Dialectology meets Typology: Dialect Grammar from a Cross-Linguistic Perspective 9783110197327, 9783110179491

Citation preview

Dialectology meets Typology

≥

Trends in Linguistics Studies and Monographs 153

Editors

Walter Bisang Hans Henrich Hock Werner Winter

Mouton de Gruyter Berlin · New York

Dialectology meets Typology Dialect Grammar from a Cross-Linguistic Perspective

Edited by

Bernd Kortmann

Mouton de Gruyter Berlin · New York

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

앝 Printed on acid-free paper which falls within the guidelines 앪 of the ANSI to ensure permanence and durability.

ISBN 3-11-017949-0 Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the Internet at ⬍http://dnb.ddb.de⬎. ” Copyright 2004 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Cover design: Christopher Schneider, Berlin. Printed in Germany.

Contents

Introduction Bernd Kortmann

1

Dialectology and typology – An integrative perspective Walter Bisang

11

Local markedness as a heuristic tool in dialectology: The case of amn’t Lieselotte Anderwald

47

Non-standard evidence in syntactic typology – Methodological remarks on the use of dialect data vs spoken language data Peter Auer

69

The typology of motion and posture verbs: A variationist account Raphael Berthele

93

Dynamic typology and vernacular universals J.K. Chambers

127

Definite articles in Scandinavian: Competing grammaticalization processes in standard and non-standard varieties Östen Dahl

147

Person marking in Dutch dialects Gunther de Vogelaer

181

A typology of relative clauses in German dialects Jürg Fleischer

211

Do as a tense and aspect marker in varieties of English Bernd Kortmann

245

Typology, dialectology and the structure of complementation in Romani Yaron Matras

277

vi

Contents

Problems for typology: Perfects and resultatives in spoken and non-standard English and Russian Jim Miller

305

Comparing grammatical variation phenomena in non-standard English and Low German dialects from a typological perspective Günter Rohdenburg

335

On three types of dialect variation and their implications for linguistic theory. Evidence from verb clusters in Swiss German dialects Guido Seiler

367

Substrate, superstrate and universals: Perfect constructions in Irish English Peter Siemund

401

The impact of language contact and social structure on linguistic structure: Focus on the dialects of modern Greek Peter Trudgill

435

Jespersen’s cycle and the interaction of predicate and quantifier negation in Flemish Johan van der Auwera and Annemie Neuckermans

453

“Gendered” pronouns in English dialects – A typological perspective Susanne Wagner

479

Population linguistics on a micro-scale. Lessons to be learnt from Baltic and Slavic dialects in contact Björn Wiemer

497

Addresses of authors

527

Language and dialect index

529

Subject index

535

Introduction Bernd Kortmann

1.

Background

1.1. Aims of the volume In the first place, this volume is an invitation! It is an invitation issued to those interested in structural variation within languages, on the one hand, and to those interested in structural variation across languages, on the other hand. The purpose of this invitation is to bring together for the first time two research traditions in the study of language variation (and change) which so far have largely worked independently of each other, to make them enrich and provide new vistas for each other. It is a fascinating additional perspective for anyone interested in the study of dialect morphology and dialect syntax to judge the observable patterns of cross-dialectal variation for individual grammatical phenomena against generalizations, hierarchies and explanations which have grown out of the study of cross-linguistic variation. Dialect data can thus be looked at in a fresh light and new questions can be asked. At the same time, dialectology has very interesting things to offer to language typology, which here is to be understood as functional typology in the Greenbergian tradition. In many domains of grammar, regional and social non-standard varieties conform to cross-linguistic tendencies where the relevant standard varieties do not. Moreover, quite a number of grammatical features and patterns of variation found in non-standard varieties are not part of the relevant standard varieties. A sizable subset of these has not even been observed in the typological literature before, whether only for the relevant language family or linguistic area, or on a world scale. Dialect data, besides serving as a rich additional data source and making a significant contribution to areal typology, are bound to help establish more fine-grained typological parameters and formulate generalizations and hierarchies which are at the same time more finegrained and more robust. Another point of fundamental importance is the following: typologists need to be careful not to neglect the differences

2

Bernd Kortmann

between the grammars of written and spontaneous spoken varieties of languages. In fact, non-standard dialects (as varieties which are almost exclusively spoken) may well serve as a crucial corrective for typological research, which is typically, especially for languages with a literary tradition, concerned with the (fairly homogeneous) written standard varieties of languages. If this spoken-written difference is not given due consideration, typologists run the risk of comparing what is comparable only within limits, and certainly only with great caution, namely ‘exotic’ languages for which they only have spoken data as opposed to typically well-described (e.g. European) languages for which they make the written standard varieties the basis of their generalizations. In sum, this volume aims to show how much dialectologists and typologists can learn from each other and that, ultimately, both typology and dialectology are bound to profit from a (functional-) typological approach to the study of dialects. To a certain extent, this volume can thus be seen as a major step in the direction of a unified account of intra- and cross-linguistic variation from a functionalist angle. To these two dimensions of variation, we must indeed add the historical dimension. It is in the study of spoken varieties, in particular, that the strict division between synchrony and diachrony becomes increasingly blurred. It is here that conservative features characteristic of previous language periods exist alongside with ongoing language change indicating, for example, new grammaticalization processes or a higher degree of grammaticalization of individual forms and constructions than in the relevant standard varieties. Such a unified research paradigm has been advocated for more than a decade by Croft (1995, 2000, 20032) under the headings of integrative functionalism and the dynamic paradigm in the study of linguistic variation. In the present volume, this paradigm is most forcefully argued for in the papers by Chambers and, above all, Bisang. 1.2. Contributors and wider context of current dialectological research The research programme underlying this volume was first advertised at the World Congress of Linguists in Paris 1997 (cf. Kortmann 1999) and subsequently implemented in a research project at the University of Freiburg (Germany) on comparative English dialect grammar in the British Isles from a typological perspective (cf. Kortmann 2002, 2003; Kortmann et al., forthcoming). It had always been intended, however, to broaden the

Introduction

3

scope of the envisaged collaboration between dialectologists and typologists beyond the study of variation in a single language. I was thus very glad when asked by the organizers of the METHODS XI conference at Joensuu, Finland (5–9 August, 2002) to organize a workshop of my own choosing. It is out of this international workshop on “Dialectology and Typology” that the current volume has grown. Ten out of the 19 contributors were invited speakers at the workshop, the remaining nine were invited to contribute to this collection as like-minded distinguished dialectologists or typologists. All authors share a pronounced interest in looking across the fence which still separates dialectology and typology and which for a long time, it seems, has been deemed too high or simply uninteresting to explore the other side of. With regard to the contributors to this volume, something else is worth noting. This volume is published at a time when we are witnessing on a broad scale an increasing interest in dialect grammar leading, above all, to the systematic collection of new material stored in large databases and computerized corpora, and to the inclusion of language-internal grammatical variation into modern syntactic theories, be they formalist (microparametric syntax) or, as in this volume, functionalist. Another fact that bears witness to the burgeoning field of dialect grammar is that, independently of each other, five major research projects on the study of dialect grammar sprang up in six European countries, all roughly beginning in the year 2000 (cf. Barbiers, Cornips and van der Kleij 2002 for a selection of work done within four of these projects). Besides the Freiburg project on English dialects mentioned above, these are projects on the syntax of Dutch and Flemish dialects (the SAND project in Amsterdam, Antwerp and Gent), Swiss German dialects (Zurich), Italian dialects (Padua), and Romani dialects (Manchester). These five projects are currently joining forces, spearheading an initiative for an international research network on European dialect syntax. It is worth noting that nine contributors to this volume belong to, or are closely associated with, four of these projects. Besides Matras (Manchester project on Romani), these are members or associates of the three projects on West Germanic dialects: de Vogelaer, Neuckermans and van der Auwera (SAND project), Fleischer and Seiler (Zurich project), as well as Anderwald, Kortmann and Wagner (Freiburg project). It is not by accident that all of these share a strong functionalist orientation. Except for the Padua project on Italian dialect syntax and the majority of the Amsterdam team of the SAND project, the relevant dialect

4

Bernd Kortmann

syntax projects are firmly grounded in the functionalist paradigm. Formalist approaches to dialect syntax are therefore only marginally represented in this volume. Where formalist theory is drawn upon, it is almost exclusively Optimality Theory, including its latest offspring Stochastic Optimality Theory (Bresnan and Deo 2001). Optimality Theory makes its strongest appearance in the paper by Seiler, who proposes ways of implementing in an OT setting the requirements for a grammar to allow for variable outputs and preference directions. 2. Coverage of dialects and languages Given what was said above about the members and associates of current projects on dialect syntax represented in this volume, it will not come as a surprise that there is a clear geographical and genetic bias with regard to the dialects and languages discussed here. Almost exclusively, it is the linguistic landscape of Europe which is at issue. The only papers which, apart from discussing European varieties, reach out in some detail to other continents are the ones by Bisang (Africa, Australia) and Kortmann (varieties of English around the world). Within Europe, there is clearly a genetic, more exactly Germanic, bias. The bulk of the papers (13) deal with standard and, overwhelmingly, non-standard varieties of West Germanic languages: Dutch and Flemish dialects (de Vogelaer, van der Auwera and Neuckermans), varieties of English (Anderwald, Chambers, Kortmann, Miller, Rohdenburg, Siemund, Wagner) and German (Auer, Berthele, Fleischer, Rohdenburg, Seiler). North Germanic dialects are investigated by Dahl. The other language families represented in this collection are Baltic (Wiemer), Slavic (Miller, Wiemer), Romani (Matras), and Greek (Trudgill). The European bias simply reflects the fact that it is in European linguistics that dialectology has a long-standing tradition and, above all, that we are currently witnessing the rise of the study of dialect grammar. The Germanic, especially the West Germanic, bias reflects the fact that there is a wealth of data for the study of dialect grammar available, especially due to the creation of computerized corpora and databases based on new large-scale fieldwork as part of the major research projects mentioned in section 1.2. Both biases are an advantage rather than a disadvantage for the ultimate objective of this volume: at best what can be presented here is a first glance at what kinds of fascinating things we can

Introduction

5

learn from and about dialects, i.e. variation within a single language, from a typological perspective, and in what ways the typologist’s ‘toolkit’ can be improved by taking into consideration intralinguistic variation. The Germanic dialects are no more than showcase exemplars demonstrating the potential of a fruitful collaboration between dialectologists and typologists. For this purpose, focusing on that genetic group of dialects which is a prime candidate for the currently best-documented and best-investigated one in Europe, if not the world, has many things to speak for itself. Beyond what has been said so far, the focus on Europe points to at least three future avenues of comparative research on dialect grammar from a typological point of view, two of which have at least begun to be explored by papers in this volume. First of all, mapping the morphological and syntactic variation across the standard and non-standard varieties of a given language should not arbitrarily be restricted to only one geographical area where such varieties are spoken, as is for example the case for the Freiburg project on English, which is restricted to the British Isles. Thus in this volume Kortmann widens the scope of his study to all varieties of English spoken around the globe, including English-based pidgins and creoles. Secondly, comparative studies of dialect syntax within one language should by all means be extended to the dialects of genetically closely related languages. To any native speaker of one of the West Germanic languages, for example, the parallels between the morphosyntax of the regional dialects of Dutch, German and English in many domains are quite striking, as is illustrated for English and German dialects in Rohdenburg’s paper. A similar broadening of the scope of comparative dialect research is called for in other language families. From broadening research on comparative dialect syntax to entire language families, it is but a small and ultimately necessary step to include the facts from the grammars of dialects into areal typologies which explore the geographical distribution of individual morphosyntactic phenomena, and possibly identify linguistic convergence areas on a smaller scale. The paper by Wiemer on the Baltic-Slavic contact area is a nice illustration of this strand of research. Bisang discusses ways in which the concept of linguistic convergence area or Sprachbund can be sharpened by including what we know about dialect continua, on the one hand, and language contact from a sociolinguistic and cognitive point of view, on the other hand. On a larger scale, it is amazing (and deplored by Miller in this volume) that, for example, dialects played a rather marginal role in such a major international project as “Typology of European Languages”

6

Bernd Kortmann

(EUROTYP; 1990–1994). The ultimate question which the inclusion of dialects in European areal typology might answer is what the ‘real’, i.e. spoken non-standard, landscape of Europe looks like: What is the nature of ‘real’ Standard Average European, which was hypothesized by Benjamin Lee Whorf almost 60 years ago and found largely confirmed by the EUROTYP project (cf. Haspelmath 2001)? Answering this and a host of related challenging questions will require a large-scale international research programme in its own right. 3. Grammatical phenomena and major issues discussed In what follows, a brief survey will be given on what the reader can expect from the papers in this volume with regard to grammatical phenomena and processes investigated, and some major issues that are repeatedly addressed. This will include a look at the different ways in which the dialectologists and typologists in this volume were inspired by work in the other field, and at observations and suggestions the authors have made with regard to what dialectologists and typologists can learn from each other. The major phenomena of dialect grammar discussed in this volume are the following: –

–

– – –

NP: nominal inflection (Wiemer); definite and indefinite articles (Auer, Dahl), gender systems (Rohdenburg, Wagner, Wiemer), clitic pronouns (Trudgill), resumptive pronouns (Rohdenburg), relative clauses and relativizers (Auer, Fleischer); VP: motion and posture verbs (Berthele), head-initial vs. head-final ordering within verb clusters (Seiler), the auxiliary-full verb distinction for the past tense forms of do/doon (Rohdenburg), number distinctions (Wiemer), reflexive markers (Wiemer), tense and aspect, especially the Perfect in English (Kortmann, Miller, Siemund) and the do/tunperiphrasis (Auer, Kortmann); agreement and lack of (especially subject-verb) agreement (Chambers, de Vogelaer); negation (Anderwald, van der Auwera and Neuckermans); complementation (Matras).

Major issues that are repeatedly addressed in this volume include the following:

Introduction – –

–

–

7

grammaticalization processes (Dahl, Kortmann, Miller, Trudgill, van der Auwera and Neuckermans); naturalness, i.e. the question whether spoken varieties exhibit more ‘natural’ grammars with a higher degree of consistency (e.g. in terms of analyticity) than (written) standard varieties do (Auer, Chambers, Kortmann); the role of language contact in shaping linguistic types and linguistic convergence areas, influencing linguistic structures of vernaculars, for example in the direction of simplification or complexification (Bisang, Chambers, Kortmann, Matras, Siemund, Trudgill, Wiemer); interesting in this respect are the diverging views by Chambers and Wiemer on vernacular universals and the part that language contact and social attitudes play in their becoming manifest; instances of redundancy in dialect grammars: e.g. multiple negation, double determination in Swedish dialects (Dahl), and the apparent functionality of redundant-seeming agreement systems, as in person marking in Dutch dialects, which is even conquering new territory (de Vogelaer).

Given the overall aims of this collection of papers, the interesting question is of course from which angle the authors tackle these phenomena, processes and major issues, and what their conclusions are with regard to how dialectology and typology can complement each other in mutually enriching ways. Apart from the fact that relevant typological studies (including methodologies; cf. especially Matras) were taken into consideration, the following concepts from language typology served as sources of inspiration for the studies by the dialectologists and sociolinguists in this volume: –

–

implicational hierarchies: the NP Accessibility Hierarchy (Fleischer, Wagner), Animacy Hierarchy (also figuring as Individuation Hierarchy; Bisang, Rohdenburg); Matras summarises his major results with regard to complementation patterns in Romani dialects in the form of implicational hierarchies and conceptual spaces. The notion of universals figures most prominently in the papers by Chambers and Siemund. Siemund examines Irish English from a universalist perspective, including universals of language contact. Chambers elaborates on vernacular universals, i.e. a small set of phonological and grammatical processes which recur in all spoken

8

– – – –

Bernd Kortmann

vernaculars of English, but which, as belonging to the language faculty, he ultimately claims to have counterparts in child language, pidgins and spoken vernaculars in all languages of the world. markedness, including local markedness (Anderwald, Matras, Wiemer); functional explanations for the observable variation across dialect grammars, e.g. iconicity (Matras, Rohdenburg); Talmy’s and Slobin’s typology in terms of satellite- vs. verb-framing languages (Berthele); Nichols’ population linguistics (applied to a micro-area by Wiemer).

Vice versa, how can typological research be enriched by the dialect data and the insights from dialectology as well as sociolinguistics and language contact presented in this volume? To start with, dialects exhibit a number of grammatical phenomena which either have not at all or only very rarely been described in the typological literature or, at least, have not been described for the relevant language (i.e. standard variety), language family, or geographical area (e.g. Europe). Examples include triple negation in Flemish, gendered pronouns in dialects of Southwest England, and do/tunperiphrasis for tense and aspect marking in the non-standard varieties of West Germanic. Furthermore, several papers offer interesting observations on areal typology (especially those by Bisang and Wiemer), to which those papers may be added which document transition zones between dialects and dialect areas on syntactic grounds (Dahl, Seiler, van der Auwera and Neuckermans). Auer makes an important contribution to the methodology of future typological studies incorporating data from non-standard varieties (what Chambers would call variationist typology). He explores which kind of spoken non-standard data – geographically restricted or non-restricted data – are most useful as a basis for syntactic typology, arguing in favour of general features with a high areal reach. A related issue, language sampling in linguistic typology and how it would be affected by including nonstandard varieties, is discussed by Bisang. More fundamentally, the authors stress time and again how important it is to distinguish between spoken and written varieties as well as between non-standard dialects/vernaculars and the standard varieties, pointing out what typology is missing out on (and how it may eventually go wrong) when ignoring these two distinctions (cf. for example Berthele, Bisang, Fleischer, Kortmann, Miller, Matras, Seiler). From an integrative perspective, in particular, various authors stress the importance of incorporating in typological theorizing what we know about the impact of society

Introduction

9

structure, social networks, social attitudes, and language contact on linguistic structures and language change (cf. especially Bisang, Chambers, Trudgill, Wiemer). To some extent, one is tempted to say that the joint warning is issued that in language typology the assumption of the homogeneity of a language community should be acknowledged as a myth, convenient as such an idealization is. The risks arising from this myth should be recognized and more attention be paid to the question whose language is studied in typological research, and to the fact that functional explanations alone cannot account for the observable patterns of language structures and language change in the world’s languages. In conclusion, the collection of papers in this volume bears witness to the fact that the collaboration between dialectologists and typologists envisaged and advocated here is one that is extremely promising, and a collaboration between partners on equal terms, with none being the junior partner of the other. 4. Organization of the volume What emerges from the overview in the previous section is that there are several ordering principles according to which the papers in this volume could be arranged. This is why, in the end, I decided to dispense with a thematic grouping, and simply arranged the papers in the alphabetical order of the authors. The only exception is the paper by Bisang. Given its programmatic character and the issues of fundamental importance which Bisang discusses with regard to the integration of dialectology and typology, this paper is ideally suited for setting the scene. Acknowledgements I would like to express my deeply felt gratitude to a number of people without whom this volume would have taken much longer to edit and produce, or may not have come about at all: the organizers of the METHODS XI conference, Markku Filppula and Juhani Klemola; Lieselotte Anderwald for her wonderful support in submitting all papers to critical, stylistic, and formal scrutiny; Werner Winter, the series editor responsible for this volume, for his prompt and helpful feedback on the draft versions of the papers; and Melitta Cocan for her tireless

10

Bernd Kortmann

communication with the authors and publishers on all formatting details, and for having drawn up the manuscript as it appears in print. Thanks a lot to all of you! References Barbiers, Sjef, Leonie Cornips and Susanne van der Kleij (eds.) 2002 Syntactic Microvariation. Amsterdam: SAND. (http://www.meertens. knaw.nl/projecten/sand/synmic) Bresnan, Joan and Ashwini Deo 2001 Grammatical constraints on variation: ‘be’ in the Survey of English Dialects and (Stochastic) Optimality Theory. Ms. Stanford University. http://www-lfg.stanford.edu/bresnan/be-final.pdf Croft, William 1995 Autonomy and functionalist linguistics. Language 71: 490–532. 2000 Explaining Language Change. An Evolutionary Approach. Essex: Pearson Education. 2003² Typology and Universals. Cambridge: Cambridge University Press. Kortmann, Bernd 1999 Typology and dialectology. In: Bernard Caron (ed.), Proceedings of the 16th International Congress of Linguists, Paris 1997. CD-ROM. Amsterdam: Elsevier Science. 2002 New prospects for the study of dialect syntax: Impetus from syntactic theory and language typology. In: Sjef Barbiers, Leonie Cornips and Susanne van der Kleij (eds.), Syntactic Microvariation, 185–213. Amsterdam: SAND. 2003 Comparative English dialect grammar: A typological approach. In: Ignacio M. Palacios Martinez, María José López Couso, Patricia Fra and Elena Seoane (eds.), Fifty Years of English Studies in Spain (1952:2002). A Commemorative Volume. 65–83. Santiago de Compostela: University of Santiago. Kortmann, Bernd, Tanja Herrmann, Lukas Pietsch and Susanne Wagner forthc. A Comparative Grammar of British English Dialects: Agreement, Gender, Relative Clauses. Berlin/New York: Mouton de Gruyter.

