), the distance between the onset of a vowel within a prominent

. Halliday (1985) and Couper-Kuhlen (1993) call this type of rhythmic unit “foot”, and speak of “foot isochrony”. I will use the term “accent isochrony” instead, to direct the attention to the fact that the phonetic basis of this type of rhythm is indeed the relatively isochronic interval between two rhythmic beats marked by pitch accents.

170 Friederike Kern

syllable to the onset of the next vowel within a prominent syllable was measured with the help of PRAAT. The first example shows a typical case of accent isochrony: Rhythm in this sequence is clearly based on accent isochrony, with the beginning of each rhythmic unit marked by a prominent syllable that constitutes a rhythmic beat. Example (1)10 404 SEM hm= /Fatih hat wieder / 0.76 sec. /5

fatih has again

/Angefangn / 0.66 sec. /3



/FAHRschule,= / 0.45 sec. /3 (faster)

started

driving school

405 / / 0.42 sec. /4

have I told you

/SAGT,

In the first part of the rhythmic pattern, three prominent syllables (FA/An/FAHR) that occur in nearly isochronous intervals constitute two very short rhythmic units (the percentage deviation between the intervals is –14 %). In the second part, the two rhythmic units, established by two more prominent syllables, are slightly shorter and thus faster but with only –7% deviation. Even though the percentage deviation between the second and the third interval is rather high (–32%) it is nevertheless perceived as part of the previously established rhythmic pattern. In this example, the first turn constructional unit’s syntactic structure is as follows: a matrix sentence (Fatih has started again) is expanded with a prepositional object (Fahrschule/driving school) in a post-positioning. However, in the post-positioning, the head of the prepositional phrase (mit der/with the) is not realized. Table 2 demonstrates the differences between the Turkish German example and a (hypothetical) colloquial German equivalent. As a result, a reduced syntactic phrase is produced that coincides with a very short rhythmic unit (FAHRschule). As noted before, the units become even shorter

10. The transcriptions follow GAT (cf. Selting et al. 2009), English translations are simplified. The slashes are placed at the beginning and end of the rhythmic units that are aligned underneath one another. The duration of the intervals and the number of syllables are also indicated after the right slashes.



Rhythm in Turkish German talk-in-interaction 171

Table 2 Turkish German

Colloquial German (hypothetical)

Fatih hat wieder Angefangen FAHRschule Fatih hat wieder angefangen mit der FAHRschule Fatih has started again driving school Fatih has started again with the driving school

in the second half of the rhythmic pattern. Together, all short units constitute a fast, staccato-like rhythm. The next example shows a similar case in point. Example (2) 353 NES mein /BRUder; / 0.36 sec. /2

my brother 354 der ist

he is

/Oben;= / 0.28 sec. /2

upstairs

355 /=Immer noch; / 0.35 sec. /3

still

356 /SCH:LÄFT; / 0.32 sec. /1

sleeps

=wEIßt du,

you know

The rhythmic pattern in Example (2), which is again based on accent isochrony, stretches over four turn constructional units (lines 353–356). It consists of four short rhythmic units, with prominent syllables at their respective beginnings. The syntactic form of these two turn constructional units forming the rhythmic pattern, and how it matches the rhythmic units, is particularly interesting. The syntactic structure can be described as follows: A left dislocation (mein Bruder) and a post-positioning (immer noch) frame the matrix sentence (der ist oben); the following sentence (schläft) is elliptical, and the connection between the two sentences is asyndetic. The overall syntactic structure is somewhat condensed, with short units, each receiving a rhythmic beat. Again, the following table (3) shows the differences between the Turkish German example and a hypothetical colloquial German one. Obviously, the syntactic structure is efficiently modified: instead of formulating one long and complex sentence, the speaker breaks up the syntactic structure into several short units and co-ordinates each unit with a rhythmic beat. Again, the emerging rhythmic pattern is based on accent isochrony: the rhythmic beats occur

172 Friederike Kern

Table 3 Turkish German

Colloquial German (hypothetical)

mein BRUder; der ist Oben; =Immer noch; SCHLÄFT; my brother; he is upstairs; still; sleeps

mein Bruder der ist immer noch oben und schläft my brother he is still upstairs and sleeps

in approximately equal intervals, with a varying number of unstressed syllables in between. The specific, somehow condensed, syntactic structure that aligns with the short rhythmic units intensifies the impression of a fast staccato-rhythm. Rhythmic patterns in Turkish German do not only go hand in hand with syntactic modifications. In many cases, we find interesting phonetic-phonological features, as seen in the following example. Example (3) 102 sie bekommt

she gets

ENDergebnis / 0.78 sec. /4



/TECHnologie ne / 0.80 sec. /5



/^VI:ER; / 0.55 sec. /1

final results technology four

103 /dAs find ich krass

that’s wicked

The example demonstrates another strategy to constitute an isochronous, accentbased rhythm: in TECHnologie, the speaker shifts the word stress from the last syllable (technolo’gie) to the first to establish another isochronous rhythmic unit. German speakers are not known to produce such shifts of word stress. Speakers with Turkish linguistic background may, on the other hand, treat word stress as arbitrary, as word stress hardly bears grammatical functions in Turkish. It is interesting to note, however, that speakers of Turkish German do not simply transfer a phonological feature from one language to another: if this was the case, the word stress would be placed on the last syllable of the word, according to Turkish rules of word accentuation. Instead, the speakers treat word stress as if it had no grammatical function in German either, so that it can be easily shifted for other purposes, such as the constitution of a rhythmic unit. Additionally, a specific syntactic structure can be observed: all prepositions are again omitted, probably for the purpose of constructing isochronous rhythmic



Rhythm in Turkish German talk-in-interaction 173

Table 4 Turkish German

Colloquial spoken German (hypothetical)

sie bekommt ENDergebnis TECHnologie  ne VIER she gets final result technology a four

sie bekommt als endergebnis in technologIe eine VIER she gets as a final result in technology a four

units. Table 4 shows the differences between the Turkish German and a hypothetical Colloquial German example. A final example demonstrates other types of phonetic and phonological variation in the context of rhythmic speech in Turkish German. The rhythmically spoken sequence consists of five rhythmic units (lines 354–360) that are slightly varied throughout the sequence. Example (4) 354 /SIE macht so-=

she does it like that

355 =ich / 0.73 sec. /4/

I

/MACH nicht=so-=

don’t do it like that

356 =e / 0.69 sec. /4/

doesn’t

/GAL;=

matter

357 =lass sie / 0.52 sec. /3/

let her

/MACHn;=

do it

358 =ich ver / 0.54 sec. /3/

I tr

/TRAUe ihr; / 0.52 sec. /3/

ust her

The first two rhythmic units contain four syllables, with one prominent and three non-prominent syllables (lines 354–356). Then, the pattern changes, and the following rhythmic units consist of one prominent and two weak syllables (lines 356–358). The intervals become shorter, in time as well as in terms of numbers of syllables per unit. In these shorter units, unstressed syllables are compressed, as it is commonly done in spoken German. In line 357, MACHn, usually a two-syllable

174 Friederike Kern

Table 5 Turkish German

Colloquial spoken German (hypothetical)

SIE macht so-= she does it like that

sie mach SO she does it like that

=ich MACH nicht so-= I don’t do it like that

ich mach NICHT so I don’t do it like that

word, is strongly compressed by the elision of the pre-final schwa. It is thus perceived as a one-syllable word. As a result, another three-syllable rhythmic unit is constructed. In the next TCU (line 358), the speaker, Filya, does not eliminate the last vowel of vertraue neither does she even reduce it to [ә], as would be expected in colloquial spoken German. Instead she produces a vowel close to [ε], so that an additional syllable is created. Again, a three-syllable rhythmic unit is constructed. As a consequence, a rhythmic pattern with equal amounts of syllables in each of the three subsequent rhythmic units is produced. Such an isometric pattern is established by use of the flexibility of the German phonotactic system that makes compressing syllables and deleting unstressed vowels possible. Primary accentuation also differs notably from German rules of accentuation. In lines 354 and 355, the placement of primary accents deviates from German: in line 354, the primary accent of the turn constructional unit is on ICH but would be expectable on so to signal the turn constructional unit‘s semantic focus (so referring to the recipient‘s previous actions that caused the argument). In line 355, the primary accent is on MACHn but would be expectable on nicht to express and communicate the contrast between Filya’s and her friend’s actions. Table 5 shows the accentuation differences between the Turkish German example and its (hypothetical) colloquial spoken German equivalent. To sum up, all examples demonstrate a wide variety of phonetic-phonological and syntactic practices that work together to establish and maintain rhythmic patterns that are accent-isochronous, and often isometric, i.e. with the same amount of syllables per rhythmic unit. However, accent isochrony is not the only way to produce a rhythmic pattern in Turkish German. I will now turn to the other rhythmic pattern that is based on different principles. 5.1.2 Second rhythmic principle: Recurrent stress pattern A second rhythmic pattern frequently found in Turkish German appears to not be established by accent isochrony. Instead, it is based on recurrent stress patterns: two (sometimes three) turn constructional units of similar length in time show identical distribution of prominent and non-prominent syllables. Such a matching



Rhythm in Turkish German talk-in-interaction 175

metrical stress pattern is the core feature of this type of rhythm in Turkish German, and it is often accompanied by lexico-syntactic parallelism and melodic repetition. Example (5) shows a case in point. Example (5) 362 er hat GAR nix, /0.75 sec. /4

he has nothing

363 er verSTEHT nix, /0.74 sec. /4

he understands nothing

In this case, both turn constructional units together form a rhythmic pattern that consists of two isochronic rhythmic units. Most importantly, both turn constructional units display identical syllable patterns, with prominent and weak syllables evenly distributed, as the representation in a metrical grid reveals.11 Example (5’): Representation of metrical structure x x x x x x x x x x x x er hat GAR nix



he has nothing

x x x x x x x x x x x x er verSTEHT nix

he understands nothing

The identical metrical pattern is easily recognizable. Again, interesting phoneticphonological features can be found. In particular, the vowel quality in the prefix /ver/, that usually contains the reduced [ɐ] as a nucleus, is changed in quality to be more close to [ε], followed by a rhotacized [ɐ]. As a consequence, the syllable is not reduced and gains more weight. It thus receives one more beat in the metrical grid above.12 The non-reduction of the syllable results in an auditive impression 11. For the presentation of metrical structures Selkirk’s (1984) system of metrical grids is chosen. See Uhmann (1992) for its adaption to German. 12. Interestingly enough, according to the DUDEN, which is used as a resource for the citation form in German, the vowel in the prefix /ver/ should be properly pronounced [εɐ]. The standard phonetic form used in colloquial German seems to be [ɐ], however. It could be argued that the

176 Friederike Kern

of a “syllable-based” rhythm: all syllables in the turn constructional unit appear to be more or less of equal length. The rhythmic structure does, however, not rely so much on identical syllable length but on identical distribution of strong and weak syllables across both turn constructional units. Moreover, the construction is, in this particular pragmatic context, syntactically and semantically odd: a supplement to the auxiliary, such as the past participle gelernt (learnt), seems to be missing. With it, the turn would be more easily comprehensible. However, with the construction a lexico-syntactic parallelism is preserved. Example (6) shows another instance of the second rhythmic principle. Example (6) 344 ich hab ihn ANgeguckt, / 1.25 / 6

I looked at him

345 ich hab mein KOPF gedreht; / 1.25 / 6

I turned my head

Example (6’): Metrical grid representation x x x x x x x x x x x x x ich hab ihn AN ge guckt

I have looked at him

x x x x x x x x x x x x x ich hab mein KOPF ge dreht

I have my head turned

Again, the identical syllable pattern of two isochronous turn constructional units, with regularly distributed primary accents, is the most salient (and indeed audible!) feature of the rhythmic pattern. And again, this is supported by syntactic and lexico-semantic parallelism. Once more, properties of the German sound system are used to achieve and maintain the metric syllable pattern: instead of marking the correct grammatical case by realizing the suffix -en either as [әn] or [n], the

Turkish German speaker uses the citation form rather than the standard colloquial form in order to prevent a phonetic reduction of a syllable.



Rhythm in Turkish German talk-in-interaction 177

speaker drops the last syllable completely (line 345). She thus produces a one-syllable word that fits well into the established rhythmic pattern. Another example of this type of rhythm is shown in (7). Example (7): 762 isch mach jetzt AUS, /89 sec./ 4

I turn it off now

763 (---) /93 sec./ 764 und ruf mal AN; /82 sec./ 4

and call me

In this example, two equally long turn constructional units are broken up by a pause that is only a little longer than any of the turn constructional units. The pause thus can be said to acquire the function of a silent rhythmic unit. A metrical grid representation shows that the weak (non-accented) and strong (accented) syllables are again evenly distributed in both turn constructional units. Example (7’): Metrical grid presentation x x x x x x x x x x isch mach jetzt AUS,

I turn off now

x x x x x x x x x x und ruf mal AN;

and call me

In this particular example, there is another interesting variation of grammatical norms compared with colloquial German: the imperative13 in the second turn constructional unit is somewhat unusually connected with ‘und’ (and) to the previous turn constructional unit that is not an imperative.14 Table 6 demonstrates 13. That it is in fact an imperative and not a coordinated paratactic clause becomes evident from a reformulation of this particular turn after a follow-up question by the recipient (rUf misch mAl auf HANdy an; call me on the mobile). 14. Conjuncts connected to each other by a copulative conjunction should belong to the same grammatical category (cf. Lang 1977).

178 Friederike Kern

Table 6 Turkish German

Colloquial spoken German (hypothetical)

isch mach jetzt AUS, und ruf mal AN; I turn it off now and call me

ich mach jetzt aus und du rufst mal an I turn it off now and you call me

the difference between the Turkish German and a hypothetical colloquial spoken German form. The Turkish German form, although untypical, allows the speaker to produce the same number of syllables in both TCUs, with the last syllable as the most prominent one in both cases. The syntactic structure is modified again to make possible the specific rhythmic pattern with identical stress distribution over two TCUs. To conclude, these examples show that other rhythmic patterns than the ones based on accent isochrony are also constructed in Turkish German. Two turn constructional units with identical metrical stress patterns as the most important feature are combined to one rhythmic pattern. Because of parallel metric syllable structures and parallel lexico-semantic and syntactic features, strong coherence is established between the two turn constructional units. 5.1.3 Turkish and German linguistic effects on rhythm in Turkish German In Turkish German, rhythmic patterns regularly co-occur with prosodic and/or syntactic variation of linguistic principles of German. In all of theses cases, the variations can be explained by the necessity of establishing a rhythmic pattern. Rhythm thus seems to be the principle that can locally govern syntactic and prosodic principles. Firstly, different accentuation rules may apply both on word and sentence level. Word stresses may be shifted, violating German morpho-syntactic rules; but such stress shifts usually occur in order to form a rhythmic pattern. Primary stress does not function as a signal of a turn constructional unit’s semantic focus, as is the rule in German. Instead, stress is regularly shifted in accordance with an existing rhythmic pattern, so that accented syllables can be perceived as rhythmic beats. In this case, the function of turn constructional unit accentuation as an information-structuring device (i.e. for signaling the semantic focus of a turn constructional unit) is abandoned in favor of the constitution and maintenance of a rhythmic pattern. As sentence stress in Turkish usually does not carry the grammatical meaning of focus signaling (cf. Kornfilt 1997), speakers may treat



Rhythm in Turkish German talk-in-interaction 179

sentence stress as a non-obligatory feature that can be re-arranged if necessary for the constitution of rhythmic units. Thus, the flexible handling of accentuation on word and sentence level leads to the construction of rhythmic patterns, and accents gain new functions in the context of rhythmic speaking in Turkish German. Secondly, the speakers use the flexibility of the German sound system creatively as a resource to establish rhythmic patterns. Sometimes, the possibility of compressing syllables with reduced vowels is exploited; on other occasions, the vowel quality is changed so that a less reduced vowel is used instead of a less reduced one. As a consequence, the syllable gains more weight and may even receive a rhythmic beat within a rhythmic pattern. Thirdly, variations of colloquial spoken German syntactic norms could also be observed. In rhythmically spoken turn constructional units, constructions typical of colloquial spoken German such as pre- and post-positionings, elliptical and asyndetic structures are often used. Additionally, prepositions and/or obligatory grammatical items are absent. Other studies routinely attribute this to the fact that prepositions do not form a word class in Turkish; temporal and spatial relations are usually realized as suffixes. A closer look at such turn constructional units reveals, however, that many of them are rhythmically organized. In theses cases, the deletion of grammatical phrases and/or prepositions serves to keep the rhythmic pattern. In particular, speakers with a native-like knowledge of Turkish may regard prepositions more easily as redundant than German monolingual speakers. 5.2

Functions of rhythm in Turkish German talk-in-interaction

So far, only the linguistic forms of rhythmic patterns have been described. It can be seen from the examples, though, that rhythm is not an omnipresent phenomenon of Turkish German talk-in-interaction but is used only at specific sequential positions and in particular conversational contexts. Its function as a contextualization cue in turn construction and sequence organization will be described in detail in the following sections. Due to space limitations, I will discuss only one example for each of the functions, but many more can be found in the data. 5.2.1 Rhythm as a rhetorical device in turn construction Rhythm is used as a rhetorical device in turn construction mainly to contextualize emphasis and insistence. Detailed analysis of the conversations at hand has shown that rhythm is particularly employed in the context of story-telling and for the production of accounts. Thus, rhythm gains an important function as a resource

180 Friederike Kern

to emphasize particular narrative events in the course of a story, or to indicate, for instance, one’s persistence on one’s standpoint. The next section presents two instances of such occurrences of rhythmic speaking. 5.2.1.1 Rhythm in story-telling.  Within story-telling, rhythm is often used to focus on turn constructional units and highlight particular narrative events. In the following sequence, Gül is telling her friend a story about her ex-boyfriend. In the story, several turn constructional units are spoken rhythmically to highlight important narrative events. Segment (1)

151 NES er sagt so;



152 wieso BIST du denn so;



153 ich MEINte so



154 =WIE bin ich=



155 =ja lass mich doch mal in ruhe; (.)



156 er sagt so;=

he says like

why are you like this I said like

what am I like

yeah well leave me alone for once he says like

-> 157 =wieso /LÄSST du /mIch nicht /kÜSsen;

why don’t you let me kiss you

158 ich SAG so;



159 wieSO denn,



160 wir sind doch gar nicht zuSAMMn,



161 SEM hm=hm 162 NES er SAGT mir nur; (---)



163 ja (--)

I say like

why then

we are not going out

he only said to me yes

-> 164 / TROTZdem /wIll ich dich /KÜSsen; (.)

still I want to kiss you

165 SEM ‘oh=’oh, 166 NES ich so bist du beSCHEUert, oder was;

I like are you mad or what



Rhythm in Turkish German talk-in-interaction 181

-> 116 wir /sInd nicht /mEhr zu/SAMMn;=

we are not going out anymore

168 =verSTEHST du das nicht, (-)



169 er sagt zu MIR so;



170 du gehörst MIR;

don’t you understand that he said to me like you belong to me

This is part of a longer narrative about Güls’ problems with her ex-boyfriend. In the segment shown above, Gül stages a dispute between her and her boyfriend by retelling a dialogue between them.15 During the narrative, Gül uses rhythmic speech to highlight the three particular instances of reported speech (lines 157, 164 and 167) to focus on them as the main arguments in the conflict. In the two turn constructional units wieso LÄSST du mIch nicht kÜSsen (why don’t you let me kiss you) and TROTZdem wIll ich dich KÜSsen (still I want to kiss you), the ex-boyfriends’ voice is imitated to present his obviously shocking request. Gül stages her parts of the dialogue also in reported speech but produces only her main argument against his demand to kiss her in rhythmic speech, namely that they are not a couple any more (line 167: wir sInd nicht mEhr zusAMMn (we are not going out anymore). Rhythmic speech is used as a resource to emphasize the turn constructional units and make them interpretable as the main events within the conflict story. As can been seen from this (and other examples), rhythm can be used as a powerful device to structure and design a story by highlighting particular narrative events. But rhythm does not only work on the level of story organization but also functions as a contextualization cue to signal emphasis, or even insistence. We will look at this particular rhetorical function more closely in the next section. 5.2.1.2 Rhythm in assertions.  Rhythm is employed in assertions, especially after speakers’ standpoints have been doubted, or hearers have neglected to respond to a strong assessment. A case in point is Segment (4) where the rhythmically spoken turn comes at the end of a long argument between the two friends (or lovers?) Aydan and Filya. The trigger of the argument was a friend’s overnight stay at Aydan’s house. Aydan suspects Filya to be jealous, but this is denied by Filya (not shown in the transcript). Aydan then produces two compliments but Filya is not sure whether Aydan is being sarcastic or not. Finally, Filya recapitulates their 15. In story-telling, direct speech is often employed for a scenic re-play of communicative events (cf. Quasthoff 1980; Günthner 2000).

