Dialectological and Folk Dialectological Concepts of Space: Current Methods and Perspectives in Sociolinguistic Research on Dialect Change 9783110229127, 9783110229110

Citation preview

Dialectological and Folk Dialectological Concepts of Space linguae & litterae

17

linguae & litterae Publications of the School of Language & Literature Freiburg Institute for Advanced Studies

Edited by

Peter Auer · Gesa von Essen · Werner Frick Editorial Board Michel Espagne (Paris) · Marino Freschi (Rom) Ekkehard König (Berlin) Michael Lackner (Erlangen-Nürnberg) Per Linell (Linköping) · Angelika Linke (Zürich) Christine Maillard (Strasbourg) · Lorenza Mondada (Basel) Pieter Muysken (Nijmegen) · Wolfgang Raible (Freiburg) Monika Schmitz-Emans (Bochum)

17

De Gruyter

Dialectological and Folk Dialectological Concepts of Space Current Methods and Perspectives in Sociolinguistic Research on Dialect Change

Edited by Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck

De Gruyter

ISBN 978-3-11-022911-0 e-ISBN 978-3-11-022912-7 ISSN 1869-7054 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. ” 2012 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Printing: Hubert & Co. GmbH & Co. KG, Göttingen 앝 Printed on acid-free paper 앪 Printed in Germany. www.degruyter.com

Table of contents

V

Table of contents

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

David Britain Countering the urbanist agenda in variationist sociolinguistics: dialect contact, demographic change and the rural-urban dichotomy .

12

Michael Elmentaler Dialectal concepts of space and linguistic variation . . . . . . . . . .

31

Sandra Hansen Dialektalität, Dialektwissen und Hyperdialektalität aus soziolinguistischer Perspektive . . . . . . . . . . . . . . . . . . . . .

48

Eivind Torgersen A perceptual study of ethnicity and geographical location in London and Birmingham . . . . . . . . . . . . . . . . . . . . . .

75

Robert Möller Der Sprachgebrauch „bei uns“ – Arealbildung in Karten des Atlas zur deutschen Alltagssprache, objektive Grenzen und subjektive Räume .

96

Johanna Vaattovaara Making sense of space – on dialect production and perception at the Finland-Sweden border . . . . . . . . . . . . . . . . . . . . . . . . 119 Philipp Stoeckle The folk linguistic construction of local dialect areas – linguistic and extra-linguistic factors . . . . . . . . . . . . . . . . . . 142 Chris Montgomery Mapping the perceptions of non-linguists in Northern England . . . 164 Christian Schwarz Konservative vs. innovative Dialektgebiete. Ein quantitativer Ansatz zu ihrer Bestimmung auf der Grundlage spontansprachlicher Daten aus Südwestdeutschland . . . . . . . . . . . . . . . . . . . . . . . . 179

VI

Table of contents

Simon Pickl and Jonas Rumpf Dialectometric concepts of space: Towards a variant-based dialectometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Benedikt Szmrecsanyi Geography is overrated . . . . . . . . . . . . . . . . . . . . . . . . 215 Tobias Streck Dialektareale in Baden-Württemberg. Dialektometrische Analysen spontansprachlicher Daten aus Südwestdeutschland und Überlegungen zum Zusammenhang von objektiven Dialektgrenzen und mentalen Raumkonzepten . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 Wilbert Heeringa and Frans Hinskens The measurement of Dutch dialect change in the sound components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 Appendix/colour maps . . . . . . . . . . . . . . . . . . . . . . . . 269

Table of contents

VII

Acknowledgements

This volume contains a collection of papers from the conference Dialectological and folk dialectological concepts of space, which was held at the Freiburg Institute for Advanced Studies (FRIAS) from November 27th – 29th 2008. We are deeply grateful to the FRIAS School of Language and Literature, which funded the conference and also the publication of this book. The concept for the conference Dialectological and folk dialectological concepts of space was a winner of the first competition of the FRIAS School of Language and Literature for Junior Research Groups planning to hold a conference at the University of Freiburg with around 15 invited speakers. The objective of the conference was to focus on the role of the concepts of “linguistic space” from a dialectological, sociolinguistic and folk dialectological point of view and to discuss their inter-relationship and relevance to variational linguistics. With an oral presentation participated Christina Ada Anders (D, Kiel), Peter Auer (D, Freiburg), David Britain (GB, Essex), Rudi Bühler (D, Freiburg), Javier Caro Reina (D, Freiburg), Michael Elmentaler (D, Kiel), Wilbert Heeringa (NL, Amsterdam), Roeland van Hout (NL, Nijmegen), Roland Kehrein (D, Marburg), Stefan Kleiner (D, Mannheim), Niina Kunnas (FIN, Oulu), Marie Maegaard (DK, Kopenhagen), Robert Möller (B, Liège), Chris Montgomery (GB, Sheffield), Jenny Nilsson (S, Göteborg), Christoph Purschke (D, Marburg), Benedikt Szmrecsanyi (D, Freiburg), Eivind Torgersen (GB, Lancaster), Johanna Vaattovaara (FIN, Helsinki) and the editors of this volume. Unfortunately not all of the oral presentations are represented with an article in this book, but the present collection contains articles covering all central subjects of the conference. We would like to thank the entire FRIAS team and our student assistants for their outstanding support in organizing the conference and preparing this book. We are indebted to the editors – Peter Auer, Gesa von Essen, and Werner Frick – for the admission of this volume into the linguae & litterae series. Sincere thanks are also given to Elin Arbin for her assistance in the formal and linguistic revision of the articles as well as to the external reviewers for their very helpful comments. Sandra Hansen, Christian Schwarz, Philipp Stoeckle, and Tobias Streck

VIII

Table of contents

Introduction

1

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck (Albert-Ludwigs-Universität Freiburg)

Introduction

In the recent past, different concepts of space have been adapted from neighboring academic disciplines (such as geography and social and political sciences) and applied to linguistic research. Major outcomes – both with respect to theoretical/methodological achievements and their application to empirical data – are collected in the two volumes Language and Space (cf. Auer and Schmidt 2010; Lameli, Kehrein, and Rabanus 2010), providing a broad, general overview of the different perspectives on language and space. This book discusses dialectological, sociolinguistic and folk dialectological concepts of linguistic space. The articles present findings mainly from empirical studies which take on these different concepts and examine how they relate to one another. The major goal of this collection is to shed light on the interrelationship between the above-mentioned aspects and their relevance to variational linguistics. In dialectology, the concept of space has always been a central issue. Due to many traditional atlas projects, the linguistic variation in space has been very well documented. In these projects the main focus was on the traditional base dialects of rural areas. Urban centers were considered non-homogeneous language areas and thus were not included in the studies in most cases. It was also common in the traditional dialectological atlases to examine only one speaker per place as representative of the dialect. Very often the informants were non-mobile, older, rural and male1 (NORM; cf. Chambers and Trudgill 1998: 29) – the oldest inhabitants of the villages. The validity of the results of this kind of traditional research is somewhat restricted, however, when it comes to describing the linguistic reality, since the data only represents the speech of a certain speaker group in a special communicative context. In traditional dialect geography, questionnaire-based methods were usually applied to elicit the dialect knowledge of the speakers. This knowledge is not necessarily identical with the informants’ actual language use. In order to get a more diversified picture of the spectrum of varieties, it is necessary to collect spontaneous speech data in different social contexts 1

Traditional German dialect atlases typically included both male and female informants.

2

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck

with several speakers per place to compare with traditional knowledge-based data. In this way, social determinants of language use and variation can be investigated. In the course of the advent of sociolinguistics in the 1960s, factors such as the age, sex, and education of the (dialect) speakers as well as the social milieu and communicative requirements in different social contexts came to the fore (cf. Mattheier 1980, 1994; Sieburg 1991). Accordingly, studies focused on the investigation of different groups of speakers and the particularly varied language in the city. For a long time traditional research in the field of dialectology and the new socio-dialectological research directions coexisted separately (cf. Britain 2010). Traditional dialectology focused primarily on the diatopic (and occasionally diachronic) dimension of linguistic variation, whereas in sociolinguistics the diastratic and diaphasic dimensions played a main role. Apart from projects such as the Mittelrheinischer Sprachatlas (‘Linguistic Atlas of the Middle Rhine’) (cf. Bellmann, Herrgen, and Schmidt 1994–2002), most sociolinguistic studies were limited to the investigation of single towns or villages (cf. Besch et al. 1981; Lenz 2003) or the variation within a city (cf. Kallmeyer 1994; Dittmar, Schlobinski, and Wachs 1986). Their results suggest that the idea of a homogeneous local speech community, as assumed in traditional dialectology, is not consistent with reality. Contact with the standard language or adjacent regional varieties, facilitated by the growing influence of the media and increased mobility, respectively, led to the majority of speakers possessing several linguistic registers and being able to use them in a context-specific way. The numerous studies that were conducted according to these two paradigms – socio-dialectological research and traditional dialectology – have provided important insights into the social stratification of language use and the geographical distribution of dialects, respectively. However, there is little knowledge about their relationship and possible mutual influences. Comprehensive areal studies incorporating the social dimension of language use are still rare. What is needed is research including two-dimensional designs (linguistic geography and the comparison of conservative and modern speaker groups) which allows researchers to make systematic statements about dialect variation and change by taking possible social factors into account. In recent years, a further aspect has gained importance: the representation and evaluation of dialects by the speakers themselves, i.e. by linguistic laypeople. One basic assumption of this subdiscipline of sociolinguistics, often referred to as “perceptual dialectology” or “folk dialectology” (cf. Anders, Hundt, and Lasch 2010; Long and Preston 2002; Niedzielski and Preston 2003; Preston 1999), is the idea that subjective classifications and attitudes of

Introduction

3

speakers can be a major motivation for variation in language use (cf. Macha and Weger 1983: 265) and therefore cause language change. Like “objective” (socio)linguistics, perceptual dialectology deals with the examination and classification of linguistic variants, but from a lay perspective. Everyday experience and evidence from the literature show different evaluations of dialects with regard to their social attractiveness, e.g. Upper Saxon being regarded as the most unpleasant dialect in the German-speaking area (cf. Anders 2010) in contrast to Northern German or Bavarian, which are rated the most pleasant dialects (cf. Eichinger et al. 2009). Another frequently pursued issue is the perception of linguistic features. Which features are salient to linguistic laypeople? Which features do they notice, and which do they talk about when asked to describe a certain dialect (cf. Preston 2010)? Which features are associated with certain varieties or groups of speakers, and do these associations correspond to “objective” linguistic findings? In addition to evaluations regarding the social attractiveness of dialects and the analysis of the salience of certain dialect features, a central research topic is how linguistic laypeople perceive and structure dialect areas (cf. Auer 2004; Preston 1993). As a methodological tool for the analysis of subjective notions about dialect geography, so-called mental maps have proven to be helpful. These are maps drawn by the informants which indicate the scope or the borders of a dialect area. The comparison of subjective and objective dialect boundaries enables researchers to identify linguistic and non-linguistic factors which could be reasons for the inclusion or exclusion of neighboring speaker groups, and thus provide indications for the driving forces behind language change. Not only the methods of data collection and the research topics have changed, but also the computational possibilities of processing large corpora of linguistic data and the visualizations of results have been refined. Methods from mathematics, statistics and geography have been adapted and applied to linguistic data sets (cf. Köhler, Altmann, and Piotrowski 2005). Various methods for the cartographic representation of “objective” and “subjective” linguistic data have been developed, especially in the field of language mapping. The volume Language Mapping (Lameli, Kehrein, and Rabanus 2010) provides an overview of the traditions of linguistic cartography, current developments and recent applications of quantitative visualization techniques. The aspects described above are discussed in this volume from different perspectives. The first part of the volume deals mainly with dialectological questions from a sociolinguistic point of view: Quantitative and qualitative analyses are presented which are based on variational linguistic methods and

4

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck

questions from sociolinguistic research on dialect change (papers by Britain; Elmentaler; Hansen). Another major part consists of the analysis of folk dialectological concepts of space. Methods of data elicitation concerning lay concepts of dialectological space are discussed as well as the relevance of the outcomes of these kinds of studies for variational linguistics (papers by Torgersen; Möller; Vaattovaara; Stoeckle; Montgomery). The second part of the volume focuses on dialectometric and quantitative methods (papers by Schwarz; Pickl and Rumpf; Szmrecsanyi; Streck; Heeringa and Hinskens). In the research described in the papers in this part of the volume, the center of interest is to analyse and visualize spatial linguistic structures on the basis of aggregated data sets, using a large number of phenomena. Starting the volume, David Britain presents a critical view of the urbanist agenda in variationist sociolinguistics. He questions the widely accepted dichotomy between urban and rural spaces that have been strongly separated in linguistic research during the past decades. The reason for this is that cities were believed to imply a much higher degree of linguistic “disorder” than is found in the countryside, a factor that led to the practice of carrying out variationist research mainly in urban settings and rarely in rural areas. Britain argues that the urban-rural dichotomy is not a productive one for explaining the root causes of linguistic change, and that there are and can be no linguistic processes that are restricted either to urban areas or to rural. The paper considers the views of human, social and economic geographers who have cautioned against the usefulness of the terms ‘rural’ and ‘urban’ because, they argue, there exist no causal social processes unique to either. It then presents evidence of the uninevitability of urbanisation, using evidence from and discussing the linguistic consequences of over half a century of net counter-urbanisation in Northern Europe and North America. Finally it argues that the causal process that helps us best understand the linguistic outcomes not just of urbanisation, but also of other demographic trends such as counter-urbanisation is language and dialect contact. Using examples from a number of urban and rural speech communities around the world, it is demonstrated that while contact may well be most obviously and vividly observed and experienced in cities, it is essentially insensitive to urban or rural locale and is not confined sociologically, demographically or epistemologically to an urban context. Michael Elmentaler argues that modern areal linguistics should pay attention to the whole spectrum of speech varieties between local dialect and spoken regional standard language, aiming to describe their variational patterns. Elmentaler describes the concepts of space in traditional dialectology

Introduction

5

and the potentials of modern research on regional varieties. Problems of constructing spatial structures in “new dialectology” are pointed out. The author discusses some of the factors to be considered if sociolinguistic aspects are to be included in areal linguistics; he also points out the potentials of a dialectology which studies speech variation on the basis of records of spontaneous conversation in an areal perspective. In the second part of his paper, Elmentaler presents the results of a perceptual experiment in which the informants had to evaluate a phonological feature (spirantization of final g) in three regions of Northern Germany. The results suggest that spatial structures depend on feature selections, the degree of abstraction in the description, and the situational context. Sandra Hansen describes the influences of sociolinguistic parameters on dialect knowledge and the occurrence of hyperdialectal forms in utterances of speakers from Southwest Germany. On the basis of socially differentiated data, measurement techniques are developed to calculate indices of dialect knowledge and indices on hyperdialectality for each speaker. Traditional data from the Linguistic Atlas of Southwest Germany (Südwestdeutscher Sprachatlas, Steger et al. 1989ff.) serve as a basis of comparison. In a further step, the influences of extra-linguistic factors on the calculated indices are tested by using statistical models. Finally, the relationship between dialect knowledge and hyperdialectality is discussed. Eivind Torgerson presents perceptual data from London and Birmingham. He used perceptual tests to investigate the uniformity of Multicultural London English (MLE) in relation to ethnicity and geographical location of speakers, as measured by London and Birmingham listeners’ ability to classify ethnicity and location based on short speech samples. The MLE features shared by most speakers in inner-city London were corroborated by small or no differences in the ethnic classification of inner-city speakers by London listeners. He has also tested listeners from Birmingham to examine to what degree MLE is also regionally neutral, i.e. if the multicultural varieties are non-regional. The results of the study show that multi-ethnic features seem to be more important than geographical factors, and that multi-ethnicity seems to be mostly associated with larger cities. Robert Möller’s paper deals with the relationship between mental borders and area formation in vernacular variation. The author discusses the elicitation methods of lay descriptions concerning local language use. He argues on the basis of a direct questioning according to the Atlas of German Everyday Language (Atlas zur deutschen Alltagssprache, Elspaß and Möller 2003ff.) that areal formation shows a striking resemblence with non-linguistic borders. This also holds for new patterns of areal distribution (new phenomena or

6

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck

areas that diverge from the dialectal ones). Even if one admits that political and administrative boundaries can still hinder language contact nowadays, it is remarkable that historical boundaries still continue to play a role for everyday speech even today. The data gathered in the sixth round of the AdA survey focuses on the informants’ localization of their own everyday speech in their folk linguistic mental map. It turns out that the areas of linguistic identification tend to have a double background inasmuch as informants refer to historical territories (and their prestigious names), yet at the same time these areas also form parts or subunits of recent political and administrative units. Johanna Vaattovaara analyzes the meaning of linguistic features that serve as collective landmarks and labels of identity. She combines qualitative findings with methods from folk linguistics. As a linguistic laboratory she chose the Tornio Valley region in the north-western part of Finland close to the Swedish border. Her paper deals with the ideological motivation behind the preservation of the h in non-initial syllables, an archaic phonological dialect feature in the Tornio Valley dialect of Finnish (e.g. saunhaan~sauhnaan; Standard Finnish and the majority of current dialects use the form saunØaan ‘into sauna’). The h serves as an index and a source of the transnational imagined community, constructed and shaped by the political history of the border and the general image of the area. Some of the central findings of a more extensive study and methodological implications are discussed in the framework of “map free” interpretation of space, currently dominant in cultural geography. Philipp Stoeckle presents findings from his work on the perception of regional dialect variation in the Alemannic speech area. The central question is: Which linguistic and extra-linguistic factors are important for the perception and structuration of dialect areas and dialect borders? For this purpose, a small sample of data from four locations was selected for a detailed analysis. First Stoeckle examines how the informants structure their own local dialect and the surrounding dialect areas geographically. In the next step, possible reasons for these subjective structurations are discussed; for this purpose, he takes a closer look at the speakers’ comments referring to both linguistic and extra-linguistic factors in the interviews. It turns out that folk strategies of structuring dialect regions include not only knowledge about linguistic features, but also – and probably more importantly – historical, political and cultural factors. This may support the observation that mental linguistic borders often coincide with administrative, confessional and political borders rather than with classifications based on findings of linguistic atlases. (For similar observations cf. Möller in this volume.)

Introduction

7

In his methodologically oriented paper Chris Montgomery discusses different ways of digitally quantifying data from “draw-a-map” tasks, using examples from his study on the perception of Northern English dialects (2006). A tracing method is demonstrated, along with older computerized methods for producing composite maps. The author contrasts the older methods with new ways of generating composite maps using the Geographical Information System ArcGIS. He discusses a number of outputs and their potential for explaining patterns in perceptual data. Various maps including additional data, such as road networks, population density and dialect areas, clearly show the advantages of the new system. Christian Schwarz presents a method of distinguishing phonologically conservative and innovative dialect areas in Southwest Germany on the basis of a large corpus of spontaneous speech data. In the first part of his contribution he describes the necessary methodological steps in data processing, resulting in a data set that can be used as input for statistical analysis and visualisation of variation in space as interpolated grid plots. In the second part results are discussed. The major outcome consists of an aggregate interpolation plot that includes variables from sixteen different etymological sound classes that can be used for demonstrating the distribution of receding phonological variables in space. The interpolation shows two conservative areas where receding forms are still widely spread. They lie within the centres of the two major dialect groups of Southwest Germany: Alemannic and Swabian. The conservative areas are separated by a broad band of intense variation between receding and innovative variants. Schwarz argues that this variation is not due to a unidirectional horizontal spread of the dominant dialect into the area of the other. Variation is rather triggered by vertical standard influence that supports any dialect form to spread out horizontally as long as it is phonologically identical or similar to the standard form. The last four texts deal with dialectometrical and related quantitative methods for a cartographic representation of language in geographical space and/or for the analysis of the relationship between language and space. All four articles were written by authors who worked with (different kinds of) aggregated data and whose methods/techniques – at least to some extent – are based on ideas of the “Salzburg School” (see the works of Hans Goebl cited below) and/or the “Groningen School” (see the works of John Nerbonne mentioned in the following). A detailed description of the history of dialectometry from Jean Séguy (see e.g. Séguy 1973) and Hans Goebl (see e.g. Goebl 1982) to John Nerbonne and others (see e.g. Nerbonne 2006, 2009; Heeringa 2004) is given in Goebl (2005, 2010) and Nerbonne (2010). Brief descriptions of the particular subfield of dialectometry as well as the authors’

8

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck

own theoretical/methodical contributions are presented in the individual articles. Simon Pickl and Jonas Rumpf sum up the development and motivation of concepts of space in classical dialectometry and argue for a variantbased dialectometry, a “bottom-up” view of language in space. They discuss methods and types of maps which are the result of a joint research project at the Universities of Augsburg and Ulm. Methods from stochastic image analysis were applied to data from the Sprachatlas von Bayerisch-Schwaben (König 1996–2009). One of the aims of the approach is to get a closer look at the spatial distributions of single linguistic features in the dialects in the region of Bayerisch-Schwaben and to find out what may have caused these distributions. The method also may reveal as yet unknown associations between linguistic variables, while at the same time demonstrating the feasibility and the potential of an alternative quantitative approach to linguistic variation in space. Benedikt Szmrecsanyi examines morphosyntactic variability in traditional British English dialects, using naturalistic corpus data from the Freiburg English Dialect Corpus (FRED). With the aid of dialectometrical and statistical methods Szmrecsanyi explores how and to what extent morphosyntactic variability is structured geographically. The study utilizes an aggregate measure of morphosyntactic dialect distances that is empirically based on the text frequencies of 57 morphological and syntactic features. With as-thecrow-flies distance, least-cost travel time, a linguistic gravity index, and dialect area membership, four language-external predictor variables are introduced. They are all tested for explanatory potency in regard to morphosyntactic dialect distances. The results of Szmrecsanyi’s analyses suggest that mere geography (in terms of as-the-crow-flies distance and least-cost travel time) is a comparatively poor predictor of morphosyntactic variability in British English dialects. Tobias Streck applies a dialectometrical analysis to a large corpus of spontaneous speech data from traditional speakers of the Alemannic dialects in southwest Germany. He demonstrates two kinds of cluster maps concerning the phonology of the traditional dialects in the 1970s/80s and discusses the areal distribution of the dialect groups. The visual interpretation of the maps suggests that the most important dialect boundary in southwest Germany (according to the dataset) runs along the former political border between the territories of Baden and Württemberg. Streck describes, among other things, a spread of the Swabian dialect into the region north-east of Lake Constance and argues that the former political structure of the area is a mental factor which strongly influences dialect geography. He therefore concludes that men-

Introduction

9

tal concepts of space that are partly based on former political-administrative territories affect linguistic behavior (cf. also Möller’s and Stoeckle’s papers). Wilbert Heeringa and Frans Hinskens study the development of koines (‘regiolects’) and other intermediate varieties outside of traditional local dialects, using phonetic transcriptions of newly collected dialect recordings of representative Dutch dialects from 20 locations in the Netherlands and the northern part of Belgium which were collected in 2007–2008. For each site, an older male and a younger female speaker were recorded, representing conservative and innovative speakers, respectively. They measure dialect change at the levels of the lexicon, morphology and the sound components. Changes in the sound components have been measured with Levenshtein distance. The authors found that the distances among dialects have significantly decreased at the level of the sound components, and that the 20 dialects have significantly converged towards standard Dutch. Dialects which were distant to standard Dutch converge more strongly to standard Dutch than dialects which were more closely related to standard Dutch. Considering dialect change, they found that the lexical level is affected most strongly.

References Anders, Christina A. 2010 Wahrnehmungsdialektologie. Das Obersächsische im Alltagsverständnis von Laien. (Linguistik – Impulse & Tendenzen, Bd. 36.) Berlin/New York: de Gruyter. Anders, Christina A., Markus Hundt and Alexander Lasch (eds.) 2010 Perceptual Dialectology. Neue Wege der Dialektologie. (Linguistik – Impulse & Tendenzen, Bd. 38.) Berlin/New York: de Gruyter. Auer, Peter 2004 Sprache, Grenze, Raum. Zeitschrift für Sprachwissenschaft 23/2: 149–179. Auer, Peter and Jürgen Erich Schmidt (eds.) 2010 Language and Space. An International Handbook of Linguistic Variation, Vol. 1: Theories and Methods. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 30.1.) Berlin/New York: de Gruyter. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994–2002 Mittelrheinischer Sprachatlas (MRhSA), 5 Bände. Tübingen: Niemeyer. Besch, Werner, Jochen Hufschmidt, Angelika Kall-Holland, Eva Klein and Klaus J. Mattheier 1981 Sprachverhalten in ländlichen Gemeinden. Ansätze zur Theorie und Methode. Britain, David 2010 Language and space: the variationist approach. In: Peter Auer and Jürgen Erich Schmidt (eds.), Language and Space. An International Handbook of Linguistic Variation, Vol. 1: Theories and Methods, 142–163. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 30.1.) Berlin/New York: de Gruyter. Chambers, J. K. and Peter Trudgill 1998 Dialectology. 2nd edition. Cambridge: University Press.