Dialectology and typology – An integrative perspective Walter Bisang Abstract Dialectologists and typologists both deal with linguistic variation and have more to say to each other than has been realised in the past. If the definition of dialect vs. language is based on extra-linguistic social factors, nothing can be predicted concerning the degree of structural variation between dialects, and thus there is no principled reason why to exclude data from dialects into typology. Findings from dialectology are of particular interest for the approach of integrative functionalism (Croft 1995) which assumes that individual speakers select the variant to be used in a particular situation according to social criteria. If this approach is extended to the level of distributional patterns of linguistic types across languages, models such as social networks (Milroy and Milroy 1985), leaders of linguistic change (Labov 2001) or the invisible hand (Keller 1990) have a lot to contribute to our understanding of how these patterns came about. Models such as these are particularly important in a situation in which typologists cannot rely on the statistical neutrality of the distribution of types within the population of the world’s languages (Maslova 2000). Studies on linguistic variation are incomplete without the incorporation of language contact and its consequences. Language contact can promote the emergence of new types and it may turn out to be a challenge for Maslova’s (2000) approach to the problem of statistical neutrality. Multidimensional scaling as developed in dialectology may be a possible way out of this problem. The integration of typology, dialectology and contact linguistics may open new perspectives for research dealing with linguistic variation and what we can learn from variation about the human brain. 1. Introduction 2. Dialectology 3. Typology 4. Fields of integration 5. Beyond dialectology and typology: Language contact 6. Conclusion

12

Walter Bisang

1. Introduction Dialectologists and typologists have both worked for a long time on linguistic variation without paying too much attention to each other. This may be due to the fact that dialectologists concentrate on social and historical motivations of variation across dialects, whereas typologists are interested in universal patterns of variation across languages and their motivation by human cognition and discourse. From such a perspective, one may indeed conclude that there cannot be much of an overlap between these two disciplines of linguistics. But this is not the whole story. As I would like to show in this paper, there is quite some potential of interaction. In more recent times, this potential has even become considerably more significant if we look at integrative functionalism as discussed by Croft (1995) and at the problem of the validity of typological universals if we don’t know whether linguistic types are distributed in a statistically neutral way in the population of the world’s languages (Maslova 2000). A simple answer to the question of what dialectologists deal with is that they deal with dialects. From a somewhat more elaborate perspective, dialectologists are interested in linguistic variation on the level of dialects and its geographical and social diffusion, its historical, political and cultural motivation and its quantification as discussed in dialectometry and multivariant scaling. Most dialectological studies cover dialects of one and the same language. Some famous examples are Georg Wenker’s (1895) Sprachatlas des Deutschen Reiches, Jules Gilliéron’s (1902–1910) Atlas linguistique de la France, Karl Jaberg and Jakob Jud’s (1928–1940) Sprach- und Sachatlas Italiens und der Südschweiz and The Survey of English Dialects conceived by Eugen Dieth and Harald Orton (1962–1978). In addition, there are also studies dealing with the diffusion of structural properties across languages such as the use of uvular /r/ from Paris up to Copenhagen, Kristiansand and Bergen (Trudgill 1974). If dialectologists deal with linguistic variation from the perspective of dialects, typologists may be said to deal with variation across languages. Typologists try to find out to what extent languages show structural variation and where they follow universal patterns. For that purpose, they set up linguistic types on the basis of empirical cross-linguistic comparison, they try to see whether there are generalizations in the sense of universal patterns of covariation, and they try to explain or motivate these generalizations (Croft 2003: 1–2).

Dialectology and typology – An integrative perspective

13

Even if we disregard the fact that dialectological studies are not always limited to dialects of one language and that typological studies may include some data from dialects of individual languages, the main problem resulting from equating dialectology with variability among dialects and typology with variability among languages is the fuzzy borderline between what makes a dialect and what makes a language. As is well-known, criteria of mutual intelligibility or structural similarity do not work. As suggested by Chambers and Trudgill (1998: 9–12), the correlation between language and dialect is rather of social nature and should be defined within the opposition pair of autonomy vs. heteronomy. An autonomous variety is perceived by its speakers as a distinct language whereas a heteronomous variety is seen as being part of a larger autonomous variety irrespective of the degree of structural divergence among the heteronomous varieties subsumed under the autonomous variety. From this perspective, Hindi and Urdu are two different languages no matter how structurally close they are, whereas the varieties of Sinitic, which are often mutually unintelligible and show a considerable degree of structural variation (for some new insights, cf. Chappell 2001), are heteronomous and thus dialects of one single language, i.e. Chinese. Since the distinction between language and dialect is to a large extent grammar-external, typologists who aim at discovering the range of attested variation may miss some of it if they concentrate on languages without taking into account the potential of structural variation hidden behind the term of “language”. The following quotation from DuBois (1985) clearly describes how easily relevant pieces of linguistic variation even within dialects may be overlooked: Volumes of so-called functionalism are filled with ingenious appeals to perception, cognition or other system-external functional domains, which are used to ‘explain’ why the language in question simply has to have a grammatical particularity that it does – when a moment’s further reflection would show that another well-known language, or even just the next dialect down the road, has a grammatical structure diametrically opposed in the relevant parameter. (DuBois 1985: 353)

Dialectology is of particular relevance for functional approaches subsumed under the term of “integrative functionalism” by Croft (1995). Departing from the existence of language-internal variants, this approach assumes that individual speakers select the variant they are going to use in a given situation according to sociolinguistic criteria, i.e. according to grammarexternal criteria. If this is true at the level of individual speakers, it is also

14

Walter Bisang

true at the level of distributional patterns of linguistic types across languages because these patterns are ultimately the result of the successful diffusion of particular linguistic types and because successful diffusion depends on social criteria. From this perspective, typologists cannot but integrate findings on the diffusion of particular language structures from dialectological and sociolinguistic studies. Cognition is only indirectly accessible to the typologist via crosslinguistic variation in the world’s languages past and present. This variation depends on social, political and cultural factors which created it. As a consequence of these factors, certain linguistic structures may be statistically under/overrepresented. It is thus not possible to draw any typological conclusions from the simple statistical distribution of a certain linguistic type or a certain covariation pair within the present population of the world’s languages (Dryer 1989). Being aware of this problem, a number of typologists (cf. section 3.3.1) provided criteria such as genetic relatedness and geographical closeness for constructing statistically balanced language samples. In more recent times, Maslova (2000) pointed out that these methods are problematic because of the question of whether the actual distribution of linguistic types across the world’s languages is statistically neutral. If one cannot say to what extent the actual situation is biased by former distributions, it is not possible to assess what the lack of a particular linguistic type means. Thus, “if a logically possible type is absent from the current language population, this can also be an ‘accident’” (Maslova 2000: 308). After these general remarks of a typologist who tries to understand possible fields of convergence with dialectology, I would like to look at the two fields and their potential of integration more concretely and in somewhat more detail. I shall start with a presentation of more or less traditional views on dialectology and typology. In section 2, dialectology will be described from three perspectives which will be taken up again in section 3, where I shall discuss typology and a number of more concrete options for interaction with dialectology. The three perspectives to be discussed in both sections are the following: defining the domains of variability (2.1 and 3.1), sources of linguistic data (2.2 and 3.2), and methods of analysis and results (2.3 and 3.3). In section 3.3, I shall look at sampling methods in typology and the impact of the dialect vs. language problem (3.3.1), at some of the findings of typological research, i.e. at implicational universals and hierarchies (3.3.2), and at typology from a geographical perspective, i.e. at linguistic areas as zones of structural

Dialectology and typology – An integrative perspective

15

convergence (3.3.3). Section 4 takes up integrative functionalism and its interaction with dialectology (4.1), and the contribution of dialectology to the problem of statistical neutrality as discussed by Maslova (2000) (4.2). As I shall briefly argue at the end of section 4.2, multidimensional scaling as practised in dialectology (2.3) may be an alternative method of coping with the problem of statistical neutrality and maybe even of assessing it to some extent. If the integrative approach starts from individuals (Croft 1995) with their cognitive equipment and their sociolinguistic behaviour, a third factor to be integrated is language contact and the whole range of its structural consequences. Language contact in general and its potential of generating new types or new patterns of covariation will be discussed in section 5. Finally, a short conclusion will be presented in section 6. 2.

Dialectology

2.1. Defining the domains of variability Data are collected from informants on the basis of a questionnaire which determines the domains of variability and can either be direct or indirect. With the direct method, the informant is asked straightforward questions for which she has to produce an adequate solution in her dialect. The indirect method tries to elicit the linguistic solution required by somehow creating a situation in which the informant is supposed to produce it. Although indirect questioning has the disadvantage of being more timeconsuming it is preferred in most dialectological studies. In addition, questionnaires vary with regard to the amount of leeway they allow to the fieldworker collecting the data. In a formal questionnaire, even if it is an indirect one, the form of the question is fixed. In an informal questionnaire, the fieldworker can ask his question in whatever form he chooses as long as it yields the desired results (for a good summary, cf. Chambers and Trudgill 1998: 21–25). 2.2. Sources of the linguistic data The main sources from which data are collected are informants. The criteria for selecting them are subject to considerable variation, although the prototypical informant in the majority of studies on dialect geography tends to be a nonmobile, older and rural male (Chambers and Trudgill 1998: 29).

16

Walter Bisang

Of course, such a selection is directly motivated by the desire to get at the authentic local dialect variety. However, given the evidence of how strongly variation depends on social factors such as age, social class, education and gender, an adequate integration of these factors by interviewing more representative sets of a population appears to be more and more inevitable. Studies concentrating on the above prototype are not very likely to cover the whole range of linguistic variety within a geographical location unless its population is very stable and conservative. Phenomena such as migration and industrialization simply call for paying due attention to the above social factors. 2.3. Methods of analysis and results Dialectological data are in most cases published on maps. Depending on the purpose of the study, the material is either transferred directly onto the map or the material is interpreted before it is integrated into the map. In the former case, the map covers the spatial distribution of the full variation concerning a particular linguistic item as collected by the fieldworker(s); in the latter case only those variables show up in the map which are interpreted as predominant within a particular area. The former type of map is called a display map and is the preferred form of language atlases, the latter type is called an interpretive map and is predominantly found in secondary studies which are geared towards more general conclusions and often take their data from display maps as their primary sources (Chambers and Trudgill 1998: 25). In both types of maps, regions which differ with respect to a particular variable of interest can be separated by lines called isoglosses. Depending on political, historical or cultural factors, different patterns of isoglosses can emerge. The more isoglosses separate a given geographical area from another the more significant that dialect area will be. Thus, bundles of isoglosses are of particular interest in dialectology and – as will be shown below (cf. linguistic areas in section 3.3.3) – in typology as well. Since maps are static and two-dimensional, while dialect variation is dynamic and multidimensional (apart from the geographical dimension there are social dimensions such as age, social class, education and gender), special techniques need to be developed for taking into account more complex dialect situations. If, for instance, the geographical diffusion of a certain variable depends on the age of the informants, a dialectologist may

Dialectology and typology – An integrative perspective

17

decide to keep the geographical aspect stable and to show the diffusion of the variable along the dimension of age by creating separate maps for each age group relevant in that particular study. The reader will then have to superimpose these maps to see the dynamics of the diffusion of that variable. Given new technological options, this situation may change in the future, but this stage does not seem to have been reached yet: “Drawing inferences from dialect maps gets more complex as the dialect situation adds dimensions. Eventually, we will have the wherewithal to create multidimensional displays involving stereoscopy or holography, but for the time being the best we can do is superimpose visual devices on flat planes.” (Chambers and Trudgill 1998: 119) Discovering the geographical distribution patterns of variables crucially depends on quantification, i.e. on suitable statistical and mathematical methods. Two such methods will briefly be sketched in this section. One method is known under the term of dialectometry as introduced by Séguy (1973) in his sixth volume of the Atlas linguistique de la Gascogne (for more details, cf. Goebl 1982, 1993). Séguy’s method ultimately consisted in establishing a dissimilarity matrix. Together with his team, Séguy counted the number of differing variables for each of the contiguous sites in his atlas. For that purpose, he divided the items of his Gascony data (more than 400 items) into different types of variables (lexical, pronunciation, phonetic/phonological, morphological, syntactic). For each of these types he first calculated a separate percentage which he then integrated into an overall percentage of linguistic distance between each of the adjacent sites of his atlas. Finally, he plotted the results of his calculations of linguistic distance on his base map. In the dialectometric method developed by Séguy, the geographical dimension is always the same irrespective of the degree of dissimilarity between the sites, i.e. each site is placed in the geographical space of the map of Gascony. The actual linguistic distance is only indicated by figures printed between the contiguous sites. In the more general approach of multidimensional scaling, the second statistical method to be discussed here, any type of quantified dissimilarity can be represented spatially by a straightforward procedure. The degree of dissimilarity between different sites, or between the results optained for different informants, to put it more generally, is graphically expressed by different proportions of spatial distance. Thus, if the dissimilarity between A and B is twice as much as between B and C, the spatial distance between A and B will be twice as large as between B and C as well. In the case of the dialectometric method

18

Walter Bisang

as introduced above, the geographical dimension is represented directly, while other potential dimensions (social, cultural, etc.) must be represented indirectly. Multidimensional scaling allows to represent any dimension directly. The advantage of this procedure is very adequately characterized as follows by Chambers and Trudgill (1998): One clear advantage of multidimensional analysis is that the investigator’s categorisations of the variables do not pre-judge the analysis or affect it in any other way. The statistical program searches the matrix for dissimilarities wherever they may be found and scales them independently for each dimension. Multivariate scaling often reveals correlations in the data that the investigator had not previously recognised. (Chambers and Trudgill 1998: 141)

The above interpretations of data concentrate either on the location of linguistic variability in the geographical space as such or on quantifying dissimilarity among different sites or individual informants be it with regard to geographical space or with regard to other domains. Another approach to variability concentrates on the covariation of structural variables as they are used by individual informants or within certain geographical or other domains. The following example from the variety of English spoken in Anniston (Alabama, USA) as described by Feagin (1979) is to illustrate this. In this variety, plural subjects often do not agree with the verb as in They was not at home (to use a constructed example; cf. also Chambers, this volume, on what he calls ‘default singulars’ in this and other non-standard varieties of English). As it turns out, the occurrence of nonagreement correlates with certain types of constituents in the subject position. The following table from Chambers and Trudgill (1998: 133) shows the lects attested in Anniston depending on whether the subject is they, a noun in the plural (NPpl), we, you (pl) or there. The (+) sign indicates that nonagreement is attested, the (ø) sign indicated that it is not. Table 1. Occurrence of nonagreement in Anniston

Lect 1 Lect 2 Lect 3 Lect 4 Lect 5

Nongreement with they NPpl ø ø ø ø ø ø ø + + +

we ø ø + + +

you (pl) ø + + + +

there + + + + +

Dialectology and typology – An integrative perspective

19

The above five lects attested in Anniston form an implicational hierarchy of the following type: (1)

they > NPpl > we > you (pl) > there

Thus, if a person from Anniston is heard using nonagreement with a certain constituent X one can conclude that she will also use it with a constituent further to the right. For instance, if a person uses nonagreement with the pronoun we in the subject position she will also use it with you (pl) and with there. As it will turn out in section 3.3.2, this approach parallels implicational universals and hierarchies. 3.

Typology

3.1. Defining the domains of variability Data are collected from descriptive grammars, text material and native speaker elicitation on the basis of semantic or cognitive concepts. Sometimes, questionnaires are set up to make sure that all the potential aspects of a concept are equally covered for each language to be analysed. Semantic or cognitive concepts form the tertia comparationis and thus the domains of variability in typology. An example of such a semantic or cognitive concept is possession defined as the relation between an entity and another entity it owns. Another famous example is hidden behind the letters S for subject, V for verb and O for object in Greenberg’s (1966) basic word-order types. The three letters of S, V and O roughly correspond to the semantic categories of agent, predicate and patient, respectively. As Greenberg (1966) points out himself, his findings on basic word order as well as his findings on the position of the genitive relative to the noun are based on semantics: If a language has verb-subject-object as its basic word order in main declarative clauses, the dependent genitive always follows the governing noun. It is here assumed, among other things, that all languages have subject-predicate constructions, differentiated word classes, and genitive constructions, to mention but a few. I fully realize that in identifying such phenomena in languages of differing structure, one is basically employing semantic criteria. (Greenberg 1966: 74)

20

Walter Bisang

The items asked for in dialectological questionnaires are, apart from straightforward instances of the lexicon, in most cases only indirectly linked to concepts, i.e. the fieldworker deals with individual morphosyntactic constructions and the variables within their expression formats. Of course, these structures ultimately reflect some concept as well but the range of structural variability seems to be narrow enough in most dialect studies for the authors of questionnaires to feel no need to depart from more abstract concept-based questions. Thus, in a study on varieties of English it seems reasonable to assume that there is a subject which agrees with the verb. Given the unlikelihood that agreement follows any typologically different patterns such as absolutive agreement (the transitive verb agrees with the patient argument rather than with the agent, cf. Dixon 1979, 1994) in any variety of English, dialectological studies such as the one by Feagin (1979) quoted in section 2.3 can operate more easily on the basis of concrete morphosyntactic structures. A typologist, by contrast, has to reckon with a much larger range of variability which a priori excludes any recourse to concrete morphosyntactic structures in a questionnaire. It will turn out below (section 3.3.1), however, that the assumption of a comparatively low degree of variability among dialects is not always justified. Thus, there is no principled reason why dialectologists may not have to reckon with more significant morphosyntactic variation which needs a cognition-based approach not only for individual lexical entries but also for morphosyntactic structures. 3.2. Sources of the linguistic data Given the global range of typological research, eliciting data directly from native informants for a large number of languages is an enormous organizational task which is probably beyond the capacities of typological projects. A solution to this problem is to construct a questionnaire and to distribute it to specialists or native speaker consultants (Dahl 1985; also cf. van der Auwera and Ó Baoill 1998). Of course, this solution is not without disadvantages. While direct elicitation itself implies problems of misrepresentations because the native speaker provides her information in an artificial sociolinguistic setting and because the linguist may unconsciously bring in her prior expectations, indirect elicitation by a questionnaire implies an even greater distance between the actual speaker of a language and the typologist, and it adds the problem of a specialized mediator who

Dialectology and typology – An integrative perspective

21

may not fully understand the intentions of the questionnaire. Texts are another source of data which are of particular relevance in typological research on discourse (Myhill 1992). Apart from the two sources of (in)direct elicitation and texts, typologists most often concentrate their research on grammars and other descriptions of individual languages. Grammars provide a relatively comprehensive overall description of a language and at least a considerable number of them are fairly well accessible. Apart from the fact that the structure of descriptive grammars can vary enormously and that the typologist may not find a detailed enough description of her research topic, many grammars tend to concentrate on a particular variety of a language as it is spoken in one village and may thus miss a lot of variability. Such a concentration is well founded from the perspective of the comprehensiveness of a grammar but it precludes any insights into the range of language-internal variability. To summarize, there is a number of sources none of which is perfect. A typologist can only try to deal with them in as responsible and as judicious a way as possible. 3.3.