182 Friederike Kern

different positions in a longer turn; this is when she starts to speak rhythmically (lines 353ff.). Segment (2) 338 AYD (-)

you are so understanding, yeah (…)

346 FIL mei meinst dus jetzt ERNST? (.)

are you serious

347 oder ve ve=verARSCHT du mich;

or are you taking the piss 348 AYD ich meins ERNST;

I am serious 349 FIL ja;

yes

350 SIEHST du, (---)

you see

351 ich hab hab noch viel verSTÄNDnis;

I even have much understanding 352 ich sag (.) oKAY,

I say okay

353 /SIE macht so-=

she does it like that

354 =ich /MACH nicht=so-=

I don’t do it like that 355 =e /GAL;=

doesn’t matter

356 =lass sie /MAchn;=

let her do it

357 =ich ver /TRAUe ihr;

I trust her

358 (--) 359 AYD HAdi lan (-)

oh no boy

360 FIL ich SCHWÖre; (---)

I swear

361 AYD VALlah, wirklich

really

362 FIL HM=hm, (---)



Rhythm in Turkish German talk-in-interaction 183

After Aydan’s first compliment (line 338), Filya first agrees but then expresses her doubt as to whether Aydan is serious about it (lines 342–343). Aydan reconfirms by producing another compliment (line 345). Filya, however, repeats her disbelief (lines 346–347). After Aydan affirms her sincerity, Filya agrees again, and then produces a rhythmically spoken account of her being so sympathetic. She contrasts Aydan’s position with her own (lines 353–354) and finally resumes her own standpoint (355–357) in an attempt to end the argument between her and her friend. Here, rhythm is used as a way to make Filya’s standpoint clear to Aydan. Sequentially, this turn present as summary of Filyas view on the matter of the argument, and is, at the same time, an attempt to finish it. And indeed, even though Aydan initially expresses her doubt about Filya’s turn, the argument is finally settled after recurrent affirmations by Filya (lines 360 and 362). 5.2.2 Rhythm as a resource for sequence organization The function of rhythm as a device to organize sequential meaning has been widely discussed in conversation-analytic inspired research (cf. Couper-Kuhlen 1993; Auer, Couper-Kuhlen & Müller 1999). The focus has been on the use of rhythm as a resource to manage transitions of turns and actions in naturally occurring talk. Rhythm in English has been described in terms of its function to signal continuity across turns of different speakers by “rhythmic integration”: endings and beginnings of turns are tied together by rhythm (cf. Couper-Kuhlen 1993: 126). Furthermore, speakers are able to project the endings of turns not only by syntactic and melodic means but also by rhythmic structures. Other studies show how rhythm is regularly used in German for the organization of closing sequences in telephone conversations (cf. Auer 1990). Rhythmic delay, or rhythmic disintegration, can be used to signal dispreferred next actions, such as disagreement where agreement would be expected (cf. Couper-Kuhlen 1993). However, not every dispreferred turn is rhythmically delayed. Sometimes, dispreferred turns are rhythmically integrated, or even start early. After announcements, the next turn is often rhythmically delayed to support the intended surprise effect of the news as “hot news”. Interestingly, in Turkish German, rhythm seems to be used as a device to signal discontinuation and discrepancy to whatever has been done before in the interaction rather than continuation or integration, especially on the level of sequence organization. Rhythmic turns are employed to signal the status of a turn as a topic-proposal when topic-change is not accomplished by stepwise transition (cf. Jefferson 1984). In other cases, rhythm serves as a device to mark an action as

184 Friederike Kern

somewhat unexpected or misplaced in the course of the sequence, or to highlight particular narrative events within a story. 5.2.2.1 Rhythm in topic-proposals.  One of the main functions of rhythm in Turkish German is to organize sequences – rhythm is used as a device to introduce a potential new topic after a sequence has come to a possible closing point. Research on sequence and topic organization in English has shown that speakers have two possibilities: they can produce topic proposals either as sequence-continuing, or as sequence-closing. On the one hand, topic proposals can close a prior sequence; in that case they are clearly marked as new beginnings (Couper-Kuhlen 2004) by prosodic features of discontinuity, such as temporal delay, high pitch peaks at the beginning of a turn constructional unit, and extra loudness. On the other hand, topic proposals that are interpreted as continuing the sequence miss discontinuity markers and often include anaphoric forms for back-reference instead. The case of Turkish German appears to be somewhat different. Here, rhythm more than any other prosodic device seems to be used to signal discontinuity: by speaking rhythmically, a turn constructional unit is made audibly different from the previous ones that do not display a distinct rhythmic structure. Other phonetic signals of discontinuity such as high pitch peaks and loudness are used less often but can also be found in combination with marked rhythmic speaking. Take a look at the following example where Sema has been telling a lengthy story about a recent driving experience. Segment (3) 401 SEM ja JA;

yeah yeah 402 du wirst dann AUCH so bald machen;>

you will also do the same soon 403 ESI inşallah; gebe gott

so God will 404 HOFfentlich;

hopefully

405 he .HE 406 SEM m=/FAtih hat wieder /Angefangn /FAHRschule,=

fatih has started again driving lessons

407 =

have I told you (…)



Rhythm in Turkish German talk-in-interaction 185

418 SEM =JA;

yes

419 (---) 420 ESI ^hm=HM;

(-)

421 SEM ich /KENN da ein /FAHRlehrer;

I know a driving teacher

422 /hAb ich dir ge(.) /ZAGT?

have I told you 423 (1.1) 424 ESI [nei:n;

no

The end of Sema’s story (line 339) is followed by a longer transition phase. After a pause (line 400) and a resumptive jaja (line 401), Sema makes an effort to open the closing sequence with a somewhat consolatory final comment (you will also do the same soon, line 402). Esin answers with an exclamation (line 403–404), and after a short laughter Sema initiates a topic-proffering turn (Schegloff 2007: 169) in line 406 m=FAtih hat wieder Angefangn FAHRschule (Fatih has started again driving school). This turn differs in its rhythmic organization from its conversational context: whereas the previous turn constructional units do not display perceivable rhythm, this one is organized in three short and distinct rhythmic units, with prominent syllables occurring in approximately equal time intervals (see also above, Section 5.1.1). With the turn, Sema offers a potential next topic, which is, however, not taken up by Esin. Consequently, the topic is closed after minimal treatment (not shown in the transcript). After another transition phase with pauses and particles (lines 418–420) Sema offers another possible topic, again in a rhythmically organized turn constructional unit: ich KENN da ein FAHRlehrer; hAb ich dir ge(.)ZAGT? (I know a driving teacher, have I told you) (line 421). Apart from rhythm, Sema uses the phrase hab=ich=dir gesagt (have I told you) in both cases as an additional feature to signal the status of the turn as topic-proffering. Many more examples show how the speakers use rhythm to signal their wish to close the preceding sequence and offer a new topic as a possible next topic. Sometimes, the speakers employ other prosodic cues such as higher pitch and faster speech, or lexical cues to disconnect their turn from the previous ones. In other instances, they make use of lower pitch and narrow pitch spans to do the same thing. In all cases, however, rhythm is the most salient feature to signal discontinuity and to contextualize a turn as a new beginning.

186 Friederike Kern

5.2.2.2 Rhythm as a misplacement marker.  Often, rhythm serves as a device to organize an abrupt change of action and thus helps to make an action interpretable as different from whatever was done before. Rhythm therefore helps to accomplish an action that is somewhat unexpected in the course of the sequence. In theses cases, rhythm functions as a non-lexical “misplacement marker” (Schegloff & Sacks 1973; Schegloff 1984), and signals discontinuity.16 The following segment, the beginning of a telephone conversation, shows a case in point. The callee Ercan uses rhythm to organize a rather rapid transition from the rejection of a (tacit) accusation to the next action, i.e. greeting. Segment (4) ((ringing)) 01 ERC JA,

yeah 02 NIM ja=ICH bins;

yeah it’s me

03 ERC ey ich wollt dich auch GRA:de anrufen;

ey I was just about to call you 04 NIM SCHWÖR mal,

swear

05 ERC ja ohne (hehe);

yeah no (hehe)

06 ich SCH:WOER auf mEine mUtter;=Alter;

I swear to my mother

07 .hhh 08

everything okay 10 (-) 11 NIM ^äh=JA:;

uhm yes

12 ERC lang nich geHÖRT; wa,

it’s been a while

16. “Misplacement markers” indicate something to be not quite at its place: “This form (i.e. a misplacement marker, F.K.) can show, among other things, that they [the speakers, F.K.] know that something other than what they are about to do is in order, or that what they are about to do is ‘out of place’ ” (Schegloff 1984: 37).



Rhythm in Turkish German talk-in-interaction 187

After the caller’s (Nimo) identification (line 2), Ercan produces a turn that functions as an answer to a tacit accusation, i.e. that Ercan never calls Nimo himself (I was just about to call you, line 3). After a routinized follow-up question (line 4) and its affirmation (line 5–6), Ercan closes the side-sequence and returns to the opening sequence (lines 8–9). This turn constructional unit is detached from the previous talk by several prosodic cues typical for new beginnings: firstly, there is a long pause that is filled with a deep inbreath (line 7). Secondly and most prominently, a slow and well­audible rhythm is used (lines 8–9). Thirdly, a change of pitch register from ‘normal’ to ‘high’ and an increase in loudness serve as further signals of discontinuity. In doing this, Ercan uses several devices to constitute a “new beginning”. Ercan thus uses accent isochrony, among other phonetic cues, to organize a quick transition from a side-sequence within the opening sequence of a telephone conversation. As can be seen from the above example, rhythm serves to signal discontinuity to whatever was done before in the course of the interaction. In some cases, speakers use rhythm when they interrupt a story to ask a question, in other cases, they employ rhythm when they switch from turn-by-turn talk to produce a specific action such as, e.g. a request. In all of theses instances, rhythm functions as a non-lexical misplacement marker that signals discontinuity via prosodic “otherness” to the prior turn constructional unit.

6. Concluding remarks In this paper, rhythmic speaking has been described as a salient feature of Turkish German talk-in-interaction. First the constitution of rhythmic patterns and their formal, i.e. prosodic and syntactic, aspects were discussed. Two types of rhythmic patterning were found in the data. The first type is based on accent-isochrony, whereas the second type is built on metrical syllable structure and complementary syntactic and lexical parallelism. Moreover, it was shown how prosodic and syntactic rules of colloquial spoken German are changed in order to establish rhythmic patterns in Turkish German. Prosodic and/or syntactic deviation is one of the main characteristics of Turkish German, often resulting in the view that Turkish German is ‘bad’ or ‘wrong’ German. It has been shown, however, that speakers of Turkish German systematically alter phonological and syntactic rules of colloquial spoken German. What is more, they often do so for a purpose: the structures are changed to produce a new rhythmic pattern or to align with an existing one. Even though phonetic-phonological and grammatical differences between

188 Friederike Kern

Turkish and German are a plausible explanation for the specific linguistic modifications speakers conduct to constitute rhythmic patterns in conversation, simply to say that speakers transfer features from one of their languages to the other does not capture the whole story. Instead, speakers use features from both languages as resources to create new linguistic principles, in this case the construction of rhythmic patterns in specific sequential contexts in talk-in-interaction. In the second part of this paper, the functions of rhythm in Turkish German talk-in-interaction were discussed. Two different types of use could be distinguished. First, rhythm is used as a rhetorical device in turn construction. In story-telling, rhythm is employed to highlight particular narrative events and mark them as important for the story. Rhythm is also employed in assertions, especially after speakers’ standpoints have been doubted, or hearers have neglected to respond to a strong assessment. Rhythm is then used as a way to signal the speaker’s insistence upon his/her own standpoint. In this case, rhythm is used “to formulate ‘last positions’, which are manifestly intended to exclude any doubt about the speaker’s commitment and are displayed as not being available for further modification, attenuation or other negotiation in continued verbal interaction” (Auer, Couper-Kuhlen & Müller 1999: 170). The second function discussed is the use of rhythm (along with other prosodic cues) to signal discontinuity in the context of sequence organization and turn transition, to indicate that something new or something other than before is in progress. With rhythmic speaking, topic proffers can be contextualized as new beginnings instead of continuations. Alternatively, rhythm can be used as a prosodic misplacement marker to indicate that the produced action is somehow “out of place”, or unexpected, in the course of the interaction. To sum up, it was discovered that rhythm, as one of the main prosodic characteristics of Turkish German, is used as a style feature to accomplish activity­related conversational tasks and thus to convey specific interactive meaning.

References Abercrombie, David. 1967. Elements of General Phonetics. Edinburgh: University Press. Androutsopoulos, Jannis. 2001. From the Streets to the Screens and Back Again. On the Mediated Diffusion of Ethnolectal Patterns in Comtemporary German. LAUD, Series A, Nr. 522: Universität Essen. Auer, Peter & Elizabeth Couper-Kuhlen. 1994. Rhythmus und Tempo in konversationeller Alltagssprache. Zeitschrift für Literaturwissenschaft und Linguistik 96: 78–106. Auer, Peter & Susanne Uhmann. 1988. Silben- und akzentzählende Sprachen. Zeitschrift für Sprachwissenschaft 7: 214–259. Auer, Peter. 1990. Rhythm in telephone closings. Human Studies 13: 361–392.



Rhythm in Turkish German talk-in-interaction 189

Auer, Peter. 2003. ‘Türkenslang’: Ein jugendsprachlicher Ethnolekt des Deutschen und seine Transformationen. In Annelies Häcki-Buhofer et al. (ed.), Spracherwerb und Lebensalter. Kolloquium anlässlich des 60. Geburtstags von Harald Burger. Basel. Tübingen: Francke, 255–264. (= Basler Studien zur deutschen Sprache und Literatur). Auer, Peter, Elizabeth Couper-Kuhlen & Frank Müller. 1999. Language in Time. The Rhythm and Tempo of Spoken Interaction. Oxford: Oxford University Press. Couper-Kuhlen, Elizabeth & Margret Selting. 1996. Towards an interactional perspective on prosody and a prosodic perspective on interaction. In Elizabeth Couper-Kuhlen & Margret Selting (ed.), Prosody in Conversation. Interactional Studies. Cambridge University Press, 11–56. Couper-Kuhlen, Elizabeth. 1993. English Speech Rhythm. Amsterdam: Benjamins. Couper-Kuhlen, Elizabeth. 2004. Prosody and Sequence Organisation: The case of new beginnings. In Elizabeth Couper-Kuhlen & Cecilia Ford (ed.), Sound Patterns in Interaction. Amsterdam: Benjamins, 335–376. Cruttenden, Alan. 1986, 1997. Intonation. Cambridge: Cambridge University Press. Dauer, Rebecca. 1983. Stress-timing and Syllable-timing Reanalysed. Journal of Phonetics 11: 51–62. Dirim, Inci & Peter Auer. 2004. Türkisch sprechen nicht nur die Türken. Über die Unschärfebeziehung zwischen Sprache und Ethnie in Deutschland. Berlin/New York: de Gruyter. Erickson, Frederick. 1992. They know all the lines: Rhythmic organization and contextualization in a conversational listing routine. In Peter Auer & Aldo di Luzio (ed.), The Contextualization of Language. Amsterdam: Benjamins, 365–398. Füglein, Rosemarie. 2000. Kanak Sprak. Eine ethnolinguistische Untersuchung eines Sprachphänomens im Deutschen. Diplomarbeit, Fakultät für Sprach- und Literaturwissenschaften der Otto-Friedrich-Universität Bamberg (unpublished). Galves, Antonio, Jesus Garcia, Denise Duarte & Charlotte Galves. 2002. Sonority as a basis for rhythmic class distinction. Speech Prosody 2002: 323–326. Grabe, Esther & Ee Ling Low. 2002. Acoustic Correlates of Rhythm Class Hypothesis. In Charles Gussenhoven & Natasha Warner (eds), Laboratory Phonology 7. Berlin etc.: de Gruyter, 515–546. Gumperz, John. 1982. Discourse Strategies. Cambridge: Cambridge University Press. Günthner. 2000. Vorwurfsaktivitäten in der Alltagsinteraktion. Tübingen: Niemeyer. Halliday, Michael. 1985. An Introduction to Functional Grammar. London: Eward Arnold. Imrie, Alexandra. 2008. Rhythm as a Resource to Generate Phonetic and Phonological Coherence in Lists. York Papers in Linguistics Series 2: 23–47. http://www.york.ac.uk/depts/lang/ ypl/ypl2issue9/YPL9_02_Imrie.pdf. 15.3.2010). Jefferson, Gail. 1984. On Stepwise transition from talk about a trouble to inappropriately next-positioned matters. In J. Maxwell Atkinson & John Heritage (ed.), Structures of Social Action. Studies in Conversation Analysis. Cambridge: Cambridge University Press, 191–222. Keim, Inken. 2001. Die Powergirls. Aspekte des kommunikativen Stils einer Migranten­gruppe aus Mannheim. In Eva Jakobs & Anneley Rothkegel (ed.), Perspektiven auf Stil. Tübingen: Niemeyer, 375–400. Keim, Inken. 2004. Kommunikative Praktiken in türkischstämmigen Kinder- und Jugend­ gruppen in Mannheim. Deutsche Sprache 32(2): 198–226. Kern, Friederike. 2008. Das Zusammenspiel von Prosodie und Syntax am Beispiel von Türkendeutsch. Habilitations­schrift, Universität Potsdam.

190 Friederike Kern

Kern, Friederike & Margret Selting. 2006a. Einheitenkonstruktion im Türkendeutschen: Grammatische und prosodische Aspekt. Zeitschrift für Sprachwissenschaft 25(2): 239–272. Kern, Friederike & Selting, Margret. 2006b. Konstruktionen mit Nachstellungen im Türkendeutschen. In Arnulf Deppermann et al. (eds), Grammatik und Interaktion. Verlag für Gesprächsforschung, 319–347. Kern, Friederike & Yazgül Simsek. 2006. Türkendeutsch: Aspekte von Einheitenbildung und Rezeptionsverhalten. In Dieter Wolff (ed.), Mehrsprachige Individuen – vielsprachige Gesellschaften. Reihe forum Angewandte Linguistik 47: 101–119. Kornfilt, Jaklin. 1997. Turkish. London/New York: Routledge. Kotsinas, Ulla-Britt. 1998. Language contact in Rinkeby, an immigrant suburb. In J. Androu­t­ sopoulos & Arno Scholz (eds), Jugendsprache – langue des jeunes – youth language. Linguistische und soziologische Perspektiven. Frankfurt/Main: Lang, 125–148. Lang, Ewald. 1977. Die Semanik der koordinativen Verküpfung. (= Studia grammatica 14). Berlin: Akademie-Verlag. Laver, John. 1994. Principles of Phonetics. Cambridge: Cambridge University Press. Low, Ee Ling, Esther Grabe & Francis Nolan. 2000. Quantitative Characterization of Speech Rhythm: Syllable-Timing in Singapore English. Language and Speech 43(4): 344–401. Pike, Kenneth. [1945] 1963. The Intonation of American English. Ann Arbor/Michigan: University of Michigan Publications. Quasthoff, Uta. 1980. Erzählen in Gesprächen: Linguistische Untersuchungen zu Strukturen und Funktionen am Beispiel einer Kommunikationsform des Alltags. Tübingen: Narr. Ramus, Franck, Marina Nespor & Jacques Mehler. 1999. Correlates of linguistic rhythm in the speech signal. Cognition 73: 265–292. Schegloff, Emanuel. 2007. Sequence organization in interaction. A primer in conversation analysis. Cambridge: Cambridge University Press. Schegloff, Emanuel & Harvey Sacks. 1973. Opening up closings. Semiotica 7: 289–327. Schegloff, Emanuel. 1984. On some questions and ambiguities in conversation. In J. Maxwell Atkinson & John Heritage (eds), Structures of Social Action. Studies in Conversation Analysis. Cambridge: Cambridge University Press, 28–52. Schwitalla, Johannes. 2001. Gesprochenes Deutsch. Berlin: Erich Schmidt Verlag. Selkirk, Elizabeth. 1984. Phonology and Syntax: The Relation between Sound and Structure. Cambridge/Massachusetts: MIT Press. Selting, Margret. 1999. Communicative Style. In Jef Verschueren, Jan-Ola Östman, Jan Blommaert & Chris Bulcaen (eds), Handbook of Pragmatics 1999. Amsterdam: Benjamins, 1–24. Selting, Margret. 2007. Interactional Stylistics and Style as a Contextualization Cue. In Ulla Fix & Andreas Gardt (ed.), Rhetorik und Stilistik. Ein internationales Handbuch. Berlin etc.: de Gruyter. Selting, Margret, Peter Auer et al. 2009. Gesprächsanalytisches Transkriptionssystem 2 (GAT 2). Linguistische Berichte 173: 91–122. Gesprächsforschung. Online-Zeitschrift zur verbalen Interaktion. http://www.gespraechsforschung-ozs.de/heft2009/heft2009.htm. 15.3.2010). Uhmann, Susanne. 1992. Contextualizing relevance: On some forms and functions of speech rate changes in everyday conversations. In Peter Auer & Aldo di Luzio (eds), The contextualization of language. Amsterdam: Benjamins, 297–336.

Constructions with Turkish şey and its German equivalent dings in Turkish-German conversations şey and dings in Turkish-German Yazgül Şimşek

University of Potsdam

An example of Turkish-German language contact and the employment of resources of Turkish in the mixed language is a construction with the dummy word şey (‘thing’). Şey occupies the position of a noun, an adverb or a verb (another verb plus the light verb yapmak ‘to do/make’) within a phrase. Following completion of the syntactic unit, which in Turkish is marked by the finite verb, the dummy word is replaced by the correct noun. The main objective of my study is to describe a similar construction of Turkish-German produced with the German equivalent dings (meaning ‘thing’). The prosodic features which I consider a distinctive cue will show how the two languages interplay on more than one structural level.

1. Introduction The aim of this study is to show the combined use of syntactic and prosodic resources of Turkish and German to form constructions highlighting and focusing on the new information contained in a speech unit. The issues referred to in studies on the language use of adolescents with a Turkish ethnic background in urban areas such as Berlin are mostly concerned with a description of the social structures of these environments and the multiethnic interaction styles used in everyday conversation to signal group and social identity. Another main thrust of research in this field is to describe rules and regularities of language contact, mainly code-switching phenomena (cf. Hinnenkamp 2005). Other research (cf. Keim 2007; Wiese 2006 and Kern 2008) shows that the relation between the two languages as it appears in the German dominant speech style goes far beyond intra- or extrasentential code-switching. Structures of the

192 Yazgül Şimşek

language other than the German spoken by adolescents appear not only to reflect a choice intended to signal identity and social boundaries but are also mixed because they are available as more effective resources to be deployed in everyday interaction between people sharing those resources. For example, both source languages of Turkish-German employ specific strategies and techniques to convey information and signal the relevance of this information within a given speech activity. In this article I will describe one such structure of spoken Turkish for the purpose of signalling the semantic focus of an utterance. These constructions are formed by one host sentence containing the lexeme şey [∫ej] (‘thing’) in the pre-verbal position that is normally reserved for new information. The lexeme is then replaced by a full lexical expression in an additional construction in post-verbal position. In Turkish-German a similar structure is observable which makes use of the German lexeme dings [dɪnɡs] (‘thing’). After a few preliminary thoughts on the methodology and database (Sections 2 and 3), I will start by briefly introducing the technique of pre and postpositioning in Turkish-German (Section 4). To follow up on the question of how the two languages interplay in conveying information, I will then analyse examples of constructions with dings and şey in both the source languages and Turkish-German (Section 5).

2. Methodology The methodology is based on (i) the theoretical approach of interactional linguistics (cf. Couper-Kuhlen & Selting 2001), since the concepts of interactional linguistics and interactional prosody research allow the combined study of interactional and linguistic phenomena, and (ii) a contrastive analysis of the structures of the two language systems within natural speech activities. The aim of this approach is to show if and how different languages cope with similar communicative situations in the same or different ways. For the following analysis prosodic phenomena are particularly important. These are all by definition linguistic structures which only become interpretable and thus functional at the suprasegmental level. Because my analysis will mainly be concerned with a description of pitch contours, I shall use the term intonation to refer to pitch movements, including the characteristic pitch movement within and at the end of a turn constructional unit (TCU). To show the relevant prosodic . TCUs are speech parts built and recognizable as a unit by grammatical and prosodic resources. They are regarded as the result of the interplay of syntactic, lexico-semantic, pragmatic, activity-type specific and prosodic devices (cf. Selting 1998).



Constructions with Turkish şey and its German equivalent dings 193

phenomena, the samples of data are transcribed according to the conventions of GAT (cf. Selting et al. 1998). The use of Turkish language elements is indicated in the transcriptions by means of a different font.

3. Database The database consists of two different corpora of spoken language: the first corpus is the data corpus of the research project Turkish-German (Türkendeutsch) in an interactional-linguistic perspective (2004–2006) funded by the German Science Foundation (DFG). The 16 to 22-year-old bilingual speakers who are members of the major group of migrants living in Berlin. To have a basis for comparison with the standard colloquial variety of German spoken in Berlin, a similar amount of data was provided by monolingual German speakers of the same age. I shall also examine a second corpus containing spontaneous speech data taken from colloquial spoken Turkish interviews and biographic narrations (24 hours) with adult speakers living in Turkey.

4. Turkish-German as a specific speech style in talk-in-interaction As pointed out by Selting (this volume), our analysis argues strongly against the idea of viewing Turkish-German as a deficient form of second language acquisition. Despite its specific social context, the linguistic structures of Turkish-German and the ability of its speakers to shift between an ethno style of Turkish-German (cf. Auer 2003) and a regionalized variety of standard colloquial German indicate a high level of adaptation to a language contact situation. The units containing the lexemes dings or şey show how a variety such as Turkish-German can make creative use of the resources of two languages at more than one structural level. Both source language systems of Turkish-German, namely Turkish and German, differ from each other syntactically above all in the forms used for syntactic . I wish to convey special thanks to Christoph Schroeder for sharing the data with me. . The term Türkendeutsch was developed exclusively to refer to the ethnic background. . The concept of Turkish-German as an ‘ethnic speech style’ as developed in the research project at the University of Potsdam (cf. Kern & Selting 2006a and 2006b and Kern & Şimşek 2006) takes into account that this form is only one out of many accessible linguistic resources. The speech units considered to be of ‘specifically Turkish-German style’ always appear as choices and functional solutions to communicative tasks.