10

Sandra Hansen, Christian Schwarz, Philipp Stoeckle and Tobias Streck

Dittmar, Norbert, Peter Schlobinski and Inge Wachs 1986 Berlinisch. Studien zum Lexikon, zur Spracheinstellung und zum Stilrepertoire. (Berlin-Forschung 14.) Berlin: Berlin Verlag. Eichinger, Ludwig M., Anne-Kathrin Gärtig, Albrecht Plewnia, Janin Roessel, Astrid Rothe, Selma Rudert, Christane Schoel, Dagmar Stahlberg and Gerhard Stickel 2009 Aktuelle Spracheinstellungen in Deutschland. Erste Ergebnisse einer bundesweiten Repräsentativumfrage. Mannheim: Institut für Deutsche Sprache. Elspaß, Stephan and Robert Möller 2003ff. Atlas zur deutschen Alltagssprache. (http://www.philhist.uni-augsburg.de/ada). Goebl, Hans 1982 Dialektometrie. Prinzipien und Methoden des Einsatzes der Numerischen Taxonomie im Bereich der Dialektgeographie. (Denkschriften, Bd. 157.) Wien: Österreichische Akademie der Wissenschaften. Goebl, Hans 2005 Dialektometrie. In: Reinhard Köhler, Gabriel Altmann and Rajmund G. Piotrowski (eds.), Quantitative Linguistik. Ein internationales Handbuch, 498–531. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 27.) Berlin/New York: de Gruyter. Goebl, Hans 2010 Dialectometry and quantitative mapping. In: Alfred Lameli, Roland Kehrein and Stefan Rabanus (eds.), Language and Space, Volume 2: Language Mapping, 433–457. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 30.2.) Berlin/New York: de Gruyter. Heeringa, Wilbert 2004 Measuring dialect pronunciation differences using Levenshtein distance. Dissertation, Rijksuniversiteit Groningen. Kallmeyer, Werner (ed.) 1994 Kommunikation in der Stadt. Teil I: Exemplarische Analyse des Sprachverhaltens in Mannheim. Berlin/New York: de Gruyter. Köhler, Reinhard, Gabriel Altmann and Rajmund G. Piotrowski (eds.) 2005 Quantitative Linguistik. Ein internationales Handbuch. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 27.) Berlin/New York: de Gruyter. König, Werner 1996–2009 Sprachatlas von Bayerisch-Schwaben. (Bayerischer Sprachatlas: Regionalteil 1.) 14 volumes. Heidelberg: Winter. Lameli, Alfred, Roland Kehrein and Stefan Rabanus (eds.) 2010 Language and Space, Volume 2: Language Mapping. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 30.2.) Berlin/New York: de Gruyter. Lenz, Alexandra 2003 Struktur und Dynamik des Substandards. Eine Studie zum Westmitteldeutschen (Wittlich/Eifel). (Zeitschrift für Dialektologie und Linguistik – Beihefte, Bd. 125.) Stuttgart: Steiner. Long, Daniel and Dennis R. Preston (eds.) 2002 Handbook of Perceptual Dialectology, Vol. 2. Amsterdam: Benjamins. Macha, Jürgen and Thomas Weger 1983 Mundart im Bewusstsein ihrer Sprecher. Eine explorative Studie am Beispiel des Bonner Raumes. Rheinische Vierteljahresblätter 47: 265–301. Mattheier, Klaus J. 1980 Pragmatik und Soziologie der Dialekte. Einführung in die kommunikative Dialektologie des Deutschen. Heidelberg: Quelle und Meyer. Mattheier, Klaus J. 1994 Varietätenzensus. Über die Möglichkeit, die Verbreitung und Verwendung von Sprachvarietäten in Deutschland festzustellen. In: Klaus J. Mattheier und Peter Wiesinger (eds.), Dialektologie des Deutschen. Forschungsstand und Entwicklungstendenzen, 413–442. (Reihe Germanistische Linguistik, Bd. 147.) Tübingen: Niemeyer.

Introduction

11

Montgomery, Chris 2006 Northern English dialects: A Perceptual Approach. Dissertation, University of Sheffield. Nerbonne, John 2006 Identifying Linguistic Structure in Aggregate Comparison. Literary and Linguistic Computing 21/4: 463–475. Nerbonne, John 2009 Data-Driven Dialectology. Language and Linguistics Compass 3/1: 175–198. Nerbonne, John 2010 Mapping aggregate variation. In: Alfred Lameli, Roland Kehrein and Stefan Rabanus (eds.), Language and Space, Volume 2: Language Mapping, 476–495. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 30.2.) Berlin/New York: de Gruyter. Niedzielski, Nancy and Dennis R. Preston 2003 Folk Linguistics. Berlin/New York: de Gruyter. Preston, Dennis R. 1993 Folk Dialect Maps. In: A. Wayne Glowka and Donald M. Lance (eds.), Language variation in North American English: Research and Teaching, 105–118. New York: Modern Language Association of America. Preston, Dennis R. 1999 Handbook of Perceptual Dialectology, Vol. 1. Amsterdam: Benjamins. Preston, Dennis R. 2005 What is folk linguistics? Why should you care? Lingua Posnaniensis 47: 143–162. Preston, Dennis R. 2010 Perceptual Dialectology in the 21st Century. In: Christina A. Anders, Markus Hundt and Alexander Lasch (eds.), Perceptual Dialectology. Neue Wege der Dialektologie, 1–29. (Linguistik – Impulse & Tendenzen, Bd. 38.) Berlin/ New York: de Gruyter. Séguy, Jean 1973 La dialectométrie dans l’Atlas linguistique de la Gascogne. Revue de linguistique romane 37: 1–24. Sieburg, Heinz 1991 Geschlechtstypischer Dialektgebrauch. Anmerkungen zu einer empirischen Untersuchung von Geschwistern in der rheinischen Ortschaft Fritzdorf. Rheinische Vierteljahrsblätter 55: 294–314. Steger, Hugo, Eugen Gabriel, Volker Schupp and Ulrich Knoop (eds.) 1989ff. Südwestdeutscher Sprachatlas. Bearbeitet von Roswitha Braun-Santa, Ewald Hall, Renate Schrambke, Hugo Steger, Bernhard Kelle, Guillaume Schiltz, Jörg Wagner, Ekkehard Felder, Markus Hundt, Christoph Maier, Volker Schupp, Erich Seidelmann, Harald Baßler. Marburg: Elwert.

12

David Britain

David Britain (Universität Bern)

Countering the urbanist agenda in variationist sociolinguistics: dialect contact, demographic change and the rural-urban dichotomy

1.

Introduction

Beginning in the 1960s, the geographical focus in dialectology shifted from an extreme concentration on the speech of rural areas, to an equally extreme focus on that found in our largest towns and cities. The linguistic lens shifted from an inspection of, for the most part, non-mobile old rural men (NORMs, following Chambers and Trudgill 1980) to an apparently much more heterogeneous sample of speakers – old and young1, male and female, working and middle class2, of different ethnic backgrounds3. This shift occurred at the same time as the incorporation into the discipline of a new analytical apparatus with which to analyse language variation and change – the linguistic variable – as well as a large number of new theoretical concepts to help us understand the linguistic diversity – apparent-time, inherent variability, change from above and change from below, etc. Since then, the variationist paradigm has been dominated by studies of cities – beginning, of course, with New York (Labov 1966), Panama (Cedergren 1973), Detroit (Shuy, 1

2

3

Though as discussed by Labov (1994), variationist sociolinguistic studies of cities have tended for the most part and until recently to only analyse post-adolescents, and avoid very elderly speakers. Work on children and adolescents is now much more common, but tends to focus on young speakers to the exclusion of postadolescents, thereby denying us a view of the whole age spectrum in the community in which those children are growing up. It is rare to see 5, 15 and 55 year olds all in the same study of the same speech community. Upper class speakers have also tended to be avoided in sociolinguistic dialectology, with few exceptions. For a good time in early variationist work, this inclusion only applied to settled migrant populations and not to ‘non-native’ or recent migrants. Work by variationists such as Horvath (1985) has reminded us of the importance of recent incomers to a community in shaping local variation and change (see also Fox 2007). All in all, then, given footnotes 1, 2 and 3, ‘urban dialectology’, for many years, gave us a rather sanitized, ‘tidy’ view of the diversity of the metropolitan speech community.

Countering the urbanist agenda in variationist sociolinguistics

13

Wolfram, and Riley 1967) and Norwich (Trudgill 1974), but extending right through to the most theoretically cutting-edge work of today in Detroit (Eckert 2000), London (Cheshire, Kerswill, Fox and Torgersen 2011), Cairo (Miller 2005), Glasgow (Lawson 2011), Beijing (Zhang 2005) and Montreal (Sankoff and Blondeau 2007), for example. Meanwhile, in a good number of the relatively few variationist works where rural locations have been selected, their selection is precisely motivated by the perceived conservativeness of the rural in contrast to the urban. A good example is Tagliamonte, Smith and Lawrence’s (2007) work on a number of very rural British dialects. They argue “one of the most informative constructs which can shed light on the origins and development of languages is the relic area. Such areas, because of their peripheral geographical location and/or isolated social political circumstances, tend to preserve older features” (2007: 90–91). The study of cities has been so central to the variationist enterprise that this approach to studying variation and change is indeed often called ‘urban sociolinguistics’ or ‘urban dialectology’. Here, I argue that there is nothing essentially urban about variationist sociolinguistics, insisting, firstly and rather obviously, that the approaches to fieldwork methodology, the analytical techniques applied to corpora of collected data, and the theoretical assumptions made are all equally applicable in rural areas, but secondly that there is no a priori reason why patterns of dialectological variation in the countryside should fundamentally differ from those found in our largest cities (see, further, Britain 2009). Important, here, is the recognition – a fairly central pillar of sociolinguistics – that we should put speakers at the centre of our theorisation of language, and if speakers move and come into contact with other speakers, the likely outcomes of that contact are typologically similar, but are insensitive to urban or rural location.

2.

Urbanisation and the urban fetish in sociolinguistic dialectology

That sociolinguistic approaches to dialectology have become fixed upon the urban is perhaps not surprising. Given that one of the central aims of early variationism was to highlight the orderly heterogeneity of the speech community, finding order in the hustle and bustle, the bright lights, and the apparent disorder of the big city was always going to make for a more dramatic and powerful demonstration of the sociolinguistic approach than finding such structure among the NORMs of the deep countryside. Our perceptions of urban areas as complex, diverse, ‘where it’s all happening’ (and of rural areas as a “quiet tranquil even backward place where nothing interesting happens” (Woods 2011: 35)) have no doubt helped to drive the fascination

14

David Britain

with the city in sociolinguistics. The urban turn in sociolinguistics also came at a rather opportune time from the perspective of global demographic developments. Cities in the Western world had seen dramatic urbanization especially since the Industrial Revolution, with poor rural farm workers moving to the (apparent) economic opportunities that employment in towns and cities offered. A good deal of the urban population increase, though, can be put down to the improvements in healthcare and sanitisation that led to a significant decrease in infant and maternal mortality. In the developing world, urbanisation gathered pace later, but even more dramatically than in the West, with the number of cities of over one million inhabitants increasing by 50 % between 1975 and 2000 (Eisinger 2006: 268). From the point of view of ‘urban sociolinguistics’, researchers were able to tap into the consequences of almost two centuries of urban growth in the West, and the rapid expansion of such growth in the developing world. So, as well as concentrating almost exclusively on the sociolinguistics of urban centres, perhaps not surprisingly a number of variationist sociolinguists deliberately focussed on the sociolinguistic dynamics of urbanisation as a process (Bortoni-Ricardo 1985; Nordberg 1994; Kerswill 1994). These works examined the consequences of the contact of rural dialects in the city, the (variable) adoption of urban forms by rural migrants and the loss of rural dialect forms, thereby examining issues of koineisation, second dialect acquisition and speech accommodation, themes common to contact dialectology more generally (Trudgill 1986). Some urban dialectologists, however, took this focus on the linguistic outcomes of urbanisation one step further, suggesting there were linguistic outcomes specific to cities (e.g. Bulot 2002; Bulot and Tsekos 1999; Calvet 1994; Messaoudi 2001, 2002), with the term ‘linguistic urbanisation’ (‘urbanisation linguistique’) entering the sociolinguistic discourse in some quarters. These scholars argued that urbanisation was inevitable, ongoing and a distinct social and, therefore, linguistic process: Pourquoi la ville? Lorsqu’on observe les taux d’urbanisation des différents pays du monde, on se rend compte que la ville se dresse à l’horizon de notre histoire immédiate comme un inévitable destin. Partout les ruraux se précipitent vers les fausses promesses de la cité, vers ses lumières, vers l’espoir d’un travail plus lucrative. Et cette convergence de migrants vers la cité a sa contrepartie linguistique (Calvet 1994: 10) … cette réalité plurilingue de la ville nous mène dans un premier temps à trois thèmes … la ville comme facteur d’unification linguistique, la ville comme lieu de conflit de langues et la ville comme lieu de coexistence et de métissage linguistique (Calvet 1994: 11) … mais comment démontrer qu’un lieu – la ville – et une function – la véhicularité – ont des effets comparables sur des langues différentes? (Calvet 1994: 11) … la ville produit aussi des formes linguistiques spécifiques, des parlers urbains (Calvet 1994: 13) … La sociolinguistique ur-

Countering the urbanist agenda in variationist sociolinguistics

15

baine ne peut pas se contenter d’étudier des situations urbaines, elle doit dégager ce que ces situations ont de spécifique, et donc construire une approche spécifique de ces situations (Calvet 1994: 15) … La ville est d’une part … l’avenir de l’humanité. La ville occupe dans l’espace européen une place de plus en plus importante … la ville est la quintessence du plurilinguisme, elle draine les différentes situations linguistiques du pays … telle une pompe, la ville aspire du plurilinguisme et recrache du monolinguisme, et elle joue ainsi un rôle fondamental dans l’avenir linguistique de la région ou de l’État (Calvet 1994: 129–130). [Why the city? One needs only to look at rates of urbanization in different countries around the world to realize that the city represents an inevitable outcome of our recent history. People from rural areas everywhere are lured by the false promises of urban life, by its bright lights and the hope of better paid work. And this coming together of migrants to the city has linguistic consequences. The reality of urban multilingualism leads us, in the first instance, to three themes: the role of the city in linguistic unification, the city as site of language conflict and the city as site of language mixing and language co-existence. But how can a setting – the city – and a function – communication via a common medium – be shown to have similar effects on different languages? … the city also produces specific linguistic forms, urban dialects … Urban sociolinguistics cannot be content to study urban contexts, it must tease out what is specific about these contexts and build a specific approach to these contexts … The city is, on the one hand, the future of humanity. It consumes an ever-increasing part of the European landscape … the city embodies multilingualism, drawing on the range of linguistic situations from the surrounding area … like a pump, the city breathes in multilingualism and spits out monolingualism, and thus plays a fundamental role in shaping the linguistic future of the region or state].

The language used here is quite emotive, and Calvet relies frequently on the word ‘spécifique’ to hammer home his idea that there really is something new in what is being found as a result of urbanisation – linguistically he proposes semantic transparency and the morphological and grammatical simplification, common outcomes in contact situations (Trudgill 1986).

3.

What’s so special about the city?

It is the idea of ‘urbanisation as special’ that I seek to critique here4. In order to do this, I will examine firstly (and briefly) fieldwork and analysis methods, to highlight the scope of the approach across both the urban and the rural domain. Secondly, I will present the views of human, social and economic geographers – after all, urbanisation falls squarely within their domain of enquiry – many of whom cast a great deal of caution on the usefulness of the 4

For a critique of a British example of the ‘linguistic urbanisation’ approach, see Britain (2009).

16

David Britain

terms ‘urban’ and ‘rural’ because, they argue, there are no causal social processes which are unique to either. Thirdly, I question the idea of urbanisation as inevitable, presenting evidence of counterurbanisation trends, as well as other mobilities in Western societies. And finally, I argue that the causal process which helps us better understand not just the linguistic outcomes of urbanisation, but also of other demographic trends such as counterurbanisation and suburbanization is language/dialect contact. As we will see, contact is insensitive to urban or rural locale. It just happens that such contact is perhaps more vividly and frequently experienced and perceived in cities, consequently helping to explain why it and urbanisation may have become confused. But radical dialect contact can occur in highly rural areas as well as in cities and the outcomes are typologically the same (Trudgill 1986). Hence linguistic urbanisation should be seen as a conceptual dead-end, as it relies on theoretically unsustainable notions of the city as well as on an only partial awareness of the demographies of contemporary Western societies. We can begin with the assertion that sociolinguistic methods of variation analysis can be applied to both urban and rural locations – indeed Labov, the founder of modern variation studies in linguistics demonstrated as much in the very early days of the subject by first applying the sociolinguistic toolkit to rural Martha’s Vineyard in Massachusetts (Labov 1963) and then to urban New York City shortly thereafter (Labov 1966)5. The range of techniques applied to the task of speaker sampling, to the careful definition of the social parameters important for a study in that place at that time, to the collection of informal recordings of vernacular speech, to the application of the linguistic variable and variation analysis to the resulting data are all equally achievable in rural as well as urban communities. There is nothing methodologically urban about these approaches, nor, in principle, is such research in one more complex, difficult, expensive or time-consuming, etc., than in the other. Geographers’ perspectives on the urban and the rural have changed over time. The fetishisation of the urban – seeing urbanity itself as the cause of social change – was found across the social sciences up until and right through the 1960s and early 1970s. This then came under attack initially by theorists such as David Harvey (e.g. 1973) and Manuel Castells (1977), among others, and has continued since. To make the point about the theorisation of the urban, the economic geographer Doreen Massey uses as an example the rise of urban unemployment in Britain in the 1980s, suggesting that ‘what was happening in the inner cities – the loss of industrial jobs – had more to do 5

See Britain 2009 for a further discussion of the New York and Martha’s Vineyard studies in this context.

Countering the urbanist agenda in variationist sociolinguistics

17

with industry than with the cities … the cities were at the sharp end of what was to become a much more general process of deindustrialisation … the fact that they were at the sharp end had more to do with the character of their industry … than with cities’ characteristics as locations’ (Massey 1985: 11 – emphasis in original). So it is argued that while ‘urban’ and ‘rural’ conjure up very distinct images in people’s minds (e.g. Woods 2011: 16–49; Hubbard 2006: 59–94), there are no social processes that are unique to either. Some social, economic, political and cultural phenomena might be found more often, more intensively, more obviously or more extremely in urban areas, and others more so in rural areas, but these are simply tendencies and do not represent unique and exclusive traits of urban or of rural contexts. Instead it is argued that there are certain causal social processes which simply have geographically uneven consequences, affecting rural areas more than urban or vice versa. Important, therefore, is to identify the causal processes, and examine how they played out in urban and rural locales, rather than endowing the urban (or the rural) with causal powers that it didn’t have. Geographers became rather forthright in their rejection both of urban causality and of the helpfulness more broadly of the terms ‘urban’ and ‘rural’. So, for example: To isolate some characteristic of the city that might distinguish it from its nominal counterpart, the rural, is extremely difficult … what is needed is urban scholarship that takes the city seriously as an object of study without lapsing into environmental determinism (Hubbard 2006: 1, 3) In everyday usages, terms like ‘city’ and ‘town’ are unproblematic and convey clear spatial images of built-up areas and hence the form is intrinsic to the definition … Theoretical analyses of capitalist urbanisation have shown that there is no material basis for this definition; the relations and activities found in such areas are not unique to them and so what seems analytic and necessary in everyday thought is shown to be synthetic and contingent by theory” (Sayer 1985: 58, emphasis in original) “The rural is a messy slippery idea that eludes easy definition and demarcation” (Woods 2011: 1) “The broad category ‘rural’ is obfuscatory … since intra-rural differences can be enormous and rural-urban similarities sharp (Hoggart 1990: 245) … the designation ‘rural’ is for many researchers merely a symbol of interest in small settlements or in particular kinds of economic activity: it is not a statement about unique causal properties (Hoggart 1990: 246) … I do not mean by this that there are no differences between (most) rural and urban places, but rather that, in the main, these are generated by the uneven presence of some known causal factor X, as opposed to either rurality or urbanity. The obvious follow-up point is that for theory to progress we should focus on X” (Hoggart 1990: 251).

18

David Britain

In addition, sociologists and geographers countered the idea that the city was somehow a distinctive entity by highlighting the consequences of the progression of capitalism in Western societies: It is my contention … that with the development of capitalism, the city has ceased to be a sociologically significant unit in Western societies (Saunders 1985: 76) … There can be no social theory of the city in contemporary industrial capitalist societies (Saunders 1985: 79). The social processes resulting in the production of the city were not distinctly urban, but endemic to capitalist society (Hubbard 2006: 40). The development of capitalism has not led to the consolidation of the institutions of the city, but rather to its eradication as a distinct social form (Giddens 1991: 148).

In the geographical literatures, the focus on urban and rural has shifted, therefore, examining the ways in which causal mechanisms have affected social and economic development and change in rural and urban areas, studying heterogeneity, conflict and diversity in both town and country as well as examining the ideologies of urban and rural and their interaction with social change. Much work has emphasised representations of urban and rural, the way in which both city and countryside are imagined in popular discourses. Woods argues, for example, that ‘rurality is understood as a social construct – that is as an imagined entity that is brought into being by particular discourses of rurality that are produced, reproduced and contested by academics, the media, policy-makers, rural lobby groups and ordinary individuals. The rural is therefore a ‘category of thought’ (Woods 2011:9). Further inspection of the linguistic urbanisation approach reveals more difficulties: Firstly, characterisations presented as being elements of urbanisation and urban living are in fact simply characteristics of late modernity: mobility-as-routine, acceleration of the pace of life and so on: L’urbanité signifie une accélération des processus de vie, de la mobilité des individus, de l’innovation des structures et des institutions et d’une nouvelle différenciation sociale (Erfurt 1999: 9). [With urban living comes an acceleration of everyday life, individual mobility, structural innovation and new social divisions]. Nous entendons par urbanisation “… un processus à travers lequel la mobilité spatiale vient structurer la vie quotidienne” (Rémy and Voyé 1992: 10)” (Bulot 1999: 24). [By urbanisation we mean a process through which spatial mobility comes to shape everyday life].

Again, these traits might (but might not) be more extreme in the city, but they are not restricted to urban contexts.