Methods of analysis and results

3.3.1. Sampling methods in typology and the impact of the dialect vs. language problem The languages analysed in a typological study are often part of a sample which needs to meet certain criteria to avoid the overrepresentation of certain linguistic structures. The two most important criteria are the genetic relatedness and the geographical closeness of the languages selected. Both of them imply social factors. The relevance of genetic relatedness for language sampling is founded on the assumption that there is a certain structural coherence among languages derived from a historically more or less homogeneous speech community which then was split up into a set of different speech communities. The relevance of geographical closeness departs from language contact among speakers of different languages and its impact on structural convergence. Other criteria may be cultural similarities among the speakers of the languages and, last but not least, the accessibility of adequate grammatical descriptions. The last criterion should not be underestimated given the fact that only a fraction of the world’s 6.000 to 7.000 languages are adequately described.

22

Walter Bisang

Rijkhoff et al. (1993) and Rijkhoff and Bakker (1998) discuss two types of samples depending on the questions a typological research project wants 1 to address : 1. 2.

Variety Sample: This sample is used to discover the full range of possible variation for expressing a given cognitive-semantic area in the world’s languages. Probability Sample: This sample is used for finding correlation pairs (typological features that form a pair such as verb-object [VO]/objectverb [OV] and Preposition/Postposition; cf. Dryer 1992).

The papers by Rijkhoff et al. (1993) and Rijkhoff and Bakker (1998) provide criteria for setting up variety samples. They present a matrix which shows how many grammars need to be selected from the 17 language families of the world as distinguished by Ruhlen (1987) and from Pidgin and Creole languages, language isolates and unclassified languages for constructing a sample of x languages, whereby x has to be minimally 30. The rather complex calculation of this matrix is based exclusively on genetic relatedness which is taken as an indicator of linguistic diversity, i.e. larger genetic distance implies greater linguistic diversity. The authors do not deny the relevance of geographical closeness for the calculation of linguistic diversity but they argue that migration and variable geographical distances between languages depending on population density are difficult to quantify adequately. As they put it, there is simply no better alternative in spite of the controversial discussion of many genetic affiliations in the literature than to calculate potential linguistic diversity exclusively on the basis of that criterion. After all, their aim is simply to provide a practicable tool for constructing a variety sample on the basis of the knowledge available at the moment. The most elaborate method of constructing a probability sample is developed by Dryer (1992). The purpose of his typological study is to see what binary word-order sequences covary with the order of verb-object and object-verb, respectively. For that purpose, he does not look at individual languages. His basic unit is the genus, i.e. a group of languages which can be roughly compared to a subfamily of the Indo-European family (e.g. Germanic, Indo-Aryan, Indo-Iranian, Romance, etc.). Dryer (1992) lists 252 genera. The languages within a genus are assumed to share a relatively large number of typological properties. As soon as a particular word-order type is attested in at least one language of a genus, that genus will get the

Dialectology and typology – An integrative perspective

23

value 1 for that type. Finally, the 252 genera are divided among six large linguistic areas of continental size (Africa, Eurasia, Southeast Asia and Oceania, Australia and New Guinea, North America, South America). With this division it is possible to see whether a certain type is a world-wide phenomenon or whether it is local. A pair of parameters is said to be a correlation pair of VO/OV and thus of typological relevance only if the two values covary equally through all six large areas. With this method, Dryer (1992) tries to integrate the genetic factor and he avoids the distortion of his statistics by the impact of geographical closeness. As was pointed out above, the correlation between language and dialect is of social nature. From a structural perspective, this implies that it is not possible to draw any conclusions about the degree of structural divergence between communities whose varieties are called languages and between communities whose varieties are called dialects. From a statistical perspective, this has its impact on both methods discussed above, the method of constructing a variety sample as well as the method of constructing a probability sample. Assumptions about how many varieties are dialects and how many of these are subsumed under one language may have an impact on the size of language families and thus on the overall statistical balance of a sampling method such as the one suggested by Rijkhoff et al. (1993) and Rijkhoff and Bakker (1998). The question of how much structural variation is to be expected within a language may in principle also represent a problem for Dryer’s (1992) method of setting up a probability sample, since not all the types actually present within a genus containing that language may show up in the statistics. The problem of how to distinguish dialects from languages and its statistical consequences will be briefly discussed on the basis of the Gbe languages (West Africa) as described by Lefebvre (1998: 59). The problem of seemingly enormous typological differences within closely related dialects/languages will be illustrated by the example of (non)configurationality in Warlpiri and Jiwarli (Austin and Bresnan 1996). If we take Man and Dalby’s (1987: 62) list of Gbe languages, we find the following ten language names (in alphabetical order): Ajagbe, Cigbe, Ewegbe, Fongbe, Gengbe, Gungbe, Kogbe, Phlaphera, Wem(gbe and Xwèdagbe. Each of these languages is supposed to be divided into several dialects. On the other hand, we have Capo’s (1984: 168) description of Gbe as a dialect cluster: ... [t]here is mutual intelligibility between dialects that are contiguous, e.g. Ewe and Gen, Gen and Aja, Aja and Fon, etc.; but the degree of mutual

24

Walter Bisang intelligibility is related to ‘geographical’ distance, e.g. although there is some mutual intelligibility between Ewe and Aja, it is less than between Ewe and Gen, and there seems to be none between Ewe and Fon which are the extreme ends of the dialect cluster. (Capo 1984: 168)

If we further assume that the Gbe languages are not the only instance within the Kwa family in which quantification of dialect vs. language is vague, it is easy to imagine that the actual number of separate statistical items assumed for this family (and for many other families in the world) is subject to considerable leeway. As a consequence, methods whose statistical balance is based on actual numbers of languages within families (and subfamilies) are to be looked at with caution. Warlpiri (Australia) attracted the attention of generative linguistics because of its nonconfigurational structure as analysed by Hale (1983). A year later, Jelinek (1984) argued that configurationality is still maintained in Warlpiri. Her argument was based on the fact that there is an obligatory auxiliary which must be cliticized to the last element of the first constituent of a finite clause. Since the person markers of this auxiliary, illustrated by =ka-pala (present-subject/3rd person/dual) in (2), have the status of arguments in Jelinek’s analysis, their corresponding noun phrases are adjuncts. Thus, configurationality still operates at the level of the auxiliary in Warlpiri, the free word-order pattern of the corresponding noun phrases is a result of their adjunct status. (2)

Warlpiri (Austin and Bresnan 1996:217): Kurdu-jarra-rlu =ka-pala maliki wajili-pi-nyi child-DUAL-ERG PRES-SUBJ:3rd dog:ABS chase-NONPST wita-jarra-rlu. small-DUAL-ERG ‘Two small children are chasing the dog. / Two children are chasing the dog and they are small.’

Jelinek’s (1984) argumentation crucially depends on the presence of pronominal clitics. If there were no person marking in the auxiliary Warlpiri would have to be analysed as a nonconfigurational language. From a theoretical perspective, this makes a lot of difference. From the perspective of mutual intelligibility, however, this difference seems to be negligible. Jiwarli, a closely related language/dialect discussed by Austin and Bresnan (1996) presumably lost its pronominal clitics in the course of history and should thus be analysed as a nonconfigurational language in

Dialectology and typology – An integrative perspective

25

terms of Jelinek (1984). Apart from the fact that Austin and Bresnan (1996) present good arguments against the pronominal-argument model as proposed by Jelinek (1984) even in Warlpiri, the case of Warlpiri and Jiwarli is a good illustration of how a look at “just the next dialect down the road” (DuBois 1985: 353, cf. section 1) provides a rather different picture to the theoretical linguist. Moreover, the same case also shows that what looks like a major structural difference from a theoretical perspective at a given time may only be of minor relevance from the perspective of speakers of closely related languages and dialects. Saying this, I do not want to exclude a priori the existence of major structural differences between closely related varieties. At the same time, I think it is safe to argue that if such instances are observed it is necessary to look very carefully at the theoretical foundations in order to see whether they can really justify why these differences must be of major relevance. 3.3.2. The findings of typology: Implicational universals and hierarchies Departing from a cognitively-defined area or from a pragmatically-defined area, typologists look at the morphosyntactic expression formats attested in a sample of languages. Within these expression formats, there is a certain number of structural types. Croft (1990: 38) defines a type as “a particular structural feature associated with a particular construction in a particular 2 language”. Once the range of attested types for a given cognitive or pragmatic area is evaluated, it is possible to see whether there are universal patterns of covariation among the presence or absence of certain types. In the rest of this section, I shall briefly look at implicational universals as introduced by Greenberg (1966) and at hierarchies. From this perspective, it will then be possible to discuss the covariation patterns of structural variability as they are used by individual informants or within certain geographical or other domains, as illustrated by Feagin’s (1979) work presented in section 2.3 above. Research on implicational universals reveals dependencies between two logically independent parameters (with one parameter usually consisting of two types in the above sense, or values). In a considerable number of cases, it turns out that only three of the logically possible four combinatorial language types are attested in a statistically significant way in cases based on two binary parameters. The following example from Hawkins (1983, universal IX’) is to illustrate this. If one takes the two word-order

26

Walter Bisang

parameters of the relative position of the genitive (G) to its nominal head (N) and the position of the relative clause (Rel) to its nominal head (N) it turns out that the combination of relative clause in front of the noun (RelN) and genitive following the noun (NG) is not attested. In a tetrachoric representation, this yields the following picture followed by Hawkins’ (1983) formulation of the universal: Table 2. Tetrachoric representation of Hawkins’ (1983) universal IX’

NRel RelN (3)

NG + -

GN + +

Hawkins (1983): Universal (IX’): NG NRel If in a language the genitive follows the noun, then the relative clause follows likewise.

Implicational universals can be combined to chains of implicational hierarchies. One such hierarchy is the animacy hierarchy as introduced by Silverstein (1976; cf. also Dixon 1979: 85). The factors integrated into that hierarchy are person, referentiality and animacy. In (4), I provide this hierarchy in a general form as presented in various textbooks (cf. eg. Croft 2003: 130). Given the complexity of how these factors can interact within individual languages, there are several specifications of this hierarchy. (4)

Animacy hierarchy: 1st person / 2nd person pronouns < 3rd person pronouns < proper names < nouns [+human] < nouns [+animate] < nouns [-animate]

This hierarchy plays an important role in a number of grammatical phenomena such as number distinctions in nouns, pronouns and other grammatical categories, assignment of gender (cf. Rohdenburg, this volume, on Low German dialects), split ergativity, direct-inverse marking and the coding of direct objects (differential object marking). In section 2.3, the occurrence or nonoccurrence of plural agreement on the verb in individual lects of Alabama English was discussed in terms of a hierarchy (Feagin 1979). For the sake of convenience, this hierarchy is repeated as (5):

Dialectology and typology – An integrative perspective

(5)

27

they > NPpl > we > you (pl) > there

Although this hierarchy also makes a distinction between pronouns and common nouns, it obviously deviates from the animacy hierarchy under (4). The third person pronoun is higher than the first and the second person and the common nouns are situated between the pronouns. This deviation may have something to do with the fact that Feagin’s (1979) hierarchy is based exclusively on plural subjects, but it may also be due to the fact that it is the result of an individual language situation which is not based on a statistically more or less balanced cross-linguistic sample. In any case, the range of validity of hierarchy (5) differs from that of the animacy hierarchy. While hierarchy (5) is attested in a number of varieties of English all over the world (cf. Chambers, this volume), but does not seem to be at work beyond English, the animacy hierarchy in (4) seems to be of universal relevance (with the caveat of what follows from Maslova 2000 as discussed in section 4.2) and shows up in a number of different grammatical structures. As is shown by Chambers (this volume, section 10), hierarchy (5) is not the result of universal principles but rather the result of “language-specific idiosyncrasies”. As for the animacy hierarchy, it is based on cognitive saliency. Nominal expressions are most easily processed when their referents are accessible in the context situation (1st and 2nd person). Given the general character of the animacy hierarchy, it is not limited to individual constructions – it rather affects the properties of a larger number of grammatically different constructions. 3.3.3. Typology and geography: Linguistic areas as zones of structural convergence The geographical dimension is of central importance in dialectology (cf. section 2.3 on maps). In typology, it is not only relevant as a criterion for constructing a language sample and for excluding the overrepresentation of certain structures due to contact-induced convergence, it is also a central issue in areal typology as introduced by Trubetzkoy (1928) on the phenomenon of the Sprachbund. Since Trubetzkoy’s programmatic definition of a linguistic area, many more definitions have been suggested in the literature. I will only quote the one by Emeneau (1956): A linguistic area is “an area which includes languages belonging to more than one family but showing traits in common which are found not to belong to the

28

Walter Bisang

other members of (at least) one of the families.” (Emeneau 1956: 16 note 28) Apart from the most famous linguistic area of the Balkan (Sandfeld 1930), there is a number of other candidates such as the Baltic (Stolz 1991) or the Circum-Baltic area (Dahl and Koptjevskaja-Tamm 2001), the Ethiopian area (Ferguson 1970, 1976; Zaborski 1991; Tosco 2000; Crass and Bisang, forthcoming), the South Asian area (Masica 1976), and the area of East and mainland Southeast Asia (Bisang 1996). None of these linguistic areas is discussed uncontroversially. Given the different definitions of what makes a linguistic area, this comes as no surprise. In my view, it is extremely hard, if not downright impossible, to come up with a clear-cut definition of a linguistic area. Apart from such obviously arbitrary criteria as the number of languages or the number of features to be involved in the definition of a linguistic area, it is not possible to define the relevant features with absolute exactness. This experience is confirmed by such theoretically divergent theories as Radical Construction Grammar (Croft 2001) and generative grammar. Croft (2001) points out that it is impossible for systematic reasons to compare constructions across languages. Generative linguists point out that what looks structurally identical at first glance may be considerably different from the perspective of Universal Grammar. From a more general perspective, one has to ask what would be the advantage of having a clear-cut definition of a linguistic area. In my view, there are two reasons that may justify the existence of such a definition: – The possibility of drawing further typological conclusions from the existence of a linguistic area. If x is a linguistic area, it must have certain other structural properties. – If the notion of a linguistic area defined in some way allowed for some concrete conclusions concerning social and/or historical properties of a speech community. Since a clear-cut definition is highly problematic and since it is unlikely that such a definition can contribute too much to the above two criteria, it is most likely that we will have to give up the concept of a linguistic area as introduced by Trubetzkoy (1928). With this I do not want to deny the existence of zones of convergence due to contact (such as the Ethiopian zone of convergence or East and mainland Southeast Asia), but I want to point out clearly that one is unlikely to find a uniform and discrete

Dialectology and typology – An integrative perspective

29

definition for these zones in terms of their structural properties. If one starts drawing a map with isoglosses for each of the relevant properties across the languages (and the dialects, provided information is available) it is most likely to get similar patterns as in dialectology. Thus, a more open approach analogous to geographical dialect continua plus sociolinguistic/ sociohistorical and cognitive motivations seems to be more to the point than a rigid, scientifically unfounded definition of what a linguistic area may be. A good example illustrating the complexity a linguist has to reckon with in a zone of structural convergence is Ethiopia. This zone is divided into a large number of small geographical zones of contact in which EthioSemitic, Cushitic, Omotic (all of them Afro-Asiatic) and Nilo-Saharan languages exert influence on each other. These small contact zones further condense into a larger zone of convergence which may well be called a 3 linguistic area. The dynamics within this large zone is only partially understood. The usual factors such as social networks, maxims of linguistic behaviour, etc. are certainly relevant. One may also safely assume the existence of continua as we know them from dialectology, that is, continua in which not all the speakers of the different varieties are actually in contact. The following quotation from Appleyard (1989) is an excellent characterization of the Ethiopian zone of convergence: It is perhaps too simplistic a statement to say ... that the presence of parallel systems in Ethiopian Semitic and Cushitic means that the former must have borrowed from the latter. What has to be taken into account here is that the languages of the horn of Africa, Semitic and Cushitic, share a common development and an infinitely complex network of mutual influence even where it cannot be assumed that their speakers have ever been in direct contact with one another. (Appleyard 1989: 304)

Many studies of concrete local contact situations are badly needed for a better understanding of the dynamics of this large zone of convergence. In this context, the use of maps would certainly help visualizing the geographical distribution of the relevant linguistic structures. However, the effort to get at all the relevant data on the languages involved – some 80 in the case of Ethiopia – seems to be by several dimensions greater than with dialect atlases if the same degree of detail is to be aimed at. Given the fact that many languages are poorly described or not described at all in many of the more complex zones of convergence, a lot of blank spaces still wait to be filled in.

30

Walter Bisang

In some other cases, where there is a large number of languages in a small area, it will be necessary to develop special representational techniques. This is for instance the case in some areas of Thailand where different languages are spoken at different altitudes depending on the time of immigration (for some more information, cf. McKinnon and Bhruksasri 1983). The speakers who came first take the lowest parts and so on until the latest speakers have to take the most mountainous areas. If this process of settlement extends over a large enough area, the map maker will face the problem of how to put all the relevant details onto a low-scale map whose size is still somehow manageable. In more recent times, a project with the title of World Atlas of Language Structures (WALS) (Comrie, Dryer, Gil and Haspelmath, forthcoming) creates maps showing the world-wide distribution of 120 structural features in 150 to 400 languages (with a core sample of 200 languages). These maps cannot provide the same degree of detail and the same degree of language density as is the case with dialects in dialectological maps. Thus, areas which look homogeneous with regard to the distribution of a certain feature may turn out to be much more complex if more languages from that area were integrated. Nevertheless, collecting data in this way is the only way to find out more about the geographical diffusion of patterns, about large linguistic areas and about structural properties which are extremely timestable. 4.