194 Yazgül Şimşek

subordination as well as in terms of word order. In terms of information structure, Turkish follows the principle of linearity: Information is packaged in linear order and conveyed successively (cf. Johanson 1971). By contrast, German contains more complex structures such as the verb bracket (see Rehbein 1995 for a contrastive analysis of finite verb position and word order in Turkish and German). Furthermore, in spoken German prosodic means such as the placement of primary accents within a speech unit are crucial in signalling the semantic focus of a TCU. Similarly to its source languages, Turkish-German has developed constructions with pre and postpositions which use syntax and prosody to act as linguistic cues for units conveying information (cf. Kern & Selting 2006a and b).

5. German dings and Turkish şey in conversation In Turkish and German the words şey and ding are nominal elements with a similar meaning. The Duden (1996) defines dings as ‘a subject, object or organisation the speaker is not able to name, does not remember or does not know, or does not choose to say’ (cf. Duden 1996: 346). It therefore becomes clear that the lexeme dings represents a kind of ‘fill-in’ lexical item. In a similar way the Turkish lexeme şey is described as a nominal which is ‘used in unclear meaning as a replacement of names, of a material, an object, a word/utterance, a state of affairs etc.’ (cf. TDK 1998: 2089). However, these definitions make no mention of the different functions the lexemes can fulfil in natural conversation and their high level of dependence on the respective context of use. As is also the case with other structural elements, tag-questions for instance, Turkish-German makes more use of dings than standard colloquial German does, as illustrated in Table 1. The results show that Turkish-German uses the form dings at a rate that is almost four times higher than that of standard colloquial German. Those speakers of Turkish-German who already show a high tendency to integrate resources from Turkish into their speech activities also use the functional equivalent şey: for example there were 56 instances of şey in recordings (4 hours) with a high tendency Table 1.  Quantity of dings in German and in Turkish-German Use in five hours of talk

German speakers

Turkish-German speakers

dings (thing) ding (thing) dinger (the things) dingsda (the thing there)

25 14  3  1

160    6    4   1



Constructions with Turkish şey and its German equivalent dings 195

towards Turkish language and only 2 instances of şey in recordings (4 hours) with virtually no use of Turkish language. In the following sections of this paper I wish to go beyond a quantitative description by using samples of natural conversation to illustrate how and for what communicative purposes both lexemes are used in the source languages. In the final section of the analysis (Section 5.3) I will then concentrate on the use of şey in specific formats of Turkish-German in order to compare this with the findings of Kern & Selting (2006b) who have already analysed the constructions with dings in Turkish-German. 5.1

Constructions with dings in German

The data from monolingual German speakers contains various forms of dings including ding/dings/dingsda/dinger (thing/thing/that thing there/things). Ding/ dings can be combined with a definite or indefinite article or can bear gender assignment. In the most cases ding/dings occupies the position of a noun integrated within the syntactic structure. Four more or less differing formats were found in the data: I: repair in question-answer sequences II: announcement of a new topic III: a flexible repair format of self-initiated self-repair IV. not replaced dings as a semantic joker In format I the lexeme ding/dings stands for an expression the speaker is querying. The co-participant provides the problematic expression as an answer. In format II dings, normally prosodically separated by its different speech rate, functions as a kind of projection before a new topic is introduced (cf. Couper-Kuhlen 2004: for a definition of new beginnings). The other formats, III and IV, are exemplified in the Excerpts (1) and (2). III: (1) dings in GER (dtm02_Ste1, dings8) 01 Ste: ich will nämlich mit n KUMpe:l- (-) I want well together with a friend 02 .hh also mit dem da in sachsen ANhalt-= I mean with the one from sachsen anhalt=

. For a definition of such elements of speech as ‘misplacement markers’ see Schegloff (1984: 37).

196 Yazgül Şimşek

03 =will ik [((blows his nose)) I want 04 And: [ha; -> 05 Ste: dings hier (--) so=n kOmisches FAHRzeug zusammenbauen; thing here put together/construct somekind of weard vehicle 06 da ist quasi so wie ein (STRAND/STAND) buggy, it is something like a (beach/standing) buggy

Dings is an integral part of the syntactically complex unit the speaker is already starting to construct in line 03. Because of the interruption of his verbalisation the speaker is not able to continue immediately with the intended explanation of what he is planning to construct with a friend begun in line 03. Therefore dings occurs in line 05 as the signal of uncertain verbalisation. The discourse marker “hier” and the pause after dings are additional cues pointing to a self-initiated self-repair. In Example (2) the utterance in question contains dings playing the role of a semantic joker (cf. Kern & Selting 2006b: 336), occupying the position of the new information without making it explicit. IV: (2) ding/dings in GER (dtm02_Ste5, ding12) 01 Ste: und dann machen wir dat ähm machen wir uns aus wann 02 ich denn mal runter komm(e) mit=dem TRÄcker; (-) and then we will make it agree upon a time where I occassionally come down with the tractor 03 Kar: JA; yes -> 04 Ste: sEtzt du dich HIN auf das ding(s);= you will sit down on the thing 05 =und dann zIeh ich (dich/die) HOCH; (-) and then I will pull you/them up 06 Kar: JA; yes

In this excerpt there is no hint to a word search or a kind of a reference problem. The speaker simply uses ding/dings as a common expression without a replacement. The referent may in such instances be clear to the participants because of their shared knowledge.



Constructions with Turkish şey and its German equivalent dings 197

The most important conclusion which can be drawn from the analysis of the monolingual German data is that dings is most commonly used in repair sequences, mainly with prosodic cues referring to a kind of verbalisation problem (hesitation markers and pauses). There are no instances of use pointing towards a strategy employing dings as a highlighting technique. The samples of Turkish-German that I refer to in the next section contain additional forms and functions to those described for the Berlin variety of standard colloquial German. 5.2

Constructions with dings in Turkish-German

As mentioned above, standard colloquial German has a technique of postpositioning constituents after their host sentences (cf. Auer 1996; Selting 1993 and 1994) to signal the semantic focus. These sentences take the primary accent on one constituent while the postpositioned element carries another primary accent, a secondary accent or none at all (cf. Uhmann 1997). In comparison with these structures, the constructions with dings in Turkish-German always have a primary accent on dings in the host sentence and on the replacement of dings. In their analysis of these constructions, Kern & Selting (2006b: 330f.) found the following five formats, variable in syntax and sequential organization but not in accent assignment: (3) proceeding structure with dings + postpositioned replacement a. dings in the position of a noun + nominal replacement -> 01 Eli: was hatn denn deutsch mit (.) DINGS zu tun; what do german(skills) have to do with thing -> 02 HANDschrift; handwriting

b. dings in the position of a noun + replacement by a phrase -> 01 Sem: oder soll ich zum DINGS gehen; or shall I go to thing -> 02 zum: HANdyladen; to a mobile phone shop

c. dings in the position of an adverb + replacement by an adverbial compound -> 01 Gül: mir kam das VOLL an’ VOLL so DINGS vor;=ja, to me it seemed quite quite so thing; yes 02 (1.3) -> 03 KOmisch vor; weird

198 Yazgül Şimşek

d. dings in the position of a verb + replacement by a verb -> 01 Esi: du muss immer DINGS machen; you have to do thing -> 02 SPEIchern; (--) save

e. dings and postpositioned replacement as a repair format -> 01 Sem: und dann sind ( ) und aslı zu DINGS gegangen; and then ( )and aslı went to thing -> 02 ä:h: SAG schon; say it now 03 (1.2) 04 Esi: he he he to the center

The first four structures are used as unique formats to focus new information. The fifth sample represents a variable repair format, where dings is replaced by the correct word. Kern & Selting analyse the use of dings in the host sentences as a kind of ‘turn-holding device’ which makes a continuation and a replacement strongly relevant because without a replacement the sentence is not semantically complete. The focus expression is thereby postpositioned and highlighted as the most important information by means of word order and accent assignment. These constructions also exemplify one very important structural difference between the source languages of Turkish-German: while German incorporates accentuation as a tool to signal semantic focus, Turkish seems to use other linguistic means to achieve similar goals. In Turkish-German, the constructions with dings seem to point to a combined use of prosodic and syntactic structures from both languages. In the following sections I will concentrate on the syntactic and sequential structures of constructions with şey. I will pursue further the idea that the extensive use of constructions with dings in Turkish-German may really be an influence of Turkish.

. The accent assignment rules of Turkish, accentuating at word level and as a rule on the last syllable (for rules of accentuation in colloquial spoken Turkish cf. for example Swift 1962 and Levi 2005), may be transferred to the utterance level in Turkish-German. This entails the application of the different accent assignment rules of both languages in Turkish-German to produce the accent on the very last word of the utterance and postpone it in the form of a postposition. This way the new information (the semantic focus) is moved from the host sentence into the postposition where it represents the semantic focus of the entire construction. . Mainly word order variations cf. Şimşek & Schroeder (2010).



5.3

Constructions with Turkish şey and its German equivalent dings 199

Constructions with şey in Turkish

As in the case of dings in German, the lexeme şey (originating from Arabic) in Turkish has no specific semantic property. Depending on the context and in combination with the numeral bir (one) it can mean nothing/something. Morphologically and syntactically, şey is treated as a nominal and can be marked with case and possessive suffixes. In such instances it takes the place of an obligatory constituent. If the structure is one of postpositioning equivalent to those with the German dings as described above, the replacement must also be marked with equal suffixes. This specific word order strategy of Turkish places şey in the pre-verbal focus position where normally the rheme, the new information of the sentence, is to be found. Erguvanlı-Taylan (1984: 60) argues that şey is the most prominent constituent of the host sentence while the postpositioned replacing element cannot be prominent in terms of prosody. By contrast, the postposition is neither accentuated nor prominent in any other form. Analysing specific constructions of colloquial spoken Turkish, Schroeder (2002) comes to the conclusion that these constructions with şey in focus position are the result of a tendency to convey information in a linear order. By means of this “linear order principle”, colloquial spoken Turkish avoids complex subordinating structures, mainly structures equivalent to the relative constructions of German, while keeping the preferred SOV word order. This way the information can be verbalised in a linear order in contrast to those structures of spoken language where the word order is reversed. The şey construction allows the speaker to withhold the new information and move the element carrying this information to the right while setting the frame for its conveyance with şey in the position where it must normally occur. The postpositioned replacement fills that frame set by means of the host sentence with şey. Following up on this technique of ‘splitting up information’, I have found a fair amount of constructions in the spontaneous speech data of colloquial spoken

. Colloquial spoken Turkish has developed a specific verb-first construction type in spoken discourse although it is an SOV language (cf. Banguoĝlu 1959: 534 for a description of “devrik cümle” (the reversed sentence), cf. also forms of topicalisation in Erguvanlı-Taylan 1984: 34ff.). For further discussions on word order and information structure in Turkish see also Göksel (1994). . Kern & Selting use the descriptive term “upshot-first-detail-later” to refer to the functions of the dings constructions in Turkish-German discourse (see Kim 2001 for the development of the description “upshot-first-detail-later” for Korean as another language of agglutinative typology conveying all relevant information before the finite verb).

200 Yazgül Şimşek

Turkish that are similar to those postpositioning structures of Turkish-German using German dings (cf. Example (3a–d) above): (4) Şey-constructions in colloquial spoken Turkish (TR-Int.-C.S) a. Șeyde oturdum- kirada; I stayed at thing, by paying rent b. Iki sene askerlik yaptım şeyde- Giresun’da; Two years I did my military duty in thing, in Giresun. c. Șey harbı çıktı- yunan harbı çıktı; The thing war came up, the Greek-War came up.

The first two constructions (a) and (b) above are short units of conversation using the syntactic structure with the postpositioned replacement of şey being marked with the same case suffixes. In the third construction (c), however, the structure seems to be a more complex and hence different one although şey is not suffixed. This complexity is due to the fact that şey is used for only the first part of a compound yunan harbi (Greek War). This means that both elements of the compound are to be replaced in the postposition.10 A classification of şey as a discourse marker often found in Turkish linguistic literature refers to a second function. Özbek (1998 and 2000), describes şey as a discourse marker used to ‘introduce new topics’ (“topic raising”) and signal “thought organising”. Even putting aside the problems of using such a broad and unspecific term as discourse marker, Özbek‘s descriptions leave questions about the formal characteristics of şey in these different functions unanswered, although the terms seem mainly refer to şey used in repair sequences. The most important question about şey used in repair sequences is: what are the contributions of şey to the speech activity and context? If there are no specific features of this kind of şey then why does the speaker not choose to use a simple hesitation marker if he/she simply wishes to gain enough time for verbalisation? Before I discuss these questions in Section 5.4 and analyse şey as a syntactic focusing strategy and signal used in repair sequences of Turkish-German, I would like to present two examples of use in the colloquial spoken Turkish language and explore the formal characteristics and functions in discourse organization in addition to the descriptions provided by Özbek (1998 and 2000) in order to enable subsequent comparison with the formal characteristics of Turkish-German.

10. The repetition of the verb as carried out in this case is not obligatory but this way the whole structure is repeated and potential misunderstanding becomes impossible.



Constructions with Turkish şey and its German equivalent dings 201

Sequential and prosodic characteristics of şey as a focusing strategy and flexible repair format in colloquial spoken Turkish In Example (5), where it is integrated into the syntactic structure of a host sentence and replaced by a postpositioned element, şey represents an example of the syntactic focusing strategy of colloquial spoken Turkish: (5) şey and focusing by means of postpositioned replacement (TR-Asi-1a) 01 Asi: ((laughing)) resmini de gösterdim;

I showed also a picture of (her) 02 sema bayıldı tabii;(--) sema was naturally thrilled 03 öbür kızım; my other daughter 04 Int: evet; yes 05 Asi: ondan sonra neyse; then however 06 nerde şimdi; where is (she) now -> 07 şeyde kalıyo-= (she) is staying at thing -> 08 =yuvada; (---) at the children’s home 09 geldik aldık bunu, we came and took her back

In line 07 the structure in question shows no prosodic signs of a verbalisation problem. There are no pauses, hesitation signals or other formal cues indicating a difficulty on the part of the speaker to find the right words. Şey itself has a falling-rising contour, the typical pitch movement it has when it is suffixed (here it carries the locative case marking) and thus syntactically integrated. The following verb is attached to the same F0 level without any break. The final pitch of the host sentence is perceptible as level-intending. The postpositioned replacement, on the other hand, displays a continuous fall in pitch, a characteristic of such a structure that results in its perception as an immediately attached but separate prosodic unit. Functionally, the focus of the whole utterance is on new information. With the şey construction the speaker is returning to the verbalisation of a former dialogue she broke off in line 05. This means that the information contained in units

202 Yazgül Şimşek

07 and 08 is important in terms of the ongoing narration and therefore presented by means of a highlighting structure. The repair sequence in Example (6) reveals a verbalisation problem even before the use of şey: (6) şey in a repair sequence (TR-Mül-1a) 01 Mül: arabanın içinde sekiz kişi var; (2,1)

there are eight people in the car ((a friend of Mül enters the room)) -> 02 o tam e:h (---) eh şeye eh (.) ne ne yoku exactly there thing what (hill) 03 hac=ha=hacı m:=m ney, (--) hac=ha=hacı what 04 yes 06 Int: ha:;=evet evet; o:h yes yes

In line 02 the speaker Mül, engaged in a narrating activity, is searching for a name she cannot remember. The long pause in line 01 is not necessarily due to an upcoming problem but to the nonverbal actions taking place during the ongoing narration because another person is entering the room. The first indication of a problem is the hesitation marker in line 02, followed by a pause which also acts as an implicit device for repair. After a second pause the speaker produces şey marked with the suffix for dative case. It can be argued that şey is intended as the constituent of a syntactic structure, since Mül is searching for the name of a place where the narrated event took place. Following omission of the probably intended structure with şey, Mül makes her search more explicit by using the wh-element ne (‘what’). The acoustic characteristics of such repair units show clearly visible differences compared with the sample shown in Example (5) above. The hesitation signals before and after şey are produced with a very low pitch range, as is the whole unit. After the first hesitation signal there is an upstep. This triggers the perception that a new speech unit is beginning. The pitch movement on şey is a slightly but steadily falling one which contrasts with the previous example where şey has a falling-rising contour. In this prosodic shape, şey is not projecting a continuation and is not perceived as a highlighted element, but instead as a signal of a verbalisation problem due to the surrounding hesitation markers with a very small pitch range.

Constructions with Turkish şey and its German equivalent dings 203



The examples demonstrate that the following formal characteristics must be taken into account when verifying the functions of şey in colloquial spoken Turkish natural discourse: Table 2.  şey constructions in colloquial spoken Turkish Position of şey

Prosodic shape

Function

Icomplete syntactic structure, şey can take case + poss. suffixes, hesitation markers

Pauses No significant pitch contour, falling  pitch, low pitch range, lengthening

Signalling of repair

Pre-verbal position in host  sentence + postpositioned replacement şey can take case + poss. suffixes

şey: Significant falling-rising pitch contour, high pitch range, no lengthening Replacement: Separate intonation unit immediately  attached, clustered, falling final pitch Entire utterance: şey at the highest pitch peak, no prosodic break between postposition and host  utterance

Focusing strategy: highlighting new information

There is an effective interplay of syntax and prosody in the use of the lexeme şey that results in a unique linguistic source to be embedded into various sequential environments. Having clarified the two contexts in which şey is used in colloquial spoken Turkish, in the next section I wish to analyse similar samples of Turkish-­German. This analysis may reveal important information that helps to assess the role of dings constructions in Turkish-German and the role of both languages in the creation of such new structures. 5.4

Constructions with şey in Turkish-German

The main concern of analysis in this section will be to work out the differences between those structures where şey constructions serve to ensure information delivery and those where it is placed in repair sequences and accordingly signals some kind of verbalisation problem. To this end I wish to illustrate the uniqueness of the focusing strategy in providing speakers of Turkish-German with a frame in which to insert their constructions in German using the German lexeme dings. The deployment of focusing constructions with Turkish şey is as common in Turkish-German as it is in spoken language of Turkey. According to my data observations speakers of Turkish-German use şey constructions according to the

204 Yazgül Şimşek

language resources of colloquial spoken Turkish. A mixed use of lexical resources from German and Turkish does not occur in these constructions so that speakers make use of dings when their utterances are mainly produced in German and use şey only within the scope of Turkish utterances.

Şey as a focusing strategy I will start by analysing examples of the focusing strategy, beginning with Example (7) where şey occupies the position of a noun replaced by a postpositioned noun. (7) Şey in Turkish-German (tkdtw08_Esi1, tk-dings31) 01 Esi: künye alıcam; .hhh (--)

I am going to buy a bracelet 02 başka ((clears her throat)) (1.2) other than that/in addition to that/something more -> 03 -= actually I wanted also to let make thing you know -> 04 =kaset; (---) a cassette 05 Mer: ne kase[ti, what kind of cassette 06 Esi: [mit so ausgesuchten LIEdern falan; with chosen songs something like that

In terms of word order the construction is similar to colloquial spoken Turkish – here the pre-verbal focus position is shared with the focus particle da (‘also’). If the replacement consists only of a substantive element as is the case here then the postpositions are clustered structures forming a separate intonational unit. The host structure has a level final pitch. Şey is perceptually more prominent than the following verb because it is produced at a higher F0 level while the following verb and postpositioned replacement have a constant pitch movement with no peaks. Constructions of this format always have falling final pitch on the postposition. The entire construction takes on all the characteristics that are typical of the function as a technique for highlighting new information and signalling a high degree of relevance. The speaker Esin is reporting to her friend about her plans to buy a present for her boyfriend on Valentine’s Day. In line 01 she verbalises the first element on her list of things to do or buy. In line 02 she attempts to continue the list but breaks off and clears her throat. The fact that Esin continues with a higher speech rate in line 03 indicates that the break is not caused by a word search or other verbalisation problem. The need to focus on new information



Constructions with Turkish şey and its German equivalent dings 205

with the construction she uses here with şey seems to be more important for the ongoing discussion. In summary, this means that the difference between the focusing and repair functions comprises a specific interplay of syntax and prosody. If there are no prosodic cues indicating a verbalisation problem and in addition a syntactically postpositioned replacement of şey takes place that normally consists of a nominal carrying the same case or possessive suffixes then the entire utterance with şey comprises a focusing strategy.

Şey in a specific Turkish-German continuation-projecting format In Example (8) the whole sequence can be considered to be a typical strategy of Turkish-German that makes use of short prosodic units in talk (cf. Kern & Selting 2006a). In addition, the speaker switches into German. Both strategies present the dialogue in a more highlighting and dramatic manner which in this case also underlines the importance of the topic. (8) şey in Turkish-German (tkdtw02_Mel1, tk-dings4) 01 Mel: bei uns ist überKRASS alter; with us it is superwicked mate 02 (2.8) 03 ((sighs)) -> 04 .hh bi şeyde, (--) at thing -> 05 geçen yakalandım, (--) newly I have been caught -> 06 BE vau ge, ((Transportation Organisation of Berlin)) -> 07 wurd isch erWISCHT; (-) I got caught 08 Ela: (ä) hast keine FAHRkarte oder was; you don’t have a ticket or what

In view of the formal differences between this case and the previous example, it is noticeable that the very short unit is the host structure. Bi şeyde forms a separate intonational unit with a rising final pitch. The speech parts in lines 04–07 form a coherent structure: each unit with the same prosody in lines 04–06 has a rising final pitch. This strategy uses prosodic means to present all the short units as belonging together. Dividing the information into small pieces makes the topic of conversation dramatic.

206 Yazgül Şimşek

Şey itself has a falling-rising pitch contour. Although the structure is syntactically not explicit, this contour is similar to the intonation of şey in focusing constructions found in the data of colloquial spoken Turkish. The şey unit signals the beginning of a new topic in particular in view of the fact that its morphological form bears the locative case suffix. The projection is successful because not even the pause after şey causes the speaker to loose her turn. The prosodic structure of short units and the perceptually very prominent contour of şey may make code-switching easier in such instances because the odds of loosing one’s turn are small. The entire construction might be seen as a repair although interestingly the replaced element is not şey itself: the entire units in lines 05 and 06 are replaced by lines 06 and 07. I therefore consider the example as a reduplication of semantically equivalent structures and not as an instance of repair. The language boundaries are identical to the prosodic boundaries11 and the switch is unproblematic, thereby sparing the need for repair. Some functional characteristics of şey seem to be comparable with those of dings in cases where dings is produced with a level or rising final pitch contour and signals the beginning of a new topic. According to my findings from the data, both lexemes can be used to project continuation but in this case şey more strongly projects a syntactic continuation. Dings, on the other hand, functions more overtly as a turn-holding device with a following unit that is not syntactically bound.

Şey as a focusing strategy embedded in a repair sequence In Example (9) a repair sequence can be assumed. However, a more detailed analysis reveals that the function of şey is to package new information in a construction that is embedded into a specific sequential environment with a specific communicative purpose. (9) şey in Turkish-German (tkdt09_Nec3a, tk-dings68) 01 Nec: baksana,

-> 02 03 -> 04

look , (.) (from/for) thing nasıl diyim; (.) how shall I say it o biraz şey böyle; he is somehow thing so

11. For the preference of bilingual Turkish-German speakers to switch between language resources at specific syntactic boundaries cf. for instance Kallmeyer et al. (2002).