Countering the urbanist agenda in variationist sociolinguistics

19

Secondly, the idea that urbanisation is ‘un inévitable destin’ is undermined by solid evidence, again from late modern Western societies, of population deconcentration, demographic shifts away from cities. Champion (2001b) highlights counterurbanization as an important process of the past half century in a number of advanced capitalist societies, with remoter rural areas seeing the largest demographic growth in percentage terms, and large metropolitan areas shedding population most rapidly (see also Champion 1989, 1998, 2001a, 2005a, 2005b; Oswalt and Rieniets 2006, and, for example, Phillips 2005 for work on rural gentrification (conversion of old barns, disused chapels and school buildings and the like)). In the British context, this counterurbanization trend of an overall shift of population from the urban to the rural has affected all age groups except the 16–24 year olds (leaving the countryside to seek advanced education and training and chase the bright lights of the city), and is, on average, both Whiter and more middle class than the population as a whole, involving ‘the wealthiest and most skilled occupational categories … professional and managerial workers’ (Champion 2001a: 43–44), though, of course, as Halfacree (2008) reminds us, these general trends hide much diversity in the counterurbanising population. Net outmigration levels (urbanfrural minus ruralfurban) from London and the six other largest English cities to the rest of the country averaged around 80 000 per year between 1971 and 1996, around 100,000 per year between 1997 and 2003, and reached 145 000 in 2004 (Champion, Coombes, Raybould and Wymer 2007: 7) – almost three million more people leaving the city for the countryside than moving in the opposite direction in less than 25 years. And this outward exodus of professional urbanites has triggered a massive reorientation of personal mobilities, as many of these millions of migrants to the countryside commute, drive to new shopping malls, and work in the highly mobile tertiary sector of the economy (see, for example, Champion, Coombes and Brown 2009, Findlay, Stockdale, Findlay and Short 2001). We are only now beginning to learn about the dialectological consequences of this counter-urbanisation. In the more sparsely populated rural areas, counter-urbanisation seems to be triggering dialect levelling and the loss of traditional dialect features. Piercy (2006), for example, demonstrated the dramatic loss of non-prevocalic /r/ in rural Dorset in the South-West of England. Once the ‘heartland’ of rhoticity in England, the young people in Piercy’s survey, both in rural areas of Dorset as well as urban, barely produced one token of non-prevocalic /r/. Similarly, in Eastern England, counterurbanisation has caused both local variant attrition and, in some cases, the adoption of standard forms among the local, non-migrant popu-

20

David Britain

lation (Britain, in preparation). The traditional vernacular forms of the area, such as the lack of third person present tense marking on verbs (‘he go’, ‘she love’), the absence of the palatal glide after non-coronals (‘few’ [fu]; ‘music’ [muzik]; ‘puke’ [puk]; ‘beauty’ [buʔi]; ‘huge’ [hu]) and the use of [ə] as opposed to [i] in unstressed syllables (e.g. ‘dishes’ [diʃəz]; ‘bucket’ [bkəʔ]; ‘wanted’ [wɒnʔəd]; ‘running’ [ɹnən]) are, for example, declining in use across apparent time. Other rural areas in East Anglia are experiencing a decline in [ευ] forms of /au/ (‘mouth’ [mευθ]; ‘out’ [ευʔ]) in favour of [aυ]), the use of flapped variants of /t/ (Amos 2007, 2009), as well as a steady decline in the use of ‘what’ as a relative marker in favour of ‘that’ (‘I ate the cake what she bought from Tescos’). It has also led to dramatic demographic decline in some of Britain’s largest cities: both Manchester and Liverpool have seen their populations roughly halve since 1931. Urbanisation is being reversed, then, in some economies, and, according to Oswalt and Rieniets, should begin to decline globally by the end of this century: between 1950 and 2000, more than 350 large cities experienced […] significant declines in population. In the 1990s, more than a quarter of the world’s large cities shrank. Their number is continually increasing, even though urban growth will continue to dominate in decades to come. An end is in sight, however: somewhere between 2070 and 2100, the world population is expected to reach its zenith, with the process of urbanization largely coming to a close (Oswalt and Rieniets 2006: 6).

We know from work on the sociolinguistic typologies of change in contexts of dialect isolation that the opposite trajectories are found to those in contexts of contact (see Trudgill 2002): complexification and conservatism appear to be the norm in isolated speech communities. We have come to expect to find such trends, however, in rural areas. In the US, Schilling-Estes and Wolfram (1999), researching the depopulating Smith Island in the Chesaspeake Bay of Maryland find that those residents who have remained on the island, despite the strong economic motivations to leave, are ‘concentrating’ the local dialect, moving it in directions that are actively diverging the local variety from those on the mainland. Counterurbanisation has meant that Smith Island-like communities are likely to be rare in rural (esp. Southern) England, but the statistics on urban depopulation raise the thought-provoking possibility that divergence and dialect concentration might be features of urban areas with similar demographic trends to those found on Smith Island. Some work investigating large British cities that have seen considerable demographic shrinkage suggests that dialect concentration may indeed

Countering the urbanist agenda in variationist sociolinguistics

21

be an appropriate way to understand certain linguistic changes in progress in these urban areas. Kevin Watson’s research in Liverpool, for example, has highlighted how this large city of the English North-West has largely resisted participation in many of the rapidly diffusing changes currently underway in British English (e.g. the glottalisation of oral stops, especially /t/), and has even accelerated some divergent linguistic changes. So, for example, the use of [h] for /t/, almost entirely restricted to Liverpool, has expanded from contexts following short vowels in small monosyllabic function words (e.g. at and what) to polysyllabic non-function words where the /t/ is preceded by an unstressed vowel (e.g. biscuit, bucket, chocolate) (Watson 2006: 59). This typology of linguistic changes has been determined then, not by rural status or by urban status, but by demographic circumstances which lead to lack of contact – wherever such circumstances are found. Since the late 20th century, other mobilities have increased in scale and intensity in the late modern Western world, too, which have, in many cases, promoted or encouraged counter-urbanisation tendencies. These include (see, further, Britain 2010), beyond an increased frequency of migration: the routinisation of long-distance commuting; the expansion in uptake of higher education; increases in the use of both public and private transport; a shift from primary and secondary to tertiary sectors of the economy (which are consequently less tied to the location of natural resources, and are driven by capitalist economics to seek out the cheapest locations in which to operate); an increase in mobile and flexible working as a result of technological developments and employment legislation; geographical reorientations of consumption behaviours; and the increasing elasticity of family ties. People are ‘on the move’ (Cresswell 2006), then, and not always or even mostly in the direction of cities.

4.

Contact as causal

Sociolinguistics has always held dear the idea that it is speakers in their speech communities that should be central to our theorisations of language change (see, especially, Milroy 1992, 2003). If we follow this line of argumentation, we should (simply) examine the interactions that speakers engage in as they pursue their daily and life-time projects wherever that might be – city or hamlet, in New York or ‘in the middle of nowhere’. Central to the interactions not only in those contexts of urbanisation proposed by the linguistic urbanisationists (amongst others), but also in contexts of counterurbanisation and those of the other mobilities mentioned above, is face-to-face contact with speakers whose dialects are liable to be distinct from one’s own,

22

David Britain

possibly radically distinct. Studies in contact dialectology have demonstrated clear typologies of change, linguistic outcomes which recur when dialect contact takes place, whether that contact be in a large metropolis or a small village deep in the countryside, in a central networked Western city, or on the most remote island in the world. These changes (see, for example, Trudgill 1986; Kerswill 2002), include simplification and the loss of redundancy, transparency, levelling, reallocation and the emergence of hybrid interdialect forms. Examples of typologically similar outcomes of dialect contact, in both urban and rural contexts, can be seen in Table 1 below. Table 1: Urban and rural contexts of dialect contact and their linguistic consequences: Levelling

Simplification

Interdialect

Reallocation

Urban

Rural

– Milton Keynes, England (Kerswill and Williams 2000); – Corby, England (Dyer 2002); – Newcastle, England (Watt 2002); – Cairo, Egypt (Miller 2005) – Amman, Jordan (Al-Wer 2003); – Kristinestad, Finland (Ivars 2005); – Bergen, Norway (Kerswill 1994) – Reading, England (Cheshire 1982, Trudgill 1986) – Jakarta, Indonesia (Wouk 1999); – Takatsuki/Osaka/Kyoto, Japan (Long 2001); – Ban Khlong Sathon, Thailand (Prompapakorn 2004)

– Falkland Islands (Sudbury 2000); – Fens, England (Britain 2001); – East Devon and West Somerset (Jones 1998); – Palau in the Western Pacific (Matsumoto and Britain 2003) – Ocracoke, United States (Schilling-Estes and Wolfram 1994); – Zeeland, Netherlands (De Vogelaer and RoozeStouthamer 2007) – Fens, England (Britain 2001); – Tristan da Cunha (Schreier and Trudgill 2006) – Fens, England (Britain 1997); – Mauritius (Domingue 1981); – Rural Flanders, Belgium (Taeldeman 1989).

Watt (2002), for example, points to evidence of dialect levelling in urban Newcastle in North-East England. There he finds evidence of the quite dramatic obsolescence of some traditional local variants of the /ei/, /ou/ and / / vowels (in ‘gate’, ‘goat ’ and ‘nurse’ respectively) in favour of variants found more widely across the North of England. Andrea Sudbury’s examination of

Countering the urbanist agenda in variationist sociolinguistics

23

the English of the Falkland Islands (FIE) (2000), however, also showed dramatic degrees of levelling between the input stage (largely varieties of English from south-west England and Scotland, but also from other parts of the South of England) and present-day FIE, despite the fact that, in the world, only Greenland has a sparser population density than the Falklands. The following features characteristic of the input varieties have been levelled away, or are rare relic forms: rhoticity, the devoicing of initial fricatives, periphrastic do, the velar fricative /x/, the /w – / distinction, Scottish vowel length, double modals and the South-West’s distinctive pronoun system (see Sudbury 2000: 202–206 for more details). Al-Wer (2003) reports on simplification – the process by which, through dialect contact, the new variety becomes more regular, having fewer grammatical categories (such as gender, case, honorifics), fewer person/number inflections and/or fewer lexical exceptions and complex constraints on variation than the varieties in the original contact mix – in urban Amman, the capital city of Jordan. Amman is especially interesting because it is a relatively new city – a deserted Roman site at the beginning of the 20th century, it is now home to over 1.5 million. Its population has also highly mixed origins, including both Jordanians and Palestinians, both urban and rural. Al-Wer discusses in detail the simplification in Ammani Arabic of the 2nd and 3rd person plural pronominal suffixes. Traditionally, Jordanian dialects a) had distinct masculine and feminine suffixes, and b) the input dialects showed a variety of different forms of the feminine suffix. In Amman, dialect contact has triggered the simplification of this system such that the gender distinction has been neutralised for both 2nd and 3rd person suffixes and, also, there is now ‘a more regular phonetic correspondence’ (2003: 65) between the two – ‘-kum’ for 2nd person, ‘-hum’ for 3rd. Meanwhile, in rural Zeeland in the South-West of the Netherlands, De Vogelaer and Rooze-Stouthamer (2007) show how contact has led to the simplification of the 3rd person singular masculine and feminine pronouns. They show that the subject v oblique distinction for masculine (hij v hem) and feminine (zij v haar) has been simplified, with the former oblique forms hem and haar now serving both subject and oblique functions. The authors state that not only does this represent simplification, it also represents a divergence from Standard Dutch (2007: 77). Interdialect forms – forms which exist in none of the input dialects, but which can be seen as the linguistically intermediate result of imperfect convergent accommodation between them – have been reported in both urban and rural locations too. Trudgill (1986: 63) reports work by Rekdal (1971) examining long-term accommodation by migrants to the Norwegian capital

24

David Britain

city of Oslo. She found that the migrants produced hybrid forms, found in neither the original migrant dialects nor those of Oslo. For example, ‘to work’ is reported as /jub/ in the original migrant dialect, /jɔbə/ in Oslo and an interdialectal /jubə/ among the now-Oslo resident migrants. The interdialectal form combines the first vowel from original migrant dialect, but the schwa-ending disyllabic form of the Oslo variety. Meanwhile, Schreier and Trudgill (2006) provide an example of interdialect from Tristan da Cunha in the South Atlantic, one of the remotest inhabited islands in the world, with a population of around 300. They report the, for English highly unusual, emergence on Tristan of // in unchecked position, in words such as ‘fur’ [f]. They argue this is an interdialect form that emerged as a result of contact between Scottish dialects in which // can occur before /r/ (e.g. /fr/ ‘fur ’), and English non-rhoticity (2006: 127). Long (2001)’s examination of the Japanese dialect of Takatsuki, a suburb that sits between urban Osaka and Kyoto, provides a fascinating example of contact-induced reallocation, the refunctionalisation of two or more variants from the contact feature pool to serve novel functions in the emergent new variety. He contrasts the different suffixes used in Osaka and Kyoto for the regular negative (i.e. ‘do not’) and the potential negative (i.e. ‘can not’). In Kyoto these are –ahen and –ehen respectively, and in Osaka –ehen and –arehen. Kyoto’s potential negative, then, is the same as Osaka’s regular negative. Long shows how in intermediate suburbs such as Takatsuki, speakers fuse the two systems, selecting the Kyoto regular negative and the Osaka potential negative form, thereby preserving the distinction between the two negation types, which would have been lost had the orientation of the reallocation been reversed. But such reallocation also occurs in highly rural areas. In my own research on the East Anglian Fens (Britain 1997), I showed how reallocation provided the most plausible account for the emergence of a Canadian-Raising-like allophonic distribution for /ai/ (as in ‘like’, ‘nice’, ‘side’, ‘buy’). Variants of /ai/ with open nuclei [ɑi – ɑ], e.g. [nɑʔtɑm] ‘night time’ from the Western Fens and variants with central nuclei [əi], [nəiʔtəim] from the East came into contact as a result of post-reclamation in-migration to produce an allophonic system sensitive to the voicing of the following segment in the central area of the Fens, with [əi] before voiceless consonants and [ɑi – ɑ] elsewhere [nəiʔtɑm]. Intensive dialect contact, then, can lead to koineisation – levelling, simplification, interdialect and reallocation, but this is not restricted to the urban areas that we perhaps associate most with contact, mobility and diversity.

Countering the urbanist agenda in variationist sociolinguistics

25

If the contact conditions are right, such processes can take place in the world’s largest cities, yes, but also in some of the world’s most remote and sparsely populated communities too.

5.

Conclusion

I’ve argued here that ‘rural’ and ‘urban’, concepts deemed by social geographers to be ‘obfuscatory’ (Hoggart 1990), are of little theoretical importance in helping us to understand the nature of linguistic change. These geographers have forcefully argued that there are no causal social processes which affect urban areas but not rural, or vice versa, and no categorical qualitative social, cultural or economic differences between these two geographical contexts. Urbanisation, then, has (only) a contingent effect on sociolinguistic practices – it has created the context for contact. Contact, in contrast, stands out as a causal force which acts in typologically similar ways wherever it occurs. I have shown here examples from both large urban centres as well as from some of the world’s most sparsely populated speech communities that demonstrate the common outcomes of dialect contact regardless of where that contact takes place. We have become used to expecting radical dialect contact in our large metropolitan areas, perhaps partly because dialectologists of late have focussed their efforts on examining change in cities, ‘where it’s all (apparently) happening’, and taken the spotlight away from rural areas (where it had been firmly fixed in the days of the traditional dialectologist). But what has become clear is that while contact may well be most obviously and vividly felt and observed in cities, it is not confined sociologically or epistemologically to an urban context. The evolution in demographic trends in Western societies reminds us that urbanization is not an inevitable path for human settlement. Counterurbanization and rural gentrification trends force us to look to the countryside (as well as the city) as a place of sociolinguistic contact, diversity, conflict and change and, if present demographic developments continue, rural areas will become increasingly important loci of dialect contact in the next decades. These developments also suggest we should be open to different types of change in depopulating, shrinking urban centres, perhaps even the sorts of change we have come to expect in more isolated communities. So as geolinguists interested in language change, rather than always heading for the city, we should instead chase contact (or lack of it). If we only look in urban areas, demographics suggest that, in the years to come, we might be missing more and more of the picture.

26

David Britain

References Al-Wer, Enam 2003 New dialect formation: the focusing of -kum in Amman. In: David Britain and Jenny Cheshire (eds.), Social Dialectology: In honour of Peter Trudgill. 59–67. Amsterdam: Benjamins. Amos, Jennifer 2007 Wadda Boo’ıful Place: an analysis of the variables (ju) and (t) in Mersea Island English. Colchester: University of Essex MA thesis. Amos, Jennifer 2009 How’s about that? A sociophonological analysis of (au) in Mersea Island English. Paper presented at the Third International Conference on the Linguistics of Contemporary English, University of London, July 2009. Bortoni-Ricardo, Stella 1985 The urbanization of rural dialect speakers. Cambridge: Cambridge University Press. Britain, David 1997 Dialect Contact and Phonological Reallocation: ‘Canadian Raising’ in the English Fens. Language in Society 26: 15–46. Britain, David 2001 Welcome to East Anglia!: Two Major Dialect ‘Boundaries’ in the Fens. In: Jacek Fisiak and Peter Trudgill (eds.), East Anglian English. 217–242. Woodbridge: Boydell and Brewer. Britain, David 2009 ‘Big bright lights’ versus ‘green and pleasant land’? The unhelpful dichotomy of ‘urban’ v ‘rural’ in dialectology. In: Enam Al-Wer and Rudolf de Jong (eds.) Arabic dialectology: In Honour of Clive Holes on the Occasion of His Sixtieth Birthday. 223–248. Leiden: Brill. Britain, David 2010 Supralocal Regional Dialect Levelling. In: Carmen Llamas and Dominic Watt (eds.) Language and identities. 193–204. Edinburgh University Press. Britain, David (in preparation). The linguistic consequences of counter-urbanisation. Manuscript. Bulot, Thierry (ed.) 1999 Langue urbaine et identité. Paris: L’Harmattan. Bulot, Thierry 2002 La double articulation de la spatialité urbaine: “espaces urbanisés” et “lieux de ville” en sociolinguistique. Marges linguistiques 3: 91–105. Bulot, Thierry and Tsekos, Nicolas 1999 L’Urbanisation linguistique et la mise en mots des identités urbaines. In: Thierry Bulot (ed.), Langue urbaine et identité. 21–34. Paris: L’Harmattan. Calvet, Louis-Jean 1994 Les voix de la ville: introduction à la sociolinguistique urbaine. Paris: Éditions Payot et Rivales. Castells, Manuel 1977 The urban question: a Marxist approach. London: Arnold. Cedergren, Henrietta 1973 The interplay of social and linguistic factors in Panama. Unpublished PhD dissertation. Cornell University. Chambers, Jack and Trudgill, Peter 1980 Dialectology. Cambridge: Cambridge University Press. Champion, Tony 1989 Counterurbanisation in Britain. The Geographical Journal 155: 52–59. Champion, Tony 1998 Studying counterurbanisation and the rural population turnaround. In: Paul Boyle and Keith Halfacree (eds.), Migration into rural areas. 21–40. Chichester: Wiley. Champion, Tony 2001a The continuing urban-rural population movement in Britain: trends, patterns, significance. Espace, Populations, Sociétés 2001/1–2: 37–51. Champion, Tony 2001b Urbanization, suburbanization, counterurbanization and reurbanization. In: Ronan Paddison (ed.), Handbook of Urban Studies. 143–161. London: Sage.

Countering the urbanist agenda in variationist sociolinguistics

27

Champion, Tony 2005a The counterurbanisation cascade in England and Wales since 1991: the evidence of a new migration dataset. Belgeo: Revue Belge de Géographie 2005:1/2: 85–101. Champion, Tony 2005b Population movement within the UK. In: Roma Chappell (ed.). Focus on People and Migration. 91–114. London: ONS. Champion, Tony, Coombes, Mike, Raybould, Simon and Wymer, Colin 2007 Migration and socioeconomic change: A 2001 Census analysis of Britain’s larger cities. Bristol: Joseph Rowntree Foundation/Policy Press. Champion, Tony, Coombes, Mike and Brown, David 2009 Migration and longerdistance commuting in rural England. Regional Studies 43: 1245–1259. Cheshire, Jenny 1982 Variation in an English dialect: a sociolinguistic study. Cambridge: Cambridge University Press. Cheshire, Jenny, Kerswill, Paul, Fox, Sue and Torgersen, Eivind 2011 Contact, the feature pool and the speech community: the emergence of multicultural London English. Journal of Sociolinguistics 15: 151–196. Cresswell, Tim 2006 On the Move: Mobility in the Modern Western World. London: Routledge. De Vogelaer, Gunther and Rooze-Stouthamer, Clasien 2007 Sociolinguistic typology and dialect levelling: a case from Zeeland Dutch. Phrasis 47: 67–88. Domingue, Nicole 1981 Internal Change in a Transplanted Language. Studies in the linguistic sciences 4: 151–159. Dyer, Judy 2002 ‘We all speak the same round here’: dialect leveling in a ScottishEnglish community. Journal of sociolinguistics 6: 99–116. Eckert, Penelope 2000 Linguistic variation as social practice: the linguistic construction of identity in Belten High. Oxford: Blackwell. Eisinger, Angelus 2006 Urbanization/Urbanisation. In: Phillip Oswalt and Tim Rieniets (eds.), Atlas of Shrinking Cities/Atlas der schrumpfenden Städte. 28–29. Ostfildern: Hatje Cantz Verlag. Erfurt, Jürgen 1999 Préface. In: Thierry Bulot (ed.), Langue urbaine et identité. 7–14. Paris: L’Harmattan. Findlay, Allan, Stockdale, Aileen, Findlay, Anne, and Short, David 2001 Mobility as a driver of change in rural Britain: An analysis of the links between migration, commuting and travel to shop patterns. International Journal of Population Geography 7: 1–15. Fox, Sue 2007 The demise of Cockneys? Language change in London’s ‘traditional’ East End. Unpublished Ph.D. dissertation. University of Essex. Giddens, Anthony 1991 The consequences of modernity. Cambridge: Polity Press. Halfacree, Keith 2008 To revitalise counterurbanisation research?: Recognising an international and fuller picture. Population, Space and Place 14: 479–495. Harvey, David 1973 Social justice and the city. London: Arnold. Hoggart, Keith 1990 Let’s do away with rural. Journal of Rural Studies 6: 245– 257. Horvath, Barbara 1985 Variation in Australian English: The sociolects of Sydney. Cambridge: Cambridge University Press. Hubbard, Phil 2006 City. London: Routledge. Ivars, Ann-Marie 2005 Town and country: when dialect meets standard in urban environments: the case of Finland Swedish. Linguistics 43: 1049–1071. Jones, Jason 1998 Phonological and lexical change in the dialects of East Devon

28

David Britain

and West Somerset 1948–1995. Unpublished PhD dissertation. University of Wales Swansea. Kerswill, Paul 1994 Dialects converging: rural speech in urban Norway. Oxford: Clarendon Press. Kerswill, Paul 2002 Koineization and accommodation. In: Jack Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.) The Handbook of Language Variation and Change. 669–702. Oxford: Blackwell. Kerswill, Paul and Williams, Ann 2000 Creating a new town koine: children and language change in Milton Keynes. Language in Society 29: 65–115. Labov, William 1963 The social motivation of a sound change. Word 19: 273– 309. Labov, William 1966 The social stratification of English in New York City. Washington, D.C.: Center for Applied Linguistics. Labov, William 1994 Principles of Linguistic Change: Volume 1: Internal Factors. Oxford: Blackwell. Lawson, Robert 2011 Patterns of linguistic variation among Glaswegian adolescent males. Journal of Sociolinguistics 15: 226–255. Long, Daniel 2001 An interdialectal negation system in Japanese. Journal of Social Sciences and Humanities (Jimbun Gakuho) 320: 1–24. Massey, Doreen 1985 New directions in space. In: Derek Gregory and John Urry (eds.), Social relations and spatial structures, 9–19. London: Macmillan. Matsumoto, Kazuko and Britain, David 2003 Contact and obsolescence in a diaspora variety of Japanese: The case of Palau in Micronesia. Essex Research Reports in Linguistics 44: 38–75. Messaoudi, Leila 2001 Urbanisation linguistique et dynamique langagières dans la ville de Rabat. In: Thierry Bulot, Cécile Bauvois, and Philippe Blanchet (eds.), Sociolinguistique urbaine: Variations linguistiques, images urbaines et sociales. 87–98. Rennes: Presses de l’Université de Rennes. Messaoudi, Leila 2002 Le parler ancien de Rabat face à l’urbanisation linguistique. In: Youssi Abderrahim, Fouzia Benjelloun, Mohamed Dahbi and Zakia IraquiSinaceur (eds.) Aspects of the Dialects of Arabic Today: Proceedings of the 4th Association Internationale de Dialectologie Arabe conference. 223–233. Rabat: Amapatril. Miller, Catherine 2005 Between accommodation and resistance: Upper Egyptian migrants in Cairo. Linguistics 43: 903–956. Milroy, James 1992 Linguistic Variation and Change. Oxford: Blackwell. Milroy, James 2003 On the role of the speaker in language change. In: Raymond Hickey (ed.), Motives for Language Change. 143–157. Cambridge: Cambridge University Press. Nordberg, Bengt 1994 The Sociolinguistics of Urbanization: The Case of the Nordic Countries. Berlin: Mouton de Gruyter. Oswalt, Phillip and Rieniets, Tim 2006 Introduction/Einführung. In: Phillip Oswalt and Tim Reiniets (eds.), Atlas of Shrinking Cities/Atlas der Schrumpfenden Städte. 6–7. Berlin: Kulturstiftung des Bundes. Phillips, Martin 2005 Differential productions of rural gentrification: illustrations from North and South Norfolk. Geoforum 36: 477–494. Piercy, Caroline 2006 “Mixed with others it sounds different, doesn’t it?”: A quantitative analysis of rhoticity from four locations in Dorset. Unpublished MA thesis. University of Essex.