Fields of integration

4.1. Integrative functionalism and dialectology Croft (1995) draws a distinction between external functionalism and integrative functionalism. External functionalism rejects the assumption of a specific innate language acquisition device, but it looks at grammar as an arbitrary and self-contained system. Thus, competition-based models of grammar such as the one by Haiman (1983, 1985) see grammatical rules as the product of a selection based on conflicting motivations, which in the case of Haiman are iconicity and economy. Since this selection operates within the boundaries of grammar and does not take any grammar-external factors into account, approaches of external functionalism take the selfcontainedness of grammar for granted. If one looks at grammar from the level of an individual’s grammar and the variability within that grammar, it turns out that the selection of grammatical structures is also determined by

Dialectology and typology – An integrative perspective

31

grammar-external factors, i.e. it is not self-contained: “In other words, the systematicity of an individual’s variable grammar is tied to the individual’s interactions in different social settings; in this sense the grammar is not self-contained.” (Croft 1995: 518) The same problem also applies to functionally-oriented typology as it is discussed in this paper: One of the important missing elements of integrative functionalism is an integrated model of all the dynamic aspects of language, including the (variable) grammar. Typological functionalists discover crosslinguistic universals and propose as explanations for them principles of cognition and discourse that are based in the human mind and its interaction with its environment. But the mechanisms that lead from the internalized cognitivesocial principles to the diversity of linguistic form in the world’s languages have hardly been explicated. (Croft 1995: 520)

The sociolinguistic findings derived from a detailed dialectological analysis of linguistic variability have a lot to contribute to our understanding of the link between large-scale processes of language change and grammatical variability in the grammars of individuals and the factors that determine which variant those individuals will chose under particular social circumstances. In the rest of this section, I shall briefly sketch three models of how social mechanisms are involved in the diffusion of language change. Each of these models belongs to what Croft (2000) calls “utterance-based theories” and each of them can be seen as a contribution to a more general integrative approach which is still to be developed. As it will turn out at the end, there is some interdependency between some of these models. The social-network model as presented by Milroy and Milroy (1985) (also cf. Milroy 1992) accounts for the diffusion of linguistic structures in terms of weak and strong ties (Granovetter 1973). Strong ties between individuals are characterized by frequent and reciprocal contacts for a number of different reasons (e.g. sports, friendship, child care, shopping, etc.) as we typically find them in communities where everybody knows everybody else. Weak ties are less frequent, not necessarily reciprocal and they are usually due to only one reason. As it turns out, changes diffuse via weak network structures. Weak ties can be established more easily and they are much more frequent. As soon as a change has spread through enough networks with weak ties it will diffuse into strong network structures through locally influential people who want to maintain their position of leadership.

32

Walter Bisang

Labov (2001) looks at diffusion from the perspective of the leaders of linguistic change. In his view, “[t]he social network effects are not the largest, but they add essential information to the description of the leaders of linguistic change” (Labov 2001: 341). In his study of linguistic change and variation in the city of Philadelphia, he shows very convincingly that women have a different style of social interaction than men and that it is the style of the former which favours the spreading of linguistic change. Thus, the leaders of linguistic change are women whose social position and behaviour is described as follows by Labov (2001: 360): 1. 2. 3. 4. 5.

The leaders are women; men play no significant role. The highest concentration of leaders is in the groups centrally located in the socioeconomic hierarchy; that is, leadership forms a curvilinear pattern. The leaders are people with intimate contacts throughout their local groups, who influence first people most like themselves. The leaders are people who are not limited to their local networks, but have intimate friends in the wider neighborhood. These wider contacts include people of different social statuses, so that influence spreads downward and upward from the central group.

Labov’s (2001) model is derived from a very sound and thorough case study and it seems that it can be transferred to other instances as well. Whether it can be claimed to be of universal value or whether one has to reckon with culturally-based differences remains to be seen. Keller’s (1990, 1994) invisible-hand model of language change departs from the purposes of individual speakers and their maxims of language use. Two of these maxims are quoted for the sake of illustration: 4

(6)

Humboldt’s maxim (Keller 1990: 132) Speak as you think that the other would speak if he were in your position.

(7)

Keller’s phatic maxim (Keller 1994: 100) a. Talk like others talk. b. Talk in such a way that you are not misunderstood.

On a microlevel, individual speakers act according to certain maxims in order to attain certain purposes. If a large number of individuals follow at least partially similar purposes, the causal consequence of these individual purposeful actions on a macrolevel is a new phenomenon which cannot be

Dialectology and typology – An integrative perspective

33

predicted from the knowledge of the purpose and the maxim(s) involved. Thus, linguistic structures are invisible-hand products. They are selected by maxims of language use which determine what structural properties are to be diffused within a language community. Keller’s (1990, 1994) invisible-hand model and the other two models are not fully independent. The density of network structures can well be seen as the product of purposeful communicative actions of individuals and their purposes. The impact of leaders of linguistic change on their environment would be impossible without a large enough number of individuals who want to be identified as belonging to that environment. This behaviour follows LePage and Tabouret-Keller’s (1986) maxim of ‘talk in such a way that you are recognized as a member of the group you wish to identify with’, which is their interpretation of the more general maxim of (7a) ‘talk like others talk’. 4.2. Is the distribution of types in the present population of languages statistically neutral? – Contributions from dialectology Typological universals are based on the methodological assumption that the linguistic types in a sample be mutually independent. To grant this, the construction of a language sample requires due consideration of such factors as genetic relatedness or geographical distance (cf. section 3.3.1). Although this is a necessary prerequisite for discovering significant typological generalizations, it is not sufficient because this method does not allow to “distinguish between general distributional universals and accidental statistical properties of the current language population” (Maslova 2000: 307). The traditional typological procedure sketched above can only yield reliable, statistically significant results if the linguistic types have reached stationary distribution, i.e. a random distribution which is not biased by former typological properties of a given language or family: In other words, for the notion of distributional universal to make sense, it is required that after some (albeit possibly long) period of time, the probability for a language to belong to a certain type does not depend on the type which this language (or its ancestor) happened to have at t0 [the time of 10.000 years ago which is assumed to be the initial time for statistical reasons; W.B.] (Maslova 2000: 310)

The factors modifying the distribution of types within the population of languages at a given time are language death and language birth. The

34

Walter Bisang

precondition for a stationary state to arise is that the types of a typology are strongly connected in the sense that any type can develop into any other type (Greenberg 1995: 146–147; Maslova 2000: 312). If this were not the case, the statistical bias of a zero probability for a language at some given initial time to change from a type X into a type Y will last for ever. Given the lack of information on the death and birth of languages and the types attested in them and given the lack of enough diachronic data on type shifts, the question of whether the distribution within the present population of languages is statistically neutral cannot be answered at the moment, although it seems to be rather unlikely. If statistical neutrality is unlikely, the methods of sampling as discussed in section 3.3.1 are not sufficient. Consequently, Maslova (2000: 328–329) suggests a method which integrates estimates of transition probability, i.e. the probability that a language shifts from one type to another type within a given period of time. The basic procedure of how to obtain transition probabilities is described in the following quotation: It is clear ... that if we want to estimate the probability of a shift Ti –> Tj we must compare two quantities: the number of languages which have undergone this shift and the number of languages which have retained type Ti or shifted to another type within the same time interval. Accordingly, in order to estimate a transition probability p(Ti –> Tj) for some time interval t, one would need a sample of languages which can be assumed to have been in a state Ti t years ago. The current frequency of type Tj in this sample would give an estimate of the transition probability p(Ti –> Tj). (Maslova 2000: 329)

In my view, this procedure is based on a tacit assumption clearly stated by Croft (1995): “Social factors influencing the spread of change are independent of functional utility; thus, they select innovated variants randomly. The random effects of socially determined diffusion of variants should even out across a large enough sample.” (Croft 1995: 524) The crucial question concerning the reliability of Maslova’s (2000) procedure is whether social factors really do not have any influence on the selection of innovations. If we look at individual languages and/or dialects as discrete items in which certain structures change over time it seems safe to assume that structural innovations are the result of purely cognitive or discourse-related factors. But as soon as innovations are due to contact (also cf. section 5), the question of whether a speech community adopts a new type may be due to social factors such as speakers’ attitudes towards the languages or dialects involved in a given contact situation or due to

Dialectology and typology – An integrative perspective

35

maxims of language use such as those presented in (6) and (7). If type shifts depend on social factors, the time interval within which they take place is arbitrary in the sense that there is no cognitive or discourse motivation which may allow generalizations about what tends to be a short-term shift and what tends to be a long-term shift. Consequently, the assumption that it is possible to calculate fixed transition probabilities from simply looking at the rate of shifts from a type A to a type B in a given sample within a certain time interval may turn out to be problematic if the percentage of contact-induced shifts within that sample is relatively high. It is needless to say that the further back we go in time, the more difficult it will be to estimate that percentage. A relatively promising method is the method of multidimensional scaling as developed in dialectology (section 2.3). In principle, it should be possible to depart from individual features or types attested in the world’s languages and to calculate dissimilarities between individual languages or dialects or sites in a more abstract sense. If the distribution of these dissimilarities, however one may have to calculate them in detail, turns out to covary in a statistically significant way with knowledge about migrations or particular forms of settlement from fields such as history, archaeology or genetics, this would certainly be of considerable interest. To the extent to which significant covariation can be found, we may even get some vague ideas about the degree of statistic neutrality. 5. Beyond dialectology and typology: Language contact In the previous sections I have tried to point out some instances in which it is necessary to integrate dialectology and typology for discovering universal patterns of linguistic variation and for obtaining a better basis from which to reconstruct motivations from cognition and discourse. The relevance of dialectology for typology has a lot to do with the latter’s closeness to the individual, since any typological research which takes the integrative approach seriously ultimately aims at the individual, its cognitive equipment and its linguistic behaviour. If this is true one may ask whether there are more sources of gaining new insights into the linguistic potential of the individual beyond the integration of findings from dialectology into typology. In my view, language contact is at least one more source.

36

Walter Bisang

Language contact gives rise to forms of structural variation unknown in dialectology, i.e. variation in small populations down to the level of individuals, and it provides insights into socially extreme situations which go beyond the comparatively stable and settled situations in Europe or the United States which are the main focus of dialectological interest. Pidgin and Creole languages arise in socially extreme situations of contact as in the case of the Atlantic creoles in West African trade settlements (e.g. McWhorter 1997). Instances of this kind are not available in dialectological studies. Mixed languages (Bakker and Mous 1994; Thomason 1997), i.e. languages which take the lexicon from one language and the grammar from another (e.g. Media Lengua in Ecuador with Spanish lexicon and Quechua grammar, Muysken 1997), provide a form of variation which does not show up in dialectology. Similarly, code switching, i.e. “the rapid succession of several languages in a speech event” (Muysken 2000: 1), produces a kind of variation pattern traditional dialectology does not come across. Of particular interest in this context are stable situations in which code-switching is an integral part of the everyday communicative behaviour of a language community, i.e. in which two languages coexist in certain well-distinguishible forms of language mixing whose use depends on context (social context as well as topic of discussion). Karelian with its three different forms of Russian-Karelian code-switching as described by Sarhimaa (1999) is an interesting case in point. From a typological perspective, language contact provides insights into cognitive processes of structural convergence at the level of bilingual individuals (Silva-Corvalán 1994) and at cognitive processes of second language acquisition and negotiation operative in language shift, in extreme forms of language maintenance as well as in pidgin and creole languages (on maintenance and shift, cf. Thomason and Kaufman 1988; on second language acquisition and negotiation, cf. Thomason 2001: 142–148; on negotiation in pidgins and Creoles, cf. Kouwenberg 1992). Language contact also shows how structures unattested in the languages involved emerge from the interaction of substrate transfer and the reduction of markedness (simplification) (cf. McWhorter 1997 on the properties of copula constructions). In my view, the impact of this process is strong enough to challenge Bickerton’s (1981, 1999) bioprogram hypothesis in the sense that, if it exists at all, it “will be important only where the structures of the substrate languages do not coincide substantially” (Thomason and Kaufman 1988: 165).

Dialectology and typology – An integrative perspective

37

Of particular interest for typology are instances of language contact which are the source of new covariation patterns if languages with particular types meet (also cf. Bisang 2001). To conclude this section, I shall look at the case of Bayso (Cushitic: East Cushitic: South Lowland East Cushitic: Omo-Tana). Bayso is an OV language in which the adjective precedes the noun (AdjN) and the genitive follows the noun (NGen). The covariation of OV with (NGen and AdjN) is excluded by Greenberg’s universal 5: (8)

5

Greenberg’s Universal 5 (1966: 110; also cf. Hawkins 1994: 320) : If a language has dominant SOV order and the genitive follows the governing noun, then the adjective likewise follows the noun.

If we look at the word-order patterns of other languages around Bayso, we can distinguish two groups of languages called Highland East Cushitic 6 (HEC) and Lowland East Cushitic (LEC) by Tosco (1994: 416–417) : (9)

Highland East Cushitic (HEC): Lowland East Cushitic (LEC):

OV with AdjN and GenN OV with NAdj and NGen

Given the two word-order patterns in (9), the extraordinary pattern of Bayso is most likely due to its position in a contact zone between Highland East Cushitic (AdjN and GenN) and Lowland East Cushitic (NAdj and NGen): “It seems safe to assume that Bayso is currently moving from the N Mod order of its closest relatives (the other Omo-Tana languages) to Mod N – very possibly under pressure from its neighbors, the premodifying HEC languages. Influence from HEC on the Bayso Adj N order is suggested by Banti (1988: 243).” (Tosco 1994: 432) The covariation of OV with (NGen and AdjN) is not limited to Bayso, it is also attested in Tigre (Ethiosemitic; Tosco 1998), where it is the product of a language-internal development. The emergence of a cliticized prenominal article supported the movement of adjectives and relative clauses into the prenominal position and at the same time precluded an analogous movement of the genitive. As Tosco (1998) shows, this development can be accounted for in terms of Hawkins’ (1994) model of parsing. Given the local relevance of Tigre, this word-order pattern was even able to trigger deviations from consistent left-branching patterns in some of its neighbouring Cushitic and Nilo-Saharan languages (Tosco 1998: 360–364).

38

Walter Bisang

As one can learn from the cases of Bayso and Tigre, individual structures and covariation patterns may slip through the grid of language sampling. Language-internal as well as contact-induced language external factors may give rise to rarely attested structures. Under adequate sociolinguistic situations nothing prevents these structures from spreading into other languages and becoming more prominent. Thus, what is a “quirky” case in the distribution pattern of linguistic types at one period of time may theoretically become a dominant pattern in another period of time. If this scenario is realistic at least in some instances, it is questionable to what extent “[t]he random effects of socially determined diffusion of variants should even out across a large enough sample” (Croft 1995: 524; cf. section 4.2). 6. Conclusion The integrative approach ultimately links the microlevel of the individual to the macrolevel of the distribution of types in the world’s languages. If we want to know more about the range of the linguistic potential of humans, each variant is relevant. In this sense, variety matters. The integrative perspective evokes our suspicion that more might have happened in the past than we can see in the structures of today’s languages. Given the appropriate sociolinguistic conditions plus extensive migration, what has been a “quirky” case in a language sample at some time in the past may have become a “normal” case in a language sample of today (section 5). The uncertainties about the validity of typological generalizations brought up by recent discussions (section 4.2, Maslova 2000) stimulate alternative methods of constructing samples based on criteria beyond genetic relatedness and geographical closeness. This leaves the question of what is the status of all the universals obtained by the traditional methods as described in section 3.3. It may be interesting to hear the conclusion of a formal linguist who critically looked at typology and its findings (Newmeyer 1998: 297–364): A more difficult foundational question is whether, given the current distribution of available languages, any attainable typological generalization is of a level of significance to require explanation by linguistic theory. Much of this chapter [Newmeyer’s chapter 6; W.B.] has been devoted to casting doubt on the significance of the generalizations that

Dialectology and typology – An integrative perspective

39

have been achieved in language typology. However, the best strategy is to assume, for want of evidence to the contrary, that the more robust-seeming of them are valid and to pursue whatever combination of formal and functional explanatory mechanisms seem appropriate to explaining them. (Newmeyer 1998: 364)

I share Newmeyer’s (1998) basic optimism as far as the “more robustseeming” generalizations are concerned. But as I tried to show in this paper, another promising way of coming up with “more robust” generalizations is the integration of typology, dialectology, and sociolinguistic models of diffusion plus findings from contact linguistics. Notes 1. 2. 3.

4. 5. 6.

I will not discuss the random or convenience sample. This type of sample does not take into account any form of statistical balance. In his second edition, Croft adapts this definition to his Radical Construction Grammar approach by stating that “a structural type is represented by a particular construction in a particular language” (Croft 2003: 42). Tosco (2000) argues against the existence of a linguistic area in Ethiopia. Given the methodological impossibility of defining a linguistic area concisely, it seems more interesting to look at concrete instances of contact induced changes and their diffusion. In Humboldt’s own version, the maxim is expressed as follows: “Es darf also niemand auf andere Weise zum Andren reden, als dieser, unter gleichen Umständen, zu ihm gesprochen haben würde.” (from Keller 1990: 132). Also cf. Hawkins (1994: 320): “... if NGen then NAdj in postpositional languages (equivalently: if AdjN then GenN). This generalization holds exceptionlessly for the Expanded Sample of Hawkins (1983: 67–68).” This terminology reflects word-order patterns, it does not reflect genetic relatedness. Genetically, the languages involved belong to the following two branches of East Cushitic: South Lowland East Cushitic (Omo-Tana [Bayso, Arbore, Elmolo, Dasenech, Somali, Rendille], Saho-Afar) and Highland East Cushitic (Hadiya, Kambata, Sidamo, Gedeo/Darasa, Burji). The HEC languages consist of the Highland East Cushitic family plus the Saho-Afar branch from the South Lowland East Cushitic family. The LEC languages consist of the South Lowland East Cushitic family minus its Saho-Afar branch.

40

Walter Bisang

References Appleyard, David 1989 The relative verb in focus constructions: an Ethiopian areal feature. Journal of Semitic Studies 34: 291–305. Austin, Peter and Joan Bresnan 1996 Non-configurationality in Australian aboriginal languages. Natural Language and Linguistic Theory 14: 215-268. Auwera, Johan van der and Dónall P. Ó Baoill (eds.) 1998 Adverbial Constructions in the Languages of Europe. Berlin: Mouton de Gruyter. Bakker, Peter and Maarten Mous (eds.) 1994 Mixed Languages. 15 Case Studies in Language Intertwining. (Studies in language and language use 13). Amsterdam: Institute for Functional Research into Language and Language Use. Banti, Giorgio 1988 Adjectives in East Cushitic. In: Marianne Bechhaus-Gerst and Fritz Serzisko (eds.), Papers from the International Symposium on Cushitic and Omotic Languages. Cologne, January 6–9, 1986, 205– 259. Hamburg: Buske. Bickerton, Derek 1981 Roots of Language. Ann Arbor, MI: Karoma. 1999 How to acquire language without positive evidence: What acquisitionists can learn from creoles. In: Michael DeGraff (ed.), Language Creation and Language Change, 49–74. Cambridge, MA: The MIT Press. Bisang, Walter 1996 Areal typology and grammaticalization: Processes of grammaticalization based on nouns and verbs in east and mainland South East Asian Languages. Studies in Language 20: 519–597. 2001 Areality, grammaticalization and language typology. On the explanatory power of functional criteria and the status of Universal Grammar. In: Walter Bisang (ed.), Language Typology and Universals, 175–223. Berlin: Akademie-Verlag. Capo, Hounkpati B. C 1984 Elements of Ewe-Gen-Aja-Fon dialectology. In: François de Medeiros (ed.), Peuples du golfe du Bénin: Aja-Ewé, 167–178. Paris: Karthala. Chambers, Jack K. and Peter Trudgill 1998 Dialectology. 2nd edition. Cambridge: Cambridge University Press.