Constructions with Turkish şey and its German equivalent dings 207



-> 05 06

Lockerloose

şakalaşıyo-

he is making fun

The frequent switches between Turkish and German are not the only formal devices indicating a formulation problem. The first switch in line 02 forms an intonational unit with final rising pitch and a falling-rising contour on the accentuated dings, a contour which can also be found on şey (see Example (8) above) and forming a short prosodic unit that projects continuation. The utterance produced in Turkish in line 03 is the first explicit indication of a formulation problem. This means that the units in lines 02 and 03 form a repair format together. The German lexeme dings takes on a special role within this structure as an ‘announcing’ but not a repair or replacement element. The şey structure in line 04 is also not a repair element. It must be classified as a focusing strategy, with the şey construction taking on specific prosodic features. Neither of the units with dings and şey contains prosodic breaks, pauses or hesitation signals. Just as in the first examples of this section, the pitch peak is on şey before falling during the remainder of the intonational unit. Syntactically şey occupies the position of an adverb and is replaced by an adverbial qualification in the following unit. In this case the replacement takes on a complex prosodic structure. The replacement is not only an adverbial qualification forming a postpositioned unit but there is an additional unit in line 06 which could also be interpreted as a replacement of şey because the units 05 and 06 are presented as equivalent structures by means of prosody (with no changes in the speech rate and no boundary signals). The function of the whole şey structure is to signal the end of the verbalisation problem and focus on the new information. The tendencies of Turkish-German to create short prosodic units and switch between language systems accordingly prove to be useful techniques. The following samples deal with the repair format and in particular the formal prosodic indications of such uses of şey. The Turkish-German data revealed two groups of repair sequences, classified as explicit and non-explicit repair formats.

Şey as a flexible repair format: Non-explicit repair Example (10) below illustrates the non-explicit repair structure. ‘Non-explicit’ means that there are no lexical idioms referring to a speaker’s problem before, during or after the repaired unit.

208 Yazgül Şimşek

(10) şey in Turkish-German (tkdtw02_Mel1, tk-dings8) 01 Mel: aber wir essen für zwei äh FÜNF euro oder so; but we eat for two äh five euros and alike 02 Ela: hm; 03 Mel: on euro gidiyo yani; (-) ((clicks her tongue)) (--) hence ten euros are gone -> 04 şey (ben)’ ↑iflaza (giricem/gidicem); thing (I) am (going into/going to enter) bankruptcy 05 Ela: {va de} du verdIenst auch GELD dabei; wa, you are getting money during this aren’t you

The speaker intends to make a résumé and says in line 04 that the money she spends on lunch is going to draw her into bankruptcy. This utterance shows none of the formal features described for şey in a repair format of colloquial spoken Turkish and in particular no pauses and hesitation signals. The only prosodic cue is a glottal stop produced after şey (ben). This causes the speaker to start again with the rest of the utterance at a higher F0 level. This way a clearly perceptible prosodic break separates the structure into two parts, the first part being repaired with the rest of the structure.

Şey as a flexible repair format: Explicit repair The next sequence illustrates the format of an explicit repair. In cases such as the following example the şey unit contains lexical and prosodic cues both before and after the use of şey that point to a problematic utterance in need of repair. (11) şey in Turkish-German (tkdtw02_Mel1, tk-dings5) 01 Ela: ha bist du HINgegangen,

yes did you go there 02 Mel: ja isch bin (.) HEUte hingegangen, (---) yes I went there today -> 03 bi de arkadaşımın (.) ä:h kaĝı äh şeyi äh (---) and also the pa of my friend äh thing äh -> 04 diese:s DINGS (.) was HEISST das; this thing what is it called 05 Ela: be VAU ge ausweis;= BVG identity card 06 Mel: =ha=ha; yes



Constructions with Turkish şey and its German equivalent dings 209

In this case there is a bundle of devices acting as typical markers of repair. The answer in line 03 projects continuation by means of its rising final pitch. The pause after this utterance is the first indication of a problem and at least points towards hesitation. The switch into Turkish in the next line also serves as such an indication. In this line the micro-pause after “bi de arkadaşımın” is the first sign of a search for words, followed by a hesitation marker and numerous breakups including şey surrounded by hesitation markers. In line 04 where the speaker asks her co participant for help the failed verbalisation becomes a self-initiated other-repair.12 It is remarkable about şey in this sequence that it carries a case suffix even though it is surrounded by hesitation markers. It seems as if şey is intended as a syntactic constituent. Furthermore, şey forms part of the problematic structure. The formal characteristics pointing to usage as a repair signal are the additional use of hesitation markers, pauses, changes in speech rate and the fundamental frequency level, glottal stops and the lengthening of şey itself. All these prosodic cues can be combined variably, but the sole use of one of these devices can also be perceived as a signal of a problem. Accompanied by missing case suffixes and prosodic boundary signals, these cues serve to indicate a formulation problem.

Şey + postpositioned replacement and repair format in one unit The fact that constructions with şey can take various forms in natural interaction is exemplified in the next example. In this sequence many of the characteristics described above with regard to focusing and repair initiation come together in one single intonational unit. (12) şey in Turkish-German (tkdtw08_Esi2, tk-dings53) -> 01 Mer: ben şey(e) (-) bu aralar eşyalarımı hep şeyden alıyom- (.)

I to thing in recent time I am buying my belongings at thing -> 02 ↑TELLY wear; at telly wear 03 (1.8) 04 Esi: ha:=a; yes

12. For a detailed explanation of repair formats cf. Jefferson (1972 and 1974) and Schegloff et al. (1977).

210 Yazgül Şimşek

The first use of şey has a falling-rising pitch contour. This contour was also observed in Example (8), where şey was classified as a unit projecting continuation with the contour serving as a technique for highlighting the announced topic. German dings is also used with a similar contour in some cases (cf. Example (9), line 02 above). The following pause and the downstep to the next speech parts are features similar to those of repair sequences. The second use of şey is not only prosodically but also syntactically embedded into the unit and occupies the place of the object. The F0 is high on şey but falls continuously thereafter. The replacement forms a small intonational unit of its own and is postpositioned. The pitch features favour a functional classification in terms of the information focusing strategy. Furthermore, the entire structure is of topical relevance within the conversation: there is a ‘topic-related change of focus’ because it is the speaker Esin who is talking about herself. With the şey construction, Merve aims to shift the focus of the conversation to her own person. The analysed data of sequences such as those shown here lead us to conclude that prosodic, lexical and syntactic characteristics must be taken into account in order to decide whether şey is embedded in a speech unit with the purpose of focusing new information or whether it is a “device” of repair initiation or the focusing şey-construction itself is being repaired.

6. Summary and conclusions The analysis of utterances containing the dummy word dings and its equivalent şey in Turkish-German as well as in both the source languages German and Turkish leads to the following two assumptions: 1. Both lexemes are functional in discourse in terms of the organisation of repair sequences although it would be an inappropriate generalisation to classify them as signals of repair because this would result in overseeing cases in which they are part of a repaired structure. 2. Both lexemes are used in constructions that can be identified as semantic focusing strategies. Such strategies serve the needs of information conveyance in a more specifically dramatic and outstanding way. Such constructions of Turkish-German are always TCUs with a unique inner structure formed by syntactic and prosodic means in co-occurrence. The prior assumption that constructions with dings in Turkish-German are different from standard colloquial German and accordingly constitute communicative strategies comprising an ‘ethnic style’ that reflects speakers’ Turkish language



Constructions with Turkish şey and its German equivalent dings 211

Table 3.  Summary of dings-constructions in German Form of dings and host utterance in German

Function

Ex.

Prosodic

Syntactic

Sequential

no accent on dings falling pitch

ding = integrated

question-answer

stands for – an unknown expression

no accent on dings steady or rising pitch

dings + additional elements no syntactically complex unit

announcement new beginning

topic change –

dings = prosodically weak

dings = integrated into complex question-answer explanation syntactic unit + discourse marker + pause after dings

self-initiated 1 self-repair

dings = prosodically weak

dings = integrated into syntacti- explanation cally complex structure

semantic joker word

2

background is also derived from the results of the analysis of German data as summarised in Table 3. The results confirm the findings stated by Kern & Selting (2006b): In standard colloquial German, speakers do not normally replace dings but use it with a prosodically weak structure that is not accentuated to signal a topic change or query an unknown referent. If dings is replaced by a postposition, a syntactic structure is employed which is different from Turkish-German postpositioning constructions. On the other hand, the structures using the German lexeme dings in Turkish-German differ from standard colloquial German by means of accentuation of dings itself in the host structure and its replacement in the postposition (cf. Examples (3a–e) above). German being a language that focuses on new and important information through accentuation, the only possible assumption is that TurkishGerman has created a new type of construction using structures taken from both source languages: the syntax of Turkish and the prosody of German. Clear examples of this strategy are special cases of dings constructions where dings forms a collocation with the German verb machen (to do) in a manner that is exactly parallel to those instances in Turkish where şey is part of a collocation with the verb yapmak (to make/to do). The analysis of colloquial spoken Turkish natural interaction revealed some interesting formal (primarily prosodic) features as shown in Table 2 above. In constructions where şey is embedded in the syntactic structure, the whole structure serves to split information and package it into two units that are strongly related to one another in terms of syntax and prosody and accordingly form a single TCU, namely the host structure and postposition. The specific word order with

şey occupies the position of an adverb, is replaced by adverbial qualification

falling pitch contour on şey

informing

rising contour of the eliminated part, şey on a high pitch peak

whole utterance is repaired, şey is part of the repaired structure

informing

hesitation markers, pauses, glottal stops, up or cut-off or repaired complex structures downsteps, falling pitch movement, low pitch range, possible lengthening

explanation clarification

story telling

şey = a unit of its own suffixed and tied to the following unit

şey = falling-rising pitch contour no lengthening, forms a unit of its own, pause  between şey and following unit

Sequential informing story telling

Syntactic

şey = falling-rising pitch contour pre-verbal position in host sentence + no lengthening postpositioned replacement şey can take case/poss. suffixes replacement = separate intonation unit, attached, steady final pitch whole utterance = şey at the highest pitch peak, rare prosodic breaks

Prosodic

Form of şey and host utterance in Turkish-German

Table 4.  Summary of şey-constructions in Turkish-German

7

Ex.

focusing strategy: highlighting new information

signalling of a verbalisation problem

12

10 11

focusing strategy: embedded in a 9 repair sequence

signalling high relevance of topic 8 dramatic presentation

focusing strategy: highlighting new information

Function

212 Yazgül Şimşek



Constructions with Turkish şey and its German equivalent dings 213

şey occupying the pre-verbal position enables the information to be packaged in a linear order as preferred in Turkish. In these cases şey does not seem referable to as a discourse marker since discourse markers in Turkish cannot normally occupy the pre-verbal position. However, the role of şey is a remarkable one since it forms part of a communicative strategy rather than simply a semantic gap. The examples of şey utterances found in the Turkish-German data are similar to those of spoken Turkish from Turkey in terms of syntax and prosody albeit that very few examples have been presented in this paper for comparison; consider Table 4 illustrating the Turkish-German findings focused on by the analysis. The analysis of şey appearing in repair sequences revealed the high dependency on sequential and prosodic features of both şey itself and its surrounding units. In both colloquial spoken Turkish and Turkish-German, şey is not an intrinsic hesitation marker or signal of repair but an element that is part of the problematic verbalisation. The strongest argument for this conclusion is the form of the lexeme. In such instances şey is suffixed and does not contain prosodic markers of hesitation such as lengthening. Without a suffix and with variable lengthening it functions as a device for repair, hesitation or some other verbalisation problem. Cases in which şey can be classified as a discourse marker are specific prosodic and sequential structures which employ şey to elicit a reaction from the recipient as confirmation of the production of a turn with more than one TCU. It is also doubled as the signal of the beginning of a new topic. In comparable sequences German dings is also used similarly in Turkish-German and is accentuated with rising pitch and a perceptually prominent contour due to the high pitch range. The accent assignment in the Turkish-German constructions with dings where the accent is carried on dings in the host utterance is obviously an influence of Turkish. Being familiar with şey constructions where şey is the most prominent element and contains the highest pitch peak of the entire construction, speakers of Turkish-German apply the prosody of Turkish to the German construction. Although a high pitch peak is not necessarily comparable with an acoustically strong accent used in German to signal semantic focus, the bilingual speakers of Turkish-German seem to interpret high pitch as a prominent element. However, as the analysis has shown, the Turkish-German constructions with dings are influenced by but not identical to colloquial spoken Turkish constructions. In particular the accentuation of the postpositioned replacement of dings by the ‘right word’ poses a remarkable difference. The constantly falling pitch towards the end in colloquial spoken Turkish constructions marks a downgrading prosody while Turkish-German also highlights the replacement by accentuation, a strategy permitted by the accentuation principles of German. These findings suggest that Turkish-German has developed its own focusing strategy, which in turn points towards the conclusion that the specific instances of

214 Yazgül Şimşek

use of şey and dings respectively confirm the creative outcome of a language contact situation. Speakers of Turkish-German creatively apply Turkish resources not as a code-switching strategy at the linguistic level but by adopting an information packaging strategy into their predominantly German style of speaking.13 Knowledge of both source languages seems to enable the creation of new forms by applying the different strategies of the two languages in combination, which in turn results in the emergence of a new speech style that is very effective for carrying out communication within the scope of everyday interaction.

References Auer, Peter. 1996. On the Prosody and Syntax of Turn Continuations. In Elizabeth CouperKuhlen & Margret Selting (eds), Prosody in Conversation. Cambridge: Cambridge University Press, 57–100. Auer, Peter. 2003. Türkenslang: Ein jugendsprachlicher Ethnolekt des Deutschen und seine Transformationen. In Annelies Häcki-Buhofer (ed.), Spracherwerb und Lebensalter. Tübingen/Basel: Francke, 255–264. Banguoĝlu, Tahsin. 1959. Türkçenin grameri. (The Turkish Grammar). 6. Aufl. 2000, Ankara: Türk Dil Kurumu Yayınları. Couper-Kuhlen, Elizabeth. 2004. Prosody and sequence organization: The case of new beginnings. In Elizabeth Couper-Kuhlen & Cecilia E. Ford (eds), Sound Patterns in Interaction. Amsterdam: Benjamins, 335–376. Couper-Kuhlen, Elizabeth & Selting, Margret. 2001. Introducing Interactional Linguistics. In Margret Selting & Elizabeth Couper-Kuhlen (eds), Studies in Interactional Linguistics. Amsterdam: Benjamins. Duden. 1996. Deutsches Universalwörterbuch. Mannheim u.a.: Duden Verlag. Erguvanlı-Taylan, Eser Emine. 1984. The function of word order in Turkish grammar. Berkeley: University of California Press. Göksel, Aslı. 1994. Linearity, focus and the postverbal position in Turkish. In Lars Johannson (ed.), The Mainz Meeting. Proceedings of the 7th Int. Conference on Turkish Linguistics. Wiesbaden: Harrassowitz, 85–106. Hinnenkamp, Volker. 2005. “Zwei zu bir miydi?” – Mischsprachliche Varietäten von Migrantenjugendlichen im Hybriditätsdiskurs. In Volker Hinnenkamp & Katharina Meng (eds), Sprachgrenzen überspringen. Sprachliche Hybridität und polykulturelles Selbstverständnis. Tübingen: Stauffenburg, 51–153. Jefferson, Gail. 1972. Side Sequences. In David Sudnow (ed.), Studies in Social Interaction. New York: Free Press, 294–338. Jefferson, Gail. 1974. Error correction as an interactional resource. Language in Society 2: 181–199. 13. Similar research on the ethnic style of Turkish-German has likewise revealed the development of new strategies and the use of prosodic means for the presentation of events as dramatic through the creation of short TCUs in narratives as outlined by Selting (this volume).



Constructions with Turkish şey and its German equivalent dings 215

Johanson, Lars. 1971. Aspekt im Türkischen. Vorstudien zu einer Beschreibung des türkeitürkischen Aspektsystems. Uppsala: Almqvist & Wiksell. Kallmeyer, Werner, Keim, Inken, Aslan, Sema & Cindark, Ibrahim. 2002. Variationsprofile. Zur Analyse der Variationspraxis bei den Powergirls. Mannheim: IDS, , 09.04.2008. Keim, Inken. 2007. Die türkischen “Power Girls”: Lebenswelt und kommunikative Stile einer Migrantengruppe in Mannheim. Tübingen: Narr. Kern, Friederike & Selting, Margret. 2006a. Einheitenkonstruktion im Türkendeutschen: Grammatische und prosodische Aspekte. Zeitschrift für Sprachwissenschaft 25: 239–272. Kern, Friederike & Selting, Margret. 2006b. Konstruktionen mit Nachstellungen im Türkendeutschen. In Arnulf Deppermann, Manfred Fiehler & Thomas Spranz-Fogazy (eds), Grammatik und Interaktion. Untersuchungen zum Zusammenhang von grammatischen Strukturen und Gesprächsprozessen. Radolfzell: Verlag für Gesprächsforschung, 319–347. Kern, Friederike & Şimşek, Yazgül. 2006. Türkendeutsch: Aspekte von Einheitenbildung und Rezeptionsverhalten. In Dieter Wolff (ed.), Mehrsprachige Individuen-vielsprachige Gesellschaften. Frankfurt am Main: Peter Lang, 101–119. Kim, Kyu-hyun. 2001. Confirming intersubjectivity through Retroactive Elaboration: Organization of Phrasal Units in Other-Initiated Repair Sequences in Korean Conversation. In Margret Selting & Elizabeth Couper-Kuhlen (eds), Studies in Interactional Linguistics. Amsterdam: Benjamins, 345–372. Levi, Susannah S. 2005. Acoustic correlates of lexical accent in Turkish. Journal of the International Phonetic Association 35(1): 73–97. Özbek, Nurdan. 1998. Türkçe’de söylem belirleyicileri (Utterance-markers in Turkish). In Dilbilim Araştırmaları 1998, 37–47. Özbek, Nurdan. 2000. Yani, işte, şey, ya: Interactional markers in Turkish. In Aslı Göksel & Celia Kerslake (eds), Studies on Turkish and Turkic languages. Wiesbaden: Harrassowitz, 393–402. Rehbein, Jochen. 1995. Grammatik kontrastiv – am Beispiel von Problemen mit der Stellung finiter Elemente. Jahrbuch Deutsch als Fremdsprache 21: 265–292. Schegloff, Emanuel. 1984. On some Questions and Ambiguities in Conversation. In John M. Atkinson & John Heritage (eds), Structures of Social Action. Studies in Conversation Analysis. Cambridge: Cambridge UP, 28–52. Schegloff, Emanuel, Jefferson, Gail & Sacks, Harvey. 1977. The Preference for Self-Correction in the Organization of Repair in Conversation. Language 53: 361–382. Schroeder, Christoph. 2002. On the structure of spoken Turkish. Essener Linguistische Skripte 2(1): 73–90. , 21.08.2008. Şimşek, Yazgül & Schroeder, Christoph. 2010. Turkisch. In Hans-Jürgen Krumm, Christian Fandrych, Britta Hufeisen & Claudia Riemer (eds), Deutsch als Fremd-und Zweitsprache. Ein internationales Handbuch. Berlin: de Gruyter, 719–725. Selting, Margret. 1993. Voranstellungen vor den Satz. Zur grammatischen Form und interaktiven Funktion von Linksversetzung und Freiem Thema im Deutschen. Zeitschrift für germanistische Linguistik 21: 291–319. Selting, Margret. 1994. Konstruktionen am Satzrand als interaktive Ressource. In Brigitte Haftka (ed.), Was determiniert Wortstellungsvariation? Studien zu einem Interaktionsfeld von Grammatik, Pragmatik und Sprachtypologie. Opladen: Westdeutscher Verlag, 299–318. Selting, Margret. 1998. TCUs and TRPs: The Construction of Units in Conversational Talk. InLiSt 4. Potsdam.

216 Yazgül Şimşek

Selting, Margret. This volume. Prosody and unit construction in an ethnic style: The case of Turkish-German and its use and function in conversation. Selting, Margret et al. 1998. Gesprächsanalytisches Transkriptionssystem (GAT). Linguistische Berichte 158: 298–325. Swift, Lyod B. 1962. Some aspects of stress and pitch in Turkish syntactic patterns. American Studies in Altaic Linguistics 13: 331–341. Türk Dil Kurumu. 1998. Türkçe Sözlük. Turkish Dictionary. Part 1. Ankara: Türk Dil Kurumu Basımevi. Uhmann, Susanne. 1997. Grammatische Regeln und konversationelle Strategien. Fallstudien aus Syntax und Phonologie. Tübingen: Niemeyer. Wiese, Heike. 2006. ‘Ich mach dich messer’: Grammatische Produktivität in Kiez-Sprache (Kanak Sprak). Linguistische Berichte 207: 245–273.

Joint attention and cooperation in the Swedish of adolescents in multilingual settings The use of sån ‘such’ and såhär ‘like’ Lena Ekberg

Lund University

This article describes the use of the Swedish sån ‘such’ and såhär ‘like’, respectively, as a means to establish joint attention and cooperation among adolescents in multilingual settings. In the south of Sweden, in the city of Malmö, sån ‘such’ is used, primarily, as a determiner, and also a focus marker. In contrast, adolescents in Stockholm (approximately 600 kilometres north of Malmö) make frequent use of såhär ‘like’ as a focus marker. However, there is no corresponding development into a determiner. The two expressions have overlapping meaning and function. They both serve to direct the hearer’s attention to a linguistic expression. Simultaneously, sån and såhär may function as an appeal to the hearer to co-construct a referent that is being introduced by the following expression.

1. Introduction A salient feature of Swedish spoken by adolescents in multilingual contexts is the exploitation of the comparative-demonstrative expressions sån ‘such’ and såhär ‘like’ as a means of establishing joint attention and cooperation. In standard Swedish sån (the written standard form is sådan) is a pronoun with comparative and/or deictic meaning, whereas, såhär derives from standard Swedish så här, a comparative and deictic/demonstrative expression. However, sån and såhär have what appears to be a complementary geographical distribution in the language use of adolescents in multilingual contexts. In the south of Sweden, more accurately in the city of Malmö, sån ‘such’ is used, primarily, as a determiner, substituting

218 Lena Ekberg

the indefinite article en/ett ‘a/an’. However, sån is also used as a discourse particle focusing on the following expression, that is, as a focus marker. In contrast, adolescents in Stockholm (approxi­mately 600 kilometres north of Malmö) make frequent use of såhär ‘like’ as a focus marker, yet there is no corresponding development into a determiner. Structurally, sån and såhär have grammaticalized on the level of syntax (as a determiner) and discourse (as a focus marker), respectively. The common trait of sån and såhär is the primary complex meaning, comprising a comparative, as well as, a demonstrative element. These intertwined meaning components collectively function as a means to coordinate the interlocutors’ joint attention, which, according to Diessel (2006: 463) is “one of the most basic functions of human communication”. Simultaneously, the use of sån and såhär is an appeal of cooperation; the speaker “asks” the hearer to fill in the missing information, irrespective of whether it relates to content or form. Thus, the two expressions not only have a common primary semantic meaning, they also have a shared interactive function of signalling joint attention and cooperation. I base my analysis on two subsets of data, here referred to as the Malmösample and the Stockholm-sample, collected within the project “Language and language use among young people in multilingual urban settings” during 2002 and 2003. The Malmö-sample consists of spontaneous speech of seven female informants from two upper secondary schools in Malmö. These informants belonged to different peer groups, one monolingual group (consisting of three informants), and the other a bilingual group (consisting of four informants). The common denominator for the two groups was the multilingual school context. The ratio of multilingual students varied between the schools, although the informant group ratio was relatively equal, consisting of approximately 65%. Speech samples from all four informants in the bilingual group and two of the three informants in the monolingual group were included in a perception experiment in which pupils in Malmö schools were asked to identify signs of the local variant of the multiethnic youth language, Rosengård Swedish (Hansson & Svensson 2004; Bodén 2007). The stimuli from two of the informants in the bilingual group were regarded as Rosengård Swedish by a significant majority of the listeners, whereas, the stimulus from one of the informants was regarded (also . The indefinite article has two variant forms depending on the gender of the following noun, namely en (common) and ett (neuter). . The overarching goal of the project is to describe, analyze and compare the everyday language use of young people living in multilingual settings in the three largest cities in Sweden, viz. Stockholm, Göteborg and Malmö. (The project was funded by the Bank of Sweden Tercentenary Foundation 2002–2006.)