Countering the urbanist agenda in variationist sociolinguistics

29

Prompapakorn, Praparat 2004 Dialect contact and new dialect formation in a Thai New Town. Unpublished PhD dissertation. University of Essex. Rekdal, Olaug 1971 Modifisert Dialekt: En undersøkelse av to språkvarianter hos inflyttere i Oslo. Unpublished MA Thesis. Universitet i Oslo. Rémy, Jean and Voyé, Liliane 1992 La ville: vers une nouvelle définition? Paris: L’Harmattan. Sankoff, Gillian and Blondeau, Hélène 2007 Language change across the lifespan: /r/ in Montréal French. Language 83: 560–588. Saunders, Peter 1985 Space, the city and urban sociology. In: Derek Gregory and John Urry (eds.), Social relations and spatial structures. 67–89. London: Macmillan. Sayer, Andrew 1985 The difference that space makes. In: Derek Gregory and John Urry (eds.), Social relations and spatial structures. 49–66. London: Macmillan. Schilling-Estes, Natalie and Wolfram, Walt 1994 Convergent explanation and alternative regularization: were/weren’t levelling in a vernacular English variety. Language Variation and Change 6: 273–302. Schilling-Estes, Natalie and Wolfram, Walt 1999 Alternative models of dialect death: Dissipation vs. concentration. Language 75: 486–521. Schreier, Daniel and Trudgill, Peter 2006 The segmental phonology of nineteenthcentury Tristan da Cunha English: convergence and local innovation. English Language and Linguistics 10: 119–141. Shuy, Roger, Wolfram, Walt and Riley, William 1967 Linguistic correlates of social stratification in Detroit speech. Final Report: Cooperative Research Project No. 6–1347, United States Office of Education. Sudbury, Andrea 2000 Dialect contact and koineisation in the Falkland Islands: development of a Southern Hemisphere English? Unpublished PhD dissertation. University of Essex. Taeldeman, Johan 1989 A Typology of Dialect Transitions in Flanders. In: Marten Schouten and Peter Van Reenen (eds.), New Methods in Dialectology: Proceedings of a Workshop held at the Free University Of Amsterdam, December 7–10, 1987. 155–163. Dordrecht: Foris. Tagliamonte, Sali, Smith, Jennifer and Lawrence, Helen 2007 English Dialects in the British Isles in Cross-Variety Perspective: A Baseline for Future Research. In: Markku Filppula, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.), Dialects Across Borders: Selected Papers from the 11th International Conference on Methods in Dialectology (Methods XI), Joensuu, August 2002. 87–117. Amsterdam: John Benjamins. Trudgill, Peter 1974 The social differentiation of English in Norwich. Cambridge: Cambridge University Press. Trudgill, Peter 1986 Dialects in Contact. Oxford: Blackwell. Trudgill, Peter 2002 Linguistic and social typology. In: Jack Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), Handbook of Language Variation and Change. 707–728. Oxford: Blackwell. Watson, Kevin 2006 Phonological resistance and innovation in the North-West of England. English Today 22: 55–61. Watt, Dominic 2002 ‘I don’t speak with a Geordie accent, I speak, like, the Northern accent’: contact-induced levelling in the Tyneside vowel system. Journal of Sociolinguistics 6: 44–63.

30

David Britain

Woods, Michael 2011 Rural. London: Routledge. Wouk, Fay 1999 Dialect contact and koineisation in Jakarta, Indonesia. Language Sciences 21: 61–86. Zhang, Qing 2005 A Chinese yuppie in Beijing: Phonological variation and the construction of a new professional identity. Language in Society 34: 431–66.

Dialectal concepts of space and linguistic variation

31

Michael Elmentaler (Christian-Albrechts-Universität zu Kiel)

Dialectal concepts of space and linguistic variation

1.

Concepts of space in traditional dialectology and modern research on regional varieties

Traditional dialect atlases often suggest a linguistic homogeneity and a clearcut distinctness of dialectal areas which do not accurately reflect the reality of speech at that time. Dialect geographers in the nineteenth century knew very well that dialects are not only differentiated on the dimension of space but also – like all spoken varieties – with respect to social and situative factors (Bellmann 1986). Therefore they used a reductionistic method of collecting data that kept the parameters that could trigger variation constant, which was completely appropriate for the goal of defining clear-cut language boundaries. This method can be summarized as follows: (1) The social factors are kept constant following the NORM principle depicted by Chambers and Trudgill (1998: 29): The typical informant is Non-urban, Old, Rural and Male – the oldest peasant of the village. (2) Individual variation cannot occur as only one person at each place is questioned. (3) Situational and inner dialectal variation is not addressed because dialectal speech is not recorded in its context. In the course of the survey and using a questionnaire, the informants are requested to present the correct dialectal variant (in their opinion) for a given linguistic variable. This normally has the effect that only one variant is listed which the informant regards as particularly authentic, even if there are several variants in the spoken dialect. (4) The informants’ knowledge and opinions of their dialect or of particular dialectal features are not taken into consideration, at least they are not referred to when dialect areas are defined. The concepts of the so-called “new dialectology”, “research on regional varieties” and “pluridimensional dialectology” (Eggers, Schmidt, and Stellmacher 2005; Voeste and Gessinger 2006; cf. Herrgen 2001; Elmentaler 2006a, 2006b) all diverge from the traditional paradigm in at least one of these four points: (1) Social factors are taken into consideration, such as in recent Romance language atlases, in which the age, sex and social status of the informants are analyzed systematically (Radtke and Thun 1996; Blaser 2002; Quesada

32

Michael Elmentaler

Pacheco and Martínez López 2008), and in the Mittelrheinischer Sprachatlas (“Linguistic Atlas of the Middle Rhine Area”) with its differentiation of old, non-mobile and young and mobile speakers (Bellmann, Herrgen, and Schmidt 1994–2002). (2) Consultation of more than one informant in each place; this is characteristic of the six atlases belonging to the Bayerischer Sprachatlas (1996–2010). (3) Situational differences in dialectal and regional speech are taken into account, such as in the current research project Sprachvariation in Norddeutschland (SiN) (“Language variation in Northern Germany”), which describes and maps these differences on the basis of sound recordings (Schröder and Elmentaler 2009). (4) The speakers’ knowledge and attitudes are considered systematically (Gessinger 2008a, 2008b; Elmentaler, Gessinger, and Wirrer 2010). A model for the mapping of such attitudinal data can be found in the Atlas Lingüístico Guaraní-Románico sociología (ALG-R) by Thun (2002). The consideration of these sociolinguistic factors in areal linguistics certainly is a necessary step, but it means that the definition of regional varieties becomes even more difficult. On the other hand, speakers themselves are able to distinguish some dialects and regional varieties quite clearly, and they attribute to them certain evaluations and stereotypes. In regional language research, too, characteristic features not only of dialects but also regional colloquial varieties are listed which are thought to make an areal differentiation of varieties possible (Mihm 2000). How can this be reconciled? I will look at this question in my paper which is divided into two parts. In part one I would like to show by means of some examples how the differentiation of speech levels (Sprachlagen), the consideration of social and situational factors and attitudinal aspects may lead to a weakening of areal structures. In part two I will ask how, on the basis of this differentiated data base, regional language areas may be designated.1

1

My reflections are connected with the research project “Language variation in Northern Germany” mentioned above. In the course of this project which is being conducted by linguists in six Northern German universities, linguistic as well as metalinguistic data from 144 women in 18 regions in Northern Germany were collected in 2008–2010. The factors sex, age and social class are kept constant as far as possible so that the influence of the factors “regional provenance” and “situational context of speech” may be investigated individually.

Dialectal concepts of space and linguistic variation

2.

33

Problems of constructing spatial structures in “new dialectology”

2.1. First aspect: Spatial structures depend on feature selections Even in traditional dialectology, spatial structures are not simply represented but constructed. Even in the case of an almost invariant and homogenous variety, defining dialect areas is not a simple process (Chambers and Trudgill 1998: 89–103; Niebaum and Macha 2006: 80–82). Dialect areas can only be constructed by excluding features that have too small or too widespread a radius. Traditional isoglosses are therefore based on selected features from very different domains. In the case of the Low German language, for example, a consonant feature (the second sound shift) is seen as a criterion for external demarcation, while a morphological feature (the verb plural ending -et vs. -en) serves as a criterion for a rough internal differentiation, and a vocalic feature (the realization of the long e- and o-phonemes) is used for a finer differentiation of Low German dialects (Foerste 1957; Schröder 2004). If other features were taken into account, the result would be completely different spatial structures, with a more limited extension and less clear-cut boundaries. The same holds for the description of the other registers – that is, the colloquial regional varieties and the spoken regional standards; the spatial structures depend on what features are taken into consideration. The selection of features depends on the dialectologist’s aim. Not only the objective findings, but also the speakers’ perception of dialect features could serve as a criterion for the selection of the features. If one wants to create an atlas of regional colloquial varieties that corresponds to the speakers’ intuitive perception of regional language boundaries, one should mainly rely on features with a high degree of saliency. 2.2. Second aspect: Spatial structures depend on the degree of abstraction in the description An example: On an abstract level, the spirantization of the plosive consonant g, such as in [tsυx] instead of [tsuk] Zug ‘train’ or [jans] instead of [ans] Gans ‘goose’, can be seen as a characteristic (although not exclusive) feature of the Northern German colloquial varieties. This is generally true, but when we have a closer look at the contexts, some differences become apparent. In final position, g can be realized as a fricative in all regions, for example in the word [fluxtsɔyç] instead of [fluktsɔyk] Flugzeug ‘plane’. But this is not true for g in initial position; here the fricative only occurs in the colloquial var-

34

Michael Elmentaler

ieties of some regions, such as Brandenburg-Berlin or the Lower Rhine region ([jut jəbratənə jans] instead of [ut əbratənə ans] gut gebratene Gans ‘well roasted goose’) (Lauf 1996). A second example: The maintenance of short vowels from the old Low German dialects in some contexts is considered a common feature of the Northern German colloquial varieties. This feature occurs for example in monosyllabic words such as [bat] ‘bath’, [ʁat] ‘wheel, bike’, [tax] ‘day’ (instead of standard German [bat] Bad, [ʁat] Rad, [tak] Tag; cf. Elspaß and Möller 2003–present, map Tag, question No. 15a). A careful analysis shows, however, that there are strong connections to certain lexemes. Whereas [tak] Tag ‘day’ is frequently shortened to [tax] and [tsuk] Zug ‘train’ to [tsυx], this is not possible in the words [lop] Lob ‘praise’, [lit] Glied ‘limb’ or [tal] Tal ‘valley’. On the other hand, there are clear regional differences, too. Word forms like [kʁiçt] ‘(he) gets’ (standard German [kʁikt] kriegt) and [tax] Tag ‘day’ are common in the colloquial varieties of North-western Germany (Lower Rhine region up to Westphalia) as well as in the northern varieties of Schleswig-Holstein. But this is not true for the short vowels in the words [ʁas] ‘grass’, [las] ‘glass’ or [as] ‘gas’ (instead of standard German [ʁas] Gras, [las] Glas, [as] Gas), which occur only in the northern Low German area but not in the West. The case with the word [ʃpas] Spaß ‘fun’ is exactly the opposite as in the West a variant with short vowel [ʃpas] is in use, whereas the people in the northern area use the standard variant, which is probably supported by the Low German form [spɔs] Spooß that most speakers in this area know. In these cases, lexeme-based analysis leads to rather different spatial structures. The abstract statement that the colloquial varieties of Northern Germany are characterized by a tendency to spirantize g and maintain dialectal short vowels is of course still correct in a way. But when we consider special phonetic contexts or different lexemes, different regional spatial structures may occur. Thus, we come to a third problem of spatial structuring, which results from the work with spontaneous speech data. 2.3. Third aspect: Spatial structures depend on the frequency and function of variants In the case of features like [jut] ‘good’ or [las] ‘glass’, a spatial differentiation is comparatively easy because the variants are in use in some regions, whereas in others they do not occur at all, that is, they are not part of the regional language system. But this is not the normal case. Many other features are “possible” in all regional language systems, but this possibility is used to different degrees and in different ways. I would like to elucidate this with ref-

Dialectal concepts of space and linguistic variation

35

erence to the variants [dat] ‘that’ and [vat] ‘what’ with the unshifted plosive t instead of standard German s in [das] das and [vas] was. The use of these variants might be expected in the whole Northern German area, because all Low German dialects show the forms with plosive t. And indeed, there is evidence for the use of [dat] and [vat] in all regional varieties of Northern Germany. A map that only registered the occurrence or non-occurrence of these variants would show no internal boundaries but only one single, widespread area. However, this cannot be regarded as an appropriate representation of the recent language reality because there are strong differences concerning the frequency of use and the communicative function of these word forms. [dat] and [vat] are highly frequent and common variants in the regional variety of the Northwest area, especially in the Ruhrgebiet (Mihm 1997; Salewski 1998). On the other hand, these forms are relatively rare in the regional varieties north of Hamburg (Elspaß and Möller 2003–present, map das, question No. 25a), where they occur only in specific dialectal contexts, for example in Low German phrases like Wat mutt, dat mutt (literally: what must, that must = engl. ‘needs must when the devil drives’) or in the context of Low German utterances. Here, in contrast to the Ruhrgebiet area, the word forms [dat] and [vat] cannot be regarded as constitutive components of the regional variety, as here one normally uses standard German [das] and [vas]. Rather, they occur in marked contexts and refer to the Low German dialect. In the variety of the Ruhrgebiet, [dat] and [vat] are loan words whose Low German origin generally would not be recognized by the speakers. In the regional varieties of the Northern area it has to be regarded as spontaneous interference from Low German. Scholars who aim to describe recent variety areas and their extension should therefore ask how frequently, in what contexts and with which function a variant is used within a region. However, this presupposes that areal linguistic research deals with records of spontaneous spoken language. Neither the frequencies of use nor the communicative functions of variants can be elicited sufficiently by the classical method of conducting questionnaire surveys.

36

Michael Elmentaler

2.4. Fourth aspect: Spatial structures depend on the choice of the speech level A fourth problem is connected with the fact that in dialectology the field of investigation has been extended from the old base dialects to the whole spectrum of speech levels between dialect and spoken standard language. The spatial structure on the level of the traditional dialects is often another level than that of colloquial varieties or regional standard varieties. A feature such as [dat], [vat] for example is part of every Low German dialect, whereas on the level of the colloquial varieties of Northern Germany it is typical only for some regions, and when one uses regional standard language, one would probably avoid these forms. Another example is the spirantization of the final g that is realized by speakers from Schleswig-Holstein in Northern Germany with an average of 84 % in free conversation, but only 16 % in reading style. On the other hand, features like the vocalization of r in words like Karte ‘card’ ([katə], Low German [kɔ()ɐt] Koort) or gerne ‘gladly’ ([εɐnə], Low German [e()ɐn, i()ɐn] geern, giern) probably show the same pattern on all three speech levels.

3.

Possibilities of constructing language areas

When sociolinguistic aspects are included in areal linguistics, this questions the clear-cut boundaries of traditional dialect maps in several ways. Depending on what speech level and what linguistic material the inquiry is focused on, the result will be different spatial structures. These structures will also be different in relation to the age, sex and social status of the informants. Despite this variation, the speakers themselves are often clearly conscious of the existence of regional language varieties. This may be explained by the fact that there are clusters of linguistic features that can be recognized by the native speakers of a language region by intuition. In variational linguistics these connections have been proved by means of co-occurrence analysis and portrayed in implicational scales (Martin 1996; Auer 1997; Salewski 1998; Möller 2006). For example, when a speaker from the Ruhrgebiet area begins his utterance with the words [dat hasə nɔniç əzen] Dat hasse nonnich gesehn (standard German Das hast du noch nicht gesehen ‘You haven’t seen that yet’), the hearer would expect this speaker to also articulate the initial affricate consonant in pflegen ‘to care for’ as a fricative [f] and the plosive g (or [k]) in Tag as a fricative [x], and to vocalize the r in Bart ‘beard’. On the other hand, when one says Er pflegt seinen Bart jeden Tag [eɐ pflekt zainən baʁt jedən tak] a rather close-to-standard phonetic realization is expected. We do not yet know the exact steps in between. Here is a constructed example:

Dialectal concepts of space and linguistic variation Standard language

Hast du

das

gesagt?

hast du

das

əzakt

Haste

dat

gesaacht

hastə

dat

əzaxt

Hasse

dat

gesacht

hasə

dat

əzaxt

37

(‘Have you said that?’)

Substandard language

Many speakers have a fine ear for the stylistic value and the pragmatic conventions of the use of variants such as hast du – haste – hasse or gesagt – gesaacht – gesacht (Menge 1997). Moreover, they know which forms they may combine syntagmatically. It can be regarded as an important task for areal linguistics based on spontaneous speech data to find out and compare the specific cooccurrence structures for different regions. Thus, another aspect has to be considered that has already been mentioned in connection with the regional variants of the unshifted word variants dat and wat. The specific feature configuration is not the only characteristic of a regional variety; there are also the particular attitudes towards the features (Preston 2002; Lenz 2010; Elmentaler, Gessinger, and Wirrer 2010). These attitudes towards substandard variants are not only focused on general categories, for example the spirantization of g as a phonological feature; they might even be tied to specific contexts or lexemes. In their language acquisition process, speakers may learn to attach attitudes to words or constructions. Thus, their speech reflects attitudinal structures that are specific for the regional speech community. In order to determine such attitudinal structures, we developed a test that was carried out in the summer of 2008 with students at the universities of Kiel, Rostock, Bielefeld and Münster.2 Of the 426 participants in the study, 169 had grown up in Schleswig-Holstein, 184 in the Westphalian part of North Rhine-Westphalia (Nordrhein-Westfalen) and 73 in Lower Saxony (Niedersachsen) and the city-state of Bremen. In all three regions, 70 % to 78 % of the participants were female and 22 % to 30 % male (74 % female : 26 % male on average). About 93 % are between18 to 29 years old. 2

My thanks to Jürgen Macha (Münster), Andreas Bieberstedt (Rostock) and Jan Wirrer (Bielefeld) for supporting my survey. Also, I would like to thank Alastair Walker at CAU Kiel who kindly helped me with the English translation.

38

Michael Elmentaler

All participants were asked to rate examples for the spirantization of final g in 30 sentences: 1) 2) 3) 4) 5) 6) 7)

8) 9) 10) 11) 12)

13)

14)

15) 16)

Ich weiß nicht, warum er mich nicht fragt. [fraxt] ‘I don’t know why he doesn’t ask me.’ Er besteigt den zweithöchsten Berg. [bεɐç] ‘He climbs the second-highest mountain.’ Ich frage mich, ob dieser Entwurf etwas taugt. [taυxt] ‘I wonder if this outline can be of use.’ Ich gehe heute noch nicht weg. [vεç] ‘I’m not going away/out today.’ Ich wünsche dir viel Erfolg. [εɐfɔlç] ‘I wish you every success.’ Das ist ja eigentlich ein Fall von Betrug. [bətʁυx] ‘This actually is a matter of cheating.’ Wir treffen uns dann auf dem Bahnsteig. [banʃtaiç] ‘Let’s meet on the station platform.’ Ist das denn überhaupt möglich? [møçlic] ‘Is that even possible?’ Die Person wurde erwürgt. [εɐvyɐçt] ‘The person was strangled.’ Ich weiß nicht, ob das genügt. [ənyçt] ‘I don’t know if that is enough.’ Das hat mich noch lange verfolgt. [fɐfɔlçt] ‘That has haunted me for a long time.’ Das kommt davon, wenn man solange in der Sonne liegt. [liçt] ‘That’s what happens when one lies in the sun for so long.’ Ich weiß es auch nicht, er hat mir nichts gesagt. [əzaxt] ‘I don’t know either, he hasn’t told me anything.’ Das ist aber ein schönes Zeugnis. [tsoyçnis] ‘That is indeed a good school certificate.’ Das war eigentlich nicht sehr klug. [klux] ‘That was not really very clever.’ Sie haben immer gut für ihn gesorgt. [əzɔɐçt] ‘They have always taken good care of him.’

[x] after long velar vowel [ç] after liquid consonant [x] after velar diphthong [ç] after short palatal vowel [ç] after liquid consonant [x] after short velar vowel [ç] after palatal diphthong

[ç] after long palatal vowel [ç] after liquid consonant [ç] after short palatal vowel [ç] after liquid consonant [ç] after long palatal vowel

[x] after short velar vowel

[ç] after palatal diphthong

[x] after long velar vowel [ç] after liquid consonant

39

Dialectal concepts of space and linguistic variation 17)

18)

19)

20) 21)

22) 23)

24) 25) 26) 27) 28) 29)

30)

Er hat heute alle Eintragungen aus der Liste getilgt. [ətilçt] ‘Today, he deleted all records from the list.’ Das passierte damals, als er nach Bielefeld zog. [tsox] ‘That happened back when he moved to Bielefeld.’ Er hat den ganzen Vormittag staubgesaugt. [ʃtaυpəzaυxt] ‘He has vacuumed all morning.’ Sie hat jetzt ihr zweites Kind gekriegt. [əkʁiçt] ‘She has now had her second child.’ Ich glaube, dass der Turm sich nach Osten neigt. [naiçt] ‘I believe that the tower leans to the east.’ Der Ritter reitet zurück zu seiner Burg. [bυɐç] ‘The knight rides back to his castle.’ Gib ihm das Buch, wenn er dich morgen fragt. [fʁaxt] ‘Give him the book when he asks you tomorrow.’ Das Projekt ist auf einem guten Weg. [veç] ‘The project is doing well.’ [Weg = way] Keiner wusste, woran das liegt. [liçt] ‘Nobody knew what the reason was.’ [liegt = lies] Tatsächlich hat er mir nichts davon gesagt. [əzaxt] ‘Actually, he hasn’t told me anything about it.’ Nach drei Wochen hatte ich aber genug. [ənυx] ‘After three weeks I had enough.’ Es ist zwecklos, das zu leugnen. [lɔyçnən] ‘It is pointless to deny this.’ Der Detektiv war ihm lange gefolgt. [əfɔlçt] ‘The detective had been following him for a long time.’ So, ich glaube, das genügt. [ənyçt] ‘Well, I think, that will suffice.’

[ç] after liquid consonant

[x] after long velar vowel

[x] after velar diphthong

[ç] after short palatal vowel [ç] after palatal diphthong

[ç] after liquid consonant [x] after short velar vowel

[ç] after long palatal vowel [ç] after short palatal vowel [x] after long velar vowel [x] after short velar vowel [ç] after palatal diphthong [ç] after liquid consonant

[ç] after long palatal vowel

Fig. 1: Sentences for testing the situational acceptance of words with final g as a fricative (instead of standard German [k])

40

Michael Elmentaler

Fig. 2: Results for Schleswig-Holstein, answer: “never” (N=169)

The informants were instructed to pay attention to the last word of each sentence. In a questionnaire containing the 30 sentences above, informants specified which situations they would regard appropriate for the fricative g-form and whether they themselves would use it: a) “in all situations” (in allen Situationen), b) “only in casual speech” (nur in lockerer Redeweise) or c) “never” (niemals). First, let us have a look at the results for the 169 informants from Schleswig-Holstein in Northern Germany. On average, 47 % stated that they would “never” use the fricative variants. However, there were large differences between the lexemes (Fig. 2). In the case of the word variant [lɔyçnən] ‘deny’ (No. 28) (instead of standard German [lɔyknən] leugnen), 143 informants (85 %) stated that they would “never” use this form. Similar high percentages for the answer “never” can be found for the variants [εɐfɔlç] ‘success’ (No. 5) with 83 % and [əfɔlçt] ‘followed (participle)’ (No. 29) with 82 % (instead of standard German [εɐfɔlk] Erfolg and [əfɔlkt] gefolgt). On the other hand, for the variants [əzaxt] ‘said (participle)’ (No. 13), [əkʁiçt] ‘got (participle)’ (No. 20) and [liçt] ‘lies’ (No. 12) (standard German [əzakt] gesagt, [əkʁikt] gekriegt, [likt] liegt), only 7–10 % of the informants answered with “never”. Hence, the acceptability of the fricative g depends highly on the lexeme in which it occurs. For some contexts, the participants in the study stated that they would specifically avoid the substandard variants. Only one lexeme is considered acceptable “in all situations” by the majority of the informants (Fig. 3).