Dialectology and typology – An integrative perspective

41

Chappell, Hilary (ed.) 2001 Sinitic Grammar. Synchronic and Diachronic Perspectives. Oxford: Oxford University Press. Comrie, Bernard, Matthew Dryer, David Gil and Martin Haspelmath (eds.) forthc. World Atlas of Language Structures (WALS). Oxford: Oxford University Press. Crass, Joachim and Walter Bisang forthc. Einige Bemerkungen zum äthiopischen Sprachbund und ihre Relevanz für die Areallinguistik. In: Walter Bisang, Ursula Verhoeven und Thomas Bierschenk (eds.), Kultur. Sprache. Kontakt. Würzburg: Ergon. Croft, William 1990 Typology and Universals. Cambridge: Cambridge University Press. 1995 Autonomy and functionalist linguistics. Language 71: 490–532. 2000 Explaining Language Change. An Evolutionary Approach. Essex: Pearson Education. 2001 Radical Construction Grammar. Syntactic Theory in Typological Perspective. Oxford: Oxford University Press. 2003 Typology and Universals. Second edition. Cambridge: Cambridge University Press. Dahl, Östen 1985 Tense and Aspect Systems. Oxford: Blackwell. Dahl, Östen and Maria Koptjevskaja-Tamm (eds.) 2001 Circum-Baltic Languages. Vol. 1: Past and Present; Vol. 2: Grammar and Typology. [Studies in Language Companion Series No 54, 55.] Amsterdam/Philadelphia: John Benjamins. Dixon, Robert M.W. 1979 Ergativity. Language 55: 59–138. 1994 Ergativity. Cambridge: Cambridge University Press. Dryer, Matthew S. 1989 Large linguistic areas and language sampling. Studies in Language 13: 257–292. 1992 The Greenbergian word order correlations. Language 68: 81–138. DuBois, John A. 1985 Competing motivations. In: John A. Haiman (ed.), Iconicity in Syntax, 343–366. Amsterdam/Philadelphia: John Benjamins. Emeneau, Murray 1956 India as a linguistic area. Republished in: Anwar S. Dil (ed.), 1980. Language and Linguistic Area: Essays by Murray B. Emeneau, 105–125. Stanford: Stanford University Press.

42

Walter Bisang

Feagin, Crawford 1979 Variation and Change in Alabama English. A Sociolinguistic Study of the White Community. Washington, D.C.: Georgetown University Press. Ferguson, Charles A. 1970 The Ethiopian language area. Journal of Ethiopian Studies 8: 67–80. 1976 The Ethiopian language area. In: Marvin L. Bender, J. D. Bowen, R. L. Cooper und Charles A. Ferguson (eds.), Language in Ethiopia, 63–76. London: Oxford University Press. Gilliéron, Jules 1902–1910 Atlas linguistique de la France. 13 vols. Paris: Champion. Goebl, Hans 1982 Dialektometrie. Wien: Österreichische Akademie der Wissenschaften. 1993 Dialectometry: a short overview of the principles and practice of quantitative classification of linguistic atlas data. In: Reinhard Köhler and Burghard Rieger (eds.), Contributions to Quantitative Linguistics, 277–315. Dordrecht: Kluwer. Granovetter, Mark 1973 The strength of weak ties. American Journal of Sociology 78: 1360– 1380. Greenberg, Joseph H. 1966 Some universals of grammar with particular reference to the order of meaningful elements. In: Joseph H. Greenberg, (ed.), Universals of Language, 73–113. Cambridge, MA/London: MIT Press. 1995 The diachronic typological approach to language. In: Masayoshi Shibatani and Theodora Bynon (eds.), Approaches to Language Typology, 143–166. Oxford: Clarendon. Haiman, John 1983 Iconic and economic motivation. Language 59: 781–819. 1985 Natural Syntax. Cambridge: Cambridge University Press. Hale, Kenneth 1983 Warlpiri and the grammar of non-configurational languages. Natural Language and Linguistic Theory 1: 5–74. Hawkins, John A. 1983 Word Order Universals. New York: Academic Press. 1994 A Performance Theory of Order and Constituency. Cambridge: Cambridge University Press. Jaberg, Karl and Jakob Jud 1928–1940 Sprach- und Sachatlas Italiens und der Südschweiz. Zofingen: Ringier.

Dialectology and typology – An integrative perspective

43

Jelinek, Eloise 1984 Empty categories and non-configurational languages. Natural Language and Linguistic Theory 2: 39–76. Keller, Rudi 1990 Sprachwandel. Tübingen: Francke. 1994 On Language Change: The Invisible Hand in Language. London: Routledge. (Translation and expansion of Sprachwandel). Kouwenberg, Silvia 1992 From OV to VO: linguistic negotiation in the development of Berbice Dutch Creole. Lingua 88: 263–299. Labov, William 2001 Principles of Linguistic Change. Social Factors. Oxford: Blackwell. Le Page, Robert B. and Andrée Tabouret-Keller 1985 Acts of Identity. Cambridge: Cambridge University Press. Lefebvre, Claire 1998 Creole Genesis and the Acquisition of Grammar. Cambridge: Cambridge University Press. Man, Michael and David Dalby 1987 A Thesaurus of African Languages: A Classified and Annotated Inventory of the Spoken Languages of Africa with an Appendix on Their Written Representation. London/Munich/New York/Paris: Hans Zell. Masica, Colin P. 1976 Defining a Linguistic Area. South Asia. Chicago/London: University of Chicago Press. Maslova, Elena 2000 A dynamic approach to the verification of distributional universals. Linguistic Typology 4: 307–333. McKinnon, John and Wanat Bhruksasri (eds.) 1983 Highlanders of Thailand. Kuala Lumpur/Oxford/New York/ Melbourne: Oxford University Press. McWhorter, John H. 1997 Towards a New Model of Creole Genesis. New York: Peter Lang. Milroy, James 1992 Linguistic Variation and Change. Oxford: Blackwell. Milroy, James and Lesley Milroy 1985 Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339–384.

44

Walter Bisang

Muysken, Pieter 1997 Media Lengua. In: Sarah G. Thomason (ed.), Contact Languages: A Wider Perspective, 365–426. Amsterdam/Philadelphia: John Benjamins. 2000 Bilingual Speech. A Typology of Code-Mixing. Cambridge: Cambridge University Press. Myhill, John 1992 Typological Discourse Analysis. Oxford: Basil Blackwell. Newmeyer, Frederick J. 1998 Language Form and Language Function. Cambridge, MA: The MIT Press. Rijkhoff, Jan, Dik Bakker, Kees Hengeveld and Peter Kahrel 1993 A method of language sampling. Studies in Language 17: 169–203. Rijkhoff, Jan and Dik Bakker 1998 Language sampling. Linguistic Typology 2: 263–314. Ruhlen, Merrit 1987 A Guide to the World’s Languages. Vol. 1: Classification. London: Edward Arnold. Sandfeld, Kristian 1930 Linguistique Balkanique. Problèmes et résultats. Paris: Klincksieck. Sarhimaa, Anneli 1999 Syntactic Transfer, Contact-Induced Change, and the Evolution of Bilingual Mixed Codes. Helsinki: Finnish Literature Society [STUDIA FENNICA: Linguistica 9]. Séguy, Jean 1973 Atlas linguistique de la Gascogne. Vol. 6. Notice explicative. Paris: Centre national de la recherche scientifique. Silva-Corvalán, Carmen 1994 Language Contact and Change: Spanish in Los Angeles. Oxford: Clarendon Press. Silverstein, Michael 1976 Hierarchy of features and ergativity. In: Robert M.W. Dixon (ed.), Grammatical Categories in Australian Languages, 112–171. Canberra: Australian Institute of Aboriginal Studies. Stolz, Thomas 1991 Sprachbund im Baltikum? Estnisch und Lettisch im Zentrum einer sprachlichen Konvergenzlandschaft. Bochum: Universitätsverlag Dr. N. Brockmeyer. Thomason, Sarah G. 2001 Language Contact. An Introduction. Edinburgh: Edinburgh University Press.

Dialectology and typology – An integrative perspective

45

Thomason, Sarah G. (ed.) 1997 Contact Languages: A Wider Perspective, 295–365. Amsterdam: John Benjamins. Thomason, Sarah G. and Terence Kaufman 1988 Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Tosco, Mauro 1994 The historical syntax of East Cushitic: a first sketch. In: Thomas Bearth, Wilhelm J.G. Möhlig, Beat Sottas, and E. Suter (eds.), Perspektiven afrikanischer Forschung: Beiträge zur Linguistik, Ethnologie, Geschichte, Philosophie und Literatur. X. Afrikanistentag, 415–440. Köln: Köppe. 1998 A parsing view on inconsistent word order: Articles in Tigre and its relatives. Linguistic Typology 2: 355–380. 2000 Is there an ‘Ethiopian Language Area’? Anthropological Linguistics 42: 329–365. Trubetzkoy, Nikolaj S. 1928 Proposition 16. Actes du Premier Congrès International de Linguistes à La Haye, du 10-15 avril 1928, 17–18. Leiden: A.W. Sijthoff. Trudgill, Peter 1974 Linguistic change and diffusion: description and explanation in sociolinguistic dialect geography. Language in Society 3: 215–246. Van Valin, Robert D. and Randy LaPolla 1997 Syntax. Structure, Meaning and Function. Cambridge: Cambridge University Press. Wenker, Georg 1895 Der Sprachatlas des Deutschen Reiches. Marburg: Elwert. Zaborski, Andrzej 1991 Ethiopian language subareas. In: Stanislaw Pilaszewicz und Eugeniusz Rzewuski (ed.), Unwritten Testimonies of the African Past. Proceedings of the International Symposium Held in Ojrzanów n. Warsaw on 07–08 November 1989, 123–134. Warsaw: Wydawnictwo Uniwersytetu Warszawskiego.

Local markedness as a heuristic tool in dialectology: The case of amn’t Lieselotte Anderwald Abstract Standard English permits both auxiliary contraction and negative contraction for a range of auxiliary verbs. Generally, negative contraction is more widespread, and verbs which permit auxiliary contraction also permit negative contraction (but not the reverse). The exception is I’m not, where a negative contracted form (*I amn’t) is missing. While some dialects have closed this Standard English gap, this only amounts to a minority option. Why amn’t is not more widespread thus calls for an explanation, and the concept of local markedness borrowed from cross-linguistic typology helps explain this striking anomaly of Standard English and almost all other English dialects. 1. 2. 3. 4. 5.

Markedness Local markedness The investigation Cognitive explanation Conclusion

1. Markedness 1

Markedness is one of the fundamental concepts in much typological work. Greenberg (1966) makes extensive use of the concept of markedness in his Language Universals, although the concept itself is much older. It dates back to Trubetzkoy’s observations on phonology, who conceived of it simply in formal terms (absence vs. presence of a feature), and to Jakobson’s extension of the term to morphology and semantics, who used markedness as a functional category (the more usual, “normal” item being unmarked, disregarding its structural complexity), cf. Winter (1989) for a lucid disentanglement. Nevertheless, it is in typology, using both formal and functional criteria to determine markedness, where significant patterns have been established across languages, and indeed across linguistic levels. Markedness patterns have been attested in phonology, morphology and

48

Lieselotte Anderwald

syntax. In what follows, I will be mainly concerned with morphological markedness. While originally markedness was conceived of as absolute (contrasting a marked item with its unmarked counterpart), it is clear today that markedness is a relative phenomenon. That is, item A is more or less marked with respect to a certain feature than item B. Prototypical examples come from inflections (Greenberg 1966; summarized in the very useful list in Croft 1990: 92–94). For example, in the category of number, the singular is unmarked, the plural is marked. Where there are three values, the relative nature of markedness relations becomes intuitively clearer. For example, in languages with a dual, the rare dual is more marked than the plural, and the plural is more marked than the singular. Especially when we are dealing with more than two values, it is useful to depict markedness relations with the help of hierarchies. These take the opposite order from markedness relations (although there is no accepted 2 typographic convention to distinguish the two) . The hierarchies equivalent to the markedness relations mentioned in this section therefore are (1)

Hierarchies for number SINGULAR > PLURAL SINGULAR > PLURAL > DUAL

Other examples come from the area of tenses (past tense is more marked 3 than present) or polarity (negation is more marked than affirmation) . Greenberg (1966) provides several criteria to determine which in a given pair (or triple, …) of terms is marked, and which is unmarked (cf. also the grouping in Croft 1990: 70–94). Text frequency is an important factor; it has even been suggested that frequency alone can account for all markedness effects and can therefore substitute the concept of ‘markedness’ altogether (see Haspelmath 2002: 237–252). Text counts across widely different languages for example confirm that singular forms are far more frequent than plural forms at a ratio of roughly three to one (Greenberg 1966: 32; Haspelmath 2002: 238–239). Another important criterion is formal marking. This has been variously expressed as “zero marking” (frequently, the unmarked value is realized as zero vs. nonzero, cf. Greenberg 1966), more generally, “the realization of the marked value will involve at least as many morphemes as the realization of the unmarked value” (Croft 1990: 71); or simply as “shortness” (“zero expression ... is just one more manifestation of the correlation between frequency and

Local markedness as a heuristic tool in dialectology

49

shortness”, Haspelmath 2002: 239). Finally, the marked value is also characterized by higher syncretism and more regularity: the marked value typically is less differentiated (more syncretized) and more regular. Again this can be related to frequency, as only the higher frequency of the unmarked value permits irregular forms and a higher number of allomorphs to survive. Although these general assessments stem from empirical observations (and are thus necessarily descriptive, although on a more general level than descriptions of the individual languages), they have an interesting predictive component. This can best be exemplified with the help of Croft’s explicit formulation that “the realization of the marked value will involve at least as many morphemes as the realization of the unmarked value” (Croft 1990: 71). In a given language, the singular could be morphologically marked or not, and the plural could be morphologically marked or not. Of the four (2 by 2) possible combinations, markedness relations predict that only three should occur: (i) singular and plural could both be marked by zero; (ii) singular and plural could both be marked morphologically by the same number of morphemes; or (iii) the plural could have more morphemes than the singular. One combination is predicted not to occur, and even not to depict a possible human language: (iv) that the plural is marked by fewer morphemes than the singular. 2. Local markedness In a number of cases, these markedness relations are violated. There are examples – individual lexemes or groups of lexemes – where the singular is morphologically more complex than the corresponding plural. Croft cites Russian gorox ‘peas’ vs. goros½ina ‘pea’ and Turkana 0a-kì ‘ears’ vs. a-k-it ‘one ear’ (Croft 1990: 145); Haspelmath gives a number of examples from Welsh, e.g. deilen ‘leaf’ vs. dail ‘leaves’, ffäen ‘bean’ vs. ffa ‘beans’ or tywysen ‘ear of corn’ vs. tywys ‘corn’ (Haspelmath 2002: 244). These are the kinds of counterexamples that should not exist, as they exemplify scenario (iv) above and run counter to the hierarchical relations in (1). However, these exceptions are themselves quite regular crosslinguistically and they have been explained with reference to their semantics. They typically comprise objects that appear in pairs or as collectives, such as foodstuffs, small parts of plants, groups of animals, birds or people, or paired/multiple body parts. In other words, there is a

50

Lieselotte Anderwald

clear functional reason behind these reversals, because “this correlation between the category of number and the noun class represents a prototype: ... some objects are prototypically nonsingular (collective)” (Croft 1990: 66, my italics). In these marked cases, it is usual to call the marked singular “singulative”. Haspelmath stresses that in these cases, it is also the frequency that is reversed, so that “a small group of nouns tend to occur more often in the plural in all languages” (Haspelmath 2002: 243). These local areas where the markedness relation does not hold as predicted have been called local markedness, localized markedness or hierarchy reversals or indeed local frequency reversals. How local markedness can help explain a striking anomaly in the English auxiliary system will be the topic of this paper. 3.

The investigation

3.1. Standard English Standard English auxiliaries are negated by the addition of not, and the negator – in spoken language – is usually cliticized. Every auxiliary thus possesses a form with a negative clitic – indeed this cliticization is so widespread that it is sometimes regarded as inflectional (Zwicky and Pullum 1983). The one exception however is the form am, to which the negative clitic does not attach (Quirk et al. 1985: 122). Curiously, the lack of a negative contracted form for am (amn’t) in Standard English has not attracted much scholarly attention over the years until very recently (Hudson 2000; Bresnan 2001), and there is therefore no generally accepted explanation yet for this “gap”. Little has also been written about the situation in English dialects, with the exception of Francis (1985), who however concentrates on interrogatives. Although it is well known that some dialects do possess amn’t, and the regional distribution of this form is reasonably clear, alternative strategies and regional delimitations have not been extensively examined. 3.2. Negative vs. auxiliary contraction In Standard English declaratives, negative contraction (for all verbs except 4 am) exists side by side with auxiliary contraction (for some verbs). The

Local markedness as a heuristic tool in dialectology

51

overview in Table 1 shows that whereas negative contraction is practically always possible, auxiliary contraction is only possible in a subset of cases. In structural – distributional – terms, negative contraction is clearly the unmarked option, whereas auxiliary contraction is more marked. Table 1. Negative vs. auxiliary contracted verbs POSITIVE FORM

Negative form NEG contraction

am is are was were have has had do does did can could will would shall should may might must need ought

isn’t aren’t wasn’t weren’t haven’t hasn’t hadn’t don’t doesn’t didn’t can’t couldn’t won’t wouldn’t shan’t shouldn’t mayn’t mightn’t mustn’t needn’t oughtn’t

AUX contraction ‘m not ‘s not ‘re not ‘ve not ‘s not ‘d not ‘ll not ‘d not ‘ll not ‘d not -

This relation of asymmetry can be captured in the hierarchy in (2). (2)

Hierarchy for English contractible verbs NEGATIVE CONTRACTION > AUXILIARY CONTRACTION

(3)

Implicational tendency for English contractible verbs AUXILIARY CONTRACTION Š NEGATIVE CONTRACTION

52

Lieselotte Anderwald

The hierarchy in (2) does not hold for am, however. For all other verbs we can state that if a verb permits auxiliary contraction, it will also permit negative contraction, expressed by way of material implication (“if …then”) in (3). Once the distribution is stated in terms of a hierarchy, expectations pertaining to frequency are raised. If we consider negative contracted forms as relatively less marked than auxiliary contracted forms, we would also expect them to be significantly more frequent than auxiliary contracted forms. In the remainder of this paper I will concentrate on those verbs for which auxiliary contraction is possible. Apart from the present tense paradigm of BE, this therefore only leaves HAVE (both present and past), WILL/WOULD and, marginally, SHALL/SHOULD. 3.3. Standard English: The gap Standard English does not close the *amn’t gap in declaratives, and recently two explanations have been put forward. Bresnan (2001) in a stochastic model of Optimality Theory (OT) also considers some dialect data (if from a categorical, not a variationist perspective). In OT, individual grammars (e.g. for different languages, dialects or indeed different speakers) differ in the order in which the same, presumably innate, constraints apply. Constraint ranking leads to radically different output of the same input. All possible forms are generated, and the optimal form is then filtered out by the application of constraints. Bresnan’s main argument is that for the StE lack of amn’t “manifests itself by the absence of a recognized pronunciation for the pair in (60) ...:

(60)

‘ ’: ” (Bresnan 2001: 37).