Joint attention and cooperation in the Swedish in multilingual settings 219

significantly) as non speakers of Rosengård Swedish. Concerning the fourth informant, the listeners were not in agreement; a little less than half of them classified her as a speaker of Rosengård Swedish. As for the two informants in the monolingual group, both were regarded as non speakers of Rosengård Swedish by a significant majority of the listeners. (For details of the significant test see Bodén 2007.) The data from Malmö was collected with no researcher or other adult present. The informants either took part in semi-directed group conversations at school, or conducted self-recordings in situations they chose themselves, interacting with peers. (See Ekberg 2010 for more details of the Malmö-sample.) The Stockholm-sample includes data from two schools, both located in suburban areas where the majority of students have a multilingual background. The sample is made up of two subsets of data; one consists of recordings from female informants, the other from male informants. In contrast to the Malmö-sample, the Stockholm-sample does not include recordings of particular peer groups. Instead, the data was chosen on the basis of the result of a perception experiment, parallel to the one carried out in Malmö, with the purpose to find characteristic features of the local variant of multiethnic youth language in Stockholm, commonly referred to as Rinkeby Swedish. The recordings of the female sub-sample include two informants who were classified as speakers of Rinkeby Swedish by a significant majority of the listeners (see Bodén 2010). At least one of the informants participates in all of the recordings. However, other informants, who were not classified as speakers of Rinkeby Swedish, or who were not included in the perception experiment, also participate in the recordings. Regarding the male sub-sample, the picture is the same: all of the recordings include at least one informant who was classified as a speaker of Rinkeby Swedish. The data from the female informants in Stockholm include spontaneous and semi-directed group conversations with no adult present, as well as, independent retellings one by one in the presence of a researcher. The data from the male informants include presentations in front of the class, as well as, one semi-directed group conversation and one self-recording. (See Ganuza 2008 for a more thorough presentation of the data collected within the project.) From the presentation above, it should be clear that some informants show linguistic traits characteristic of the local multiethnic variety, while others do not, in particular the monolingual informants in Malmö. In the present investigation, I do not restrict myself to describe the language use of potential speakers of multi. At these group conversations, the informants were supposed to discuss subjects given by the teacher. However, in the course of the conversation the informants left the given subjects and started to talk about other things.

220 Lena Ekberg

ethnic varieties, although, these speakers play a central role as “guarantors” for the existence of a multiethnic speech style in most of the recorded situations. Instead, the overarching purpose is the way linguistic resources are used in communication in different multilingual settings to achieve the same goal: joint attention and cooperation in constructing referents. In the two following sections, I will account for the use of sån ‘such’ in the Malmö-sample (Section 2) and the use of såhär ‘like’ in the Stockholm-sample (Section 3). In the discussion in Section 4, the similarities in meaning and function of sån and såhär are in focus. It is concluded that both expressions are strongly hearer-oriented in that the speaker invites the hearer to co-construct the referent denoted by sån and såhär. 2. The use of sån ‘such’ among adolescents in Malmö 2.1

The recognitional meaning

A study of the language use of female adolescents in a multiethnic context in Malmö shows that sån has a number of extended functions in relation to the primary meaning, some appearing in spoken standard Swedish, others not. The primary meaning of sån, as a pronoun, is to compare two referents concerning type, cf. the standard use in (1) where the comparand (B) serving to identify the comparee (A) is expressed in a relative clause introduced by som ‘as’: (1) The comparative use of sån in standard Swedish.



Jag vill ha [en sån klänning]A [som Lisa har]B I want to-have [a such dress]A [as Lisa has] B

I want to have a dress such as the one Lisa has.

When the comparand is physically present or can otherwise be derived in the speech situation, the relative clause may, however, be left out: cf. (2) where there is no linguistic constituent corresponding to the comparand (B). (2) The comparative use of sån in standard Swedish; the relative clause is left out.



Jag vill ha [en sån klänning]A I want to-have [a such dress]A

I want to have a dress like that.

. This section is largely a summary of the findings presented in Ekberg (2010).



Joint attention and cooperation in the Swedish in multilingual settings 221

The meaning of sån is not only comparative but deictic, as well as, demonstrative; the speaker is pointing to B, linguistically and/or physically, in order to denote A. When B is present in the speech situation although not linguistically explicit, the deictic meaning is strengthened (cf. (2)). In multilingual settings in Malmö, sån is normally construed without a relative clause denoting a comparand: cf. (3)–(4). With closer observation, it becomes obvious that the use in (3)–(4) is not identical to the examples illustrated in (2), (3) and (4) not only lack a linguistically explicit comparand, there is also a lack of a comparand present in the speech situation. (The denoted referent, the comparee, is underlined.) (3) The use of sån in multilingual settings in Malmö (Gordana, Sabaah). 1 G: du vet jag har sån ee säng. you know I have such ee bed

you know I have this bed 2 S: mm. 3 G: du vet sån ee # med sån fjärrkontroll # you know such ee # with such remote-control #

you know with this remote control # 4 så du vet man kan so you know one can

you know you can 5 flytta upp den # åå # så man kan # move up it # and # so one can #

move it up # so you can 6 så du vet man kan flytta upp sängen å sånt du vet. so you know one can move up bed-the and such you know

you know you can move up the bed and all that you know (4) The use of sån in multilingual settings in Malmö (Märta). 1 åå vi bytte såna kort # såna typ fotokort # and we exchanged such cards # such like photo-cards #

and we exchanged these cards # like photo cards # 2 såna+ Spice Girls å Backstreet Boys such+ Spice Girls and Backstreet Boys

like Spice Girls and Backstreet Boys

. All names of the informants are pseudonyms. The transcripts are written in a modified standard orthography. See Appendix for explanations of the symbols used.

222 Lena Ekberg

In uses such as these, sån points to entities beyond the physical speech situation: mental entities that exist in the heads of the participants. In other words, the deictic/demonstrative meaning inherent in the primary comparative use has been extended to the mental domain. When using sån the speaker is referring to entities that she assumes the hearer is able to identify through shared knowledge and experience. More specifically, the speaker introduces a new referent in the current discourse that is construed as known, in other words, old to the hearer. I refer to this function as recognitional, borrowing a term that has been used to describe a particular cross-linguistic use of demonstratives in spoken language (Himmelmann 1996, 1997; Diessel 1999; Lindström 2000). Specifically, a demonstrative may be used in the initial mention of an NP, introducing a referent that is new in the current discourse, although, construed as old to the hearer, cf. (5), where the demonstrative those is used in this manner. (5) Demonstrative used in initial mention of an NP (from Himmelmann 1996: 230).

it was filmed in California, those dusty kinds of hills that they have out here by Stockton and all

A central feature of the recognitional use is, as Himmelmann expresses it, that “the speaker anticipates problems with respect to the information used in referring to a given referent” (1996: 230). This also seems true for the recognitional use of sån. Consider (3), where the speaker also by means of the expression du vet ‘you know’ signals that help is required from the hearer to construe the referent (du vet jag har sån säng ‘you know I have this bed’). Recognitional sån is therefore interactive and implies a certain degree of closeness between the participants. Hence, it functions as a positive politeness marker, such as, you know and other tags that encode the hearer’s point of view (Brown & Levinson 1987: 118ff.). 2.2

Sån as a determiner

There is another significant difference between spoken standard Swedish in general and the Malmö-sample. In the Malmö-sample, there is a strong tendency to leave out the obligatory indefinite article en/ett ‘a/an’ in a noun phrase with recognitional sån. Consider again the example in (3) where the singular indefinite NP lacks the article en. In standard Swedish, sån NP would be preceded by en, that . The term ‘recognitional’ was to my knowledge first used by Sacks and Schegloff (1979) to refer to a particular practice of person-reference. They conclude that names, being a basic sort for recognitionals, are not only used when known but may be introduced for subsequent use when not already known to the recipient.



Joint attention and cooperation in the Swedish in multilingual settings 223

is, en sån NP ‘a such NP’ or merely construed as en NP ‘a NP’. In Ekberg (2010) I argue that sån in fact is replacing the article in these cases. In the Malmö-sample there are 156 occurrences of sån in noun phrases where the indefinite article is obligatory in standard Swedish. That is, in standard Swedish these noun phrases would be construed as en (sån) NP ‘a (such) NP’. 115 of these 156 NPs lack the indefinite article, that is, the indefinite article is left out in approximately 74% of the cases. The bilingual group, which includes potential speakers of the local multiethnic variety, mentioned above, (see the Introduction), leaves out the determiner more often than the monolingual group. However, the difference is insignificant (see Ekberg 2010). The question is whether sån apart from taking the place of the indefinite article also fulfils the typical function of the article, which is, to introduce a discourse referent that is new to the hearer. One of the contexts where sån occurs without an indefinite article as part of a noun phrase is as the logical subject of a presentational construction, for example, de(t) är sån NP ‘it is such NP’, idiomatically ‘there is this NP’. This is the typical syntactic context for introducing a new discourse referent. In (6) the informant (Duhi) uses a presentational construction with sån (underlined) no less than three times to introduce the new topic, a character in a movie, which she eventually refers to as sån fågel som e gul ‘a bird that is yellow’ (line 4). (6)  Presentational constructions with sån. (Transcript from a self-recording where the informant, Duhi, tells her friends about a cartoon character in a movie.) 1 så så finns de sån ee typ tecknad figur nåt sånt # so so is it such ee type cartoon character something such #



then there is this like cartoon character or something like that #

2 nej de e inte tecknad figur # no it is not cartoon character #



no it is not a cartoon character #

3 de e sån figur du vet som e för barn # it is such character you know that is for kids #



it is a character you know for kids #

4 sån fågel som e gul så jag # such bird that is yellow so I #



a bird that is yellow so I #

The specific referent that the informant introduces is unknown to the hearer. However, the type of referent is assumed to be familiar; compare the accompanying use of typ (literally ‘type’) ‘like’ in the first presentational construction (line 1). As . In English the presentational construction is usually referred to as the there-construction.

224 Lena Ekberg

can be seen, the recognitional use in (3) and (4) introduces a referent that is new in the current discourse although old, as in familiar, to the hearer. The notion of “old to the hearer” can, however, be further specified. Consider (6) above, where sån denotes a type of referent of which it is likely that the hearer has a mental representation. This is the actual, or typical, recognitional use, parallel to the recognitional use of demonstratives in, for example, Swedish, English and German (cf. Lindström 2000; Himmelmann 1996, 1997). However, there are also instances where sån denotes a referent of which the hearer cannot have a specific mental representation. Prince (1992: 307) distinguishes a type of hearer-old referents that are “formed in such a way that we ‘feel’ we are supposed to know about the entity they represent, but in fact we may very well not”. This is exactly the case in (7) below, where the hearer cannot have a specific mental picture of the referent introduced (the baby in the bank). (7) Sån is used to introduce a referent that is new to the hearer (Gordana). 1 asså när jag var i bibl+ ee var i banken i fredags # well when I was at libr+ ee was at bank-the last Friday



well when I was at the libr+ at the bank last Friday

2 så var det sån lite[n] bebis # so was it such little baby



then there was this little baby

3 asså du vet så jag ville bara ta du vet som å # well you know so I wanted just take you know like and #



well you know I just wanted to take you know like and #

4 ta henne å krama henne take her and hug her



take her and hug her

While, sån in (3)–(4) introduces a new discourse referent that presumably is familiar to the hearer, sån in (6) introduces a type of referent that is familiar to the hearer, although, the specific referent is unfamiliar. Finally, the use of sån in (7) introduces a new discourse referent that is “completely” new to the hearer, and is elaborated in the subsequent discourse. This indicates that sån has reached (at least) the second stage in the evolution hierarchy of indefinite articles: the article is used to introduce a referent that is presumed to be unknown to the hearer, and this referent is taken up as definite in the subsequent discourse (Heine 1997: 72). As an indefinite article, sån is therefore cataphoric; it marks the newly introduced NP as an important topic in the subsequent discourse. An interesting parallel is found in spoken American English, where, according to Givón (1995: 66), the indefinite article this is used for cataphoric grounding of topical indefinite referents, that is, referents that tend to appear as subjects in the following discourse.



Joint attention and cooperation in the Swedish in multilingual settings 225

A similarity between sån and this is the demonstrative meaning. In the following, the meaning of sån when used as a determiner will be explored. As a comparative pronoun, sån relates the designated entity, the comparee, to another entity, the comparand (cf. (1) above). In contrast, the recognitional use of sån lacks a linguistically expressed, or a physically present, comparand. However, when the speaker is using sån in (3)–(4), and (6)–(7), she is referring to an imagined comparand, appealing to the shared knowledge and experience of the participants. The imagined comparand is necessary to identify the comparee, although, neither the comparand nor the implicit common knowledge of the participants can be explicitly referred to. In (7) sån is used to introduce a new discourse referent, while it simultaneously evokes a meaning of comparison. The speaker is construing the referent as “known to us” by pointing to an imagined comparand, which the hearer is encouraged to “create” in the ongoing speech situation. That is to say, the referent sån liten bebis ‘such little baby’ is construed as a type of baby known to the hearer, although the specific referent is unknown. The meaning of sån as a determiner is therefore complex. The overall indefinite meaning (sån denoting an indefinite referent) comprises a definite element arising from the conception of sån pointing to a known type to which the indefinite referent adheres. This layered meaning explains why sån simultaneously can be used to identify a referent that the hearer is familiar with, through the category it belongs to, and to introduce a new discourse referent unknown to the hearer. Concluding, sån has a recognitional meaning also when it functions as a determiner. 2.3

Sån as a discourse particle

Apart from the function as a determiner, there are other uses in the Malmö-sample that are worth noting, as they do not appear in standard Swedish. Consider the transcript in (8) where sån is placed immediately before a noun containing significant information. In this, and other examples, sån seems to function as a focalizing discourse particle, rather than a determiner or a comparative pronoun. If this is correct sån is not part of the noun phrase, which is a probable analysis of the use in (8), since sån does not agree with the following noun. (Sån appears in common gender although the following noun (busfrö) is neuter.)

. As part of the noun phrase sån must agree in gender (and number) with the head noun, cf. (i) sån bil (ii) sånt hus such.common car.common such.neuter house.neuter

226 Lena Ekberg

(8) Sån used as a discourse particle (Bodil). 1 hon ville inte ha mej där bak # she wanted not to-have me there back #



she didn’t want me to sit at the back [of the room] #

2 asså jag var sån busfrö you-know I was such little-devil



because I was such a little devil

The use of sån in (8) is similar to the one described for Swedish såhär ‘like’ (see Section 3.1), which functions as a focuser, preceding the rhematic constituent in an utterance (Öqvist 1997). Såhär as a discourse particle is not found in the Malmö-sample but interestingly it is frequently found in the project data from Stockholm. In the following section, I will account for the use of såhär among the informants in Stockholm, which will provide a basis for evaluating the similarities of sån and såhär.

3. The use of såhär ‘like’ among adolescents in Stockholm 3.1

Såhär as a discourse particle

Såhär is a fusion of a deictic expression, så här, idiomatically ‘like this’, where the first element, så ‘so’; ‘like this/that’, is a comparative-demonstrative adverb and the second, här ‘here’, a proximal demonstrative. In the recordings from Stockholm there are slight variations with regard to the pronunciations of såhär, but in the transcripts these variants are usually represented in a homogenous manner. As the expression is unstressed, it is often difficult to hear exactly how it is pronounced. In a study by Öqvist (1997) the same expression has the form sär, which gives the impression of a further phonetic reduction of the word combination så här. Nevertheless, I regard sär and såhär as orthographic variants of the same lexeme. In its primary deictic/demonstrative use, så här refers to some aspect of the speech situation, either the physical environment or the surrounding discourse: cf. du kan göra så här ‘you can do like this’, where så här typically is accompanied by a physical demonstration.

. However there may be also other lexical sources of sär, viz. så där lit. ‘so there’ (“like this”) or sån här/där lit. ‘such here/there’ (“like this/that”) (Öqvist 1997). The common denominator is that the first part is a comparative adverb or pronoun (så ‘so’ or sån ‘such’) and the second a demonstrative adverb (här ‘here’ or där ‘there’).



Joint attention and cooperation in the Swedish in multilingual settings 227

When investigating youth language in the nineties, Swedish scholars (Kotsinas 1994; Eriksson 1997; Öqvist 1997) observed that såhär was used as a discourse particle. They noted that såhär (or så här) occurred at the end of the sentence, marking the boundary between sentence or phrase constituents (Kotsinas 1994: 89;  Eriksson 1997: 129). Eriksson also noticed that såhär sometimes occurred before a rhematic constituent and, furthermore, was used to introduce a quotation (Eriksson 1997: 130). Despite the rhematic function seeming to be less important to Eriksson’s data, Öqvist (1997) finds this to be the dominating function in her study. In our data from Stockholm (collected 2002) the depiction is similar to the one given in Öqvist. The use of såhär, as a discourse particle, is one of the characteristic traits of the language use of the informants in Stockholm. The analysis of såhär is based on 348 occurrences found in the Stockholm-sample. Viewing these more closely, it appears that in most of the cases, såhär precedes a constituent that is information significant, that is, såhär is used as a focus marker. Consider the extract in (10) where the informants (Fawza and Bushra) discuss the way a young author is representing (a particular variety of) multiethnic youth language in his novel. (10) Såhär as a discourse particle (Fawza, Bushra).

1 F: å dom skulle nog såhär garva åt de(t) and they should probably like laugh at it

and they should probably like laugh at it

2 F: som såhär hu(r) han stavar å hu(r) han pratar å sådär as like how he spells and how he talks and like-that

as like how he spells and how he talks and things like that

3 B: aha # asså såhä(r) du vet man bruka(r) ju säga ah ba kolla du vet aha # so like you know one use-to course say ah like look you know

aha # like you know you usually say ah like look you know

4 B: såhä(r) dom ha(r) såhä(r) skrivi(t) såhä(r) stora bokstäver like they have like written like capital letters

like they have like written like capital letters

5 B: såhä(r) BEHÖVDE # såhä(r) stort å ö. like BEHÖVDE10 # like big å ö

like needed # like a big Å Ö

10. Bushra emphasizes the word behövde ‘needed’ in order to demonstrate that it is written with capital letters.

228 Lena Ekberg

The informant (Fawza) describes people’s reaction when they get in touch with the unconventional spelling of the author in question. She uses the unstressed såhär to focus on the verb phrase garva åt de(t) ‘laugh at it’ (line 1) and the clause hur han stavar å hur han pratar å sådär ‘how he spells and how he talks and things like that’ (line 2). As a focus marker såhär may precede any type of constituent, that is, a phrase or a bare head, or a clause. In most of the cases såhär precedes a noun or a noun phrase. The transcripts in (11)–(13) illustrate såhär preceding a noun phrase, an adjective and a preposition phrase. (11) Såhär as a focus marker preceding a noun phrase (Daniella).

men de(t) e såhär roliga ord som kommer ibland but it is like funny words that come sometimes



but it is like funny words that just come sometimes

(12) Såhär as a focus marker preceding an adjective (Bushra).

ah den va(r) såhä(r) intressant oh it was like interesting

(13) Såhär as a focus marker preceding a preposition phrase (Daniella).

oss ungdomar i typ såhä(r) från förorten us adolescents in like like from the-suburb



youth like us like from the suburbs

However, there are also instances in the Stockholm-sample where såhär occurs at the end of a sentence, having a boundary marking function, cf. (14), and where it occurs before a quotation, cf. (15)–(16). (14) Såhär as a boundary marker. (Fawza tells about the plot in a novel.) 1 ah de(t) e såhä(r) mycke(t) fantasi de(t) va(r) de(t) som va(r) dåligt # ah it is like much fantasy it was this that was bad #



oh it is like a lot of fantasy it was this, that was bad



well he ought to take something like something usual # like

2 asså han borde ta nånting såhära # såhä(r) vanligt # såhär. well he ought to-take something like # like usual # like

(15) Såhär as a quotative (Fawza).

de(t) e bara tänka såhära # it is only to-think like # the-holiday is over



you just have to think like the holiday is over



Joint attention and cooperation in the Swedish in multilingual settings 229

(16) Såhär as a quotative (Bushra).

ah såhä(r) oh like yeah whatever

The most common function of såhär in the Stockholm-sample is to focus on new and significant information. Out of 348 total occurrences, 204, or 58.6%, have this function, whereas 14.7% of the såhär-occurrences are boundary markers and 10.6% are quotatives. A much smaller amount of såhär-occurrences show as demonstratives, 7.5%, with the final 8.6% of the occurrences proving difficult or impossible to characterize in terms of a specific function. Table 1 summarizes the distribution of different uses of såhär in the Stockholm-sample. Table 1.  Uses of såhär in the Stockholm-sample Function

N

%

Focus marker Boundary marker Quotative Demonstrative Unclear Total

204   51   37   26   30 348

  58.6  14.7  10.6    7.5    8.6 100

Together with the function as a focus marker, såhär may have an interactive function, which has already been observed by Öqvist (1997). The speaker encourages the hearer to cooperate in the construction of the focused referent. Such cooperation presupposes shared knowledge and experience, and by using såhär the speaker communicates that he/she has, or wants to have, a particular social relation with the hearer (cf. Östman 1982: 169) who gives a similar description of the use of you know). The use of såhär/sär thus expresses social closeness between speaker and hearer. The extract in (17) is a telling example of how såhär simultaneously is used both as a focus marker and as an interactive appeal of cooperation. The informants, Fawza, Bushra and Daniella, are discussing a short story and what impact it has on their understanding of the principal character that the story takes place in their own neighbourhood, a suburb with a large number of immigrants. Fawza (line 1), places såhär before the information heavy constituent, nånstans nära ‘somewhere near’. When Daniella is searching for a word to characterize a young immigrant from this area (line 3), Bushra fills in with the adjective svartskallish (line 4), which is a new coinage from the Swedish noun svartskalle ‘wog’. The use of såhär is, however, not only a means of focusing, it is also a means to “hold the floor” while planning the utterance and searching for the appropriate

230 Lena Ekberg

expression. I assume, following Östman (1982: 162), that the function as a pausefiller generally accompanies the focuser function (or any other structural function of discourse particles). Essentially, såhär (line 3) is simultaneously a focus marker, a marker of cooperative speaking and probably a floor holding device. (17) Såhär as a focus marker and a marker of cooperative speaking (Fawza, Bushra, Daniella). 1 F: ah de(t) utspela(r) sej såhä(r) nånstans nära ah it takes-place like somewhere near

oh, it takes places somewhere near 2 B: xxx 3 D: ah i närheten då bli(r) man mer såhä(r) ah ah in the-neighbourhood then gets one more like ah

oh, in the neighbourhood, then you get more like arhhh 4 B: xxx svartskallish xxx wogish

There are close correspondences in English to the discourse particle såhär. In a paper from 1988, Underhill investigates the distribution and function of the English particle like, which like såhär has a primary comparative meaning. Judging from Underhill’s account, like shows a similar pattern as såhär. He draws the conclusion that the particle like functions as “a marker of new information and focus” (Underhill 1988: 234). Moreover, it marks significant information. As regards the distribution, like is placed before all kinds of constituents, most commonly a VP or an NP. A few years later, Romaine & Lange (1991) distinguish a further use of like, which is argued to be developed out of like as a focus marker, namely the function as a quotative. Following the footsteps of Romaine & Lange (1991), several scholars have regarded this development as a clear case of grammaticalization (Ferrara & Bell 1995; Tagliamonte & Hudson 1999). Additional, såhär is used both as a focus marker and, though not as often, as a quotative, as is evident from the data from Stockholm. The development of såhär is hypothesized to have followed the same track as like: from being used to focus new and significant information to introducing a quotation, that is, an unuttered thought or a “real”, more or less direct quotation.11

11. A further parallel example of grammaticalization from a focus marker to a quotative is given by Swedish ba (Eriksson 1997). Ba is a reduced form of bara with the lexical meaning ‘only; just’.