Dialectal concepts of space and linguistic variation

41

Fig. 3: Results for Schleswig-Holstein, answer: “in all situations” (N=169)

More than half of the informants (56 %) claimed that they would use the variant [əzaxt] instead of standard German [əzakt] in all situations. The variants [əkʁiçt] ‘got (participle)’ (No. 20) and [liçt] ‘lies’ (No. 12) yielded 37 % and 39 %, respectively, and for most of the variants the values are even lower. This means that most of the word forms with the fricative g are regarded as colloquial forms that should be avoided in formal conversation. The interindividual variability in accepting or rejecting such variants points to the existence of regional attitudinal structures which influence language behavior. In the other two regions investigated, we find amazingly similar structures, albeit on different levels. Figures 4 and 5 compare the results for the 169 informants from Schleswig-Holstein with those for the 184 Westphalians and the 73 informants from Lower Saxony and Bremen. First of all, it can be stated that the percentages for the Westphalian informants (white squares in Fig. 4) are in 25 of 30 cases higher than those for the people from Schleswig-Holstein (grey squares), and the percentages for the people from Lower Saxony/Bremen (black squares) are in 23 of the 30 lexemes even higher than those for the Westphalians. Thus, on the whole, the people from Westphalia and especially Lower Saxony tend to reject those forms more strongly than the students from Schleswig-Holstein (Fig. 5). This might be explained by the fact that Low German in Westphalia and in most parts of Lower Saxony (except East Frisia) is no longer in use in everyday life, whereas in Schleswig-Holstein many people still have at least passive competence in the Low German language. The stronger presence of the

42

Michael Elmentaler

Fig. 4 and 5: Results for Schleswig-Holstein (grey; N=169), Westphalia (white; N=184) and Lower Saxony (black; N=73), answer: “never”

Low German dialect in the very North of Germany might lead to a generally higher acceptance of substandard variants. Only in three cases (No. 17: [ətilçt] getilgt ‘deleted’; No. 11: [fɐfɔlçt] verfolgt ‘haunted’; No. 14: [tsoyçnis] Zeugnis ‘certificate’) do we find lower values in the Westphalian data, which means that the acceptance for the fricativized g is higher. For some word forms, the differences in frequency can be explained by the fact that they are commonly used in Schleswig-Holstein but not so much in Westphalia. This holds, for example, for some substandard variants with short vowels such as No. 10: [ənyçt] genügt ‘suffices’, No. 23: [fraxt] fragt ‘asks’ and No. 25: [liçt] liegt ‘lies’. In Westphalia, the verb forms with a long vowel ([ənyçt], [fraxt], [liçt]) are more common.

Dialectal concepts of space and linguistic variation

43

Fig. 6: Results for Schleswig-Holstein (grey), Westphalia (white) and Lower Saxony/ Bremen (black), answer: “never” (sorted by the criterion ‘phonetic context’)

Apart from these regional differences, the findings suggest that the quality of the preceding sound plays a certain role in the acceptance of the fricative realization of final g (Fig. 6). In the position after r or l, as in [bεɐç] ‘mountain’, [əzɔɐçt] ‘taken care of ’, [fɐfɔlçt] ‘haunted’ and [ətilçt] ‘deleted’ (standard German [bεʁk] Berg, [əzɔʁkt] gesorgt, [fεɐfɔlkt] verfolgt, [ətilkt] getilgt), the fricative is rejected by an average of 68 to 80 % of the informants who declared that they would “never” use these forms. In the position after long vowels, however, only 44 to 61 % reject it, and after short vowels only 34 to 48 %. This is no random distribution, for the same tendencies can be found in all three regions.3 Therefore, there are similar attitudinal structures which allow the fricative realization of the g to appear more acceptable after vowels than after liquid consonants. Here we find a striking parallel to the conditions for the fricativization of g in initial and medial position. For example in Brandenburg and Berlin, fricative g is rather frequent before and between vowels ([jut] gut ‘good’, [jəzen] gesehen ‘seen’, [ʁejən] Regen ‘rain’), but much less 3

Percentages for the answer “never”: g after r, l: Schleswig-Holstein: 68 %, Westphalia: 70 %, Lower Saxony and Bremen: 80 %; g after long vowel and diphthong: Schleswig-Holstein: 44 %, Westphalia: 52 %, Lower Saxony and Bremen: 61 %; g after short vowel: Schleswig-Holstein: 34 %, Westphalia: 46 %, Lower Saxony and Bremen: 48 %.

44

Michael Elmentaler

common before or after r and l ([jʁyn] grün ‘green’, [jlyk] Glück ‘luck’, [fɔljən] folgen ‘follow’).

4.

Conclusion

The survey about the perception and evaluation of fricativized g in three regions of Northern Germany has shown that spatial structures depend on feature selections and on the degree of abstraction in the description. Even with regard to one single phonological category as in the case of the realization of g, the results are quite different depending on the lexemes and contexts. Word forms like [əkʁiçt] gekriegt ‘got (participle)’, [əzaxt] gesagt ‘said (participle)’, [ənυx] genug ‘suffices’ or [liçt] liegt ‘lies’ are quite commonly accepted in all three regions so that, on the basis of the individual ratings, we can describe a broad Northern substandard variety that differs sharply from the standard language and the southern substandard varieties. But when we look at word forms like [klux] klug ‘clever’, [lɔyçnən] leugnen ‘deny’, [εɐfɔlç] Erfolg ‘success’ or [əfɔlçt] gefolgt ‘followed (participle)’, the picture is rather different, since in these cases the acceptance of substandard variants is much lower. Moreover, for word forms like [ənyçt] genügt ‘suffices’, [fraxt] fragt ‘asks’ and [liçt] liegt ‘lies’ (with short vowels), there are obvious differences between the regions, as the acceptance in Westphalia is considerably lower than in Schleswig-Holstein. On the other hand, substandard variants have different spatial ranges depending on the situative context that is concerned. In Schleswig-Holstein, 18 of the 30 word forms tested are regarded by a majority of the informants as acceptable in a relaxed speech situation, but only one word form seems to be appropriate for more formal situations (cf. Fig. 3). Therefore, we find relatively strict conventions about using substandard forms in the spoken varieties of Northern Germany, but the rules for their situationally and socially appropriate use are rather complex. A modern areal linguistics that does justice to the diversity of this situational and contextual variation in speech has to abandon the idea of clear-cut variety boundaries. More appropriate than a combination map with bundles of isoglosses sharply separating dialects would be a series of maps in which the spatial structures for different speech levels, different linguistic features and different types of speakers are represented.

Dialectal concepts of space and linguistic variation

45

References Auer, Peter 1997 Co-occurrence restrictions between variables: A case for social dialectology, phonological theory and variation studies. In: Frans Hinskens, Roeland van Hout and W. Leo Wetzels (eds.), Variation, change and phonological theory, 71–102. (Amsterdam studies in the theory and history of linguistic science, Series 4: Current issues in linguistic theory 146.) Amsterdam: Benjamins. Bellmann, Günter 1986 Zweidimensionale Dialektologie. In: Günter Bellmann (ed.), Beiträge zur Dialektologie am Mittelrhein, 1–55. (Mainzer Studien zur Sprachund Volksforschung 10.) Stuttgart: Steiner. Bellmann, Günter, Joachim Herrgen and Jürgen Erich Schmidt 1994–2002 Mittelrheinischer Sprachatlas, Vol. 1–5. Tübingen: Niemeyer. Blaser, Jutta 2002 Geolinguistik in Iberoamerika. Zum Stand der Forschung im deutschsprachigen Raum. Neue Romania 25: 11–30. Chambers, Jack K. and Peter Trudgill 1998 Dialectology, 2nd edition. (Cambridge textbooks in linguistics.) Cambridge: University Press. Eggers, Eckard, Jürgen Erich Schmidt and Dieter Stellmacher (eds.) 2005 Moderne Dialekte – neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD) am Forschungsinstitut für deutsche Sprache “Deutscher Sprachatlas” der Philipps-Universität Marburg vom 5.–8. März 2003 (ZDL-Beihefte 130.) Stuttgart: Steiner. Elmentaler, Michael 2006a Dialektgeografie im Wandel: Polystratische Sprachgeografie und variationslinguistisch basierte Areallinguistik. In: Anja Voeste and Joachim Gessinger (eds.), Dialekt im Wandel. Perspektiven einer neuen Dialektologie, 17–31. (Osnabrücker Beiträge zur Sprachtheorie 71.) Duisburg: Gilles & Francke. Elmentaler, Michael 2006b Sprachlagenspektren im arealen Vergleich. Vorüberlegungen zu einem Atlas der deutschen Alltagssprache. Zeitschrift für Dialektologie und Linguistik 73: 1–29. Elmentaler, Michael, Joachim Gessinger and Jan Wirrer 2010 Qualitative und quantitative Verfahren in der Ethnodialektologie am Beispiel von Salienz. In: Christina Ada Anders, Markus Hundt and Alexander Lasch (eds.), Perceptual dialectology – Neue Wege der Dialektologie, 111–149. (Linguistik – Impulse und Tendenzen 38.) Berlin/New York: de Gruyter. Elspaß, Stephan and Robert Möller 2003–present Atlas zur deutschen Alltagssprache (AdA). http://www.philhist.uni-augsburg.de/lehrstuehle/germanistik/sprachwissenschaft/ada/ Foerste, William 1957 Geschichte der niederdeutschen Mundarten. In: Wolfgang Stammler (ed.), Deutsche Philologie im Aufriss 2., überarb. Aufl., Bd. 1, 1730–1898. Berlin: Erich Schmidt. Friebertshäuser, Hans and Heinrich J. Dingeldein 1989 Neue Dimensionen areallinguistischer Forschung. Die Konzepte des “Wortatlas der deutschen Umgangssprachen”, der “Wortgeographie der städtischen Alltagssprachen in Hessen” und des “Mittelrheinischen Sprachatlas” im Vergleich. In: Wolfgang Putschke, Werner Veith and Peter Wiesinger (eds.), Dialektgeographie und Dialektologie. Günter Bellmann zum 60. Geburtstag von seinen Schülern und Freunden, 110–125. (Deutsche Dialektgeographie 90.) Marburg: Elwert. Gessinger, Joachim 2008a Dialektologie und Perzeptive Spracheinstellungsforschung. Niederdeutsches Jahrbuch 131: 133–144.

46

Michael Elmentaler

Gessinger, Joachim 2008b Ethnodialektologie und sprachlicher Wandel. In: Thomas Stehl (ed.), Kenntnis und Wandel der Sprachen. Beiträge zur Potsdamer Ehrenpromotion für Helmut Lüdtke, 57–78. (Tübinger Beiträge zur Linguistik 507.) Tübingen: Narr. Herrgen, Joachim 2001 Die Dialektologie des Deutschen. In: Sylvain Auroux, E. F. K. Koerner, Hans-Josef Niederehe and Kees Versteegh (eds.), History of the language sciences. An International Handbook on the Evolution of the Study of Language from the Beginnings to the Present, Volume 2, 1513–1535. (Handbücher zur Sprach- und Kommunikationswissenschaft 18.2.) Berlin/New York: de Gruyter. Hinderling, Robert, Werner König, Ludwig Eichinger, Hans-Werner Eroms, Horst Haider Munske and Norbert Richard Wolf (eds.) 1996–2010 Bayerischer Sprachatlas. Vol. 1: Sprachatlas von Bayerisch-Schwaben (SBS). Vol. 2: Sprachatlas von Mittelfranken (SMF). Vol. 3: Sprachatlas von Unterfranken (SUF). Vol. 4: Sprachatlas von Nordostbayern (SNOB). Vol. 5: Sprachatlas von Niederbayern (SNiB). Vol. 6: Sprachatlas von Oberbayern (SOB). Heidelberg: Winter. Lauf, Raphaela 1996 ‘Regional markiert’: Großräumliche Umgangssprache(n) im niederdeutschen Raum. Niederdeutsches Jahrbuch 119: 193–218. Lenz, Alexandra 2010 Zum Begriff der Salienz und zum Nachweis salienter Merkmale. In: Christina Ada Anders, Markus Hundt and Alexander Lasch (eds.), Perceptual dialectology – Neue Wege der Dialektologie, 89–110. (Linguistik – Impulse und Tendenzen 38.) Berlin/New York: de Gruyter. Martin, Victoria C. 1996 Modelle der Umgangssprache. Überlegungen zum theoretischen Status eines linguistischen Begriffs am Beispiel des Wiener Deutsch. Zeitschrift für Dialektologie und Linguistik 63: 129–156. Menge, Heinz H. 1997 Noch einmal von vorn? Zur Systematisierung der sprachlichen Variation im Ruhrgebiet. In: Konrad Ehlich, Wilhelm Elmer and Rainer Noltenius (eds.), Sprache und Literatur an der Ruhr, 2., erw. und überarb. Aufl., 39–55. (Schriften des Fritz-Hüser-Instituts für Deutsche und Ausländische Arbeiterliteratur der Stadt Dortmund, Reihe 2: Forschungen zur Arbeiterliteratur 10.) Essen: Klartext-Verlag. Mihm, Arend 1997 Die Realität des Ruhrdeutschen – soziale Funktion und sozialer Ort einer Gebietssprache. In: Konrad Ehlich, Wilhelm Elmer and Rainer Noltenius (eds.), Sprache und Literatur an der Ruhr, 2., erw. und überarb. Aufl., 19–38. (Schriften des Fritz-Hüser-Instituts für Deutsche und Ausländische Arbeiterliteratur der Stadt Dortmund, Reihe 2: Forschungen zur Arbeiterliteratur 10.) Essen: Klartext-Verlag. Mihm, Arend 2000 Die Rolle der Umgangssprachen seit der Mitte des 20. Jahrhunderts. In: Werner Besch, Anne Betten, Oskar Reichmann and Stefan Sonderegger (eds.), Sprachgeschichte. Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung, 2., überarb. Aufl., Zweiter Teilband, 2107–2137. (Handbücher zur Sprach- und Kommunikationswissenschaft 2.2.) Berlin/New York: de Gruyter. Möller, Robert 2006 Mögliches und Unmögliches zwischen Dialekt und Standard. Kookkurenzrestriktionen als Zugang zur Struktur regionaler Umgangssprache im Rheinland. In: Anja Voeste and Joachim Gessinger (eds.), Dialekt im Wandel. Perspektiven einer neuen Dialektologie, 101–117. (Osnabrücker Beiträge zur Sprachtheorie 71.) Duisburg: Gilles & Francke. Niebaum, Hermann and Jürgen Macha 2006 Einführung in die Dialektologie des Deutsche, 2., neubearb. Aufl. (Germanistische Arbeitshefte 37.) Tübingen: Niemeyer.

Dialectal concepts of space and linguistic variation

47

Preston, Dennis 2002 Language with an attitude. In: Jack K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 43–51. (Blackwell Handbooks in Linguistics.) Malden/Mass: Blackwell. Quesada Pacheco, Miguel Angel and Juan Antonio Martínez López 2008 El Atlas Lingüístico Pluradimensional de América Central. Estado actual del proyecto. Revista internacional de lingüística iberoamericana 6 (11): 101–109. Radtke, Edgar and Harald Thun (eds.) 1996 Neue Wege der romanischen Geolinguistik. Akten des Symposiums zur empirischen Dialektologie (Heidelberg, Mainz 21.–24. 10. 1991). (Dialectologia pluridimensionalis Romanica 1.) Kiel: Westensee. Salewski, Kerstin 1998 Zur Homogenität des Substandards älterer Bergleute im Ruhrgebiet. (ZDL-Beihefte 99.) Stuttgart: Steiner. Schröder, Ingrid 2004 Niederdeutsch in der Gegenwart: Sprachgebiet – Grammatisches – Binnendifferenzierung. In: Dieter Stellmacher (ed.), Niederdeutsche Sprache und Literatur der Gegenwart, 35–97. (Germanistische Linguistik 175/176.) Hildesheim: Olms. Schröder, Ingrid and Michael Elmentaler 2009 Sprachvariation in Norddeutschland (SiN). Niederdeutsches Jahrbuch 132: 41–68. Thun, Harald 2002 Atlas lingüístico Guaraní-Románico: sociología (ALGR-S). Con la colab. de María Pereira Jacquet, Andreas Harder, Martín Ramírez Machuca y Johanne Peemöller. (Dialectologia pluridimensionalis Romanica 2.) Kiel: Westensee. Voeste, Anja and Joachim Gessinger (eds.) 2006 Dialekt im Wandel. Perspektiven einer neuen Dialektologie. (Osnabrücker Beiträge zur Sprachtheorie 71.) Duisburg: Gilles & Francke.

48

Sandra Hansen

Sandra Hansen (Albert-Ludwigs-Universität Freiburg i. Br.)

Dialektalität, Dialektwissen und Hyperdialektalität aus soziolinguistischer Perspektive

1.

Einleitung

Um den Einfluss von soziolinguistischen Parametern auf das Dialektwissen eines Sprechers zu untersuchen, spielt die Quantifizierung der Dialektkenntnis eine wesentliche Rolle. Doch wie ist es möglich, das Dialektwissen eines Sprechers zu quantifizieren? Ein methodisches Vorgehen, welches in diesem Beitrag vorgestellt wird, ist die Ermittlung des sprecherspezifischen Dialektwissens durch eine Dialektabfrage und die anschließende Messung der Dialektalität der geäußerten Lexeme. Unter Dialektwissen beziehungsweise Dialektkenntnis verstehe ich ein aktives Sich-bewusst-sein von lautlichen und grammatischen Eigenschaften eines Dialektes in einer Dialektabfrage. Es werden folgende Fragestellungen diskutiert: – Wie lässt sich die Dialektalität einer elizitierten Äußerung messen? – Woran kann es liegen, dass manche Personen ein höheres Dialektwissen aufweisen und somit mehr über die lautliche Gestalt des heimatlichen Dialektes wissen als andere? – Wie kommt es dazu, dass manche Informanten bei der Dialektabfrage dialektalere Varianten realisieren, als man in ihrem Geburts- und Wohnort erwarten würde? Wie lassen sich diese Formen charakterisieren und gibt es bestimmte Sprechergruppen, die solche „hyperdialektalen“ Formen besonders häufig verwenden? – Gibt es einen Zusammenhang zwischen der Häufigkeit von hyperdialektal artikulierten Lauten und dem Dialektwissen der einzelnen Sprecher? Als Grundlage der Analyse dienen nach soziolinguistischen Kriterien differenzierte Daten, die im Rahmen des Projekts „Regionaldialekte im alemannischen Dreiländereck“ (Universität Freiburg)1 erhoben wurden und in Ab1

Ziel des Forschungsvorhabens ist es, die tatsächlich im Alltag von durchschnittlichen, ortsgebundenen Sprechern verwendeten Dialekte und ihre Struktur in Bezug auf den phonologischen Dialektwandel flächendeckend zu erfassen. Zu diesem Zweck wurden in insgesamt 37 Orten des oberrheinalemannischen und

Dialektalität, Dialektwissen und Hyperdialektalität

49

schnitt 2 besprochen werden. In Abschnitt 3 wird ein speziell für diese Arbeit entwickeltes Messverfahren vorgestellt, mit denen sprecherspezifische Dialektalitätswerte ermittelt werden. Mit Hilfe statistischer Modelle werden Einflüsse von bestimmten außersprachlichen Variablen auf die Dialektkenntnis der einzelnen Sprecher aufgedeckt. Abschnitt 4 beschäftigt sich mit verschiedenen Typen von hyperdialektalen Formen und der Messung von Hyperdialektalitätsindizes. Es wird überprüft, inwiefern außersprachliche Faktoren einen Einfluss auf das Vorkommen von hyperdialektalen Formen haben. Außerdem wird in Abschnitt 5 ein Bezug zu den Ergebnissen der Dialektalitätsgradmessungen hergestellt, indem überprüft wird, ob das Dialektwissen der Sprecher Auswirkungen auf die Vorkommenshäufigkeit von Hyperdialektalismen hat.

2.

Datenmaterial

In diesem Beitrag werden die auf der Karte in Abbildung 1 dargestellten Orte berücksichtigt. Das Ortsnetz besteht aus kleineren Dörfern, wie z. B. Schönenberg mit 337 Einwohnern2, und kleineren Städten, wie z. B. Schopfheim mit 18 910 Einwohnern3. Um einen diachronen Vergleich durchführen zu können, wurden die Orte in Anlehnung an den Südwestdeutschen Sprachatlas (SSA)4 ausgewählt. Ausgehend von der Hypothese, dass sich das Sprachbewusstsein unter Rückgriff auf sozialwissenschaftliche Kategorien beschreiben und erklären lassen (vgl. Löffler 2001: 33), werden in den Orten jeweils sechs Personen aus unterschiedlichen Geschlechter-, Alters- und Berufsklassen interviewt. Vier Sprecher pro Ort sind einer älteren Altersgruppe zugehörig (zwischen 60 und 70 Jahre alt) und zwei Sprecher einer jüngeren Altersklasse (zwischen

2

3

4

hochalemannischen Sprachraums (in Deutschland, in der Schweiz und im Elsass) Gewährspersonen befragt. Die Interviews bestanden aus zwei Teilen: einer Abfrage des Dialektwissens und einer ethnodialektologischen Befragung, die der phonologischen Studie in dem Projekt als spontansprachliches Material dient (zur inhaltlichen Auswertung der ethnodialektologischen Befragung vgl. den Beitrag von Stoeckle in diesem Band). Vlg. http://www.orte-in-deutschland.de/23885-gemeinde-schoenenberg-schwarzwald.html eingesehen am 21. 08. 2010. Vgl. http://www.orte-in-deutschland.de/21430-gemeinde-neuenweg.html eingesehen am 21. 08. 2010. Zur ausführlichen Beschreibungen des Projekts Südwestdeutscher Sprachatlas vgl. Steger und Schupp (1993, 1998), König und Schrambke (1999: 102–115) und Schrambke (2009).

50

Sandra Hansen

Abbildung 1: Übersicht über die Untersuchungsorte

25 und 35 Jahre alt). Zwei Sprecher aus der älteren Altersklasse (ein Mann und eine Frau) arbeite(te)n in einem handwerklichen oder landwirtschaftlichen Berufsfeld. Typische Berufe für dieses Berufsfeld sind beispielsweise Landwirt, Schreiner oder Drechsler. Die anderen zwei sowie die zwei Sprecher aus der jüngeren Altersklasse (jeweils ein Mann und eine Frau) üben einen kommunikationsorientierten Beruf aus, d. h., dass diese Personen oft einen höheren Bildungsabschluss haben und die geschriebene bzw. gesprochene Sprache in ihrem beruflichen Alltag eine wichtige Rolle spielt (vgl. Mattheier 1994: 427–428). Die Informanten aus diesem Berufsfeld sind typischerweise Verwaltungsbeamte im höheren Dienst oder Projektmanager in unterschiedlichen Bereichen.