This missing pronunciation filters out the generated forms I amn’t working from the list of possible candidates because every output must have a pronunciation. Instead, the uncontracted forms are selected in the model. That is, speakers would use I am not working for the presumably unpronounceable *I amn’t working (Bresnan 2001: 38).

Local markedness as a heuristic tool in dialectology

53

As Hudson (2000) notes, although this model can serve as an accurate description of the state of Standard English today, it does not have much explanatory power. In the case of missing pronunciations, the usual way would be to fix the pronunciation of the word, not to do without the word altogether. And indeed the awkward sequence /mn/ could be changed in several ways to make it more pronounceable: by homogenization to /m/ or /n/ (>/3OV/, >/3PV/), or by the insertion of an epenthetic vowel (>/3O‹PV/), a strategy that is in fact found in some dialects that do possess amn’t. Another important point is actual usage, which is not addressed by Bresnan. She only contrasts negative contracted forms with their uncontracted counterparts, although uncontracted forms are extremely rare in spoken language. As Table 1 has illustrated, the paradigm BE also possesses auxiliary contracted forms in the present tense. Bresnan states in a footnote that “the choice between the full verb am and the reduced ‘m is an orthogonal issue that is not addressed here” (Bresnan 2001: note 28). This choice however is far from being unimportant, as the following sections will show. Another explanation has been put forward by Hudson (2000) in his – essentially functional – framework of word grammar, a model rooted in Halliday’s systemic grammar. The main component of this grammar is inheritance: words inherit their characteristics from more general words by default. For example, the form are is a (form of) BE and inherits all those features that BE has, so that only its additional features have to be specified. Multiple inheritance is possible, and this almost inevitably leads to the possibility of conflicts between two more general forms that perhaps specify different features where none is more specific than the other, so that a no-win situation results. The technical term for this kind of conflict is a “Nixon diamond” (Nixon was a Quaker and a Republican. He inherited pacifism from “Quaker”, but non-pacifism from “Republican”). Computers typically capitulate in front of this kind of problem. Hudson argues that *amn’t suffers from a similar Nixon diamond type conflict, resulting in its non-existence. *Amn’t would both be a kind of am, inheriting the features , and a kind of aren’t, inheriting . Again, neither form is more specific than the other (aren’t crucially is not marked for person, but analyzed as the general form of negated BE), so that the (human?) computer balks and refuses to create this anomalous form. It is however not clear why amn’t, if it cannot be created regularly by the rules of inheritance, could not be stipulated – after all, the real Nixon

54

Lieselotte Anderwald

wasn’t immobilized by conflicting inheritances; as Hudson concedes, “for whatever reason, the principles of Republicanism triumphed over those of Quakerism, whereas the logic of inheritance should have left him paralyzed and unable to make decisions” (Hudson 2000: 320). This type of overriding a conflict is usually modelled by stipulation, and indeed stipulation is necessary for other phonologically irregular negative forms like won’t, 5 don’t, shan’t. Amn’t then could be stipulated, but obviously isn’t. Neither Bresnan nor Hudson however look at actual language use – apart from using a few dialect data categorically, something this paper sets out to remedy. 3.4.

The traditional dialects

3.4.1. Amn’t Perhaps the most obvious strategy of dialects faced with the Standard English gap would be to close it. The largest set of forms for amn’t can be found in Wright’s English Dialect Dictionary from 1898, where he reports forms for am with a contracted negative for a range of counties. Table 2. Forms of amn’t in (Wright 1898) County Antrim

sentence type declarative

south Scotland west Yorkshire

interrogative declarative interrogative interrogative declarative

Lancashire Shropshire

interrogative

A imin’t imnae Ym-n’ aa? Ai, a, i am‹t Am‹t ai, a, i? Am’t aw? I amma amna

alternative forms am no Aim, am, im not

arna binna bunna

Ammad amnad I

These figures confirm the first impression that in traditional dialects at the end of the nineteenth century (i.e. recording the use of speakers who will have acquired their speech in the first half of the nineteenth century) the form amn’t seems to be a Scottish/Irish and north-central phenomenon. The

Local markedness as a heuristic tool in dialectology

55

regional restriction of amn’t is also displayed in material from the Survey of English Dialects (SED), collected roughly half a century later. Forms of I amn’t are only recorded sporadically in Northumberland close to the Scottish/English border, in south Yorkshire and Lancashire (ammet), as well as throughout Derbyshire. Another small area is situated in Shropshire (amno’); one isolated occurrence comes from Warwickshire, as Map 1 6 illustrates.

Map 1. Amn’t in the Survey of English Dialects (SED)

56

Lieselotte Anderwald

Map 2. I aren’t in the Survey of English Dialects (SED)

This regional delimitation of amn’t is further confirmed by the dialect data 7 from FRED (the Freiburg English Dialect corpus) , containing texts from the 1970s and 1980s, i.e. of speakers one to two generations younger than in the SED. Instances of I amn’t here are attested for Scotland and Yorkshire. Amn’t is also still present in Northern Ireland in the 1980s, as attested in the Northern Ireland Transcribed Corpus of Speech (NITCS).

Local markedness as a heuristic tool in dialectology

57

These dialect areas have clearly closed the “gap” of the Standard English system in a regular way. Overall, however, I amn’t is not a majority pattern. Another non-standard strategy for negating I am is the all-purpose negator ain’t. Although this is not infrequently used also for I, at least regionally, it should probably be excluded from this discussion, because ain’t has neither a corresponding positive form (*ai) nor an auxiliary contracted form (*I’ai not). When using ain’t, speakers therefore have no choice of auxiliary vs. negative contraction, and ain’t for this reason does not fit into Table 1 above, which will be crucial to my argument. One final non-standard strategy deserves to be mentioned here, the use of are with I. Forms of I aren’t are again attested in the SED across the country, but they seem to constitute a rather marginal phenomenon (cf. Map 2). In the later data from FRED, I aren’t is attested in the Midlands (although here overall figures are too low to make any definite claims) and in the South (both the Southeast and the Southwest). This is a striking contrast to the SED data, which has ain’t in the Southeast and bain’t in the Southwest as the preferred forms. Possibly this new distribution indicates a change in progress; on the other hand, aren’t is a morphological form from the Standard English paradigm (in contrast to the highly stigmatized ain’t), and I aren’t is only non-standard by way of the combination with the first person singular. It is possible therefore that the presence of I aren’t indicates a different level of formality of FRED compared to the SED. Overall, however, instances from FRED are too rare to give definitive 8 answers. As a last resource, I have therefore turned to the spoken sections of the British National Corpus (BNC). 3.5. Present-day spoken language The BNC contains as one of its subcorpora a collection of unmonitored every-day speech – although this is not necessarily dialectal, it is a good 9 representation of spoken informal English. These spoken texts were collected in the early 1990s, comprise roughly five million words and are roughly geographically representative of Great Britain as a whole (cf. Aston and Burnard 1998). Forms of present tense BE here are frequent enough in practically every dialect area to allow a meaningful comparison. 10 The negated forms for I am in declaratives are distributed as follows:

58

Lieselotte Anderwald

Table 3. Negation of I am in declaratives (British National Corpus) Dialect area Ireland Scotland Wales Northeast Central North North Humberside Lancashire Merseyside Midlands West Midlands Central Midlands East Midlands South Midlands East Anglia Home Counties London Upper Southwest Central Southwest Lower Southwest Total

I aren’t 1 1 1 3

I’m not 29 81 102 105 73 32 35 100 17 40 144 149 93 36 67 239 358 28 146 71 1945

I am not 1 2 1 1 1 3 3 3 6 10 1 1 33

% aux. c. 100 98.8 97.1 99.1 98.6 100 97.2 99.0 100 100 98.0 98.0 100 100 95.7 97.6 97.3 100 99.3 97.2 98.2

Table 3 nicely shows that, contra Bresnan, uncontracted I am not is never really an optimal solution. With 33 instances out of 1981 instances (or 1.7 per cent), I am not constitutes a marginal phenomenon at best. The preferred option is clearly auxiliary contraction (I’m not) with an average of over 98 per cent and indeed in many dialect areas the only option attested. On the one hand, the clear preference for I’m not is not surprising, given the systematic absence of a negative contracted form. On the other hand, recalling the hierarchy in (2) and the frequency distribution associated with it, it raises the question of how wide-spread this preference for auxiliary contraction is. Table 4 therefore details the contraction ratios for all other verbs (again in declarative environments) – as differences between individual dialect areas are only minimal, they will be considered together here.

Local markedness as a heuristic tool in dialectology

59

Table 4. Negative vs. auxiliary contraction in declaratives (BNC SpS) Positive form

Negative form NEG contraction

is are have has had will would

isn’t aren’t haven’t hasn’t hadn’t won’t wouldn’t

% 12.4 3.2 89.7 76.3 95.2 95.7 99.9

AUX contraction

‘s not ‘re not ‘ve not ‘s not ‘d not ‘ll not ‘d not

% 86.9 95.2 10.0 23.0 3.9 2.9 0.1

Note: bold print indicates preferred options

Table 4 displays the striking asymmetry between the present tense BE forms is and are and all other contractible verbs. Just as for am, auxiliary contraction rather than negative contraction is the rule in declaratives for is and are. For all other verbs, negative contraction is the favoured pattern, and contraction ratios here are almost the reverse of those of present tense BE. These figures confirm that auxiliary contraction is not restricted to am alone. Instead, the whole present tense paradigm of BE seems to behave in this way. In other words, frequency distributions indicate that for present tense BE, the usual hierarchy seems to be reversed. In terms of text frequency, auxiliary contraction for BE is far more frequent than negative contraction, indicated in the hierarchy in (4). (4)

Hierarchy for present tense BE AUXILIARY CONTRACTION > NEGATIVE CONTRACTION

This hierarchy is the exact reversal of the hierarchy in (2): for present tense BE, auxiliary contraction is the unmarked member, and negative contraction is the exception, or more marked. From the hierarchy in (4) follows the implicational tendency in (5), again the reversal of the tendency in (3): (5)

Implicational tendency for present tense BE NEGATIVE CONTRACTION Š AUXILIARY CONTRACTION

And this implication that follows from the simple frequency distribution above covers exactly the asymmetry that we have already encountered in the distribution of structural possibilities of English: for forms of present

60

Lieselotte Anderwald

tense BE, auxiliary contraction is always possible (namely for am, is and are), whereas negative contraction is only found in a subclass of these three verb forms (namely is and are, but not am). We seem to be dealing with a hierarchy reversal here that is similar to local(ized) markedness phenomena known from the realm of number exemplified above. It is clear that re-naming the *amn’t-gap a hierarchy reversal or a phenomenon of local markedness does not in itself amount to a substantial gain in know-ledge yet. This tool (and insight) from functional typology is of interest here because of the cognitive explanation associated with it. 4. Cognitive explanation We have seen that the plural-singular hierarchy reversal can be plausibly explained through a difference in prototypes associated with a certain delimited group of nouns (namely those that are prototypically nonsingular). Applying this insight from functional typology to the distribution of verb forms in British English dialects, we would have to specify in which way present tense BE is prototypically different from all other verbs, as such a difference could then justify its exceptional behaviour. 4.1. Cross-linguistic evidence Ferguson notes that cross-linguistically, words for BE, i.e. the copula, are generally very different from other verbs. In languages that do possess a copula, “it generally differs from verbs … in certain respects, in some languages so much as to constitute a separate word class, in other languages in such a way as to belong to a distinct sub-class of verbs” (Ferguson 1996: 116). In many languages, the copula is possibly even more different from other verbs in that it can be systematically expressed by zero. Copula 11 absence seems particularly frequent in the present tense. If the copula is not completely absent, in many languages it can be variably deleted in the present tense forms. Striking examples come from many pidgins and creoles as well as some African languages (Holm 1988: 174–182). Thus, Principe Creole Portuguese and Guyanais Creole French have zero variably before noun phrases and adjectives, i.e. in typical equative copula constructions; Guinea-Bissau Creole Portuguese, Haitian Creole French, Sranan Creole English and Jamaican Creole English all have zero

Local markedness as a heuristic tool in dialectology

61

categorically before adjectives, as does the African language Mandinka, etc. Although Holm calls this a “very un-European pattern” (Holm 1988: 175), copula absence is not restricted to African languages and (arguably) African-derived creoles, it is also characteristic of an Indo-European language like present-day Russian and of Arabic. While it is also true that there are languages which do not possess a lexeme for ‘have’, this other case is different, however. The concept of ‘having’ in these languages is often expressed by a polymorphemic locative construction including the lexeme BE, for example in constructions like ‘it is at me’ (for example, in Russian, cf. also Miller this volume; in Finnish or Irish Gaelic). However, the concept of ‘having’ even in these languages cannot be expressed by zero. Although Russian both expresses ‘having’ by way of a form of BE and belongs to the group of languages that permit copula absence in the present tense, this still leaves the preposition and the pronoun as possession markers in the present tense, as (6) illustrates. (6)

u menja kljuþ at me key ‘I have (the) key’

‘Having’ then seems to indicate a concept that – unlike the concept of ‘being’ – cannot be variably deleted. Even in languages where the copula is regularly expressed, specific contexts – like telegraphese, newspaper headlines, so-called foreignerspeak or baby speak, etc. – may show copula absence systematically and indeed stereotypically (see Ferguson 1971). In addition, forms of present tense BE may be variably deleted in individual varieties of a language that possesses a copula. Again, this is so wide-spread that copula deletion is a candidate for Chambers’s vernacular primitives or vernacular universals (Chambers 1995: 243, and see the contribution by Chambers, this volume). For English, probably the most prominent example where we find this 12 phenomenon is African American Vernacular English (AAVE). It is exactly the present tense forms of BE that can principally always be deleted, resulting in sentences like He hungry. A sentence like She very nice is probably always – unambiguously – interpreted as lacking a copula.

62

Lieselotte Anderwald

4.2. Language-internal differences For English, another important difference – apart from its sheer frequency – that distinguishes present tense BE in particular from the other (contractible) primary verb HAVE, as well as the range of auxiliaries that can be contracted in English, is that it carries the least semantic content. The prototypical meaning of a verb could be described as denoting an ‘action’ of some kind (in school grammars, verbs are typically explained as being some kind of doing words, and since the advent of prototype theory it is well known that some verbs are “verbier” than others). This meaning does clearly not apply to BE, whose basic meaning can be described by ‘existence’, ‘stativity’ or ‘habituality’. Thus the OED notes for BE that the primary sense appears to have been ..., ‘to occupy a place’ (i.e. to sit, stand, lie, etc.) in some specified place; thence the more abstract [sense] was derived by abstracting the notion of particular place, so as to emphasize that of actual existence, ‘to be somewhere, no matter where, to be in the universe, or realm of fact, to have a place among things to exist.’ (OED be v. B.)

Because of this empty semantics, the presence or absence of BE does not greatly change the overall meaning of a clause, and a verbless clause is most naturally interpreted as containing a form of BE, which might be an underlying reason for the possibility of copula absence cross-linguistically. It should be noted that this minimal semantic weight of BE holds in particular for the present tense. The past tense (and all other tenses) typically behaves differently. Marking for tense alone seems to be sufficiently important for the copula to surface again. Thus contraction types in English for was and were conform to the overall pattern of all other verbs, and those languages that permit copula absence in the present tense typically have a copula in the past tense and in all other tenses (Ferguson 1996: 116). 4.3. Summary Cross-linguistic evidence as well as language-internal evidence from English indicates that the verb BE in the present tense is indeed very different from all other verbs, so that a pattern of local markedness can be motivated cognitively and, not surprisingly, typologically. In English, the local markedness takes the form of radically different contraction ratios;

Local markedness as a heuristic tool in dialectology

63

thus BE prefers auxiliary contraction in contrast to all other (eligible) verbs, where the negator is contracted instead. Local markedness can then explain the lack of a form like amn’t in most dialects of English: the hierarchy reversal shifts the ‘gap’ towards the end of the system where the presence of a form is optional. In the dialectal systems, then, there is no urgent need to fill this ‘gap’, because there simply is no gap. *Amn’t is just one more form that is not present in the system, and as such has the same status as the non-presence of an auxiliary contracted form of he does not or I must not. To close, Table 5 is an elaborated version of Table 1, indicating the unmarked contraction forms for all verbs: Table 5. Preference for auxiliary or negative contraction POSITIVE FORM am is are was were have has had do does did can could will would shall should may might must need ought

NEGATIVE FORM NEG contraction

AUX contraction

isn’t aren’t wasn’t weren’t haven’t hasn’t hadn’t don’t doesn’t didn’t can’t couldn’t won’t wouldn’t shan’t shouldn’t mayn’t mightn’t mustn’t needn’t oughtn’t

‘m not ‘s not ‘re not ‘ve not ‘s not ‘d not ‘ll not ‘d not ‘ll not ‘d not -

Note: unmarked option in bold

64

Lieselotte Anderwald

5. Conclusion The concept of local markedness or hierarchy reversals borrowed from functional typology explains not only the diametrically opposed frequency distributions in contraction options which we find consistently across Great Britain – and, incidentally, also for American English, as comparable figures from Biber et al. (1999) show – but also the fact that the *amn’t ‘gap’ is not very urgently closed by many dialect systems. This explanation hinges on the special status of BE, which was demonstrated both crosslinguistically for a range of languages and intra-linguistically for English. Ultimately, then, the *amn’t gap does simply not constitute a gap. Notes 1. 2. 3. 4. 5.

6. 7. 8.

I will not be concerned here with markedness in the generative paradigm; for a useful survey of both research traditions see Battistella (1990). For this reason I will depict hierarchies in SMALL CAPITALS, markedness relations in normal script. For an analysis how polarity affects markedness patterns in non-standard verb paradigms see Anderwald (2003). Auxiliary contraction cannot occur in interrogatives, which is why the discussion will be restricted to declarative environments. Interrogatives are discussed in more detail in Anderwald (2002: 72–92). Another point is the status of am. Because am does not possess a negative contracted form, its status is different from the other forms of BE. In word grammar, negative verbs can only be formed by adding -n’t to the corresponding positive forms, and only those forms count as positive forms that are arrived at through the subtraction of -n’t (Hudson 2000: 307). This leads to the analysis of am not as a positive form, as it clearly lacks a corresponding negative, but an inflected WHOLE which by definition prevents the generation of a negative form. This argument seems circular at best. Collected from the responses in the published basic material (responses to IX.7.10 in Orton and Halliday 1962–1964; Orton and Wakelin 1967–1968; Orton and Barry 1969–1971; Orton and Tilling 1969–1971). Financed by DFG grant no. Ko/1181/1-1 and 1181/1-2. FRED seems biased towards past tense contexts (probably due to its emphasis on Oral History material). It contains only 56 instances of negated I am – even if one only considers the distribution across the six very general dialect areas, this low number results in many empty cells. It is also well known that

Local markedness as a heuristic tool in dialectology

65

low numbers can lead to considerable distortion in the data, cf. Paolillo (2002: 42). 9. As is also well known, the majority of speakers of English are not speakers of traditional dialects, and in this sense the BNC might very well be representative of the British society in general, even if (or because) it does not contain many speakers of traditional dialects. 10. Rows are roughly in north-to-south order. 11. There seem to be some other specific contexts apart from the present tense that favour copula absence cross-linguistically (main clause, presence of subject and complement, attributive complement, third person subject) (Ferguson 1996: 116). 12. The pattern of copula deletion/absence has been used as an argument for the creole-origin of AAVE, but also as an argument against the creolist hypothesis, cf. Walker (2000). For a detailed discussion of copula absence in AAVE see also Rickford et al. (1991), Rickford (1998). Note that copula absence when applied to English also includes uses of auxiliary BE (made explicit in Labov 1969; Rickford et al. 1991).