3.2

Joint attention and cooperation in the Swedish in multilingual settings 231

The use of såhär as a matter of linguistic activity

Romaine & Lange (1991) notice a gender differentiation as regards the use of like as a quotative with the majority of their examples drawn from young women. They propose that women use like more than men because “the topics they discuss are typically talked about in an involved conversational style” (Romaine & Lange 1991: 228). As regards the more general use of like as a discourse particle, Eckert’s data (referred to in Eckert 2004) show no gender differentiation. When looking at the data from Stockholm, it appears to be quite a remarkable frequency difference between females and males as regards the use of såhär. Also here the majority of examples are from female informants. There are 273 instances of såhär in the recordings from the female informants: the F-sample, corresponding to 1.4% of the total number of tokens (19 718). In the recordings from the male informants, the M-sample, there are only 75 instances of såhär, which is 0.4% of the total amount of tokens (19 139). In order to investigate whether this frequency difference is due to situational variables, such as topics and linguistic activities, I explored the two sub-samples separately. The data of the two samples were of different types (see Section 1). The M-sample mainly consisted of presentations conducted in front of the class at school, whereas the F-sample consisted of spontaneous conversations amongst peers. One could hypothesise that the frequency difference is due to the different type of data (see above). Presentations in front of the class, which are conducted only by the male informants, are comparatively formal and are, therefore, closer to the written form of the language, where focus markers, such as, såhär practically do not occur. However, there is at least one situation where a male informant uses såhär quite frequently. This is, in fact, when Roshan (see (18)) makes a presentation of collective work about poorness in the world, in front of the whole classroom. Here, Roshan introduces the work he is going to present by giving an overview of how he and his classmates have selected the relevant information. This introduction is obviously spontaneous in the sense that Roshan is not reading from a manuscript at this point. Roshan uses såhär no less than five times; all of which appear immediately before significant noun phrases (numbered (i)–(v) in the transcript: olika mått ‘different measurements’, katastrof ‘catastrophe’, krig ‘war’, lista ‘list’ and BNP ‘GNP’. (18) Roshan presents in front of the class.

okej eh # vi ha(r) tagit  å de(t) de(t) finns flera utbildning å # hunger å krig # å så tog ja(g) #

såhär (i) olika mått # på fattigdom # styckna men ja(g) ha(r) tagit eh BNP såhär (ii) katastrof och såhär (iii) t+ eh tre av dom fattigaste länderna

232 Lena Ekberg

i en såhära12 # en såhär (iv) lista vi hade från såhär (v) BNP # som hade lägst BNP # ah ja de(t) # fattigaste landet.  okey eh # we have taken like (i) different measurements # on poorness # and it it is several numbers but I have taken eh GNP education and # hunger and like (ii) catastrophe and like (iii) war # and then I took # t+ eh three of the poorest countries in a like # a like (iv) list we had from like (v) GNP # that had the lowest GNP # ah yes the # poorest country

When Roshan proceeds with the body of his presentation, he is particularly bound to a manuscript. In spite of this dependence, he continues to place såhär before significant parts in the presentation, with an obvious abandonment of the manuscript during these instances. Together with the focusing function såhär is probably used also as a pause-filler. A presentation in front of the class is mono-logical and probably characterized by a comparatively formal language use,13 which is a comparatively different linguistic activity when the informant interacts with peers while recording themselves in their familiar surroundings at home. One such instance comes from Ekmel, an informant who made a self-recording of 75 minutes. During these 75 minutes, there are only three instances of såhär as a discourse particle. The low accounts of såhär may be explained by the fact that the recording is dominated by dialogic conversations consisting of short utterances with no narrative parts. There are no self-recordings from the female group, therefore, the possibility for a comparison is not an option. However, the recordings that consist of free or semi-directed group conversations are judged to be at least as informal as selfrecordings. (There is no teacher or adult present in the group conversations.) The instances of såhär in the group conversations of the female informants amount to 1.2%, which is significantly higher than in both the M-sample in general (0.4%) and in Ekmel’s self-recording (0.2%). However, there is another type of linguistic activity in which the amount of såhär is even higher than in the group conversations. This is when the female informants are engaged in retellings in the presence of a researcher,14 in which case, såhär amounts to 1.9% of the tokens, cf. (19).

12. Såhära is a variant of såhär. 13. Note however that the use of formal language does not exclude spontaneous (or impromptu) speech. 14. The retellings are recorded and transcribed by Natalia Ganuza for her thesis about syntactic variation in the Swedish of adolescents in multilingual urban settings (2008).



Joint attention and cooperation in the Swedish in multilingual settings 233

(19) Såhär used in a retelling (Fawza).

1 men den va(r) såhär så lite såhär för grov but it was like so a-bit like too coarse



but it was like a bit like too coarse



it was like too much I don’t know

2 de(t) va(r) såhär för mycket ja(g) vet inte # it was like too much I know not # 3 men de(t) va(r) en bra historia but it was a good story



but it was a good story



you know it was like touching I think

4 asså den va(r) såhär rörande tycker ja(g) so it was like touching think I

Obviously, the narrative monologue favours focus marking, either the monologue is a presentation in front of the class or a retelling. Thus, the difference between the F- and the M-sample as regards the discourse particle use of såhär might, primarily, be a matter of different linguistic activities, not of gender.

4. Summary and concluding discussion As should have been clear from the account in Sections 2 and 3, sån and såhär overlap in meaning and function. The primary meaning of both expressions is comparative and demonstrative. Sån denotes an entity A by pointing to an entity B, thus comparing A with B. The B-entity may be either linguistically present or physically present, or both. In extended uses, however, sån points to a B-entity that is present solely in the memory or mental imagination of the speech participants. I have referred to this function as recognitional. A further development of sån is when it is grammaticalized into a determiner, replacing the indefinite article. When functioning as a determiner the recognitional meaning is weakened, but still present. Another extension, despite not occurring frequently, is when sån is used as a focus marker, pointing at the rhematic constituent. Just as the development into a determiner can be argued to be motivated by the demonstrative/deictic meaning, the extension into a focus marker is also motivated by the demonstrative/deictic meaning. Furthermore, both are cataphoric, pointing to a following constituent. As regards såhär, it is explicitly demonstrative in its adverbial function; as a discourse particle the demonstrative meaning is transferred to the linguistic context. Såhär points at the following constituent, thus marking the content or form

234 Lena Ekberg

as significant. Relating to information structure, såhär functions as a focuser; it marks the rhematic part of the utterance. In addition to the structural function as a focus marker, såhär may have an interactive function. The speaker points at a specific referent that he/she has in mind and “asks” for the hearer’s cooperation. The invitation to the hearer to co-construct the referent or the expression denoting the referent, which is part of the usage of both såhär and sån, creates a sense of engagement and closeness, typical of an involved conversational style. Summarizing, såhär and sån have both structural and pragmatic functions. The structural function of såhär is as a focuser, whereas, sån has two structural functions; the dominating one is as a determiner, the other as a focuser. In the latter case, sån overlaps with såhär. Hence, the dominating function of såhär and sån, respectively, is quite different. Såhär is operating at the discourse (information structure) level, while sån is operating at the level of syntax. There is, however, a shared, cognitively based, function of såhär as a discourse particle and of sån as a determiner. They both serve to direct the hearer’s attention to a linguistic expression. When using såhär or sån the speaker is, metaphorically speaking, pointing to the following expression, establishing a joint focus of attention. As Langacker (2004) argues, the physical pointing gesture is the conceptual archetype underlying not only demonstratives but definite and indefinite determiners as well. Diessel (2006: 476) claims that the same psychological mechanisms underlie the use of a demonstrative to point to entities in the linguistic context (the anaphoric and discourse deictic use) as the use of exophoric demonstratives, pointing to entities in the physical world. The extension from the physical world, undoubtedly, goes further, namely from pointing to concrete entities to pointing to entities in the mental world (the recognitional use). The complementary geographical distribution of sån and såhär should be seen in light of the fact that the two expressions have an overlapping meaning and function. With closer observation, the different use of discourse particles and determiners among the informants in Malmö and Stockholm appears as a surface phenomenon. The underlying similarities are striking: adolescents in multilingual settings in the two cities make frequent use of grammaticalized expressions that are strongly hearer-oriented. The particular use of sån and såhär cannot be unambiguously associated with particular multiethnic varieties. There is a small, although not significant, difference in frequency as regards the use of sån as a determiner between the bilingual and the monolingual group; the former group includes potential speakers of the local multiethnic variety, whereas, the latter does not. A hypothesis that can be tested in future research is whether the tendency to use expressions signalling co-construction of referents and directing the listener’s attention is stronger in



Joint attention and cooperation in the Swedish in multilingual settings 235

multilingual settings than in linguistically more homogenous settings, or perhaps even stronger among speakers of multiethnic varieties than among adolescents in general. The present analysis may, therefore, provide the basis for formulating hypothesis about communicative strategies in multilingual settings, specifically the use of grammaticalized expressions to establish joint attention and cooperation.

References Ailin Svendsen, Bente and Pia Quist. (eds). 2010. Linguistic Practices in Multiethnic Urban Scandinavia. Multilingual Matters. Bodén, Petra. 2007. “‘Rosengårdssvensk’ fonetik och fonologi.” In Språket hos ungdomar i en flerspråkig miljö i Malmö, Lena Ekberg (ed.), 1–47. Nordlund 27. Småskrifter från Nordiska språk vid Lunds universitet. Bodén, Petra. 2010. “Pronunciation in Swedish multiethnolect.” In Linguistic Practices in Multiethnic Urban Scandinavia, Bente Ailin Svendsen and Pia Quist (eds), 35–48. Multilingual Matters. Botley, Simon and Anthony Mark McEnery. (eds). 2000. Corpus-based and Computational Approaches to Discourse Anaphora. Amsterdam/Philadelphia: John Benjamins Publishing Company. Branderud, Peter and Hartmut Traunmüller. (eds). 2004. Proceedings. FONETIK 2004. Department of Linguistics, Stockholm University. Brown, Penelope and Levinson, Stephen. 1987. Politeness: Some Universals in Language Usage. Cambridge, U.K.: Cambridge University Press. Diessel, Holger. 1999. Demonstratives: Form, function and grammaticalization. Amster­dam/ Philadelphia: John Benjamins Publishing Co. Diessel, Holger. 2006. “Demonstratives, joint attention, and the emergence of grammar.” Cognitive Linguistics 17(4): 463–489. Eckert, Penelope. 2004. “Language and Gender in Adolescence.” In The Handbook of Language and Gender, Janet Holmes and Miriam Meyerhoff (eds). Blackwell Reference Online. Ekberg, Lena. (ed.). 2007. Språket hos ungdomar i en flerspråkig miljö i Malmö. Nordlund 27. Småskrifter från Nordiska språk vid Lunds universitet. Ekberg, Lena. 2010. “Extended uses of sån ‘such’ among adolescents in a multilingual context.” In Linguistic Practices in Multiethnic Urban Scandinavia, Bente Ailin Svendsen and Pia Quist (eds), 49–65. Multilingual Matters. Enkvist, Nils Erik. (ed.). 1982. Impromptu Speech: A Symposium. Papers Contributed to a Symposium on Problems in the Linguistic Study of Impromptu Speech (Åbo, Finland, November 20–22, 1981). Meddelanden från Stiftelsens for Åbo Akademi Forskningsinstitut Nr. 78. Eriksson, Mats. 1997. Ungdomars berättande. En studie i struktur och interaktion. Institutionen för nordiska språk, Uppsala universitet. Ferrara, Kathleen and Bell, Barbara. 1995. “Sociolinguistic Variation and Discourse Function of Constructed Dialogue Introducers: The Case of Be + like.” American Speech 70(3): 265–290. Fox, Barbara. (ed.). 1996. Studies in Anaphora. Amsterdam/Philadelphia: John Benjamins Publishing Co.

236 Lena Ekberg

Ganuza, Natalia. 2008. Syntactic Variation in the Swedish of Adolescents in Multilingual Urban Settings. Subject-verb Order in Declaratives, Questions and Subordinate Clauses. Centre for Research on Bilingualism, Stockholm University. Gernsbacher, Morton Ann and Talmy Givón. (eds). 1995. Coherence in Spontaneous Text. Amsterdam/Philadelphia: John Benjamins Publishing Co. Givón, Talmy. 1995. “Coherence in text vs. coherence in mind.” In Coherence in Spontaneous Text, Morton Ann Gernsbacher and Talmy Givón (eds), 59–115. Amsterdam/Philadelphia: John Benjamins Publishing Co. Hansson, Petra and Svensson, Gudrun. 2004. “Listening for ‘Rosengård Swedish’”. In Proceedings. FONETIK 2004, Peter Branderud and Hartmut Traunmüller (eds), 24–27. Department of Linguistics, Stockholm University. Heine, Bernd. 1997. Cognitive Foundations of Grammar. New York/Oxford: Oxford University Press. Himmelmann, Nikolaus P. 1996. “Demonstratives in Narrative Discourse: A Taxonomy of Universal Uses.” In Studies in Anaphora, Barbara Fox (ed.), 205–254. Amsterdam/Philadelphia: John Benjamins Publishing Co. Himmelmann, Nikolaus P. 1997. Deiktikon, Artikel, Nominalphrase: Zur Emergenz syntaktischer Struktur. Tübingen: Niemeyer. Kotsinas, Ulla-Britt. 1994. Ungdomsspråk. Uppsala: Hallgren & Fallgren. Langacker, Ronald W. 2004. “Remarks on nominal grounding.” Functions of Language 11(1): 77–113. Lindström, Eva. 2000. “Some uses of demonstratives in spoken Swedish.” In Corpus-based and Computational Approaches to Discourse Anaphora, Simon Botley and Anthony Mark Mc­ Enery (eds), 107–128. Amsterdam/Philadelphia: John Benjamins Publishing Company. Mann, William C. and Sandra A. Thompson. (eds). 1992. Discourse description: Diverse linguistic analyses of a fund-raising text. Amsterdam/Philadelphia: John Benjamins Publishing Co. Prince, Ellen F. 1992. “The ZPG Letter: Subjects, Definiteness, and Information-status.” In Discourse description: Diverse linguistic analyses of a fund-raising text, William C. Mann and Sandra A. Thompson (eds), 295–332. Amsterdam/Philadelphia: John Benjamins Publishing Co. Psathas, George. (ed.). 1979. Everyday Language: Studies in Ethnomethodology. New York: Irvington. Romaine, Suzanne and Lange, Deborah. 1991. “The use of like as a marker of reported speech and thought: A case of grammaticalization in progress.” American Speech 66(3): 227–279. Sacks, Harvey and Schegloff, Emanuel A. 1979. “Two Preferences in the Organization of Reference to Persons in Conversation and Their Interaction.” In Everyday Language: Studies in Ethnomethodology, George Psathas (ed.), 15–21. New York: Irvington. Tagliamonte, Sali and Hudson, Rachel. 1999. “Be like et al. beyond America: The quotative system in British and Canadian youth.” Journal of Sociolinguistics 3(2): 147–172. Underhill, Robert. 1988. “Like is, like, focus.” American Speech 63(3): 234–246. Öqvist, Jenny. 1997. “Jamen förr i tiden så va re ju mera sär “å vicken vacker hatt du har”. Om partikeln ‘sär’s funktioner i samtal.” D-uppsats, Institutionen för lingvistik, Stockholms universitet.



Joint attention and cooperation in the Swedish in multilingual settings 237

Östman, Jan-Ola. 1982. “The symbiotic relationship between pragmatic particles and impromptu speech.” In Impromptu Speech: A Symposium. Papers Contributed to a Symposium on Problems in the Linguistic Study of Impromptu Speech (Åbo, Finland, November 20–22, 1981), Nils Erik Enkvist (ed.), 147–177. Meddelanden från Stiftelsens för Åbo Akademi Forskningsinstitut Nr.78.

Appendix Transcription symbols # short pause ## longer pause + interrupted speech xxx unintelligible sequence quoted speech [: text] standard form of a word (de) standard orthographic addition Capital letters indicate stress

Linguistic variation and linguistic virtuosity of young “Ghetto”-migrants in Mannheim, Germany* Inken Keim and Ralf Knöbl

Institute for the German Language, Mannheim

In this paper, we provide an insight into the life world and social experiences of young Turkish migrants who are categorised by German society as “social problem cases”. Based on natural conversational data, we describe the communicative repertoire of one migrant adolescent and that of his friends. Our aims are (a) to isolate those linguistic features that convey the impression of “foreignness”, and stand out among other German speakers’ features, and (b) to analyse the variability in our informants’ discursive practices – i.e. code- or style-switching, as it is commonly referred to in the literature – in order to show how variation serves as a communicative resource. Our findings show that these adolescents’ remarkable linguistic proficiency and communicative competence contrast markedly to their low educational and professional status.

1. Aim of the study The paper is based on earlier work on “German-Turkish variation and the construction of social styles of communication in young migrant groups of Turkish origin in Mannheim”, which used ethnographic, sociolinguistic, and conversation analytic methods of data collection and analysis, and was carried out at the Institute for German Language in Mannheim, Germany. This project focused particularly on those young migrants who were successful at school and who reached socially satisfactory positions in German society. But so far, very little has been written about the social and linguistic development of young migrants who are regarded by German society and German educational institutions as “school * We thank Ibrahim Cindark for his valuable and enlightening analytical comments. . For publications, see www.ids-mannheim.de/prag/soziostilistik/tuerkisch.html.

240 Inken Keim and Ralf Knöbl

failures” or as “social problem cases”. Our aim is to make a contribution to this issue, and provide greater insight into the life world and communicative practices of Turkish migrant adolescents who have grown up in a typical migrant neighbourhood in Mannheim. The data on which our analysis is based are (a) an ethnographic interview with the seventeen-year-old Murat who, as his friends join in, switches to a peer conversation with the ethnographer and (b) an interview with a TV reporter with the same young man and his friends two years after the first interview. The aim of our analysis is – to present the linguistic repertoire this adolescent displays when he is with his friends in the first conversational situation and to reconstruct the discursive, interactive, and social functions of characteristic variational patterns; and – to describe the adolescents’ conversational performances in a more formal situation two years later. After a short introduction to existing research on ethnolects (Section 2), we describe the young migrant’s life world and his educational development, as well as the phonetic and prosodic characteristics of his German (Section 3). Section 4 presents the characteristic variation patterns used by the migrant in conversation with the ethnographer as well as in conversation with his friends. In Section 5, we show that in the second type of data, some new linguistic phenomena can be observed.

2. Ethnolectal varieties of German In the last few decades, (multi)ethnolectal varieties of standard languages have developed in migrant contexts in many European cities. In Germany, Füglein (1999) was the first to describe these for Bamberg and Munich. She called these “socially marked” ethnolectal varieties “Kanak Sprak”. Dirim and Auer (2004: Chap. 6.9) observe that adolescents in Hamburg use ethnolectal varieties, young migrants as well as young Germans, and Wiese (2006) describes ethnolectal forms in Berlin. In the inner city of Mannheim, ethnolectal forms have developed as a kind of lingua franca between children and adolescents of different linguistic backgrounds. University students from Turkish backgrounds regard these forms as socially

. Füglein (1999) adopts the term “Kanak Sprak”, which was introduced by Feridun Zaimoğlu (1995), a German writer of Turkish origin. . Auer’s term (2003) for ethnolectal forms of German used by migrants of Turkish origin is “Türkenslang”.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 241

marked and call them Ghettodeutsch (“ghetto German”); German teachers call them Stadtteilsprache, (district variety), and the speakers themselves call them unser Ghettoslang (“our ghetto slang”). The (multi)ethnolectal varieties in Mannheim are simplified versions of the regionally spoken German Standard where some characteristics occur quite regularly, others rather infrequently. The following characteristics occur fairly regularly: a. Deletion of preposition and article in locatives and directives as in: isch muss toilette, isch geh schwimmbad (‘I must toilet, I go pool’). b. New constructions with the verb ‘machen’: isch mach disch krankenhaus (‘I make you hospital’ = ‘I’ll beat you up so badly that you have to go to the hospital’). c. The use of formulas such as isch schwör (‘I swear’) for confirmation and isch hass des (‘I hate it’) for a negative evaluation. d. The use of Turkish lexemes as address forms (lan, moruk, ‘man’, ‘old man’), in abuse (siktir lan, ‘fuck you, man’), and as interjections or discourse markers. e. A specific way of processing information that presupposes a large amount of shared knowledge and makes explicit only a small part of the background knowledge that is necessary for the addressee’s understanding (see Kallmeyer & Keim 2003; Keim 2008: 451ff.). f. A complex of prosodic and phonetic characteristics (Auer 2003; Kern & Selting 2006; Kern this volume; and Section 3.2 below). The following characteristics occur less frequently:10

. See Keim (2008), Part III, 2.1. . At the Hauptschule of the district, a German student answers the interviewer’s question about his speaking German in a strange way as follows: das ist eben unser Ghettoslang (“that is our ghetto slang”), see Oberle (2006). . For similar characteristics, see Auer (2003) and Wiese (2006). . The correct German forms are: ich muss auf die Toilette, ich geh ins Schwimmbad. . A correct German form would be: ich schlag dich krankenhausreif. Wiese (2006) describes in detail the ethnolectal utterance ich mach dich messer. . Siktir is a shortened version of siktitir git (‘let yourself be fucked and beat it’). 10. These features are also described by Auer (2003), Dirim and Auer (2004), and Wiese (2006).

242 Inken Keim and Ralf Knöbl

g. Deletion of the article in noun phrases and in prepositional phrases: gib mir kippe; isch war schlechteste; bevor=sch von klassenzimmer rausgeh11 (‘give me cigarette; I was worst; before I leave classroom’). h. Deletion of the pronoun, for example: wann has=du fotografiert instead of wann has=du sie fotografiert (‘when did you take a photo of her’). i. Changed gender: rischtiges tee (neuter instead of masc.), meine fuß (fem. instead of masc.). j. Other grammatical cases: wenn=sch mit ihm heirate instead of wenn ich ihn heirate (‘when I marry him’); er geht mir fremd instead of er geht fremd (‘he is unfaithful’). k. Changed word order: hauptsache lieb isch ihm instead of hauptsache ich liebe ihn (‘the main point is, I love him’). (Mulit)ethnolectal varieties have developed and stabilised in multilingual groups, e.g., in school classes, youth centres, and sports or music groups where Turkish adolescents play an important role.