Dialektalität, Dialektwissen und Hyperdialektalität

51

Als Datengrundlage der hier vorgestellten Analyse dient eine im Rahmen der Interviews durchgeführte Dialektabfrage, die zum Ziel hat, Art und Umfang der Dialektkenntnis der einzelnen Sprecher zu eruieren. Die Informanten werden in dieser Interviewsituation angehalten, die von dem Exploratoren abgefragten Sätze in den intendierten Ortsdialekt zu übersetzen. Insgesamt werden 45 verschiedene phonologische und morphophonologische Phänomene in 32 unterschiedlichen Lexemen berücksichtigt (siehe Anhang). Die Lexeme sind aus dem Fragebuch des SSA entnommen, sodass diese von 1974 bis 1986 erhobenen Daten dieser Untersuchung als Vergleichsgrundlage dienen. Die in Trägersätze eingebetteten Lexeme werden mit Fokus auf die zu untersuchenden Phänomene nach dem Internationalen Phonetischen Alphabet (IPA) transkribiert. Das Vorgehen wird an zwei Beispielen demonstriert: a) Der Trägersatz Er hat sich den Kopf angeschlagen beinhaltet das für die Untersuchung relevante Lexem Kopf. Für die Dialektgliederung des Untersuchungsgebietes sind zwei Phänomene bedeutend: die wortinitiale /k/-Frikativierung oder /k/-Affrizierung im Hochalemannischen sowie die Hebung des offenen /ɔ/ zu [o] (vgl. hierzu SSA Karte II/6.03 „o vor labialer Affrikata Kopf/Kropf“ und SSA Karte II/105.00 „Westgerm. k im Anlaut vor Vokal“). Varianten, die im vorliegenden Datenmaterial je nach basisdialektalem Bezugssystem vorkommen, sind5: [xopf], [xɔpf], [kopf] und die standardsprachliche Variante [kɔpf]. b) In dem Satz Wer zuviel trinkt, bekommt Probleme mit der Leber geht es um drei Variablen innerhalb des Lexems Leber : die Frikativierung von /b/ zu [v], die Senkung des Stammvokals /e/ zu [ε], [] oder [a] und die Dreisilbigkeit des Lexems (vgl. hierzu die Belege aus den SSA-Abfragedaten zu Frage 264/003 „Leber“). Je nach ortsspezifischer Grundmundart sind folgende Varianten im Datenmaterial belegt: [lavəʁə], [lvəʁə], [lεvəʁə], [labəʁə], [lbəʁə], [lεbəʁə], [lεvɐ], [lbɐ], [lεbɐ] und die standardnahe Variante [lebɐ]6. Folgende Realisierungsvarianten sind aufgrund von Kookkurrenzrestriktionen nicht möglich: [lavɐ], [lvɐ], [levɐ] [labɐ], [lebəʁə] und [levəʁə].

5

6

Die Aspiration von anlautendem /k/ wird nicht mittranskribiert, da es diesbezüglich keine Variation gibt und sie deshalb für die Gliederung des alemannischen Sprachraums nicht relevant ist. Der bilabiale Plosiv wird im Untersuchungsgebiet auch in der gesprochenen Standardsprache entstimmt und lenisiert artikuliert.

52

3.

Sandra Hansen

Messung der phonetischen Dialektalität sprachlicher Äußerungen

3.1. Methode Die Messungen entstehen vor dem Hintergrund eines sprachlichen Kontinuums zwischen den beiden Extrempolen Dialekt und Standard (vgl. Auer 2005). Die einzelnen Segmente der transkribierten Lexeme werden mit den Daten aus dem SSA, die als basisdialektales Vergleichskorpus dienen, verglichen. Daraus ergeben sich Differenzen, die sowohl auf qualitativen Abweichungen in Bezug auf Artikulationsart oder -ort, Phonation, Lippenrundung als auch auf Unterschiede in der Länge der einzelnen Laute basieren. Es wird gemessen, indem den relevanten Phänomenen je nach phonetischem Abstand zur grundmundartlichen Variante und zur mündlich realisierten Standardsprache bestimmte Werte zugeordnet werden: Wenn der realisierte Laut innerhalb des betreffenden Lexems der traditionellen Variante nach dem SSA entspricht, wird der Wert 1 (höchste Dialektalitätszuschreibung für einen Laut) vergeben. Entspricht der realisierte Laut dagegen der standardsprachlichen Variante, so wird der Wert 0 (niedrigste Dialektalitätszuschreibung für einen Laut) vergeben. Bei den Lauten, die sich phonetisch zwischen der standardsprachlichen und der dialektalen Variante befinden, ist die Zuordnung schwieriger. Diesen Lauten müssen Werte zugeordnet werden, die den phonetischen Differenzen der realisierten Laute zum gesprochenen Standard und zur traditionellen grundmundartlichen Variante entsprechen. Aus diesem Grund wird zunächst phänomenweise die Anzahl der realisierten Laute ermittelt, die phonetisch zwischen der Standardvariante und der basisdialektalen Variante liegen. Anschließend wird diese Anzahl in Bezug gesetzt zu dem Wert für die höchste Dialektalitätszuschreibung (1). Das geschieht, indem der Wert 1 (= höchste Dialektalitätszuschreibung) durch die ermittelte Anzahl an Varianten dividiert wird. Der Quotient dieser Berechnung ist der Dialektalitätswert, der dem betreffenden Laut zugeordnet wird. Auf diese Weise werden den Varianten eines Phänomens, die bezüglich der Dimensionen des Vokalvierecks phonetisch zwischen der traditionellen Dialekt- und der Standardvariante liegen, Werte zugewiesen, die diesen Abständen entsprechen.7 Zur Verdeutlichung des hier angewendeten Messverfahrens werden im Folgenden die Messungen der relevanten Phänomene in dem oben beschrie7

Inwiefern diese gemessenen Abstände mit dem perzeptuellen Eindruck der Informanten übereinstimmen, ist unklar und muss in weiteren Arbeiten überprüft werden.

Dialektalität, Dialektwissen und Hyperdialektalität

53

benen, abgefragten Lexem Leber am Beispiel des Ortes Endingen dargestellt. Die mündliche Standardvariante der gesprochenen Sprache in Südwestdeutschland lautet [lebɐ], die grunddialektale Form in beispielsweise Endingen ist [lavəʁə] (vgl. hierzu die Belege der SSA-Abfragedaten zu Frage 264/003 „Leber“). Ist der wortmediale Lenisplosiv [b] realisiert, so wird der Wert 0 für die Standardvariante vergeben. Realisiert der betreffende Sprecher allerdings ein basisdialektales [v], so wird diesem Laut der Wert 1 für die dialektal tiefste Variante zugeschrieben. Analog dazu wird für den Stammvokal für ein standardnahes [e] der Wert 0 vergeben. Für den basisdialektalen Vokal, der zugleich repräsentativ für den Vokal mit dem höchsten artikulatorischen Abstand ist, in diesem Fall das [a], wird der Wert 1 vergeben. Außerdem kommen im Datenmaterial zwei Varianten vor, die artikulatorisch zwischen Standardvokal und basisdialektalem Vokal liegen: das [ε] und das etwas offenere []. Gemäß dem Öffnungsgrad werden in diesem Fall die Werte 0,33 für das geschlossenere [ε] und 0,67 für das offenere [] vergeben. In Endingen werden folgende Varianten des Lexems Leber artikuliert: (1) [lεbɐ] von einer jüngeren Frau aus einem kommunikationsorientierten Berufsfeld (2) [lbɐ] von einem jüngeren Mann aus einem kommunikationsorientierten Berufsfeld (3) [lεbɐ] von einer älteren Frau aus einem kommunikationsorientierten Berufsfeld (4) [lbəʁə] von einem älteren Mann aus einem kommunikationsorientierten Berufsfeld (5) [labəʁə] von einer älteren Frau aus einem handwerklichen-landwirtschaftlichen Berufsfeld (6) [lavəʁə] von einem älteren Man aus einem handwerklichen-landwirtschaftlichen Berufsfeld Gemäß dem oben beschriebenen Verfahren bekommen die von den Sprechern artikulierten Lexeme die folgenden Dialektalitätswerte: (1) [lεbɐ]: 0,33 für die Öffnung von [e] zu [ε] + 0 für das standardsprachliche [b] + 0 für die standardnahe Zweisilbigkeit = 0,33 (2) [lbɐ]: 0,67 für die Öffnung von [e] zu [] + 0 für das standardsprachliche [b] + 0 für die standardnahe Zweisilbigkeit = 0,67 (3) [lεbɐ]: 0,33 für die Öffnung von [e] zu [ε] + 0 für das standardsprachliche [b] + 0 für die standardnahe Zweisilbigkeit = 0,33 (4) [lbəʁə] = 0,67 für die Öffnung von [e] zu [] + 0 für das standardsprachliche [b] + 1 für die basisdialektale Dreisilbigkeit = 1,67 (5) [labəʁə] = 1 für den basisdialektalen geöffneten Vokal [a] + 0 für das standardsprachliche [b] + 1 für die basisdialektale Dreisilbigkeit = 2

54

Sandra Hansen

(6) [lavəʁə] = 1 für den basisdialektalen geöffneten Vokal [a] + 1 für das basisdialektale frikativierte [b] + 1 für die basisdialektale Dreisilbigkeit = 3 Nach der Wertevergabe für jedes Phänomen in den betreffenden Lexemen werden die Werte für jeden Sprecher aufsummiert und durch die Summe aller berücksichtigten Phänomene dividiert. Auf diese Weise erhält jeder Sprecher einen eigenen Dialektalitätsindexwert. Im Gegensatz zu dem erprobten Messverfahren von Herrgen und Schmidt (vgl. Herrgen et al. 2001) wird durch die Quantifizierung der phonetischen Abstände zum traditionellen Ortsdialekt berücksichtigt, dass der Dialektalitätswert durch das Lautsystem der zugrunde liegenden Dialekte beeinflusst ist. Denn nach dem Verfahren von Herrgen und Schmidt wird der Abstand zur mündlich realisierten Standardsprache ermittelt, ohne das Lautinventar jeweiliger Basisdialekte zu berücksichtigen. So erhält z. B. eine Informantin aus Endingen, die in Lexemen mit grundmundartlich frikativiertem /b/ den Frikativ realisiert, nach Herrgen und Schmidt aufgrund des höheren artikulatorischen Abstands zur Standardsprache einen höheren Dialektalitätswert wie eine Informantin aus Münstertal, die ein für ihren Wohnort basisdialektales, nicht frikativiertes /b/ realisiert. Im Rahmen des hier angewendeten Messverfahrens würde die Sprecherin aus Endingen den gleichen Dialektalitätswert erhalten wie die Sprecherin aus Münstertal, die ebenfalls die basisdialektale Variante ihres Heimatortes verwendet. Bei der Informantin aus Endingen wäre nach Herrgen und Schmidt der entsprechende Wert für ihre basisdialektale Variante erst gar nicht klassifizierbar. Aufgrund der unterschiedlichen basisdialektalen Bezugssysteme würde dabei die Gefahr bestehen, dass man nicht die sprecherbezogene Dialektalität, sondern vielmehr die Dialektalität eines Ortes misst: die Sprecher in dem Ort Münstertal hätten in unseren Daten im Durchschnitt niedrigere Dialektalitätswerte für dieses Phänomen als die Sprecher in Endingen. Nicht nur das Bezugssystem, sondern auch die Wertevergabe bei Herrgen und Schmidt unterscheiden sich von dem in diesem Beitrag vorgestellten System: Bei Herrgen und Schmidt wird ein Unterschied in einem phonetischen Merkmal mit einem Punkt gewertet. Unterscheidet sich ein Laut in mehreren phonetischen Merkmalen, so bekommt dieser Laut einen Dialektalitätswert, der der Anzahl der Unterschiede entspricht (vgl. Herrgen et al. 2001: 2). Im Vokalbereich wird beispielsweise für einen Kontrast von einer Dimension zur nächsten ein Punkt vergeben; für eine halbe Stufe oder eine halbe Klasse ein halber Punkt (vgl. Herrgen et al. 2001: 2–3). Wenn z. B. der Standardvokal, wie in dem oben diskutierten Beispiel ein [e] ist und der realisierte, gemessene Vokal ein [ε], so bekommt der Laut einen Dialek-

Dialektalität, Dialektwissen und Hyperdialektalität

55

talitätswert von 1, der der einstufigen Abweichung im Öffnungsgrad entspricht. Wenn anstelle des standardsprachlichen [e] ein [a:] realisiert wird, so bekommt der Laut entsprechend den Abweichungen im Öffnungsgrad einen Dialektalitätswert von 2. Der Maximalwert bei vokalischen Differenzen liegt nach diesem System bei 3 Punkten, der bei konsonantischen Abweichungen bei 2 Punkten (vgl. Herrgen et al. 2001: 3–4). Ein Unterschied des hier angewendeten Verfahrens gegenüber dem von Herrgen und Schmidt liegt darin, dass durch den Vergleich zum basisdialektalen Bezugssystem ein maximaler Dialektalitätswert (1) angenommen werden kann. Basisdialektal realisierte Laute bekommen mit diesem Verfahren maximale Dialektalitätswerte unabhängig von ihrem artikulatorischen Abstand zur Standardsprache. Dies bedeutet, dass die berechneten Unterschiede für die Zwischenformen mit einer steigenden Anzahl an Varianten abnehmen. Mit diesem Verfahren wird vermieden, dass Orten, deren basisdialektale Laute artikulatorisch weiter von der Standardsprache entfernt sind, höhere Dialektalitätswerte erhalten, als Ortschaften, in denen die grundmundartlichen Realisierungen näher am Standard sind. Durch den maximalen Dialektalitätswert 1 wird vermieden, dass basisdialektal realisierte Laute, die hohe phonetische Abstände zum Standard aufweisen, überbewertet werden, indem man ihnen additiv für jede phonetische Dimension einen Punkt zuschreibt. In dem hier entwickeltem Verfahren werden durch die oben erläuterte Division (1 geteilt durch die Anzahl an Varianten) die Abstände relativiert. Aufgrund des basisdialektalen Bezugssystems können außerdem hyperdialektal artikulierte Laute identifiziert und aus den Berechnungen herausgenommen werden. 3.2. Hypothesen Da wir – wie oben beschrieben – davon ausgehen, dass die vielfältige Differenzierung sprachlicher Möglichkeiten zwischen Dialekt und mündlicher Standardsprache in der Dialektabfrage von sozialen Faktoren abhängt, sollen mit Hilfe einer statistischen Auswertung im nächsten Kapitel folgende Hypothesen überprüft werden: 1. Bezüglich der Variable Alter ist ein direktes Proportionalitätsverhältnis anzunehmen, d. h. je älter die Sprecher sind, desto höher ist die Wahrscheinlichkeit, dass sie aufgrund ihrer langjährigen Ansässigkeit im Heimatort Dialektwissen besitzen und in der Dialektabfrage höhere Dialektalitätswerte aufweisen (vgl. Niebaum und Macha 2006: 184). Wir gehen also davon aus, dass die Sprecher aus der älteren Altersgruppe im Vergleich zu den jüngeren Sprechern mehr basisdialektale als standardnahe

56

Sandra Hansen

Formen realisieren und deshalb der Dialektalitätsgrad bei den älteren Informanten höher ist als bei den jüngeren. 2. In der Literatur wird häufig davon ausgegangen, dass ein wesentlicher Parameter für das sprachliche Verhalten eines Menschen die Art der Berufstätigkeit ist (vgl. hierzu Niebaum und Macha 2006: 186). Um eine Einteilung vornehmen zu können, werden handwerkliche oder landwirtschaftliche Tätigkeiten von kommunikationsorientierten Tätigkeiten, für die meistens eine gehobene Ausbildung vorausgesetzt wird, unterschieden. Dabei gehen wir davon aus, dass bei handwerklichen oder landwirtschaftlichen Berufen keine standardsprachliche oder standardnahe Kommunikation gefordert wird, da für den Berufsalltag und den beruflichen Erfolg der Personen mit diesem beruflichen Hintergrund die standardnahe schriftliche sowie mündliche Kommunikation eine untergeordnete Rolle spielen. Im Gegensatz dazu werden bei kommunikationsorientierten Berufen meistens überregionale, schriftorientierte Tätigkeiten und/oder kommunikative Tätigkeiten auf verbaler Ebene ausgeführt. Befinden sich Personen aus diesem Berufsfeld in leitender Position, wird von diesen besondere sprachliche Flexibilität und standardsprachliche Kompetenz gefordert (vgl. Lenz 2003: 50). Demnach wird angenommen, dass sich durch den an der Standardsprache orientierten Sprachgebrauch in kommunikationsorientierten Berufen die Dialektkenntnis der Sprecher verringert und Gewährspersonen aus diesem Berufsfeld niedrigere Dialektalitätswerte in der Dialektabfrage aufweisen, als Sprecher aus einem landwirtschaftlichen bzw. handwerklichen Berufsfeld, in dem die standardsprachliche Kommunikation durch das regionale Umfeld selten gefordert ist. 3. Ergebnisse aus Untersuchungen zum geschlechtstypischen Dialektgebrauch werden in der Forschungsliteratur kontrovers diskutiert. Es zeichnen sich zwei gegensätzliche Positionen ab: Einerseits geht man davon aus, dass „die Frau […] zum Bewahren des Dialektes neige und so zum ‘Hort der Mundart’ wird“ (Sieburg 1991: 299), andererseits wird behauptet, dass Frauen zu einer höheren Anpassungsbereitschaft und zu dem sozial Höherstehenden neigen, wodurch eine größere Nähe zur Standardsprache begründet wird (vgl. Sieburg 1991: 299). In diesem Beitrag wird von einem „sozialen Geschlecht“ (Mattheier 1980: 35) ausgegangen, was bedeutet, dass geschlechtsspezifische Unterschiede durch soziale Parameter, wie z. B. durch die verschiedene Position von Frau und Mann in der gesellschaftlichen Sozialordnung, motiviert sind. So ist „die immer wieder festgestellte größere Anpassungsbereitschaft der Frau als ein Ergebnis gezielter Sozialisationsmaßnahmen […] [anzusehen], die schon im frühen Kindesalter einsetzten“ (Mattheier 1980: 35). Aus diesem Grund wird an-

Dialektalität, Dialektwissen und Hyperdialektalität

57

genommen, dass Frauen eher zur standardnahen Realisierungen tendieren und deshalb niedrigere Dialektalitätswerte in der Dialektabfrage aufweisen als Männer. 3.3. Statistische Auswertung Da klassische Varianzanalysen (ANOVAs) bei unbalancierten Designs8 schnell instabil werden, fußt die Berechnung auf einem linearen gemischten Modell (vgl. Pinheiro und Bates 2000): die Einflüsse der außersprachlichen, unabhängigen Faktoren Alter, berufliche Tätigkeit und Geschlecht sowie die Einflussstärke des Zufallsfaktors9 Geographie, in unserem Falle der Erhebungsort, werden modelliert. Das gemischte Modell bestätigt einen signifikanten Kontrast zwischen den beiden Altersgruppen: Jüngere Sprecher weisen einen signifikant niedrigeren Dialektalitätswert in der Dialektabfrage auf als ältere Sprecher (t = –6,79; p < 0,001MCMC10). Außerdem bestätigt das Modell eine Interaktion zwischen den Variablen Alter und Geschlecht (t = 2,35; p < 0,02MCMC ): Es kann gezeigt werden, dass jüngere Frauen einen signifikant niedrigeren Dialektalitätswert aufweisen als jüngere Männer, während es in der älteren Gruppe keinen Geschlechtsunterschied gibt. Beide Ergebnisse spiegeln sich in dem in Abbildung 2 dargestellten Mittelwertvergleich mit Standardfehlern wieder. Informanten aus der älteren Sprechergruppe weisen in der Dialektabfrage höhere Dialektalitätswerte auf als jüngere Sprecher. Der Unterschied zwischen den Männern und Frauen aus der älteren Gruppe ist nicht signifikant. Hypothese 3 wird damit zum Teil bestätigt: Es sind jüngere Frauen und nicht ältere Frauen, die im Vergleich zu den Männern in der jeweiligen Altersgruppe mehr Nähe zum Standard aufweisen und die grundmundartlichen 8

9

10

Die Datenmatrix ist unbalanciert, weil das Untersuchungsdesign bei den jüngeren Sprechern keine Zelle mit Sprechern aus einem handwerklichen bzw. landwirtschaftlichen Berufsfeld vorsieht. Die fehlende Sprechergruppe wurde aus pragmatischen Gründen nicht mit aufgenommen: das Interviewen von 8 Personen pro Ort hätte den Rahmen des Projekts gesprengt. Wir haben uns für jüngere Sprecher aus einem kommunikationsorientierten Berufsfeld entschieden, da diese den besten Gegenpol zu den älteren Handwerkern bzw. Landwirten darstellen. Durch die Aufnahme von Zufallsfaktoren wird die Varianz, die von diesen Faktoren hervorgerufen wird, vor der Berechnung der Einflüsse fester Faktoren „herausgerechnet“, da unsere Untersuchungsorte nur einen Teil der geographischen Abdeckung repräsentieren, ist die Aufnahme als Zufallsfaktor gerechtfertigt. Da die Angabe von Fehlerfreiheitsgraden für gemischte Modelle im Gegensatz zu klassischen ANOVAs nicht möglich ist, wird der Signifikanzwert p über MCMCSimulationen ermittelt (vgl. Bates und Maechler 2010).

58

Sandra Hansen

Abbildung 2: Durchschnittlicher Dialektalitätsgrad in Abhängigkeit von Geschlecht und Altergruppe (Effekt Alter t = –6,79; p < 0,001MCMC; Interaktion Alter* Geschlecht t = 2,35; p < 0,02MCMC )

Realisierungen in der Dialektabfrage vermeiden. Sie haben damit ein niedrigeres Dialektwissen als die jüngeren Männer und die älteren Sprecher. Die Dialektalitätswerte der jüngeren Männer sind signifikant niedriger als die der älteren Sprecher; dieses Ergebnis lässt sich durch den oben diskutieren Alterseffekt erklären. Bei den älteren Sprechern zwischen 60 und 70 Jahren hat das Geschlecht keinen Einfluss auf den Dialektalitätswert bzw. auf die Dialektkenntnis. Es sind demnach vor allem die jüngeren Frauen, die sich an der Standardsprache orientieren. Dieses Ergebnis lässt darauf schließen, dass – wie vermutet – soziale Parameter für dieses Sprachverhalten verantwortlich sind. Gerade die jüngeren Frauen zwischen 25 und 35 befinden sich meistens inmitten ihrer beruflichen Karriere und müssen sich – im Gegensatz zu den älteren Frauen – in ihrem beruflichen Umfeld noch beweisen, um Aufstiegschancen zu haben. Der Unterschied zu den gleichaltrigen und gleich qualifizierten Männern ist durch die vermehrt in kleineren Ortschaften noch häufig vorhandene unterschiedliche Position der Frau in der Gesellschaft zu begründen.

Dialektalität, Dialektwissen und Hyperdialektalität

59

Die zweite Hypothese bezieht sich auf den Einfluss der Art der beruflichen Tätigkeit. Wie in Kapitel 3.2 erläutert, wird angenommen, dass Gewährspersonen aus einem kommunikationsorientierten Berufsfeld eine niedrigere Dialektkenntnis aufweisen als Sprecher, die einen handwerklichen oder landwirtschaftlichen Beruf ausüben. Diese Hypothese kann durch das statistische Modell nicht bestätigt werden. Es lassen sich keine Einflüsse durch die berufliche Tätigkeit auf den Dialektalitätsgrad und somit auf das Dialektwissen der Sprecher feststellen. Ausgehend von der Hypothese, dass ein geringes Dialektwissen zu Unsicherheiten bei der Dialektabfrage führt und deshalb Vorraussetzung für die Produktion von hyperdialektalen Varianten ist (vgl. Besch 2000: 188; Lenz 2005: 86), soll im nächsten Kapitel untersucht werden, ob Sprechergruppen mit einer niedrigeren Dialektkenntnis zu hyperdialektalen Formen neigen. Zuvor wird auf die terminologische Abgrenzung von sog. Hyperdialektalismen eingegangen.

4.