References Anderwald, Lieselotte 2002 Negation in Non-Standard British English. (Studies in Germanic Linguistics 8.) London/New York: Routledge. 2003 Non-standard English and typological principles: The case of negation. In: Günter Rohdenburg and Britta Mondorf (eds.), Determinants of Linguistic Variation, 507–529. Berlin/New York: Mouton de Gruyter. Aston, Guy and Lou Burnard 1998 The BNC Handbook: Exploring the British National Corpus with SARA. Edinburgh: Edinburgh University Press. Battistella, Edwin L. 1990 Markedness: The Evaluative Superstructure of Language. Albany, NY: State University of New York Press. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad and Edward Finegan 1999 Longman Grammar of Spoken and Written English. Harlow: Longman. Bresnan, Joan 2001 Explaining morphosyntactic competition. In: Mark Baltin and Chris Collins (eds.), Handbook of Contemporary Syntactic Theory, 11–44. Oxford: Blackwell.

66

Lieselotte Anderwald

Chambers, Jack K. 1995 Sociolinguistic Theory: Linguistic Variation and its Social Significance. Oxford: Blackwell. Croft, William 1990 Typology and Universals. Cambridge: Cambridge University Press. Ferguson, Charles A. 1971 Absence of copula and the notion of simplicity: A study of normal speech, baby talk, foreigner talk and pidgins. In: Dell Hymes (ed.), Pidginization and Creolization of Languages, 141–150. Cambridge: Cambridge University Press. 1996 Sociolinguistic Perspectives: Papers on Language in Society, 1959– 1994. Ed. By Thom Huebner. Oxford: Oxford University Press. Francis, W. Nelson 1985 Amn’t I, or the hole in the pattern. In: Wolfgang Viereck (ed.), Focus on: England and Wales, 141–152. (Varieties of English Around the World 4.) Amsterdam/Philadelphia: John Benjamins. Greenberg, Joseph H. 1966 Language Universals. The Hague/Paris: Mouton. Haspelmath, Martin 2002 Understanding Morphology. (Understanding Language.) London: Edward Arnold. Holm, John 1988 Pidgins and Creoles. Vol. I: Theory and Structure. Cambridge: Cambridge University Press. Hudson, Richard 2000 *I amn’t. Language 76: 297–323. Labov, William 1969 Contraction, deletion, and inherent variability of the English copula. Language 45: 715–762. OED 1994 Oxford English Dictionary on CD-ROM. Oxford: Oxford University Press. Orton, Harold and Wilfrid J. Halliday (eds.) 1962–1964 Survey of English Dialects. Vol. 1: The Six Northern Counties and the Isle of Man. Leeds: Arnold. Orton, Harold and Martyn F. Wakelin (eds.) 1967–1968 Survey of English Dialects. Vol. 4: The Southern Counties. Leeds: Arnold. Orton, Harold and Michael V. Barry (eds.) 1969–1971 Survey of English Dialects. Vol. 2: The West Midland Counties. Leeds: Arnold.

Local markedness as a heuristic tool in dialectology

67

Orton, Harold and Philip M. Tilling (eds.) 1969–1971 Survey of English Dialects. Vol. 3: The East Midland Counties and East Anglia. Leeds: Arnold. Paolillo, John C. 2002 Analyzing Linguistic Variation: Statistical Models and Methods. (CSLI Lecture Notes 114.) Stanford, CA: CSLI Publications. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A Comprehensive Grammar of the English Language. Harlow: Longman. Rickford, John R. 1998 The Creole origins of African-American Vernacular English: Evidence from copula absence. In: Salikoko S. Mufwene, John R. Rickford, Guy Bailey and John Baugh (eds.), African-American English, 154–200. London/New York: Routledge. Rickford, John R., Arnetha Ball, Renée Blake, Raina Jackson and Nomi Martin 1991 Rappin on the copula coffin: Theoretical and methodological issues in the analysis of copula variation in African American Vernacular English. Language Variation and Change 3: 103–132. Walker, James A. 2000 Rephrasing the copula: Contraction and zero in Early African American English. In: Shana Poplack and Sali Tagliamonte (eds.), The English History of African American English, 35–72. Oxford: Blackwell. Winter, Werner 1989 Markedness and naturalness. In: Olga Mišeska Tomiü (ed.) Markedness in Synchrony and Diachrony, 103–109. (Trends in Linguistics Studies and Monographs 39.) Berlin: de Gruyter. Wright, Joseph 1898 The English Dialect Dictionary. Oxford: Oxford University Press. Zwicky, Arnold M. and Geoffrey K. Pullum 1983 Cliticization vs. inflection: English n’t. Language 59: 502–513.

Non-standard evidence in syntactic typology – Methodological remarks on the use of dialect data vs spoken language data Peter Auer Abstract If it is right that (written) standard varieties do not always represent the best empirical basis for typology, which non-standard data should we turn to in order to enhance our chances to capture type-representative regularities (consistency) in a more effective way? The present paper addresses this question by discussing dialect data and data from spoken language as two alternative answers. On the basis of examples from German it is suggested that general features of spoken (dialect and non-dialect) syntax need to be distinguished from geographically restricted (dialectal) and geographically non-restricted features of non-standard syntax. While the first offer the best alternative to written standard language in order to capture typological traits of German as a whole, dialects can provide insights into typological syntactic generalisations when taken as linguistic systems in their own right. 1. 2. 3. 4. 5.

Introduction Types of syntactic variation: Oral vs. areal The usability of dialect features for syntactic typology The unity of variable phenomena in syntax Conclusion

1. Introduction It has long been recognised in linguistic typology that standard varieties may not always represent the best empirical basis for cross-language generalisations. Thus, the recent interest in dialect syntax from a typological perspective (cf. Kortmann 1999, 2002) reflects a widespread assumption that non-standard language data display a degree of structural regularity or consistency which is absent from the standard varieties due to explicit regulations and codifications in the latter, which inhibit ‘natural’ language change (cf. Stein 1997). To give a simple example: although

70

Peter Auer

German, English and Dutch started from roughly the same degree of syntheticism in their earliest history, the grammars of Standard English and Standard Dutch have moved on to a much more analytic stage over the centuries than the grammar of Standard German. One likely reason for which the ‘drift’ towards analytic constructions has been impeded in High German is deliberate language planning by intellectuals (Sprachpflege, cf. von Polenz 1994: 254). Arguably, this language planning has not extended to the dialects so that non-standard forms of German can be expected to be more in line with ‘natural’ tendencies than the standard variety (and therefore, in this case, with the drift towards analyticity). And in fact, many dialects of German have moved further toward analyticity than the standard variety (and lost, for instance, the genitive as well as the dative/accusative or accusative/nominative distinction altogether). But this line of reasoning can also be challenged; there can be no doubt that certain dialects exhibit ‘crazy’ structures which are extremely idiosyncratic and unlikely to reflect (at least directly) type-conforming regularities of German in toto. To stick to the example, some dialectal features of German, such as the inflected complementisers of Bavarian or the inflected ‘depictive’ adjectives/participles in Swiss German (and elsewhere; see below, ex. [19]), are even more synthetic than those of the standard. In fact, there is an explicit argument in dialectology (propagated among others by Andersen (1988) and more recently Trudgill, e.g. this volume) according to which the isolation of speech communities leads to ‘unnatural’ (highly marked) grammatical or phonological structures, while 1 levelling results in more ‘natural’ (unmarked) structures. It may therefore be argued (contra the use of dialect data) that it may be more rewarding to look at spoken language in general, including spoken standard languages; arguably, spoken standard varieties are the result of some kind of levelling and may therefore reflect ‘natural’ tendencies more directly than dialects (particularly those which have a very restricted reach). Research on dialect syntax would be replaced by research on spoken vs. written syntax in order to find the proper empirical basis for syntactic typology. Incidentally, the problem does not seem to have existed for our dialectological forefathers. Traditional dialectology assumed a simple equation of dialect and spoken language, and of standard language and written language. As a consequence, linguists included syntactic features into their descriptions of dialect syntax which modern research would regard as typical of spoken language in general, such as prolepsis,

Non-standard evidence in syntactic typology

71

anakoluthons, contaminations, syntactic breakoffs, a tendency to use paratactic instead of hypotactic constructions etc. Equally, the first serious publications on spoken German of the time around 1900 (such as Wunderlich 1894) resorted mainly to dialect speech to make their point. There is a simple explanation for this conflation of medium and areality; this is the empirical fact that at least until the late 19th century, the standard language did not exist in any but a written version for most people in Europe (cf. van Marle 1997). In this sense, the ‘modern’ view according to which geographical (areal) and medial variation (spoken/written) need to be kept distinct as two different dimensions of linguistic variation is nothing but a somewhat belated recognition of the sociolinguistic facts: in particular, of the fact that standard varieties started to be used for oral communication in the 20th century in addition to the dialects. Against the background of these two opposing positions, i.e. dialects as a residue of ‘crazy’ syntax and dialects as the vanguard of natural ‘drift’, the question to be discussed in this paper is this: which non-standard data should we turn to in order to enhance our chances to capture typerepresentative regularities (consistency) absent from the written standard? The model which I would like to propose (and which is realistic at least for the German language area) would concede that although dialect syntax is always spoken syntax, there is a certain number of features of dialect syntax which are not found in the syntax of the spoken standard variety and which therefore cannot be reduced to orality. syntax of standard language = syntax of written language

dialect syntax = syntax of spoken language Figure 1. Traditional view

spoken syntax (standard plus dialect) Figure 2. Radical alternative

written syntax (only standard)

72

Peter Auer syntax of spoken standard

syntax of written standard

medium area

dialect syntax (almost always spoken syntax) Figure 3. Compromise

Dialect syntax is, according to this model, a subset of oral syntax which is defined by a restricted geographical reach. 2. Types of syntactic variation: Oral vs. areal Following this model, we need to distinguish between at least three sources of variation in syntax: Type A: General syntactic features of spoken language, i.e. the structural consequences of orality: Evidence for type A features is that they occur everywhere, in spoken standard German as well as in the dialects. Among them, we would count prolepsis, break-offs or contaminations,. Type B: Geographically restricted syntactic features: Syntactic phenomena which occur in a restricted part of the German language area (neither in the standard variety nor in all dialects) are dialect features sensu stricto. Numerous examples will be discussed below. Type C: Non-dialectal non-standard features : Types A and B do not exhaust the possibilities of variation in syntax. There are syntactic phenomena which are observed in all German dialects but not in the spoken standard. By definition, they cannot be called dialect features since they do not show areal distribution; since they do not occur in spoken standard German, they are therefore non-dialectal non-standard features.

Non-standard evidence in syntactic typology

73

For instance, in a handbook of Low German, Stellmacher (22000: 196– 198) discusses at length the following ‘typical’ syntactic features of this variety: (1)

periphrastic doon ‘to do’ wenn Lüüd nah’n Goddesdeenst nach Huus gahn doot, when people after the mass (to) home (to-)go do denn kümmt dat männigmal vör, dat se ok oever de then comes that sometimes up, that they also about the Predigt sprĊken doot. sermon speak do. ‘when people go home after mass, then it happens sometimes that they talk about the sermon.’

(2)

double negation a. He is keen Buur nich. He is no farmer not. ‘He is no farmer.’ b. He het keen Geld un keen Tüüg un keen He had no money and no thing and no gonnicks nich. at all nothing not. ‘He had no money, not a thing, nothing at all.’

However, if we leave the northernmost dialects of German and look into grammatical descriptions of the southernmost dialects, we encounter the same constructions which now figure as typical of Bavarian or Alemannic. For instance, in Zehetners grammar of Bavarian (1985), we find: (3)

dua-periphrasis (Zehetner 1985: 150–151) B Muadda duad koocha. The mother does cook ‘Mother prepares/is preparing the meal.’

(4)

double negation (Zehetner 1985: 149) nia neamdd ghoiffa. Mia h od To-me has never nobody helped. ‘Nobody ever helped me.’

74

Peter Auer

And for High Alemannic, Weber’s grammar of the Zurich dialect gives the following examples: (5)

(6)

tue-periphrasis (Weber 1948: 249–250) Mer tüend grad z Morgen we do-1.Pl. right-now to-morning ‘We are having breakfast.’

ässe. eat

double negation (Weber 1948: 270) er tuet käm Mäntsche nüüt z läid. he does no one-DAT man-DAT nothing for harm ‘He doesn’t do any harm to any (no-) body.’

In fact, both constructions seem to occur all over Germany, but they are sanctioned in spoken (as well as, of course, in written) standard German and will not be found there, unless as an interference with the dialectal forms. (Abraham’s contention [2002:11] that tun-periphrasis is restricted to Upper and Low German is wrong, has his many examples from Middle German, e.g. [2002: 16-20] clearly show.) In both double negation and tunperiphrasis, there can be no doubt that this absence is due to a conscious process of purification of the language in which it was purged of its supposedly illogical or superfluous aspects, as several studies have shown (see Langer 2001 on periphrastic tun). So what basically seems to be a ‘natural’ tendency is kept from spreading into the most prestigious spoken variant by a deliberate act. Other examples of C-type syntactic variation are the analytic expression 2 of possession as in (7)

analytic possessive construction (fabricated) dem Vater sein Haus the-DAT father his-NOM house ‘father’s house’ 3

or apo koinu (“pivot”) constructions such as (8)

apo koinu construction (fabricated) das hat so geschneit it has so-much snowed ‘it’s been snowing so much today!’

hat das heut! has it today

Non-standard evidence in syntactic typology

75

There can be no doubt that the classification of a given syntactic feature as belonging to type A, B or C can change over time; this is what processes of standardization are made of, but we equally observe processes of destandardisation, and, of course, of geographical spread to the whole of a language area. Whether a given feature is suppressed in the spoken standard or not is clearly an ideological issue in which issues of ‘good’ and ‘pure’ language play a central role. For reasons which are only in part grammar-internal, certain phenomena of spoken German are therefore more subject to sanctions both in writing and speaking than others. There is, however, a sense in which universal features of orality (such as those related to self-repair) seem to be less salient and therefore less exposed to normative pressure than those which are more restricted to German or to the Germanic languages such as do-support or analytic possessives. However, structural considerations cannot explain why, for example, prolepsis (‘left dislocation’) should be less salient and therefore less 4 sanctioned than apo koinu constructions. 3. The usability of dialect features for syntactic typology The discussion so far suggests that it is type B phenomena which are problematic in syntactic typology since they can represent ‘crazy’ structures surviving in isolated communities only (i.e. because nobody needs to accommodate speakers of these varieties). However, type B variation obviously is a gradual phenomenon. There are examples of pervasive type B phenomena which occur in a very large area of German, such as the almost total replacement of the synthetic preterite by the socalled Perfekt tense (the analytic past tense). The area is basically the same in which word-final schwa was apocopated in Early New High German. (Schwa apocope is therefore the traditional explanation for the Upper German loss of the preterite.) Maurer’s somewhat idealised map drawn on the basis of the Deutsche Sprachatlas shows the respective isogloss running 5 north of the river Main. Another large-scale dialect phenomenon is the southern raising of elements of the predicate group in subordinated clauses: (9)

6

verb raising in subordinated clauses a. Std.Germ.: wie wir heim(ge)kommen sind ... no raising: when we home-come have ... ‘when we came home yesterday ...’

76

Peter Auer

b.

Upper German: raising (i): wie when raising (ii): wie when

wir we wir we

sind have heim home

heim(ge)kommen ... home-come ... sind (ge)kommen ... have come ...

These phenomena, to be sure, are unlikely to represent ‘crazy’ rules, given their spread in a large and certainly not peripheral part of the German language area. Other widespread German syntactic dialect phenomena (in descending order of their geographical extension) are the following: (i) relative clauses introduced by the uninflected locative question word wo, which may or may not be supported by a case and number marked pronoun der/die/das etc., found throughout the Middle and Upper German language area (and perhaps elsewhere; cf. Fleischer, MS): (10)

dialectal relativisers a. Alemannic (Freiburg/Br; Günther 1967) di kindhait, woo mer bewussd midèrlèbt hed, ... the childhood which (lit. where) one consciously lived has, ... Std.: die Kindheit, die man bewusst miterlebt hat ... ‘the childhood which we consciously lived through ...’ b. Bavarian dea bua, dea wo da voan gsessn iis ... the boy who (lit that where) there in-front sit has ... Std.: der Bub, der da vorn gesessen ist/hat ... ‘the boy who was sitting there in front ...’

(ii) the Upper and Middle German use of definite article with proper names and kinship terms and the postpositioning of the first name to the family name (which may receive case – genitive – marking): (11)

7

Upper and Middle German family names (fabricated) der Auer(s) Peter the SURNAME (-GEN) FIRSTNAME Std.: Peter Auer

(iii) the Upper and Middle German emphatic double article in constructions of the type DET ADVB DET ADJ (N):

Non-standard evidence in syntactic typology

(12)

77

Upper and Middle German double indefinite articles (here: 8 fabricated Zürich German) e psunders e liebi Frau a particularly a nice woman Std.: eine besonders liebe Frau ‘a particularly nice woman’ 9

(iv) the indefinite article with mass nouns: (13)

High German mass nouns with indefinite article (fabricated) heute kommt noch ein Schnee today comes PART a snow Std: heute kommt noch Schnee ‘there’ll be snow today’

However, syntactic dialect features can also have a more restricted range and coincide by and large with one of the large German dialect areas. For instance, Glaser (1995) has recently suggested that the zero partitive is a 10 typical feature of Alemannic: (14)

Alemannic zero partitive constructions (Glaser 1995: 69) a. Std. German, fabricated dialogue A: ich hätte gern Kirschen – habt ihr welche? have-2.Pl. you some ‘I would like some cherries – do you have any?’ B: nimm Dir welche! take-IMP you-DAT some! ‘take some!’ b. Alemannic (Oberschefflenz), same fabricated dialogue kh(^6( – h(d^ Ø? A: i heet g((^n have-2.Pl. Ø! B: nem d(^(^ take-2.Sg. you-DAT

Topicalisation in subordinated clauses (‘movement’ to pre-complementiser position) seems to be a feature of Bavarian dialects in general (although it occurs in some adjoining areas as well, such as in Thuringia and Middle Franconia):

78

(15)

Peter Auer

Bavarian topicalisation in subordinated clauses (Zehetner 1985: 150) a. with conditional wenn miassn! Da Vadda wann des daleem h(d the father if it live had must ‘If father had lived to see that’ (negative implication) b. with complementizer dass gmoand. An Fümfa dàs-e griag, h(d-e ned a five that-I get, had-I not thought ‘That I would get a ‘five’ (= mark), I would not have thought.’