3. Life world and linguistic development of the “Ghetto” adolescents – as exemplified by Murat Our ethnographic research was carried out between 2000 and 2004 in an inner city district of Mannheim, an industrial town of 320,000 inhabitants in southwestern Germany. Over 21% of Mannheim’s population are migrants,12 most of them of Turkish origin. The district under investigation, traditionally a working class district, has a migrant population of over 65%; inhabitants of the district as well as outsiders call it a “migrant ghetto”. The children’s first contact with German, usually, is in kindergarten with their teachers. But since in some educational institutions of the district up to 100% of the children have a migrant background, the chances of obtaining sufficient Standard German input are low. As a consequence, the children develop a morphologically and lexically reduced German variety mixed with elements from other languages. When they start school, they usually do not have a high degree of competence in Standard German. This situation has serious consequences for the children’s educational careers.

11. In correct German: gib mir eine Kippe; ich war die Schlechteste; bevor ich aus dem Klassenzimmer rausgeh. 12. In October 2004, 67,000 migrants from 177 nations were living in Mannheim, most of them in inner city districts; see Mannheimer Morgen, 26.10.2004, Nr. 249, p. 20.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 243

One of the findings of our ethnographic research is that migrant children develop different socio-cultural orientations and communication practices depending on their school careers.13 Children who have the opportunity to go to the Gymnasium or the Realschule (10–20% of an age group) – both types of schools are situated outside the district – enter at the age of ten German educational worlds where migrants form quite a small minority. But most children go to the Hauptschule. This has implications for a ten-year-old child, since the Hauptschule is situated in the “ghetto” territory, and so s/he will be growing up in an environment and in peer groups where German-Turkish mixing or highly marked ethnolectal German varieties are the normal codes of interaction. When the adolescents leave the Hauptschule at the age of 15 with low marks or without any qualifications, as 25–30% of the students do, they have almost no chance of obtaining a professional qualification.14 These youths typically develop an antieducational attitude, see themselves as “losers”, and align with other “ghetto” peer groups, where members are proud to be school failures. 3.1

Murat, a typical “Ghetto”-youth

At the time of the data collection, Murat was 17 years old, unemployed, and looking for a job. Murat was born in the migrant district and went to primary school there as well as to the Hauptschule. At the Hauptschule, he found most school subjects boring, and life after school was much more interesting. The boys roamed the streets, played computer games at the internet cafe, hung about in the playgrounds, or watched videos at Murat’s house. They started to skip lessons and get poor marks, and left school with very poor grades. The central interest in Murat’s life is his clique with its (partly illegal) activities. The boys see themselves as “Turks” and they are determined to marry only a Turkish girl (from the district or from their parents’ home regions). Since they cannot imagine returning to their parents’ home country, they decide to live in Mannheim. Murat is eloquent, speaks German fluently, and has a wide linguistic repertoire including Standard German, Turkish and German-Turkish mixings, as well as ethnolectal forms.

13. For a detailed description, see Keim (2008: Part I). 14. Only 16% of the graduates from a Hauptschule in Mannheim receive professional training; see Mannheimer Morgen, Nr. 281, 03.12.2004, p. 20.

244 Inken Keim and Ralf Knöbl

3.2

Phonetic features of Murat’s German

The following description of phonetic features is based on data derived from the interviews with Murat. We focus on features that remain salient in the variety orientated to German Standard forms, which is the variety predominantly used in the interviews. The standard of reference for the description and classification of salience (in the non-technical sense of “foreignness”) is the spoken Standard German of the Mannheim area. Murat pronounces after front vowels consistently as a palato-alveolar [∫].15 Although this feature is salient with reference to Standard German (with the palatal fricative [ç] in this position), it is not salient in relation to the regionally spoken Standard of Mannheim. But in contrast to typical Mannheim forms, Murat’s articulation of the prepalatal fricative is stronger (more forte), longer in duration, with a higher degree of lip-rounding, and more palatal than dialectal [ɕ]. A further difference to regional forms is apparent in the realisation of short front vowels preceding the prepalatal fricative, especially the vowel in the pronoun (‘I’) and the intensifier (‘really’), which Murat uses frequently. He articulates /e/ and /i/ before not as tense vowels, which would conform to the regional Standard, but as lax, centralised vowels. This is striking because Murat tends to articulate short vowels in other positions in conformity with the regional norm as tense vowels. This indicates that the ‘ideal’ Standard German vowel system with the correlation between openness and duration of vowels is dissolved (see also Auer 2003). With this, the distinction between short and long vowels is levelled out, which is still perceptible in the Mannheim vowel system. Frequently, Murat deletes /t/ in final position. He does so not just in the frequently occurring function words (third person of ‘to be’) or (‘not’), where t-deletion is part of conventionally spoken Standard German, but he tdeletes almost consistently and independently of the lexical and morphological element involved. In cases where Murat pronounces final /t/, the articulation is without strong release and aspiration of the plosive. The affricate [ts] is reduced to /s/, e.g. in (‘to’), and in general, the articulation of obstruents is more voiced. In relation to the Mannheim reference forms, voiced /s/ (even in final position) and plosives are salient. In particular /d/ and /b/ positioned between vowels are articulated with a very long phase of obstruction followed by a soft release involving a high degree of vibration of the vocal cords. The tendency not to strengthen consonants (e.g. [z] and non-aspirated /t/) in syllable-final positions does not correspond to the process of German Auslautverhärtung. Voiced and 15. The so-called ‘coronalisation’ is also documented as the only consistently occurring feature in the data of the “Turkish Power Boys” (Tertilt 1996, analysed in Auer 2003).



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 245

lenis obstruents positioned before morpheme boundaries have bearings on the constitution of junctures between words. Murat tends not to segment word-units with clear boundaries; in addition to the tendency not to devoice final consonants, vowels at the beginning of morphematic units tend to set in weakly, without glottal stop (see also Auer 2003). The place of articulation of /a/ is often lower and more central than a Standard German /a/. Even in accented positions, the /a/ tends to have the quality of a deep-central schwa [ɐ]. This means that the a-sound both in (potentially) stressed positions and in unstressed positions is levelled out not only in terms of quantity (duration) but also in quality (timbre). Murat’s pronunciation of /r/ essentially conforms to the regional and general conventions of spoken Standard German. Like locals without a migrant background, he pronounces the /r/ that does not occur in syllable-final position almost consistently as a uvular fricative.16 At the end of syllables, /r/ is vocalised.17 To be more precise, Murat does not vocalise but simply deletes the /r/ consistently in positions following /a/, e.g. in: [a(:)baɪt]), which conforms to common Standard German conventions. He also deletes /r/ after /ε/ and rounded vowels, e.g. [lε:liŋ] or [vu(:)dn̩], which is salient with reference to Standard German articulation but not with reference to regional Standard forms. Real vocalisation, the production of as a vowel, takes place in positions following the front vowels /i/ and /e/, when the syllable carries the word-accent, so that / are articulated as diphthongs [iɐ]/[eɐ]. Neither deletion nor vocalisation of /r/ is salient in the Mannheim system. But what is striking, and different from the regional forms is the prominence – above all based on duration – of the vowel generated, especially in the position of unstressed , , e.g. in [nomalɐvaɪzә]. The prominence of syllables, which are not – according to Standard German conventions – stressed, or, more generally, the fact that there are no structural provisions for the reduction of linguistic forms, is a basic feature of Murat’s German. On the one hand, non-reduced syllables in unstressed positions result in ‘strange’ accent structures at the level of word-stress, e.g. by Murat’s unreduced articulation of the vowel in the prefix . On the other hand, the accent structure at the level of the intonational phrase and especially rhythm are moulded by unreduced forms, in particular by the fact that Murat scarcely uses articulatory

16. In the data analysed, there are two exceptional cases; in both cases, the /r/ is pronounced as an apical flap in the syllable-initial consonant cluster . 17. The only exceptional apical production of syllable-final /r/ is used when Murat is articulating a Turkish first name of a friend of his (Erkan).

246 Inken Keim and Ralf Knöbl

reduction forms of function words (weak forms in the sense of Kohler 1995; see also Kern this volume). This feature of missing reduction forms is the basis of a very salient ‘ethnolectal’ phenomenon. It is the fast beat of the ethnolect in the sense of the orientation to a syllable-timed rhythm. This means that typically it is not the time intervals between accentuated syllables that tend to be isochronous, but the intervals between the onset of single syllables.18 The iso-syllabic rhythm has implications for focus projection. Especially at the phrase level, Murat shows ‘unmotivated’ or random markings of focus domains in the sense of accents that do not correlate with the information structure of the utterance, e.g., accents marking functional elements such as prepositions or pronouns instead of content elements. At the phrase level, a recurrent intonation contour is salient. The tonal progress of Murat’s most typical contour rises to an early peak on the second or third syllable of the intonation phrase and the fundamental frequency is kept on a relatively high level with a low pitch range.19 Murat applies the contour like a default form, which in many cases seems not to blend with the information structure of his propositions – that is, without giving intonational cues that help to contextualise the utterance. The interpretation of the information structure is even more difficult because there is hardly any support from the pause structure. Within the high level contours, Murat tends to articulate relatively long units without clear boundaries and pauses. Most segmental and prosodic features are interrelated and accomplish a holistic ‘Gestalt’, which is seen as conspicuous in relation to the German system of reference. From the perspective of Gestalt theory, one could say that ‘the whole’ of Murat’s ‘foreign accent’ is probably more than the sum of single features.

4. Linguistic variation and communicative practices of the seventeen-year-old Murat The interview with Murat took place at the district’s central playground and lasted for almost two hours.20 Murat describes his family’s situation, his educational career as a poor student, crucial experiences at school, his life with his friends, and 18. See Auer and Uhmann (1988) for a discussion of the “hypothesis of isochrony” (Aber­ crombie 1967: 96f.). 19. See Keim and Knöbl (2007) for a description and examples of this recurrent intonation contour. It is comparable to hat-contours (see von Essen 1964 or Peters 2005). 20. Murat knew the interviewer and agreed immediately to the interview.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 247

his ideas about his future life. In the course of the interview, three of Murat’s friends join in. In this conversational phase, Murat talks about a fight that he had with another adolescent. In the course of the whole conversation, Murat displays a wide range of variation. Specific variation patterns that serve specific interactive and social functions can be differentiated. In order to demonstrate Murat’s variational skills, we have selected maximally contrasting conversational phases: – the beginning of the interview with the ethnographer where Murat presents his professional situation; this is the most formal part of the interview (Section 4.1); – the welcoming of his friends with characteristic forms of peer-group communication (Section 4.2); and – the narration of a fight addressed to the ethnographer as well as to his friends (Section 4.3). 4.1

Beginning of the interview: Murat’s professional situation

Murat starts with the statement that he has no employment. Despite his poor Hauptschule qualification, he got an apprenticeship as a motor mechanic, but he did not like the company. In this part of the narration, Murat uses near-standard German, and his speech displays characteristic features of spoken German:21 (a) the post-positioning of structural parts, and (b) strings of main clauses with verbfirst position. a. Post-positioning of complex structural parts: – also=sch bin jetz siebzehn * hab mit fünzehn die hauptschule beendet- * (…) mit=nem schleschtn durschschnitt- (…) drei komma vier- (‘well now I am seventeen * finished the Hauptschule at fifteen * with a poor average grade * three point four’) – → hab abba trotzdem ne ausbildungsstelle gefundn als kfz mechaniker in Ludwigshafen- * bei einem türkn- ← (‘but nevertheless got an apprenticeship as a motor mechanic in Ludwigshafen * at a Turk’s’) The post-positioned elements are treated as independent prosodic units separated by a pause, or as units that are integrated at the prosodic and syntactic level into the previous structure. According to Kern and Selting (2006), these accentuation structures in post-positionings are typical forms of spoken German where the

21. For a detailed overview of characteristics of spoken German, see Schwitalla (2006).

248 Inken Keim and Ralf Knöbl

preceding structures always carry the main (sentence) accent. The single construction units and the preceding structures, as well as the post-positionings, all end with level intonation.22 This, too, is a common means in oral narrations and complex descriptions of producing a coherent global structure. b. Strings of main clauses with verb-first position: – also=sch bin jetz siebzehn ← ** hab mit fünfzehn die hauptschule beendet* (…) → hab abba trotzdem ne ausbildungsstelle gefundn (‘well, now I am seventeen ** finished the Hauptschule at fifteen * […] but nevertheless got an apprenticeship’) These features, too, are characteristic of oral German narrations. Stringing together main clauses with verb-first position is described by Günthner (2006: 99ff.) as “uneigentliche Verbspitzenstellung” and as “dichte Konstruktionen”. They are, especially, used in those parts of oral narratives that call for highly detailed and intensive descriptions. Murat’s use of these features is motivated by his task, which is to present in detail his development from school age up to his current situation. The semantic agent is established from the start (=sch or isch, ‘I’) and projected into the succeeding narrative sequences. The use of parallel syntactic structures and level pitch function as central elements for the production of narrative coherence and give Murat’s presentation a dense and intensive quality. In his last utterance, there is also post-positioning: (‘I did not like it very much there, the company’). When the ethnographer asks why he did not like the company, Murat gives two reasons. Thereby, he even comes very close to written German: – the first reason is that he felt exploited: weil isch bissl zu zu musste un=die überstundn wurdn nisch bezahlt (‘because I had to work too too hard and the overtime was not paid’); – the second reason is that the actual conditions of his apprenticeship did not correspond to the legal regulations. At first, he refers to these regulations: → norma“lerweise darf ja ein lehrling * nischt mehr wie siebeneinhalb stundn am tag arbeitn↓← (‘normally, a trainee is not allowed to work more than seven and a half hours a day’); and then, when comparing his actual working conditions with them, he emphasises their illegal quality: un isch hab über ze“hn stundn am tag gearbeitet↓ * un manschmal sogar über zwö“lf stundn↓ (‘and I had to work more then ten hours a day and sometimes even more than twelve hours’).

22. See Peters (2005: 122f.) for a description of level-pitch-contours.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 249

As an inevitable consequence, Murat terminated his contract (syntactically connected by the particle da = ‘then’, louder voice and strong accentuation): (‘then I handed in my notice’). This segment consists of complex syntactic structures without the characteristic structures of spoken German described above, such as post-positionings and main clauses with verb-first position. Of course, there are also features of spoken German, such as word search (zu zu vie“l, ‘too too much’), elisions and clitical units (un=die), as well as the southern German form bissl (‘a bit’). But all together, this way of speaking is clearly oriented towards a literal presentation.23 But on the phonetic and prosodic level, all the features described above (3.2) that signal “foreignness” occur: – tense articulations of short vowels in [be(|)endet] (‘finished’), [nomalɐvaɪzә] (‘usually’); – lax realisation of ‘long’ vowels: [lε(:)liŋ] (‘trainee’); – reduced front vowels preceding alveo-palatal articulations of : [ә∫:] (‘I’, transcribed as isch); – levelling out of vowel quantity: [∫ule] (‘school’);24 – prominence of syllables in unstressed positions – on the one hand, caused by vocalisation, as in [ɐbɐ] (‘but’), and on the other hand, in cases without vocalisation such as [be(|)endet] (‘finished’); in all these cases, the vowels in stressed and unstressed position are qualitatively and quantitatively levelled out; – t-deletion: [mo:na], [gәkʏndɪ∫] (‘month’, ‘terminated my contract’); – iso-syllabic orientated rhythm, salient accent structure and focus marking.

23. For example, in the utterance norma“lerweise darf ja ein lehrling * nischt mehr wie siebeneinhalb stundn am tag arbeitn↓ the final plosive in (‘not’) is not deleted – which can be considered as orientation to written Standard forms in contrast to conventions of the spoken Standard, where the /t/ in is regularly deleted. 24. In the example , the vowel /u/ – which is supposed to be long and carrying the word-accent – continues for 0.042 seconds; it is shorter than the second vowel at the end of , which is supposed to be in an unstressed position but continues for 0.053 seconds.

250 Inken Keim and Ralf Knöbl

4.2

Two greeting rituals: Regular forms vs. peer group forms

When Murat’s friends arrive, two greeting rituals are performed: (a) the ritual between Murat’s friend Serat and the ethnographer (01/05); and (b) the ritual between Murat and Serat (06/09):25 01 SE: hallo gut hello fine 02 IN hallo: ** na wie geht=s dir↑ hello well, how are you 03 IN: oder ihnen →ich muss ja junger mann I should say ‘Sie’ and call you a young 04 IN: 05 SE: 06 MU:

|sagn← LACHT| gut↑ man LAUGHS fine mhm jayes |alles klar↑| nerden gelyon everything ok where do you come from

07 SE: Heddesheim Edeka * işim Heddesheim my Edeka job 08 MU: n=aptın what did you do 09 MU: ja- un was geht |dreiun|zwa|nzigs|ta↑| and what’s up twenty-third

a.  Ritual between Serat and the ethnographer (01/05) Serat opens the ritual with the formula hallo (‘hello’), and the ethnographer (IN) answers with the second part hallo: (‘hello’), thereby manifesting that she sees herself as the addressee of the informal greeting ritual. Then, she continues with a further greeting formula na wie geht=s dir (‘well, how are you’, 02), characteristic of rituals between good acquaintances. This is answered by Serat with a 25. For transcription conventions, see Appendix. Note that transcription lines are organised like in a musical score. Scores are separated by a continuous line. So, lines 01 and 02, and lines 04, 05 and 06 form a score. In a score, successive contributions are represented from left to right. So, in line 01, after SE has finished his utterance hallo, IN responds with hallo: ** na wie geht=s dir. Instances of simultaneous talking are presented one below the other in the score and marked by |, as for instance in lines 04 and 06. The Turkish elements are in bold letters.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 251

corresponding part: gut- (‘fine’). Then the ethnographer repairs the addressing pronoun du to the more official Sie and addresses Serat as a ‘young man’ (03). With this playfully modalised self-repair, she defines the relation to Serat as more formal. The greeting ritual ends with the ethnographer’s reformulation gut↑ (‘fine’, 04) and Serat’s further confirmation ja- (‘yes’, 05). b.  Ritual between Murat and Serat (06/09) The welcoming between the adolescents is quite different. It is initiated by Murat with the greeting formula alles klar↑ (‘everything ok’, 06) characteristic of welcoming among (male) friends. Serat gives a short, positive reply (mhm, 05), spoken in a low voice. Then, with a switch into Turkish, Murat asks for information about his friend’s latest activities: nerden gelyon- (‘where do you come from’, 06), which is answered by Serat’s naming a village near Mannheim: Heddesheim (07). To Murat’s second request: n=aptın- (‘what did you do’, 08), Serat answers in Turkish, too: Edeka * işim (‘my Edeka job’, 07). Then, Murat changes the topic and asks about a future event, thereby switching into ethnolect: ja- un was geht dreiunszwanzigsta↑ (lit.: ‘what goes twenty third’, meaning ‘what’s happening on the twenty-third’, 09). The characteristic ethnolectal feature of this construction is the use of the temporal adverb without preposition and article (dreiunszwanzigsta instead of am dreiundzwanzigsten).26 The formula has the format ‘was geht + X’, where X can be either an address form (such as lan or Mann in ‘was geht Mann’), a local adverb (such as jungbusch in ‘was geht Jungbusch’) or a temporal adverb as in was geht dreiundzwanzigsta. The welcoming ritual between the adolescents consists of minimal utterances with just as much information as is necessary for continuing the exchange. The interlocutors switch from colloquial German into Turkish and then into ethnolect. The exchange is produced in a plain, reduced, and monotonous way of speaking. In contrast to the greeting ritual between Serat and the ethnographer, where especially the ethnographer’s phrases have an expressive quality, the two adolescents perform a welcoming interaction that is characterised by short utterances with minimum content, code-switching, and a reduced way of speaking. They perform a peer-group interaction thereby presenting themselves as “cool Turkish ghetto-boys”.27

26. Furthermore, in dreiunszwanzigsta↑ (‘twenty-third’) the ethnolectal feature of prominent vowels instead of /r/ in word final is very salient. 27. For the migrant male youth’s self-characterisation as cool, see Keim (2008: Part I, Chap. 4).

252 Inken Keim and Ralf Knöbl

4.3

Narrative of a fight: Double addressing and linguistic variation

As one of the friends pulls brass knuckles from his pocket, ‘fighting’ becomes the new topic. When the ethnographer asks if Murat is often involved in fighting, he at first vehemently rejects such an idea, but when Serat contradicts him, he admits that he had a fight quite recently. This confession induces the ethnographer to ask for further details: the reason for the fight was Murat’s former girlfriend who left him for another young man. This young man provoked the fight. When the ethnographer asks for the outcome of the fight, Murat – prompted by his friend Serat – has to confess that he ran away. As an explanation for this exceptional behaviour, Murat starts the fight narrative. It consists of various segments differing on the prosodic as well as on the syntactic and phraseological level. We have chosen the first segments of the narrative for a detailed analysis of the interactive and discursive functions of the linguistic variation, as well as for a description of Murat’s wide linguistic repertoire. a.  Background information explaining his running away (528–532) Shortly before the fight took place, Murat had broken his foot and was in no condition to fight:28 528 MU: 529 MU:

also * bevor wir gekämpft habn * eine * well before the fight took place * one nei“n * ja“ ei“ne woche davor wurd isch no yes one week before, I had an

530 MU: operiert↓ |hie“r | * >e=dinge=ding< (‘eh=thing’, 530), spoken in a soft voice. According to Kern and Selting’s (2006) analysis, ding or dings as a substitute or placeholder for names, nouns, adverbs, and verbs is characteristic of the ethnolectal German that the authors call “Türkendeutsch” (ibid.: 336). Typically, the element replacing the placeholder ding is produced in a prosodically independent postposition. In most instances of our data, the proform ‘ding’ is accentuated. It occurs predominantly in event descriptions with high involvement, and is followed by a short pause; the post-positioned supplement is produced as an independent prosodic segment.29 In comparison to this structure, the use of ‘ding’ in line 530 does not have the function of a placeholder but rather the function of a hesitation signal, >e=ding< being produced with a lower voice, faster, and without accent, and the postpositioned specification follows without a pause. Thus, di“ng * (with accent and following pause) in the function of a placeholder does not belong to the formulation pattern which Murat uses in background descriptions addressed to the ethnographer; in fact, it does not occur in background sequences in the entire conversation. b.  Start of the fight narrative (533–536) The start of the fight narrative leads ‘in medias res’. Murat starts his next utterance with the temporal/modal adverb auf einmal and changes to a dramatising formulation pattern: 533 MU: auf einmal↑ * un dann halt bin isch at the same time, and then, I just go out of

29. Murat uses di“ng * as a placeholder for a noun, e.g., in isch bin in=s di“ng- * in=s büro von dem- (‘I went to the thi“ng * to the office of the-’), and a placeholder for a quotation: dann hat mein lehrer gesagt di“ng- * eigentlisch will isch ja jedem helfen […]. (‘then my teacher said thing * usually I want to help everyone […]’).