Hyperdialektalismen und Messung der Hyperdialektalität

4.1. Hyperdialektalismen Eine Hyperform ist das Ergebnis einer normwidrigen sprachlichen Annäherung eines Sprechers an eine angestrebte Zielvarietät bzw. Prestigevarietät (vgl. Wurzel 1976: 178 und Lenz 2005: 75). Nach Lenz handelt es sich um „‘fehlerhafte’ Varianten, die aus dem defizitären Regelwissen ihrer Sprecher abgeleitet werden können“ (vgl. Lenz 2005: 75–76). Hierbei gibt es je nach intendierter Zielvarietät (Dialekt oder Standardsprache) zwei Formen: Hyperdialektalismen und Hyperstandardformen. Die Analyse von Hyperformen ist in mehrfacher Hinsicht interessant (vgl. hierzu auch Lenz 2005: 77): Es stellen sich die Fragen, wie die phonologische Struktur von Hyperformen zu erklären ist, wie sie linguistisch motiviert sind und ob sie soziolinguistisch erklärt werden können. Eine weitere Forschungsfrage ist, welche Systemrelevanz den Hyperformen zuzuschreiben ist, d. h. ob sie eher zufällig auftreten oder unter bestimmten Bedingungen in das Sprachsystem aufgenommen werden (vgl. Wurzel 1976: 177). Da die Gewährspersonen während der Dialektabfrage sehr bemüht sind, die dargebotenen Sätze in den intendierten Ortsdialekt zu übersetzen, kommen in dem vorliegenden Datenmaterial keine Hyperstandardformen vor. Aus diesem Grund wird es in den weiteren Ausführungen ausschließlich um den Status von Hyperdialektalismen gehen, die immerhin in 4 % aller Fälle vorkommen (319 von 8887 Phänomenrealisierungen). Als hyperdialektal

60

Sandra Hansen

werden alle Formen angesehen, die einen höheren phonetischen Abstand zur Standardvariante aufweisen als der Basisdialekt im betreffenden Ort. Es lassen sich zwei Arten von Hyperdialektalismen finden: Phänomenrealisierungen, die in der realisierten Form in dem betreffenden Lexem nie auftreten und keinem historischem Lautwandelprozess zuzuschreiben sind (Typ I) und Varianten, die aus dem traditionellen Dialekt anderer, benachbarter Orte übertragen werden (Typ II). Ca. 4 % aller hyperdialektal realisierten Phänomene entstehen nach Typ I (n = 12). Demzufolge sind mit Abstand die meisten Hyperdialektalismen (ca. 96 %; n = 307) Entlehnungen aus benachbarten Orten. Bevor der Einfluss der soziolinguistischen Variablen auf das Vorkommen von hyperdialektalen Varianten in der Dialektabfrage diskutiert wird, werden im Folgenden zwei Beispiele für die unterschiedlichen Typen vorgestellt: Ein Beispiel für Typ I lässt sich an der Variation des Diphthongs [ɔə] am Beispiel des Lexems Mäuse (Trägersatz: Hoffentlich sind da nicht noch mehr Mäuse) zeigen. Der standardsprachliche Diphthong trifft in den hier untersuchten Dialekten – je nach historischen Bezugslauten11 – auf [i] bzw. [i] oder [y] bzw. [y]. Die Karte in Abbildung 3 zeigt die basisdialektale Verteilung nach dem SSA (vgl. die Belege aus den SSA-Abfragedaten zu Frage 302/006 „Maus“ / „Mäuse“ und SSA-Karte II/26.00 „iu vor dentalem Reibelaut Mäuse“). In den Orten nördlich von Malsburg und Todtmoos ist die basisdialektale Variante ein entrundeter, geschlossener, vorderer Monophthong: [mis] oder [mis]. In den Gebieten südlich von Malsburg und Todtmoos ist die grundmundartliche Vokalvariante gerundet12: [mys] oder [mys]. Die traditionellen Varianten nach dem SSA werden von den meisten Sprechern in der Dialektabfrage der aktuellen Daten aus dem REDI-Projekt auch als solche realisiert und sind demzufolge im Dialektwissen der Informanten vorhanden. Zwei Sprecher aus dem Gebiet, in dem basisdialektal ein ungerundeter Monophthong realisiert wird, nennen jedoch die hyperdialektale Variante [miəs]13. Diese basisdialektfremde Variante ist mit dem Prozess der neuhochdeutschen Monophthongierung zu erklären, die u. a. den Laut [iə] in [i] veränderte und vom alemannischen Dialekt nicht angenommen wurde (vgl. Wiesinger 1983a). So ist beispielsweise der mittelhochdeutsche Diphthong [iə] in den Lexemen lieber, nie, Kriese, sieden, fliegen, bieten und schießen in den ale11

12 13

Zum lautgeschichtlichen Hintergrund der neuhochdeutschen Diphthongierung vgl. Wiesinger (1983a). Zum lautlichen Hintergrund der Entrundung vgl. Wiesinger (1983b). Die beiden Sprecher, die die hyperdialektale Variante realisieren, sind ein älterer Mann aus Oberried mit einem handwerklichen Beruf und ein jüngerer Mann aus Bombach mit einer kommunikationsorientierten beruflichen Tätigkeit.

Dialektalität, Dialektwissen und Hyperdialektalität

61

Abbildung 3: Realisierung basisdialektaler Monophthonge und Entrundung in dem Lexem Mäuse der SSA-Abfragedaten (vgl. die Belege aus den SSA-Abfragedaten zu Frage 302/006 „Maus“ / „Mäuse“ und SSA-Karte II/26.00 „iu vor dentalem Reibelaut Mäuse“)

62

Sandra Hansen

mannischen Basisdialekten erhalten (vgl. SSA Karte II/33.00). Die Regel, die die Sprecher unbewusst in diesem Fall anwenden, könnte laiensprachlich formuliert folgendermaßen lauten: „Manchmal entspricht dem standardsprachlichen [i] im Dialekt ein [iə]“. Die hyperdialektale Realisierung beim Lexem Mäuse kommt dadurch zustande, dass die Regel im falschen, nämlich einem schon dialektalen Kontext angewendet wird. Ein Beispiel für einen Hyperdialektalismus nach Typ II bieten die Aussprachevarianten des standarddeutschen offenen, gerundeten, hinteren Vokals [ɔ] in dem Lexem Wolke (Trägersatz: da kommt eine große, schwarze Wolke). In der folgenden Beispielanalyse geht es um die Hebung des Vokals zu [o] oder [υ].14 Die Karte in Abbildung 4 verdeutlicht die basisdialektale geographische Verteilung der realisierten Varianten des Vokals nach dem SSA (vgl. Belege aus den SSA-Abfragedaten zu Frage 310/008 „eine Wolke“). In sechs der 16 Orte15 wird traditionell die stark gehobene Variante [υ] realisiert. Diese Orte sind in der Karte mit drei eng beieinander liegenden schwarzen Linien versehen. Der geschlossene, gerundete, hintere Vokal [o] wird basisdialektal in zehn der untersuchten Orte artikuliert. In WaldshutTiengen sind nach dem SSA zwei Varianten belegt, weshalb basisdialektal hier zwei Varianten angenommen werden (die stark gehobene Variante [υ] ist eine suggerierte Form (vgl. hierzu die Kommentare zu den Belegen aus den SSA-Abfragedaten zu Frage 310/008 „eine Wolke“). Diese Orte sind in der Karte mit einem schwarzen Balken markiert. In dem durch einen gestrichelten schwarzen Balken gekennzeichneten Ort, Schönenberg, liegt basisdialektal das dem Standard entsprechende offene [ɔ] zugrunde. Anhand der Karte lässt sich gut erkennen dass die stark gehobenen hinteren Vokale eher an den Grenzen zur Schweiz und zum Elsass hin artikuliert werden. Der geschlossene Vokal [o] und der artikulatorisch dem Standard nahen, offene Vokal [ɔ] kommen basisdialektal eher im Landesinneren des Erhebungsgebietes vor. Die aktuellen Daten aus der Dialektabfrage aus dem REDI-Projekt zeigen eine Veränderung in der geographischen Verteilung der Vokalvarianten: Wie auf der Karte in Abbildung 5 dargestellt, wird der gehobene Laut [υ], der von allen drei Varianten artikulatorisch am weitesten vom Standard entfernt ist, nicht mehr nur in sechs, sondern in 12 Ortschaften genannt. 14

15

Zur Vokalhebung in den deutschen Dialekten vgl. Wiesinger (1983c); zur Vokalhebung im Rahmen der Vokalspaltung vgl. Seidelmann (2006). Es sind 16 und nicht – wie oben beschrieben – 17 Orte, da aus Schopfheim keine Belege im SSA-Datenmaterial für das Lexem „Wolke“ vorhanden sind.

Dialektalität, Dialektwissen und Hyperdialektalität

63

Abbildung 4: Realisierung der Hebung von [ɔ] in dem Lexem Wolke der SSAAbfragedaten (vgl. Belege aus den SSA-Abfragedaten zu Frage 310/008 „eine Wolke“)

64

Sandra Hansen

Abbildung 5: Realisierung der Hebung von [ɔ] in dem Lexem Wolke der REDI-Abfragedaten von mindestens einem Sprecher pro Ort

Dialektalität, Dialektwissen und Hyperdialektalität

65

In sechs Orten (Herbolzheim, Elzach, Opfingen, Schönenberg, Holzen und Hasel) werden von mindestens einem Sprecher die basisdialektalen Laute [o] und [ɔ] durch diesen hyperdialektalen Laut [υ] ersetzt. In Waldshut-Tiengen wird der im SSA suggerierte Laut [υ] in den aktuellen Daten von zwei Sprechern artikuliert. Insgesamt realisieren 28 Sprecher die gehobene Variante [υ]. 11 dieser Sprecher stammen aus Orten, in denen basisdialektal ein weniger stark gehobener Laut belegt ist.16 Es scheint so, dass die stark gehobene Variante [υ] heute im Dialektwissen der Sprecher verankert ist. Ob sich diese Anhebung auch in der Spontansprache finden lässt, muss in weiteren Studien überprüft werden. 4.2.

Messung der Hyperdialektalität

4.2.1. Methode Um das Vorkommen der Hyperdialektalismen zu quantifizieren und ihre mögliche soziolinguistische Motivation herauszufinden, werden für jeden Sprecher Hyperdialektalitätsgradmessungen durchgeführt. Analog zu den in Kapitel 3.1 beschriebenen Dialektalitätsgradmessungen erfolgt eine Berechnung von Hyperdialektalitätsgraden mit anschließender Überprüfung der Einflüsse von außersprachlichen Sozial-Variablen auf die ermittelten Grade. Durch den Vergleich der basisdialektalen Variante aus dem SSA ist es möglich, die hyperdialektal realisierten Phänomene zu ermitteln: Waren die Abstände eines realisierten Phänomens zum Standard in den aktuellen Daten höher, als in den Daten des SSA, wird dieses als Hyperform klassifiziert. Anschließend werden die Hyperdialektalismen aufsummiert und durch die Anzahl der Phänomene pro Sprecher dividiert. Das Ergebnis ist ein relativer Hyperdialektalitätsgrad pro Sprecher. 4.2.2. Hypothesen Es wird angenommen, dass Sprechergruppen mit einem niedrigen Dialektalitätsgrad bei der Dialektabfrage und einem damit verbundenen geringen Dialektwissen mit dem Dialekt weniger gut vertraut sind und deshalb höhere Hyperdialektalitätswerte aufweisen als Sprechergruppen, die bei der Dialektabfrage höhere Dialektalitätswerte und somit eine höhere Dialektkenntnis haben (vgl. Besch 2000: 188; Lenz 2005: 86). Bezogen auf die soziolinguisti16

Die anderen Sprecher realisieren entweder die standardnahe Variante [ɔ] oder Formen, die zwischen dem stark gehobenen Laut [υ] und dem Standard liegen.

66

Sandra Hansen

schen Variablen Alter, Art der beruflichen Tätigkeit und Geschlecht werden in Abhängigkeit der Ergebnisse der Dialektalitätsgradmessungen in Kapitel 3.3 folgende Hypothesen aufgestellt: 1. Jüngere Sprecher nennen mehr hyperdialektale Formen als ältere Sprecher und haben somit einen höheren Hyperdialektalitätsindexwert, da sie signifikant niedrigere Dialektalitätswerte in der Dialektabfrage aufweisen. 2. Der berufliche Hintergrund hat keinen Einfluss auf die Vorkommenshäufigkeit von Hyperdialektalismen, da sich die Dialektalitätsindizes der Sprecher aus einem handwerklichen bzw. landwirtschaftlichen Berufsfeld von denen aus einem kommunikationsorientierten Berufsfeld nicht signifikant unterscheiden. 3. a) Männliche Gewährspersonen der jüngeren Altersgruppe haben einen geringeren Hyperdialektalitätswert als jüngere Frauen, da sie über ein signifikant höheres Dialektwissen verfügen. b) Bei den älteren Sprechern hat das Geschlecht keinen Einfluss auf das Vorkommen von Hyperdialektalismen, da es keine signifikanten Unterschiede bezüglich der Dialektalitätswerte in der Dialektabfrage gibt. 4.2.3. Statistische Auswertung Um die oben aufgestellten Hypothesen zu testen, wurde aus den gleichen methodischen Gründen wie bei der statistischen Auswertung der Dialektalitätsgrade in Abschnitt 3.3 ein gemischtes Modell gerechnet, indem die Einflüsse der festen Faktoren Alter, berufliche Tätigkeit und Geschlecht sowie die Einflussstärke des Zufallsfaktors Erhebungsort auf die Hyperdialektalitätsgrade getestet werden. Aufgrund der unterschiedlichen linguistischen Erklärungsweisen für Typ I und Typ II erfolgt für jeden Typ eine getrennte Durchführung der statistischen Tests. Formen nach Typ I treten zwar recht selten auf, werden aber von allen Sprechergruppen in gleichem Ausmaß verwendet. Das gemischte Modell zeigt keine sprechergruppenspezifische Variation. Bezüglich der Hyperdialektalismen nach Typ II bestätigt das Modell einen signifikanten Kontrast (t = 2,20; p < 0,03MCMC) für die Variable Alter. Wie auch der in Abbildung 6 dargestellte Mittelwertvergleich mit Standardfehlern verdeutlicht, weisen jüngere Sprecher signifikant höhere Hyperdialektalitätsgrade auf als ältere Sprecher. Hypothese 2 kann ebenfalls bestätigt werden: die statistische Auswertung zeigt keinen signifikanten Unterschied, der auf die Art der beruflichen Ausübung zurückzuführen ist.

Dialektalität, Dialektwissen und Hyperdialektalität

67

Abbildung 6: Durchschnittlicher Hyperdialektalitätsgrad in Abhängigkeit von der Altergruppe (Effekt Alter t = 2,20; p < 0,03MCMC )

Bezüglich der Variable Geschlecht kann ebenfalls kein signifikanter Kontrast festgestellt werden: Weder bei den älteren, noch bei den jüngeren Sprechern hat das Geschlecht einen Einfluss auf den Hyperdialektalitätsgrad. Es sind also nicht, wie erwartet, jüngere Frauen mit durchschnittlich niedrigeren Dialektalitätsgraden in der Dialektabfrage, die aufgrund ihres geringeren Dialektwissens mehr Hyperdialektalismen als jüngere Männer verwenden. Hypothese 3a kann somit nicht bestätigt werden: Jüngere Männer, die einen signifikant höheren Dialektalitätswert und somit eine höhere Dialektkenntnis aufweisen, verwenden genauso viele hyperdialektale Formen, wie Frauen aus der jüngeren Altersgruppe. In der älteren Sprechergruppe ist – wie in Hypothese 3b angenommen wird – kein Effekt des Geschlechts auf den Hyperdialektalitätsgrad festzustellen.

68

5.

Sandra Hansen

Dialektwissen und Hyperdialektalismen: Diskussion

Die Ergebnisse in Abschnitt 4 zeigen, dass es keinen eindeutigen soziolinguistisch motivierten Zusammenhang zwischen Dialektalitätsgrad und Hyperdialektalitätsgrad in der Dialektabfrage gibt. Hyperdialektalismen nach Typ I weisen keine sprechergruppenspezifischen Effekte auf. Da diese Formen zuvor noch nie belegt wurden und sie in unserem Datenmaterial mit 12 Belegen selten vorkommen, ist anzunehmen, dass solche Formen durch die Interviewsituation der expliziten Dialektabfrage zu erklären sind. Hyperdialektale Formen nach Typ II werden einerseits zwar signifikant häufiger von jüngeren Sprechern, die im Vergleich zu älteren Sprechern eine geringere Dialektkenntnis aufweisen, verwendet, andererseits gibt es keinen geschlechtsspezifischen Unterschied in der Verwendung bei den jüngeren Sprechern, obwohl jüngere Frauen weniger Dialektwissen aufweisen als jüngere Männer. Zur Übersicht sind in Tabelle 1 die durchschnittlichen Dialektalitäts- und Hyperdialektalitätsgrade pro Sprechergruppe aufgeführt (Dia.Grad = durchschnittlicher Dialektalitätsgrad; Hyper.Grad = durchschnittlicher Hyperdialektalitätsgrad): Tabelle 1: Durchschnittliche Dialektalitäts- und Hyperdialektalitätsgrade pro Sprechergruppe Sprechergruppe

Hyper.Grad Dia.Grad

jung-kommunikationsorientiert-weiblich

0,042

0,761

jung-kommunikationsorientiert-männlich

0,043

0,802

alt-kommunikationsorientiert-weiblich

0,035

0,877

alt-kommunikationsorientiert-männlich

0,031

0,861

alt-handwerklich/landwirtschaftlich orientiert-weiblich

0,029

0,879

alt-handwerklich/landwirtschaftlich orientiert-männlich

0,035

0,878

Die Schattierungen spiegeln die oben beschriebenen signifikanten Unterschiede wider (Dialektalitätsgrad: Haupteffekt in Bezug zur Altersvariable, Interaktion bezüglich der Faktoren Alter und Geschlecht; Hyperdialektalitätsgrad: Alterseffekt). Ältere Sprecher haben im Durchschnitt niedrigere Hyperdialektalitätsgrade und höhere Dialektalitätsgrade in der Dialektabfrage als jüngere Sprecher: Frauen aus der älteren Sprechergruppe mit einem handwerklichen- oder landwirtschaftlichen Beruf weisen beispielsweise mit einem durchschnittlichen Hyperdialektalitätsgrad von 0,029 den niedrigsten Wert auf. Gleichzeitig ist der berechnete durchschnittliche Dialektalitätsgrad für

Dialektalität, Dialektwissen und Hyperdialektalität

69

diese Sprechergruppe der höchste (0,879). In Bezug auf die Variable Alter kann somit ein Zusammenhang zwischen Dialektwissen, welches sich in einem niedrigeren Dialektalitätswert widerspiegelt, und der Häufigkeit von hyperdialektalen Formen angenommen werden. Die durch das Modell vorhergesagte Interaktion bei der Berechnung der Einflüsse auf die Dialektalitätsgrade lässt sich im Aufkommen von Hyperdialektalismen nicht wiederfinden: In der Abfrage des Dialektwissens verwenden die jüngeren Frauen aus einem kommunikationsorientierten Berufsfeld mit einem durchschnittlichen Grad von 0,761 signifikant weniger dialektale Formen als jüngere Männer aus dem gleichen Berufsfeld und das, obwohl sie – gemessen an dem Aufkommen von hyperdialektalen Formen – genauso dialektsicher sind wie ihre männliche Vergleichsgruppe. Anders ausgedrückt weisen jüngere Männer aus einem kommunikationsorientiertem Berufsfeld höhere Dialektalitätswerte in der Dialektabfrage auf, verwenden aber fast gleich viele hyperdialektale Formen wie die weiblichen Vertreter dieser Alters- und Berufsgruppe. Hyperdialektalismen nach Typ II werden von 93,14 % der Sprecher (95 von 102 Informanten) mindestens einmal genannt. Die meisten hyperdialektal artikulierten Laute kommen mehrfach vor und haben – wie in Abschnitt 4.1 erläutert – das Potential sich auszubreiten. Diese Art von Hyperdialektalismen werden – wie im letzten Abschnitt gezeigt wurde – vor allem von jüngeren Sprechern aus einem kommunikationsorientierten Berufsfeld genannt, die von allen interviewten Sprechergruppen die höchste berufliche und private Mobilität aufweisen. Die höhere Mobilität dieser Sprecher könnte dazu führen, dass diese mehr mit anderen, naheliegenden Dialektsprechern in Kontakt sind und die basisdialektalen Varianten anderer Orte in den eigenen Dialekt übernommen werden.

6.

Zusammenfassung

In diesem Beitrag konnte gezeigt werden, dass außersprachliche Variablen (erwartungsgemäß) einen Einfluss auf das Dialektwissen von Sprechern haben: jüngere Sprecher mittleren Alters weisen eine geringere dialektale Sprechlage in der Dialektabfrage auf als ältere Sprecher. Außerdem bestätigt das angewendete statistische Modell eine Interaktion der Variablen Alter und Geschlecht: jüngere Frauen orientieren sich in der Dialektabfrage mehr an der Standardsprache als jüngere Männer. Dies gilt jedoch nicht für Frauen der älteren Sprechergruppe. Bezüglich der Analyse der Hyperdialektalität können zwei Typen ermittelt werden: Formen die vorher im lokalen Dialekt nicht vorhanden waren und Realisierungsvarianten, die in anderen Ortschaften schon belegt sind und eine Tendenz aufweisen, sich auszubreiten. Die

70

Sandra Hansen

statistische Ermittlung von soziolinguistischen Einflussfaktoren auf den letztgenannten Typ zeigt, dass jüngere Sprecher mehr hyperdialektale Formen verwenden als Informanten der älteren Sprechergruppe. Die Vermutung, dass Sprecher mit niedrigen Dialektalitätsgraden in der Dialektabfrage und deshalb niedrigerem Dialektwissen mehr Hyperdialektalismen realisieren, kann nicht eindeutig belegt werden. Es sind zwar die Informanten aus der jüngeren Sprechergruppe, die niedrige Dialektalitätswerte in der Dialektabfrage und ein höheres Aufkommen an hyperdialektalen Formen aufweisen, allerdings bildet sich die Interaktion zwischen Alter und Geschlecht bei der Einflussberechnung außersprachlicher Variablen auf den Dialektalitätsgrad nicht in dem Aufkommen von Hyperdialektalismen ab: jüngere Frauen aus einem kommunikationsorientierten Berufsfeld mit den signifikant niedrigsten Dialektalitätswerten in der Dialektabfrage weisen keine signifikant höheren Hyperdialektalitäswerte auf als ihre männlichen Vertreter aus der gleichen Alters- und Berufsgruppe. Inwiefern dieses Ergebnis durch Dialekt- und/oder Ortsloyalität, Mobilität oder weiter soziale Parameter zu erklären ist, wird sich in nachfolgenden Studien zeigen.

7.