254 Inken Keim and Ralf Knöbl

534 MU: 535 MU: 536 MU:

rau“s↑ * hab noch krü“ckn- * un dana“ch the house, am still walking on crutches and * hat=a gemeint di“ng↑ * n=kampf mit mir↓← un=isch anxious to fight with me, and I

The sequence un dann halt bin isch rau“s↑ * hab noch krü“ckn- (‘and then, I just go out of the house, am still walking on crutches’, 533/534) is characterised by the use of the historic present tense, short syntactic structures, and verb-first position in the second part. This formulation pattern is characteristic of a dramatised portrayal at the climax of a narrative. But when he starts to present the action of the adversary, there is a change of focus and perspective: un dana“ch * hat=a gemeint di“ng↑ * (‘and then, he said thing’, 534/535). Instead of the rival’s quote as it is projected by the quote introduction and the placeholder di“ng↑ *, he sums up the adversary’s motive: n=kampf mit mir↓← (‘he just was very anxious to fight with me’, 535/536). With this, Murat switches to a formulation pattern characteristic of background information. That means that he repairs the early start of a dramatising description by filling in background information and explaining why he accepted the fight in spite of his serious injury. c.  Explanation of motives (537–542) Then, Murat gives two reasons for accepting the challenge to fight in spite of his injury: first, he had to stand up for his honour; second, he took precautions and, for his safety, took a club along:30 537 MU: 538 MU: 539 MU: 540 MU: 541 MU: 542 MU:

un=isch konnt net sagn isch kann jetz net and I could not say I am not fit to fight right now weil mein fuß gebrochn is odda so * bin because my foot is broken or something like that isch trotzdem hingegangn obwohl mein fuß noch I went there even though my foot still was zusammngenäht war * da hab=sch halt zur sischerheit sewn up and bandaged, but I took a club with me ein schlagstock mitgenomm↑ * →falls es for my safety, in case something really wirklisch schiefgehn sollte odda so← should go wrong or things like that

30. This is a shortened version of the original transcription.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 255

Murat presents his accepting the challenge as completely natural, downgrading it (537/539) by hedges and a matter-of-fact way of speaking. Thereby, he characterises himself as a young man of honour who has to accept the challenge of a rival, even though he is not sufficiently fit for a fight. Then, he admits that he did not march off to fight unprotected (540/542). These utterances are characterised by complex syntactic structures, hypotaxes with the conjunctions weil, obwohl and falls, as well as the conditional past tense in schiefgehn sollte. This sequence, where Murat explains his motives for accepting the fight challenge, comes close to German Standard and is elaborately formulated. d.  Description of the fight (543–549) After this background information, Murat leads back to the description of the events by presenting the next action of his adversary. With this, he switches again into a dramatising formulation pattern: 543 MU: 544 MU:

der is halt auf misch drauf↑ hab=sch he throws himself on me I pulled out schlagstock rausgezogn * hab=sch ihm vom the club I grabbed him by

545 MU: hals so gepackt=nach hintn hab the collar backwards I 546 IN: oh * aha * 547 MU: 548 MU: 549 MU:

isch den gepa“ckt↑ ** ↓ (…) ham den bissl zusammngetre“tn↑ * on him, kicked him around a bit and beat him up

As the adversary jumps on Murat, he draws the club out of his pocket and grabs the adversary by the collar. At this moment, Murat’s brother and his friend, who by chance are passing by, grab the adversary and throw him down because they want to help the injured Murat. Portraying the fight, Murat again uses the linguistic devices of a dense narrative description: progressing in small steps, sharpening of the main point, and using main clauses, short syntactical units with ligatures and deletion of sounds, and, above all, there are grammatical deviations, such as the deletion of the article

256 Inken Keim and Ralf Knöbl

in the noun phrase schlagstock (544)31 and the wrong preposition in the prepositional phrase vom hals so gepackt (544/545, instead of am hals). In the research literature, such features are described as characteristics of ethnolectal German. It is interesting that Murat uses them here, in the dense narrative descriptions. The verb-first positionings (with inverted subject pronoun hab=sch, 543, 544, 545) in this segment as well as in Murat’s other conversation material are not ethnolectal features. They can be found in narratives of the dialect-speaking Mannheim population as well. At the climax of the narrative, unexpectedly, two new characters appear, Murat’s brother and his brother’s friend, who help him fight against the adversary. This formulation (sin=die dann glei auf ihn drau“f↑← ‘they jumped on him’, 548/549) as well as the result of their help (ham den bissl zusammengetre“tn↑ * ‘kicked him about a bit and beat him up’, 549) are in the dense narrative pattern, too: short syntactical constructions, verb-first positioning, ligatures, and series of intonation units with level pitch at the end of high-level intonation contours.32 These features contrast clearly with the formulation pattern that is characteristic of background descriptions with the more complex syntactic structures, no verbfirst positioning and clearer articulation.

Summary The narrative consists of two formulation patterns that occur in different structural parts: – dense foreground presentations with detailed event portrayals progressing in small steps are achieved by short syntactical units (main clauses), verb-first positioning, sometimes the use of the historic present tense, ligatures, and, especially, ethnolectal grammatical features; – background presentations, such as explaining of motives or giving reasons for actions, are realised by complex syntactical structures (causal, conditional, and concessive clauses), usually without grammatical deviations. Murat displays a high narrative competence, presenting a well constructed and highly convincing narrative. Since ethnolectal grammatical features only occur in foreground presentations, the question of their functionality arises. This question

31. The same noun phrase was produced with an article in the previous background presentation da hab=sch halt zur sischerheit ein schlagstock mitgenommn (‘but I took a club with me for my safety’, 540/541). 32. Furthermore, in this segment, an exceptional apical articulation of /r/ (in the word-initial consonant cluster , 549) occurs, which intensifies the ethnolectal impression.



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 257

can be answered if one compares the narrative that is addressed to the ethnographer as well as to his friends with other narratives where only the ethnographer is present. In one of these narratives, Murat describes a confrontation with his teacher that was crucial for his further school career. In this narrative, all the characteristics of dense descriptions are displayed (strings of short syntactical units, verb-first positioning, historic present tense, self- and other-quotations). But there are no ethnolectal grammatical features. Comparing these two types of narrative gives rise to the suggestion that the use of ethnolectal forms in the fight narrative depends on the specifics of the actual situation, namely Murat’s addressing the ethnographer and his peers at the same time. Murat has to cope with two different interactional demands; he has to take into account the different background knowledge of the ethnographer, who knows nothing about the events, and his friends, who do know about them (that is made clear at the beginning of the narrative). And above all, he has to cope with two different social demands: establishing an intimate relationship with his friends and a more distant social relationship to the ethnographer. These tasks are performed by the use of different stylistic means. In the greeting ritual among friends (see above 4.2), the intimate relations of the peer group are indexed by German-Turkish mixing and ethnolectal forms. In the narrative, ethnolectal forms occur only in the presentation of the fight, especially in those parts which are also addressed to his friends. More complex and elaborate linguistic forms occur in the presentation of background information necessary for the ethnographer’s understanding of Murat’s motives.33 That is, more complex and elaborate linguistic forms used vis-à-vis the outsider of the group signal a friendly but socially distanced relationship.

5. Murat at nineteen: New linguistic and communicative forms Two years later, Murat displays some new linguistic and communicative forms, namely Mannheim dialect, as well as prosodically and phonetically unmarked Standard German, in a similar situation. It is again an interview situation, but it is much more formal and requires a high degree of linguistic reflection. A reporter from SWR3 (a regional TV station) wants to produce a TV report about the migrant youth group to which Murat belongs and asks the adolescents about their language use in front of the cameras. The interview starts with the reporter asking whether “Kanaksprak” – the communicative behaviour of a migrant youth as it is constructed by the mass 33. Some of this background information is elicited by the interviewer, and she reacts to it; see the complete version of the transcript in Keim and Knöbl (2007).

258 Inken Keim and Ralf Knöbl

media and performed by comedians such as Erkan and Stefan34 – was echt oder show (‘real or show’). The group members ask Murat to answer for them. Murat evaluates the comedians’ productions negatively, and he states very explicitly that the performances of the comedians are nothing but an unsuccessful attempt to imitate young migrants’ linguistic practices because they cannot speak Turkish: 58 MU: 59 MU: 60 MU: 61 MU:

also wi“ssn sie- ** →wenn man zwei drei wörter well listen if one knows two or three words tü“rkisch kann und dann versucht irgendwie in Turkish and tries to imitate the die kana“knsprache zu machn← * dann ge“ht language of the kanakas that does des net↓ ** man muss schon was * drauf habm not work one really has to know Turkish

62 MU: auf türkisch↓ ja“ ** very well yes 63 ÜM: ja“ ** yes 64 XM: >hm * stimmt< hm that’s right

Murat calls the young migrants’ way of speaking “Kanakensprache” (“the language of the kanakas”), and, in contrast to the term “Kanaksprak” used by the reporter in order to refer to the ethnolectal construct of the media, he defines “Kanakensprache” as German-Turkish mixing. His friends agree with him (58/64). When defining the group’s way of speaking, Murat uses a complex syntactic structure (if-then format), as well as prosodically and phonetically unmarked Standard German, e.g. reductions of unstressed function words and matching the accent structure with the information structure (58/61). The ideological background to Murat’s self-presentation is the stylised portrayal of young male migrants in the mass media (films and comedies) with the following features: they are uneducated, simple-minded and clumsy, they like to show off, they have rough manners, and they speak “Kanaksprak”, an artificial, (multi)ethnolectal German.35 Murat 34. The comedians “Erkan und Stefan” produce comedies in which they present the adventures of foolish young migrants speaking “Kanaksprak”, an ethnolectal construction by the media. See Keim (2003) for an analysis of the playful use of the “kanak”-media figures in young migrants’ natural conversations. 35. See Androutsopoulos (2007) for forms of media stylizations of young migrants; for a short description of “Kanaksprak”, see Keim (2003).



Linguistic variation and virtuosity of young “Ghetto”-migrants in Mannheim 259

contrasts his group with this social category by using a maximally contrasting way of speaking, namely a complex Standard German without a foreign accent. In doing this, he symbolises his disaffiliation from the category of a “Kanaksprak”speaker as well as his affiliation with German natives.36 After this performance, Murat emphasises his position by des=s stimmt * da hab=sch voll reschd (‘that’s right, there I am totally right’, 66/67). Together with this laughingly produced affirmation, he switches from Standard German to a way of speaking that contains elements of youth language (voll as an intensifying particle) as well as elements of the Mannheim dialect, such as the tense articulation of the vowels in voll reschd (‘totally right’, 66–67). Dialectality becomes even more salient in Murat’s reaction to the reporter’s following question: 69 RE: also krass alder würde nicht stimmen↓ so, ‘krass alder’ would not be right 70 MU: nää no 71 ÜM: nei“n no 72 MU: →kra“ss alder des is doch=kä but ‘krass alder’ is not Turkish 73 MU: 74 ÜM:

tü“rgisch← after all des khört nischt zu uns that does not belong to us

When the reporter asks if one of the most salient stereotypes of the “Kanaksprak” of the media, the interjection krass alder, is taken from young migrants’ speech, the young men protest vehemently: nää (‘no’), nein (‘no’) and des khört nischt zu uns (‘that does not belong to us’). In his rejection faster< ↑ (.) (0.6) Stress hhh Eng: Bold:

overlapping speech louder volume than surrounding utterances lower volume than surrounding utterances uintelligible speech parts I am uncertain about my comments prolongation of preceding sound faster than surrounding utterances local pitch raise short pause timed pause stress laughter breathe English translation lines in extracts Utterances pronounced with prosody characteristic of the urban youth style spoken in contemporary Copenhagen

Developing multiethnic youth language in Helsinki* Heini Lehtonen

University of Helsinki

This paper discusses the phenomena of youth language in multiethnic suburbs in Eastern Helsinki. The paper describes how adolescents make use of multiethnic social reality and the linguistic resources of their surroundings, i.e. different languages as well as varieties of Finnish. The possibility of the emergence of new multiethnic youth varieties or styles (Quist 2008) is discussed. The paper is mainly based on a sociolinguistic study conducted in 2002–2004 (Lehtonen 2004) and the data consist of participant observation, recorded interviews and other audio and video recordings. In the data, phenomena such as stylisation and language crossing (Rampton 1995) and a continuum from code-switching to language mixing (Auer 1999) can be observed.

1. Introduction Most sociolinguistic studies about (multi)ethnic urban youth language are performed among the 2nd or the 3rd generation of immigrants. This article sheds light on linguistic variation among adolescents in eastern Helsinki in a situation where only the first generation of immigrants has settled down in the neighbourhoods. Among the multiethnic suburban youth, different languages and ethnic varieties are used as interactional resources for stylistic purposes and mixing * I wish to thank Jyrki Kalliokoski, Hanna Lappalainen, Marja-Leena Sorjonen, Mia Halonen, Salla Kurhila, Marja Kokkonen, Liisa Raevaara and Johanna Vaattovaara for their help, advice and comments at various stages of writing this article. I would also like to thank the editors of the volume as well as the anonymous reviewers for their helpful comments. . By multiethnic youth language (cf. multietnisk ungdomspråk by Fraurud & Bijvoet 2004: 11) I here refer to all varieties used by adolescents for creating and expressing a multiethnic ingroup identity, or to linguistic practices and youth styles that make use of the local multilingual resources.

292 Heini Lehtonen

immigrant languages with Finnish functions as a way of showing solidarity to the multiethnic peer group. New ways of speaking develop to demonstrate the hybridity of the social contexts (cf. Coupland 2007: 107). The following example introduces two teenage boys who came to Finland as children: Mohamed from Somalia and Tarkan, a Kurd from Turkey. In the interview extract they describe how they use Turkish and Somali expressions when interacting with each other in Finnish. The excerpt is translated from Finnish by the author. (1) (Interview 2002. Mohamed (M, Somali, male), Tarkan (T, Turkish Kurd, male), interviewer H.) 01 H: do you sometimes, for fun, use each other’s languages? 02 M: yeah. 03 T: yes usually every- when I speak Finnish with him I mix- I mix it with Somali. 04 H: really? eh he 05 T: yeah, eh he like for example I say hey come here, ehm like in Somali and 06 H: yeah 07 T: and, let’s go eat 08 H: mm. 09 T: in Finnish, like that. 10 H: oh. 11 M: yes. 12 H: what does it- what does it feel like when he speaks Somali? 13 M: well it is totally normal that I- I’m used to him speaking Somali and I speak 14 Turkish, too. 15 H: when do you speak Turkish? 16 M: for fu- sometimes, for fun when I feel like it I- I say some word and so 17 H: yes 18 M: he answers, tries to answer in Somali or something. 19 H: well how does it feel when he speaks Turkish? 20 T: it is, I think, normal because I’m used to it.

Immigration to Finland accelerated as late as in the 1990s, and Tarkan and Mohamed can be seen as typical representatives of the children who came with the first larger immigration wave. I conducted this interview with them in 2002. As my surprised reaction in line 4 suggests, at that time even a student of linguistics was not familiar with the ways of speaking among the multiethnic youth in Helsinki. For Tarkan and Mohamed mixing expressions of immigrant languages with . Hybridity refers to a mix or a collage of signs or symbols from different social or cultural contexts that together create new social meanings (cf. Hinnenkamp 2003). On the term hybridity in general see Bhabha (1994).



Developing multiethnic youth language in Helsinki 293

Finnish had already become a normality (lines 13 and 20), but Finnish sociolinguists have yet not widely researched the linguistic practices that are typical for multiethnic and multilingual urban settings of late modernity. The recent formation of multiethnic and multilingual suburbs in eastern Helsinki allows us to follow the early stages in the development of multiethnic urban youth styles. In this article I will demonstrate the ways in which multiple linguistic resources are used by the multiethnic youth of eastern Helsinki. Before specifying the aims of the article, I shall describe the current linguistic and sociodemographic situation in Helsinki and the changes that have led to it. 1.1

Multilingualism in Finland and Helsinki since the 1990s

Since the beginning of the 1990s, the socio-demographic situation in Finland has gone through extensive changes (Statistics Finland 2008a): Whereas the number of foreigners in the year 1989 was slightly over 20 000, at the end of 2007 the number was 132 708, which is approximately 2.5 percent of the whole population (5.3 millions). Of these foreign citizens, 15 147 people were nevertheless born in Finland. At the end of 2007, 172 928 people (approximately 3 percent of the whole population) spoke a mother tongue other than Finnish, Swedish or Sámi (Statistics Finland ibid.). New ethnic groups have brought with them new languages which were formerly fairly unknown to Finns. A comparison of the statistics from 1990 with recent statistics shows the radical changes that have taken place over 18 years (Nuolijärvi 1991, 2005; STV 2003; Statistics Finland 2008b): At the beginning of 1990, the only non-European languages among the ten most spoken languages in Finland were Vietnamese and Arabic. Today, among the ten most spoken languages are also Somali (with over ten thousand speakers), Chinese and Kurdish (both with over five thousand speakers). The number of Arabic speakers has grown from less than a thousand speakers to over seven thousand, and the number of Vietnamese speakers from a little over one thousand to almost five thousand. The number of Russian speakers has grown from three thousand to over 45 000, the number of Estonian speakers from only 740 to over 20 000. Most of the immigrants have settled down in the capital area, in Helsinki and in the surrounding cities. The population of Helsinki at the end of 2007 was 568 . http://www.stat.fi/til/vaerak/2007/vaerak_2007_2008-03-28_tie_001_en.html 17.12.2008. . People with citizenship other than Finnish. . http://www.tilastokeskus.fi/til/vaerak/2008/vaerak_2008_2009-03-27_kuv_005_fi.html 14.6.2009.

294 Heini Lehtonen

531 people, of which 6.4 percent were foreign citizens and 9.0 percent spoke a mother tongue other than Finnish (84.9%) or Swedish (6.1%) (City of Helsinki). In certain parts of eastern Helsinki the number can rise to over 20 percent of the population (City of Helsinki Urban Facts). The number of immigrants in Helsinki is still fairly small compared to many cities in Europe, but considering the rapid change in the socio-demographic reality, especially in eastern Helsinki, it is worth asking whether and in what ways new kinds of multiethnicity and multilingualism have had an impact on everyday interaction in different social environments, such as in schools. 1.2

Data and aims of the article

The article is based mainly on a study (Lehtonen 2004, cf. also 2006a, 2006b, 2008) carried out in 2002–2004. The data were collected in two Helsinki junior high schools in 2002–2003. For approximately two months I attended lessons in the schools, observed the interaction around the schools, interviewed 13 pupils from different ethnic backgrounds and recorded or videotaped 6 pupils in different situations in the lessons as well as on breaks during their school day, e.g. on the school yard and in the school cafeteria. The interviews were also recorded. In this article the analysis of these older data will be supplemented by observations on the data from my current, larger research project. Between 2006 and 2009, I carried out ethnographic field work among multiethnic youth in schools. The new data consist of field notes, field diaries and recorded interviews as well as videos and tape recordings recorded by both the researcher and the adolescents themselves. Most of the adolescents I have been working with came to Finland as children or they belong to the second generation of an immigrant family. They were 13–18 years old at the time of the data gathering. The aims of the article are (1) to describe some of the linguistic resources that are used in a multiethnic eastern Helsinki school, (2) to identify the sociolinguistic variation and stylisation that makes use of the multiethnic and multilingual social reality, and (3) to discuss the possibility of certain linguistic phenomena to serve as a starting point for generally recognisable forms of multiethnic youth language, e.g. a Helsinki multiethnolect. . http://www.hel.fi 17.12.2008. . http://www.hel2.fi/tietokeskus/julkaisut/pdf/07_02_20_vuori_hki_alueittain_2006.pdf 17.12.2008. . On the term multiethnolect see Clyne (2000); Quist (2000); as well as Fox et al. and Freywald et al. (this volume).



Developing multiethnic youth language in Helsinki 295

2. Resources in interaction The pupils of a multilingual school are exposed to several languages every day, and different languages are used as resources for interaction (see Example (1) and Section 2.1). By linguistic resources I mean any bits of language(s) that the social contexts offer for speakers to use (cf. Quist 2005: 203–206; Blommaert 2008). Even in multilingual surroundings, linguistic diversity comprises much more than just the variation between different languages. There is inner variation in every individual language. The variety of an immigrant language spoken at home may differ from the standard taught in the lessons at school. For instance, there are several varieties of Somali spoken in Finland. Also, several varieties of Finnish are present in the everyday life of the pupils. Every dialect and variety as well as register or style, e.g. hip hop style, can function as a resource, and combinations of different resources form new ways of speaking; new styles and varieties (cf. Eckert 2005: 24; Auer 2007: 12). In the following sections I will illustrate the ways in which different linguistic resources are combined in interaction. In Section 2.1, I will show how lexical items of the Somali language can have different functions when brought into a predominantly Finnish conversation by different people. In Section 2.2, I will focus on the inner variation of the Finnish language. I will first explore different varieties of Finnish that the pupils in the schools are exposed to, and will then show how immigrant adolescents make use of different resources of Finnish in creating huono suomi, a humoristic stylisation of a kind of learner language. 2.1

Crossing and lexical mixing

Many of the adolescents I have interviewed say that most of their friends know some Somali words, since the Somalis have a large presence in many schools. If I ask what kind of words the pupils know or recognise, the word waryaa ~ woraa10 ‘hey you (to a male person)’ is often mentioned. Waryaa also is the Somali word that is most frequently used by non-Somalis in my data, and the first word of Somali I learnt when I started gathering my data. The use of waryaa or other features

. In Finland, a pupil with a mother tongue other than Finnish or Swedish has the right to receive lessons in his/her home language. In Helsinki, lessons for 40 different immigrant languages are organised in schools. 10. The established written form of the lexeme is waryaa, but in speech and in my data its pronunciation varies. One common spoken form in my data is woraa, and this form occurs in some of my examples. When I refer to the lexeme in general, I refer to it as waryaa.

296 Heini Lehtonen

of the Somali language by someone who actually does not speak Somali can in my data often be analysed as an act of stylisation (Quist 2005: 72–74; Rampton 2006: 224–226), a playful imitation of another’s linguistic variety. Stylisation makes use of social styles and varieties that are recognisable in the community, creating a stereotypical image of a certain social group. Rampton (1995) has analysed different instances where adolescents use a linguistic variety that is not associated with their own ethnic or social group, and he refers to linguistic practices of this kind as language crossing. The term voicing, that in Coupland’s (2007: 114) words “refers to how a speaker represents or implies ownership of an utterance or a way of speaking”, and Bakhtin’s idea of polyphony are related to stylisation and crossing (Rampton ibid.: 221; cf. Quist & Jörgensen 2007): with stylisation, a speaker can bring a foreign voice into his speech and show that the utterance or the way of speaking he is using might not entirely “belong” (Rampton 1998: 291) to him. Sometimes, however, it seems waryaa has become more or less a shared word among friends from different ethnic backgrounds, and it no more represents a clearly foreign voice in the speech of non-Somalis (Section 2.1.2). 2.1.1 Styling the Somali voice In the following Example (2) I am having a rather informal conversation with Tarkan and Mohamed and their Finnish-born classmate Jaakko. Tarkan and Mohamed have only known Jaakko for nine months. They spend a lot of time together at school, but Jaakko does not spend his free time with Tarkan and Mohamed’s multiethnic group of friends. My interviews and recordings with Tarkan and Mohamed have revealed that they mix Somali and sometimes Turkish words into Finnish in their interaction, and in Example (2), I am curious to know whether Jaakko has also learnt some Somali or Turkish words. The Somali items used by Jaakko are highlighted in bold face. Transcription symbols are explained in Appendix. (2) (A discussion between the researcher and the boys in 2003. Mohamed (M, Somali, m) and Jaakko (J, Finn, m) describe an incident where Jaakko suprisingly was calling Mohamed names in Somali. Tarkan (T, Turkish Kurd, m) also takes part in the conversation.) 01 J: joitain sanoja [oon kyllä oppinu.

I did learn some words.

02 M: [£yhden sana.£ ((yskii))



£one word.£ ((coughing))

03 H: millasia< millasia< millasia

what kind of< what kind of< what kind of