Anhang: Tabellarische Aufstellung der abgefragte Lexeme und Phänomene

Lexeme

Phänomene

Kopf

– /k/-Verschiebung; Vokalhebung

gesundes

– Flexion – Schwatilgung im Präfix

höher

– f[x] – Entrundung /ø/

Zeitung

– Diphthongierung mhd. /i/ – Suffixvariation [υŋ]f[ik]

bleiben

– /b/-Frikativierung – Diphthongierung mhd. /i/ – /n/- Tilgung

kannst

– – – –

trocknen

– /k/-Verschiebung – Hebung von /ɔ/ – /n/-Tilgung

/s/-Palatalisierung /n/-Tilgung /t/Tilgung /k/-Verschiebung

Dialektalität, Dialektwissen und Hyperdialektalität

71

Lexeme

Phänomene

Mäuse

– Diphthongierung mhd. /y/ – Schwatilgung – Entrundung /y/

Vöglein

– Entrundung /ø/ – Diminutivendung – Dreisilbigkeit

Wolke

– /k/-Verschiebung – Hebung von /ɔ/ – Schwatilgung

Schaufel

– – – –

gekommen

– /k/-Verschiebung – Tilgung Präfix/Kurzform

sieben

– /b/-Frikativierung – Reduktion /i/ – /n/- Tilgung

Tage

– Schwatilgung – Verdumpfung oder Umlautung /a/

Woche

– Hebung von /ɔ/ – Schwatilgung

Feste

– /s/-Palatalisierung – Schwaöffnung oder -tilgung – Vokalöffnung /ε/

gelb

– Dehnung /ε/ – Vokalöffnung oder -schließung /ε/ – /b/-Tilgung

Latten

– Vokaldehnung /a/ – /t/-Lenisierung – /n/- Tilgung

auch

– /x/- Tilgung – Hebung oder Verdumpfung von /aυ/

sehen (1. Person Plural und 3. Person Plural)

– Vokalöffnung od -schließung /e/ – Suffixvariation -/əd/

sagen (1. Person Plural)

– – – –

Diphthongierung mhd. /u/ Dreisilbigkeit Endung -/lə/ Palatalisierung /u/ f /y/

Umlautung oder Verdumpfung /a/ Reduktion /a/ /n/- Tilgung Suffixvariation -/əd/

72

Sandra Hansen

Lexeme

Phänomene

hast (2. Person Singular)

– /s/-Palatalisierung – Umlautung /a/

haben (1. Person Plural und 3. Person Plural)

– Umlautung oder Hebung von /a/ – Kurzform [han] oder [hεn]

gehabt (Partizip von haben)

– – – –

getrunken

– Tilgung Präfix – /k/-Verschiebung oder –Tilgung – /n/- Tilgung

geschrieben

– – – –

Füße

– Schwatilgung – Entrundung /y/ – Monophthongierung mhd. /yə/

geschwollene

– Schwatilgung im Präfix – Hebung von /ɔ/

(hat nähen) müssen

– Entrundung /y/ – /n/- Tilgung – Monophthongierung mhd. /yə/

Leber

– /b/-Frikativierung – Dreisilbigkeit – Öffnung von /e/

Schwatilgung im Präfix Umlautung /a/ /k/-Verschiebung /t/-Tilgung

Schwatilgung im Präfix /b/-Frikativierung Reduktion /i/ /n/- Tilgung

Literatur Auer, Peter 2005 Europe’s Sociolinguistic Unity, or: A Typology of European Dialect / Standard Constellations. In: Nicole Delbecque, Johan van der Auwera und Dirk Geeraerts (Hrsg.), Perspectives on Variatio, 7–42. (Trends in Linguistics; 163.) Berlin, New York: Mouton de Gruyter. Bates, Douglas und Martin Maechler 2010 lme4: Linear mixed-effects models using S4 classes. R package version 0.999375–33. URL http://CRAN.R-project.org/package=lme4 Besch, Werner 2000 Variantentyp Hyperkorrektion. In: Annelies Häcki Buhofer (Hrsg.), Vom Umgang mit sprachlicher Variation. Soziolinguistik, Dialektologie, Methoden und Wissenschaftsgeschichte. Festschrift für Heinrich Löffler zum 60. Geburtstag, 187–193. (Basler Studien zur deutschen Sprache und Literatur. 80.) Tübingen [u. a.]: Franke

Dialektalität, Dialektwissen und Hyperdialektalität

73

Herrgen, Joachim, Alfred Lameli, Stefan Rabanus und Jürgen Erich Schmidt 2001 Dialektalität als phonetische Distanz. Ein Verfahren zur Messung standarddivergenter Sprechformen. Marburg: Dokumentenserver der Philipps-Universität Marburg [http:// archiv.ub.uni-marburg.de/es/2008/0007/pdf/dialektalitaetsmessung.pdf (eingesehen am: 21. 08. 2010)]. International Phonetic Association 2001 Handbook of the International Phonetic Association. A guide to the use of the international phonetic alphabet. Cambridge u. a.: Cambridge Univ. Press König, Werner und Renate Schrambke 1999 Die Sprachatlanten des schwäbisch-alemannischen Raumes – Baden-Württemberg, Bayerisch-Schwaben, Elsaß, Liechtenstein, Schweiz, Vorarlberg. (Themen der Landeskunde, Bd. 8.) Bühl/Baden: Konkordia. Lenz, Alexandra N. 2003 Struktur und Dynamik des Substandards. Eine Studie zum Westmitteldeutschen (Wittlich/Eifel). (Zeitschrift für Dialektologie und Linguistik. Beihefte 125.) Stuttgart: Steiner. Lenz, Alexandra N. 2005 Hyperdialektalismen und Hyperkorrektionen. Indizien für Varietätengrenzen. In: Alexandra Lenz und Klaus, J. Mattheier (Hrsg.), Varietäten – Theorie und Empirie, 76–95. (VarioLingua 23.) Frankfurt/Main [u. a.]: Peter Lang. Löffler, Heinrich 2001 Dialektologie. Eine Einführung. Tübingen: Narr. Mattheier, Klaus J. 1980 Pragmatik und Soziologie der Dialekte. Einführung in die kommunikative Dialektologie des Deutschen. Heidelberg: Quelle und Meyer. Mattheier, Klaus J. 1994 Varietätenzensus. Über die Möglichkeit, die Verbreitung und Verwendung von Sprachvarietäten in Deutschland festzustellen. In: Klaus J. Mattheier und Peter Wiesinger (Hrsg.), Dialektologie des Deutschen. Forschungsstand und Entwicklungstendenzen, 413–442. (Reihe Germanistische Linguistik 147.) Tübingen: Niemeyer. Niebaum Hermann und Jürgen Macha 2006 Einführung in die Dialektologie des Deutschen. 2. Auflage. (Germanistische Arbeitshefte 37.) Tübingen: Niemeyer. Pinheiro, José C. und Douglas M. Bates 2000 Mixed-Effects Models in S and S-PLUS. (Statistics and Computing Series.). New York: Springer. Schrambke, Renate 2009 Sprache und Region: Der Südwestdeutsche Sprachatlas als Forschungsinstrument und Ausgangspunkt für vergleichende Teilstudien. In: Lioba Keller-Drescher und Bernhard Tschofen (Hrsg.), Dialekt und regionale Kulturforschung. Traditionen und Perspektiven einer Alltagssprachforschung in Südwestdeutschland, 67–98. (Studien & Materialien des Ludwig-Uhland-Instituts der Universität Tübingen, Bd. 35.) Tübingen: Tübinger Vereinigung für Volkskunde. Seidelmann, Erich 2006 Mhd. o, ö und das leere Fach. Zur sogenannten „Vokalspaltung“ im Alemannischen. In: Hubert Klausmann (Hrsg.), Raumstrukturen im Alemannischen. Beiträge der 15. Arbeitstagung zur alemannischen Dialektologie, 53–59. Schloss Hofen, Lochau (Voralberg): Neugebauer, W. Sieburg, Heinz 1991 Geschlechtstypischer Dialektgebrauch. Anmerkungen zu einer empirischen Untersuchung von Geschwistern in der rheinischen Ortschaft Fritzdorf. Rheinische Vierteljahrsblätter 55: 294–314. Steger, Hugo und Volker Schupp (Hrsg.) 1993 Einleitung zum Südwestdeutschen Sprachatlas I. Marburg: Elwert. Steger, Hugo und Volker Schupp (Hrsg.) 1998 Einleitung zum Südwestdeutschen Sprachatlas II. Marburg: Elwert.

74

Sandra Hansen

Wiesinger, Peter 1983a Diphthongierung und Monophthongierung in den deutschen Dialekten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke und Herbert Ernst Wiegand (Hrsg.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung. Band 2, 1076–1083. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 1,2.) Berlin, New York: Walter de Gruyter. Wiesinger, Peter 1983b Rundung und Entrundung, Palatalisierung und Entpalatalisierung, Velarisierung und Entvelarisierung in den deutschen Dialekten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke und Herbert Ernst Wiegand (Hrsg.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung. Band 2, 1101–1105. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 1,2.) Berlin, New York: Walter de Gruyter. Wiesinger, Peter 1983c Hebung und Senkung in den deutschen Dialekten. In: Werner Besch, Ulrich Knoop, Wolfgang Putschke und Herbert Ernst Wiegand (Hrsg.), Dialektologie. Ein Handbuch zur deutschen und allgemeinen Dialektforschung. Band 2, 1106–1110. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 1,2.) Berlin, New York: Walter de Gruyter. Wurzel, Wolfgang Ullrich 1976 Adaptionsregeln und heterogene Sprachsysteme. In: U. Wolfgang Dressler und Oskar E. Pfeiffer (Hrsg.), Phonologica. Akten der dritten Internationalen Phonologie-Tagung. Wien, 1.–4. 9. 1976, 175–182. (Innsbrucker Beiträge zur Sprachwissenschaft 19.) Wien: Becvar.

A perceptual study of ethnicity and geographical location

75

Eivind Torgersen (Sør-Trøndelag University College)

A perceptual study of ethnicity and geographical location in London and Birmingham

1.

Introduction

Findings from London show emerging multi-cultural varieties of English (Cheshire et al. 2008b; Kerswill, Torgersen and Fox 2008). Speakers having different ethnic backgrounds, but whose main functional language is English, appear to converge on a number of phonological variables. In particular, the diphthongs face, price and goat have monophthongal or near-monophthongal realisations in the region of [e], [a ] and [o], respectively. A study in Birmingham has found similar vowel variants for these diphthongs used among non-Anglo speakers of different backgrounds (Khan 2006). These vowel qualities are almost identical to those found in London. This begs the question of whether the multi-cultural varieties in the UK are non-regional and whether listeners are unable to hear a difference between multicultural varieties of English spoken in the two cities. Hence, there would be a difference between multicultural varieties and more traditional urban varieties of English in terms of the degree of geographical variation, which is a major marker of different varieties. While traditional dialectology has largely been concerned with geographical differences (Foulkes and Docherty 1999: 5; Trudgill 1990), urban dialectology examines additional variables such as social class and gender. Multicultural varieties may be at an extreme end where geographical space (Britain 2010) becomes blurred. Traditionally, London and Birmingham have very different local accents, although they share some vowel features. London, situated in the south, has [ɑ] in bath and a split in the foot and strut set. Birmingham has [a] in bath and may have the same vowel in foot and strut. Both accents have traditionally shifted vowels in relation to the RP quality for some diphthongs: [ɑi] or [ɔi] in price, [i] in face and [υ] in goat (Wells 1982). However, recent developments in south-east England have seen young speakers preferring un-shifted vowel qualities. Dialect levelling has been observed in the London periphery, where Reading and Ashford now have very similar short vowel systems (Torgersen and Kerswill 2004). Un-shifted price and goat with qualities in the region of [a ] and [əυ] have been found in Read-

76

Eivind Torgersen

ing and Milton Keynes (Kerswill and Williams 2005). Levelling across Britain could result in similarities between London and Birmingham accents where previously diphthong shifted vowels are becoming un-shifted. Since dialect levelling centres around large cities, for example in Tyneside (Watt 2002), both large cities’ accents should still be distinct, but dialect contact between speakers of different ethnic backgrounds within (and possibly between) the cities may complicate the picture.

2.

Findings in London and Birmingham

Contrary to expectations from the literature, the dialect levelling observed in south-east England seems not to originate in London. A large-scale sociolinguistic study of two London boroughs, Hackney in inner London and Havering in outer London, showed only limited diffusion of features from centre to periphery. Instead, there was dramatic, mostly ethnic-minority and male-led innovations, particularly on the phonetic level, but also for some grammatical features (Cheshire et al. 2008b). The face, mouth, price and goat vowels had monophthongal qualities which were limited to nonAnglo speakers and Anglo speakers in high-density multi-cultural friendship groups (Kerswill et al. 2008). goose and strut had peripheral qualities of which again the non-Anglo speakers were the most frequent users (Cheshire et al. 2008a). In addition a number of consonantal and grammatical features were not found outside inner London (Cheshire and Fox 2007; Cheshire et al. 2008a; Cheshire and Fox 2009; Gabrielatos et al. 2010). Findings in Birmingham show differences between Anglo and nonAnglo speakers. In Khan’s (2006) study of price and goat, the older informants used traditional diphthong-shifted variants for price and goat, whereas young speakers frequently use un-shifted variants. Differences were observed between speakers of white British, Pakistani and Caribbean background. Whereas the white British young speakers favour the non-localised and non-shifted variants for goat, which was interpreted as resulting from dialect levelling, Pakistani and Caribbean speakers used backed and monophthongal variants similar to the ones observed in London. White British teenagers used a traditional diphthong-shifted variant in price, but less than the elderly speakers, and they also use the non-shifted variants. The Pakistani and Caribbean teenagers used the shifted variants to lesser extent than the white British teenagers, preferring the un-shifted variants. Khan argues that the monophthongal [o] in goat is an innovative variant with roots in the Pakistani and Caribbean communities. The changes in Birmingham are seen by Khan as generated within the city, the use of these

A perceptual study of ethnicity and geographical location

77

variants being a marker of ethnic orientation, which is a sensitive variable for the Pakistani and Caribbean speakers but not for the White British speakers. Such large differences were not found between different ethnicities in inner London. There were still differences between Anglo and non-Anglo speakers, but the largest difference was mainly between Anglo speakers not belonging to multi-cultural friendship networks and non-Anglo speakers. Birmingham had a difference between Anglo and non-Anglo speakers who were mainly of Afro-Caribbean and Pakistani origins. It may be that speakers just adopt Asian English vowels: monophthongal variants are also found for diphthongs in Indian English (Gargesh 2008). Pakistani English may have a mixture of diphthongal and monophthongal realisations for these vowels, but level of education has a strong effect on the realisation (Mahboob and Ahmar 2008). A differentiation between speakers of Asian origin and white speakers has also been found in Blackburn. Blakeley (2008) found monophthongal variants of this type for speakers of Asian background, but variants that largely resembled local features for white speakers. Monophthongal variants for canonical diphthong are also found in Caribbean English, especially Jamaican English (Wells 1982; Devonish and Harry 2008). Having monophthongal variants for diphthongs is shared with outer-circle varieties of English more widely (Kachru 1986). The reason why no great differences have been found between ethnicities in inner London may be due to the very large number of ethnicities and language backgrounds represented in the people residing there. In many boroughs, this, we presume, leads to relatively high inter-ethnic contact. This means that some of the dialect levelling changes happening in the south-east, such as goat-fronting, foot-fronting, the use of an RP-type [aυ] in the mouth lexical set and a number of consonant changes, cannot be traced back to London but are better understood within the context of post-World War II socio-historical changes in London and its hinterland, in particular out-migration and mobility, which findings in Milton Keynes and Reading demonstrate (Williams and Kerswill 1999; Kerswill and Williams 2005). Some of these features have been identified as youth norms (Williams and Kerswill 1999; Foulkes and Docherty 2000). These are consonant features generally considered to be diffusing across the country from the south: T-glottaling and TH-fronting (Stuart-Smith 1999; Williams and Kerswill 1999; Stuart-Smith and Timmins 2006) and the use of labiodental /r/ (Foulkes and Docherty 2000). Another feature is L-vocalisation, which is the use of a variant in the region of [o ] for dark /l/ ([]). L-vocalisation is tradi-

78

Eivind Torgersen

tionally a London feature. Hudson and Holloway (1977) found working class boys to have the highest degree of L-vocalisation with 71 % [o ], followed by working class girls with 51 %. Middle class speakers had less than 20 %. L-vocalisation is also common in other parts of England, and is spreading (Britain and Johnson 2007). Access to a common youth culture and possibly influence from television has arguably increased the amount of L-vocalisation in Glasgow. Originally the Glasgow dialect has L-vocalisation after low back vowels, but now there is also L-vocalisation after high back vowels (StuartSmith, Timmins and Tweedie 2006). We may say these consonant features are all representative of white informal speech styles in the south-east. Research shows that speakers of other ethnicities may not take part in these changes, or they may use other features. In Birmingham, Khan (2006) found that while Afro-Caribbean and White British speakers had overwhelmingly [f] for /θ/, there was persistence of [θ] for /θ/ among young Pakistani speakers. Differences between non-Anglo and Anglo speakers were also found for the choice of vowel qualities (see above). Ethnic heritage was found to be the most important predictor overall for choice of variant. For the Pakistani and Afro-Caribbean speakers, ethnic identity, as measured by self-ascription, was stronger than either social network or attitude towards Birmingham. For Anglos, the strongest predictor of use of ethnic variants was social network. This shows that ethnic background and social network influence youth language, and we may get a series of interdependent youth varieties. We find similar results in London for vowels. While the Anglo speakers without multicultural friendship networks largely had more traditional or levelled vowel qualities, the non-Anglo speakers used different innovative qualities. The innovations include the already mentioned monophthongal qualities for diphthongs (Kerswill et al. 2008), extreme fronting of goose and backing of strut (Torgersen, Kerswill and Fox 2006; Cheshire et al. 2008a). There were also effects of social network: Anglo speakers with largely multiethnic friendship networks may have vowel qualities that resemble those of non-Anglo speakers. Consequently, not all non-Anglo speakers participate in the young white speaker-led levelling in southern England. On the other hand, non-Anglo speakers may be in the lead in some, but not all changes.

3.

Multicultural London English

Cheshire et al. (2008b) argue that the set of innovative phonological and grammatical variants in London, which are shared by speakers of different backgrounds, constitute a variety or varieties which they label Multicultural

A perceptual study of ethnicity and geographical location

79

London English (MLE). Membership of a dense multi-cultural friendship group increases the number of MLE features. The ongoing research of which a part is reported here focuses on the characterisation of this ‘multiracial vernacular’ of London and examines if it is ethnically neutral and if there are any differences within MLE between areas in London. Given the hypothesis that Multicultural London English is ethnically neutral, it is of interest to identify individuals of differing ethnic backgrounds whose accents do not reveal ethnic background in perception and how local traditional accents potentially interact with ethnicity in perception. To answer these questions we carried out perceptual testing in London and Birmingham. While there is little previous research on perception of regional accent and ethnicity in the United Kingdom, the area has seen some studies in the United States. We turn to these now.

4.

Perception studies of ethnicity and regional background in the USA

Thomas (2002) concludes that although the ethnicity of a speaker can be identified in perception tests, it is not clear what features are used by the listeners in its identification. Listeners most likely use multiple cues to identify speakers’ ethnicity as tests focusing on specific cues, such as prosody, vowel quality and F0, have not produced clear results. Studies of the interaction between ethnicity and geographical region pose the question of whether there is a lack of regional variation in ‘ethnic’ speech. African American Vernacular English (AAVE) phonology is largely non-regional and it is largely unaffected by ongoing sound changes in the USA (Labov 1994; Edwards 2008). It shares some features with (white) dialects in the southern USA, but listeners have problems correctly identifying the regional background of AAVE speakers, which may be due to regional levelling of AAVE (Wolfram and Thomas 2002). If black speakers come from an area where black speakers share features with white speakers, they are often identified as white (Wolfram 2000). Research also shows that middle class African American speakers align their vowel systems to correspond with largely local (white) vowel systems (Jones and Preston forthcoming). Several factors have been found to affect dialect labelling. Clopper found effects of gender, linguistic experience and mobility on accuracy of identifications. Listeners were able to assign similarity ratings to speakers from the same dialect areas. Same-sex pairs of speakers from the same regions were rated more accurately than different-sex pairs of speakers (Clopper, Levi and Pisoni 2006). Mobile listeners were more accurate in identifying accent re-

80

Eivind Torgersen

gion than non-mobile listeners (Clopper and Pisoni 2004). More experienced listeners, i.e. those having lived in different dialect regions, were more accurate than listeners who had not (Clopper and Pisoni 2006).

5.

Perception studies in the UK

Studies in the UK have largely focused on accuracy of regional classification. Williams, Garrett and Coupland (1999) found that listeners were more accurate if the speakers were from the same region as the listeners. There was also an effect of experience: school teachers were more accurate than students, which was attributed to increased linguistic and travel experience. Such familiarity is found to be of great importance. Evans and Iverson (2007) show that to achieve familiarity with another accent one needs long-term accommodation with speakers of that accent. Adank, Evans, Stuart-Smith and Scott (2009) found differences in perceptual accuracy between familiar and unfamiliar accents in noisy listening conditions. There were higher processing costs, signalling that increased cognitive effort is required, for unfamiliar accents than for familiar accents in noisy conditions. No such effects were observed in quiet listening conditions. Heselwood and McChrystal (2000) found that male speakers of Panjabi background were perceived as more Panjabi-sounding than females by phonetically trained listeners. Production analysis confirmed differences between the male and female speakers in the realisation of stop consonants and vowels, but there was no difference in the level of Panjabi influence on the actual realisation of the stop consonants between males and females. It was probably a holistic judgement by the listeners that deemed the males to be more Panjabi-sounding. Kerswill and Williams (2002), using a holistic accent identification task, argue that when voices are perceived as more distinct they are also identified more accurately, when for example particular features were not found in the other speech samples. It is probably the distinctiveness of particular features that is indicative of the distinctiveness of the accent as a whole and because listeners from towns with little geographical mobility, belonging to focused speech communities, are more accurate at identifying voices from that particular town because of slow rate of language change. A local network and family ties may therefore be factors in the level of accuracy. This means that if the speech sample includes familiar local features, it will be identified more accurately by a listener who is local to the area. A speech sample with familiar accent features from another area may also be easier to identify. We can deduce from this that an introduction of ‘new’, unfamiliar features to the accent makes the identification task

A perceptual study of ethnicity and geographical location

81

harder, regardless of whether the speech sample and listener come from the same area.

6.

Methodology

We refined the holistic accent identification task used by Kerswill and Williams (2002). We hypothesise that inner London (Hackney) and Birmingham speakers will be easier to identify correctly than speakers from outer London (Havering). We also hypothesise that it will be more difficult to identify correctly Anglo speakers in multi-ethnic networks than Anglo speakers without such networks because of the presence of fewer traditional accent features. Non-Anglo speakers from London and Birmingham should be more difficult to identify correctly than Anglo speakers since the non-Anglo speakers may share a number of linguistic features such as the monophthongal qualities for face, price and goat. Of particular interest is the possible effect of friendship network on the classification of the speaker’s ethnicity and geographical location. The speakers had previously been assigned a network score based on information they provided in the interviews (Cheshire et al. 2008a): 1 2 3 4 5

all friends of same ethnicity as self up to 20 per cent of a different ethnicity than self 21–40 per cent of a different ethnicity than self 41–60 per cent of a different ethnicity than self 61–80 per cent of a different ethnicity than self

Speakers with high network scores may be harder to identify correctly because of fewer ‘ethnically-marked’ linguistic features. Twelve speakers were selected, representing four Anglo speakers from Hackney (low and high friendship network scores), two Anglo speakers from Havering (the two speakers with the highest friendship network score) and six non-Anglo speakers from Hackney representing a variety of ethnic backgrounds (Table 1). Also included were four speakers from Birmingham, two Anglos and two non-Anglos (with Afro-Caribbean backgrounds). Short passages were extracted from interviews, each around 10 seconds long. The utterances did not contain any ‘ethnicity-specific’ information or references and were not controlled for specific linguistic variables such as the vocalic or consonantal variants described in Section 2. This was because the extracts were taken from sociolinguistic interviews and the most important issue was to have a fairly fluent stretch of speech without interruptions

82

Eivind Torgersen

Table 1: Speakers included in the London listening test Megan Andrew Laura Ryan Sulema Kirsty Grace Dom Amjad Chris Kelly Dale Male Female Male Female

Hackney Hackney Hackney Hackney Hackney Hackney Hackney Hackney Hackney Hackney Havering Havering Birmingham Birmingham Birmingham Birmingham

Anglo Anglo Anglo Anglo Non-Anglo (‘other’) Non-Anglo (‘Asian’) Non-Anglo (‘black’) Non-Anglo (‘other’) Non-Anglo (‘Asian’) Non-Anglo (‘black’) Anglo Anglo Anglo Anglo Non-Anglo Non-Anglo

3 (mainly Anglo network) 3 (mainly Anglo network) 5 (multiethnic network) 5 (multiethnic network) 5 (multiethnic network) 5 (multiethnic network) 5 (multiethnic network) 4 (multiethnic network) 5 (multiethnic network) 5 (multiethnic network) 2 (mainly Anglo network) 2 (mainly Anglo network)

from other speakers. The alternative would have been to use prepared read speech which would have had a negative impact on the naturalness we tried to preserve. The number of trials had to be limited because of the young age of the listeners. This is a weakness of the study, but it was compensated for by having a large number of listeners. The soundfiles were played to individual listeners who were equipped and fitted with headphones who could play the file repeatedly, before selecting type of ethnicity and location (forcedchoice). Finally, they could move on to the next trial.

7.

London listeners

All listeners were from inner London and were aged 12 or 17 (N=68). They gave their overall impression of each utterance by classifying the speaker as ‘white’, ‘black’, ‘Asian’ or ‘other’. Previous research had found an effect of a multiethnic friendship network on production: Anglo speakers with a highly multiethnic network had vowel and consonant qualities intermediate between Anglo speakers with Anglo networks and non-Anglo speakers. The options for geographical location were ‘London’, ‘Essex’, ‘Birmingham’ and ‘Manchester’. The ‘London’ category should be assigned to the Hackney speakers and ‘Essex’ to the speakers from Havering. We chose to label the Havering speakers as ‘Essex’ because our fieldwork showed that Havering was an unfamiliar geographical concept. Havering is located next to Essex and was also formerly part of the county. Statistical analysis of correct identi-

A perceptual study of ethnicity and geographical location

83

Figure 1: Identification of speaker ethnicity, Havering voices (Anglo speakers)

fications of ethnicity and geographical location using a binary bivariate regression modelling showed no listener effects (age, sex and ethnicity of the listener) on making correct identifications, but significant effects of speaker on listeners’ classifications. Only the ‘London’ and ‘Essex’ responses were included in the statistical analysis. Figure 1 shows that Anglos from Havering with network score 2 are overwhelmingly identified as ‘white’, i.e. if the network score is low it is easier to identify ethnicity correctly, a difference that was significant (p