Finally (and contrary to a widespread belief that syntactic dialect features always have a wide reach and contrast with phonological and morphological features in this respect), there are syntactic dialect features which only occur in relatively small areas, such as the use of the verb geben ‘to give’ as an auxiliary to form passives (instead of Std. German werden) in Southwest Moselle Franconian (16) (cf. Girnth 2000: 137ff) or verb duplication in High Alemannic (17) (cf. Lötscher 1993, Schmidt 2000): (16)

geben passive in Southwest Moselle Franconian (example from a settlers’ variety in Rio Grande do Sul, Brasil) un wie se wolld heirat (.) is de mann umbracht geb. and as she wanted to-marry is the man killed give ‘And when she wanted to marry (him), the man got killed.’

(17)

verb duplication in High Alemannic (here: Luzern; cf. Werlen 1994) a. gönd er go (gen) ärrne? go you go- (to) harvest? ‘Are you going harvesting?’ b. s food scho aafo räägne. it starts already start (to) rain ‘It’s already starting to rain.’

Verb duplication is relatively widespread in High Alemannic for the verb 11 ‘to go’ (17a) but areally very restricted for ‘to begin’ (17b). Other examples of low reach syntactic features are the following: (i) analytic dative marking by locative prepositions such as im or am in parts of the High Alemannic area (central Switzerland and some small areas in

Non-standard evidence in syntactic typology

79

southwestern Germany) – a more analytic construction as compared to the standard language and the regional dialects which is nonetheless 12 geographically isolated: (18)

analytic (prepositional) dative (High Alemannic, ex. from Löffler and Besch 1977; cf. Seiler 2002) im Lehrer > in dem ‘to the teacher’/ ‘teacher DAT’

(ii) the marking of subordination in complement phrases by verb-first syntax (instead of dass/ob as complementisers otherwise), which is also restricted to Swiss High Alemannic and possibly some small areas outside (Lötscher 1997: fn. 8): (19)

verb-first dependent clauses in High Alemannic (Lötscher 1997) ich find s guet, Ø het er chönne choo I find it good, have you can come ‘I think it’s good that you have been able to come’

(iii) reflexive constructions such as in (20) which only occur in a part of Middle Franconian: (20)

Ripuarian reflexive adjunct middles (from Cornips and Corrigan, 13 forthcoming) Der sal sengt sich legt. the hall sings REFL easily ‘This hall has good acoustics (lit. ... sings well).’

In sum, the distinction between Type B (areal phenomena, dialect features) and Type C (pervasive non-standard features) is gradual. For typology this means: the greater the geographical reach of a syntactic feature, the more likely it is to reflect a ‘natural’ (type-relevant) tendency of German. Type B syntax would therefore not be discarded entirely if we look for typologically relevant data on German as a whole, but highly restricted phenomena would be treated with caution as they may represent local idiosyncrasies. Obviously, this does not mean that they are of no interest to general linguists, and even typologists when taken in their own right.

80

Peter Auer

4. The unity of variable phenomena in syntax Up to now, I have used the notion of a ‘syntactic feature’ very loosely to refer to syntactic variability in German. However, the unity of these features is not always easy to establish. In fact, the distinction between my A, B and C features crucially depends on the way in which dialect features are defined. Three problems are particularly important for the relationship between dialect syntax in the narrow sense of the word and the syntax of spoken language. These are variation in meaning, variation in syntactic (including lexical) context (internal syntax) and variation in the syntactic construction itself (external syntax). 4.1. Semantico-pragmatic unity The same syntactic structure may occur in various dialects but receive a different semantico-pragmatic interpretation. Periphrastic tun can illustrate the problem. Although the use of tun-periphrasis as a carrier for the present tense as well as the subjunctive mood (cf. Kortmann, this volume) may be found in all dialects, many handbooks and publications argue that the periphrastic construction is not semantically equivalent to the synthetic 14 construction in other cases (cf. Abraham 2002 for an overview); however, it is not clear, whether the difference in meaning is the same throughout the German-speaking area. This is particularly true for aspect marking and emphasis by periphrasis. Thus, Stellmacher (22000: 198) offers the following function for the periphrastic do-construction in Low German: “Verstärkung einer Aussage” (intensification), and (particularly in northern and eastern Low German) marking of “Aspekt” (present progressive). For Zürich, Weber (1948) also mentions intensification, but additionally lists the mitigating use of the periphrastic construction in questions and commands (cf. his example tüend ietz uffpasse! lit. ‘do pay attention now!’, but better translated as ‘please pay attention now’). According to Merkle (1975: 76) dua-periphrasis in Bavarian also has a mitigating effect in imperatives (duàds need fräch weàn! ‘don’t be cheeky!’); but for Eroms 15 (1984), it also marks progressive aspect. In sum, although there is some overlap, aspectual marking is not claimed for High Alemannic, the mitigating use is not claimed for Low German, and intensification is not claimed for Bavarian. The question arises whether we are dealing with the same phenomenon; structurally, there can be no doubt that we are, but if we

Non-standard evidence in syntactic typology

81

require strict semantic equivalence, we probably are not. From this it follows that the periphrastic structure clearly is of type C (a pervasive nonstandard phenomenon), while its semantics (and pragmatics) may be more regionalised (located more towards type B on the C/B-continuum). Another variable syntactic phenomenon of German which formally is of type A but in which there is semantico-pragmatic variation of the B-type is the use of declarative verb-first syntax in main clauses in spoken German which violates the requirement of standard written German to fill the prefront field (cf. Auer 1992 for details). In spoken standard German just as in the dialects, verb-initial clauses can be of two kinds. In the first case, all obligatory nominal complements of the verb are present but positioned after the finite verb, (21, aĺ). The expletive pronouns es and da, used in standard written German to fill the front field in such a case, are avoided. In the second case, a noun phrase is textually recoverable but is not marked by an anaphoric pronoun, i.e. the construction is ‘elliptical’ (21, bĺ): (21)

[conversational data from Selting 1995] D: aso MEIN: hausarzt hat soFORT gemerkt you see my family doctor immediately noticed dass ich rauche; (-) that I am a smoker; der hat mich ABgehört un hat gesacht he sounded me and he said RAUchen sie? do you smoke? aĺ meint ich JAA, mean-PRET-1SG I yes I said yes, aĺ meint er JA; mean-PRET-3SG he yes he said yes; bĺ HÖRT man. hear-PRES-3SG one one can hear it.

The a-clauses could be turned into standard written German by inverting the order of finite verb and pronoun (ich meinte ‘ja’/er meinte ‘ja’), which is not possible in the b-clause (*man hört) where the object noun is lacking (man hört es).

82

Peter Auer

There are strict pragmatic conditions on the use of these verb-first clauses in spoken German. They occur either in narratives (as in the a-lines of the preceding extract), or they orient backward to the preceding utterance, expressing some kind of stance towards its contents, i.e. commenting, elaborating or evaluating it. The latter is the case in the b-line where the doctor comments on the answer given by the patient in the previous line. However, there are some regional data which contradict this analysis, such as the following: (22)

ĺ

[from Auer 1993] (H. has just announced to his brother M. that he will have to undergo surgery soon) H: (mein nu) GUT; mein (in/de) einer hat de(s) lieber JETZT (I mean) o.k.; I mean ( ) one has it rather now als (–) SPÄter n[e, than (–) later right? M: [als SPÄter ne, than later right? H: des=s RIChtig ne=aber (-) du WEISST ja ne, that’s right isn’t it=but you know don’t you, M: m, H: kommt des Alles zu sonner ZEIT wieder wo=s mir GAR nicht (-) geNEHM is ne, that comes all at a time again when I don’t like it at all right,

This V-initial clause sounds awkward to many speakers of German since it does not express an event-clause in a story and since the speaker introduces a new topical aspect into the conversation, i.e. the utterance is forward oriented. The interesting fact in the framework of the present discussion is that the speakers who seem to have generalised the verb-initial construction to contexts beyond those in which we find them everywhere in German, come from certain restricted areas (particularly the Ruhrgebiet). Although the issue needs further investigation, this could suggest a dialectal extension of a more widespread syntactic construction. Both periphrastic tun and verb-initial clauses show that the areal reach of the syntactic construction itself and that of its semantico-pragmatic interpretation need not coincide. If we want to capture the most general structural tendencies, which may reflect some kind of ‘natural’ tendency,

Non-standard evidence in syntactic typology

83

we would like to be able to group all instances of do-periphrasis or of verbinitial clauses together despite these discrepancies. This seems justified as long as the variable meanings (or pragmatic functions) are related to each other in a plausible way. This holds for the verb-initial clauses, since the Ruhrgebiet variant simply relinquishes the pragmatic constraints valid elsewhere. In the case of the periphrastic tun, things are more difficult. In particular, it is difficult to find a common denominator between the mitigating use of tun and the aspect-marking use. 4.2. Syntactic unity (external syntax) The second problem connected to establishing the unity of a syntactic phenomenon is syntactic variation in the use of a given construction. For instance, periphrastic tun is in most cases clearly a feature of non-standard German (type C), but there are some syntactic contexts in which the construction can/must be used in spoken standard German as well (type A). Cf. the case of the topicalisation (fronting) of infinitives, for instance in contrastive constructions of the type in (23): (23)

tun-periphrasis in (spoken) standard German (fabricated) glauben tu ich’s nicht, ... aber hoffen believe-INF do-1SG I it not ... but hope-INF (tu ich ’s) schon. (do-1SG it) PART ‘I hope it, but I don’t believe it.’

This construction is found in spoken standard German as well as in the dialects. Another example is the so-called Rhenish progressive, which in many ways is not at all a feature of the Ripuarian dialects but rather a type-A feature of spoken German in general (with partial acceptance even in written German). The construction, which consists of the copula verb 16 sein plus the (former?) preposition am followed by the infinitive, fills a gap in the German temporal-aspectual system since it expresses imperfectivity/ progressivity. Thus, a sentence like (24) will clearly be acceptable to every speaker of German:

84

(24)

Peter Auer

‘Rhenish’ progressive (fabricated example) Ich bin grad am (Suppe-)Kochen, ruf doch bitte I am just-now at (soup-)cook-INF, call PART please später nochmal an. later again PREFIX. ‘I’m cooking, can you call me back later please?’

This status does not even change when an indefinite object noun (such as am Suppekochen) is incorporated into the infinitival construction. What is Rhenish then about the ‘Rhenish’ present progressive? It is above all17 the fact that in the Ripuarian dialects, the construction has expanded into syntactic contexts where it cannot be used elsewhere. For instance, most varieties of spoken German do not permit limiting adverbials together with the am-progressive, while the Ripuarian dialects do (cf. Ebert 2000): (25)

i.

The (true) Rhenish progressive, (fabricated example) Anne war zwei Stunden alleine am Spielen. Anne was two hours alone at play-INF ‘Anne was playing alone for two hours.’ 18

More important, the Ripuarian (as well as some other ) dialects allow nonincorporated indefinite and even definite noun phrases to precede the am & infinitive-construction: (25)

ii.

The (true) Rhenish progressive, (fabricated example) ich bin grad die Suppe am Kochen. I am just the soup at cook-INF ‘I’m just cooking the soup.’

Here, a syntactic contextual constraint is relinquished and the construction becomes generalised. For typology, variability of this kind would not seem to be a problem as long as the generalisation can in itself be explained in a satisfactory way. 4.3. Formal unity (internal syntax) In some cases, the question arises as to whether we are dealing with one, two or more syntactic features. For instance, when used with proper nouns,

Non-standard evidence in syntactic typology

85

the analytic possessives of example (7) show internal structural variation between northern Ruth ihr Kleid and southern der Ruth ihr Kleid. In this case, it is justified to treat both under one heading since there is evidence that the definite article before proper names is a syntactic variable in its own right which has nothing to do with the expression of possessivity. Another instance of a dialect feature which may be treated in a unified way although it consists of more than one type of deviation from the standard are the relativisers wo and der wo mentioned above (cf. [13]). On the other hand, there may be cases of structurally closely related dialect phenomena which resist integration into one syntactic variable. A case in point (analysed in detail by Bucheli, MS) may be the depictive markers found in restricted areas of High Alemannic (Appenzell), which can be argued to be a fundamentally different kind of phenomenon from the depictive adjectives/participles which occur in other parts of High Alemannic (Wallis) and seem to be related to copredicative agreement, a remnant of Old High German morphology (as in är ischt alt-e for std. er ist alt ‘he is old’): (27)

depictives in High Alemannic (from: Bucheli, MS) a. depictive markers in Appenzell du moscht d=milch abe waam-e trink-e! you must the-milk(SG) (but) hot-DEPIC drink-INF b. depictive adjectives elsewhere (here: Wallis) dü müoscht d=milch de heiss-i triich-u you must the-milk(f.Sg.) (then) hot (f.Sg.) drink (INF.) meaning in both a and b: ‘you must drink the milk (while it is) hot.’

While both contructions seem to be highly similar, Bucheli (MS) argues that they represent fundamentally different phenomena since in (a), the adjective marked by the depictive suffix does not agree with the ‘governing’ noun (the suffix is formally that of the masculine, while milch is feminine); in the (b)-case, there is agreement (-i is the feminine suffix). Beyond the question of defining syntactic variables, it seems wise to define structural domains of German syntax in which syntactic variation occurs, rather than listing syntactic phenomena individually. In such a way, it may be possible to identify the stable and variable parts of German syntax at large. A good example is the infinitive complement which is introduced by (um) zu in standard German. The construction is avoided in many if not all Upper and Middle German dialects, which points to a

86

Peter Auer

structural domain of high variability. The more preferred structures are those in (28): (28)

infinitival complements in German dialects (data from Russ 1990) a. für .... zu + INF (Moselle-Franconian) 't war eng Hetz fir d'Aarbecht fäerdeg ze kréien it was a hurry for the work ready to get Std.: es war eine Hetze, die Arbeit fertig zu bekommen ‘It was a rush to get the work finished.’ b. zum (> zu dem) + INF (Bavarian) habds vui zum doÃaÃ? have-you much to-the do Std.: habt ihr viel zu tun? ‘Do you have a lot of work (lit: to work)?’ c. mit + INF (Upper Saxonian) wan fängsd n wider an mit aarweedn? when start-you PART again PREFIX with work+INF Std.: wann fängst du denn wieder an zu arbeiten? ‘When do you start to work again?’ d. zero (Alemannic) mir fanget etz aa Ø schaffe we start now PREFIX work-INF Std.: wir fangen jetzt an zu arbeiten ‘We now start to work.’

These four non-standard forms are united by their function of adjoining an infinitive to a main verb. Their geographical distribution is complex; they 19 often co-occur in the same area. It seems nevertheless justified to group these constructions together since they point to a variable position in German syntax. The individual solutions chosen by particular dialects are much more restricted in geographical reach than the phenomenon of nonstandard, zero or prepositional marking of infinitival complement. 5. Conclusion In this paper, I have discussed some of the questions related to choosing data, especially spoken data, for syntactic typology. We started out with the suspicion (or even assumption) that data from the written standard variety

Non-standard evidence in syntactic typology

87

may be non-optimal for assigning language to types since they are at least in part the result of conscious codification over many centuries. Several alternatives have been discussed; in particular, I have raised the issue of oral vs. areal data sources. The idea of dialect data as providing access to more type-representative tendencies of German may be misleading in some cases, since some dialect phenomena represent highly idiosyncratic developments which may be relevant and, indeed, of great value for linguistic theory, but do not contribute to the investigation of German as a whole from a typological perspective. Preference would therefore be given to general features of spoken and/or non-standard language which have a high areal reach. Notes 1.

2. 3.

4. 5.

6.

The empirical evidence for this position is controversial; Bavarian surely would not qualify as a peripheral dialect of a closed network community of dialect speakers. Anderson’s example, the epenthetic consonants (as in Ripuarian, cf. Cologne /ring/ for Std. /rain/ ‘Rhein’, i.e. ‘velarisation’ of nasals) are also problematic in this respect. Here, V2 syntax is only possible in concurrence with prolepsis: während der Klaus, der kommt bestimmt ‘while Claus, he surely comes’. Despite the numerous publications on dialect syntax which list this phenomenon as typical of the respective area (such as, among many others Zehetner 1985: 149–150 for Bavarian), the phenomenon clearly also occurs in spoken standard German. See map 4.77 in Eichhoff (2000). The northern area shows the use of the analytic possessive in spoken northern standard German; in the Low German dialects, the construction would be even more frequent. Cf. Scheutz (1992) for details. The pivot or koinon (so geschneit in ex [8]) is part of the ‘left’ and of the ‘right’ construction. The left construction may be incomplete, the right construction is always complete. The example represents just one particular case, i.e. that of a mirror-image construction. Whether the fourth logically possible type, i.e. syntactic features of spoken standard German only, exists, remains an open question. There are few syntactic innovations in spoken standard German which have no basis in the dialects. Peripherally relevant would seem to be the case of new conjunctive adverbials which become grammaticalised in the standard variety, such as von daher in the sense of deshalb ‘therefore’ or insofern in the sense of deshalb ‘therefore’. They seem to have their basis in spoken registers (not in writing). Certain genres of spoken standard German may also have syntactic features of

88

7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

17.

18. 19.

20. 21.

Peter Auer their own; one example is the use of prolepsis with weak (instead of strong) resuming pronouns which is typical of TV commentators and reporters (die Wahlen, sie [instead of: die] werden bald eine Entscheidung bringen ‘the elections, they will soon bring a decision’). I.e. in Upper German (Oberdeutsch) and parts of the Middle Rhine area; the respective map by Friedr. Maurer is reproduced in König (2001: 163). Again based on the Deutscher Sprachatlas, i.e. Wenker sentence 24; see the same map by Maurer as reproduced in König (2001: 163). A map of this feature’s areal distribution may be found in Eichhoff (2000, map 4.76). The construction is used today in Bavarian and East Franconian, as well as in High Alemannic; cf. Henn-Memmesheimer (1986: 117). Cf. Glaser (1993) and Henn-Memmesheimer (1997) who give examples from Bavarian, East Franconian (Nuremberg) and High Alemannic. According to Henn-Memmesheimer (1997), the zero realisation is also found in the Palatinate area, i.e. outside Alemannic in the adjoining Rheno-Palatian dialect. In the Swiss German examples, the doubled verb form is a shortened form of the full verb, i.e. gönd ĺ go, foot ĺ fo. In the rest of the German-speaking area, the dative is marked synthetically on the determiner: dem Lehrer seine Hose. The exact geographical distribution is not given by the authors; the construction is nowadays hardly accepted any longer. Some publications take this position; cf. Bernhardt (1903); Lötscher (1983: 107–108). There can be no doubt that many of the periphrastic constructions can be explained in purely formal terms. As Weber (1948) points out for High Alemannic, rhythmical considerations can play a role as well. Rohdenburg (2002) even suggests a phonological explanation for Northern Low German. According to Russ (1990), do-periphrasis also occurs in Swabian and in Highest Alemannic to mark progressive aspect, cf. Swab. se deen ässe ‘sie essen gerade’, Bosco Gurin zwei Maanut tiä-wär gar gghein Sunnu ggsee ‘zwei Monate tun wir gar keine Sonne sehen’. I cannot go into the debate here of whether am still is a preposition. In addition, there is a structural difference since the Ripuarian dialect also allows double prepositional marking of the type ich bin grad am die Suppe am Kochen (‘I’m just cooking the soup’). For further details on the Rhenish progressive, cf. Ebert (2000) and Bhatt and Schmidt (1993). Werlen (1994: 70) gives examples such as i bi-n am rüebli rüschte ‘I am about to clean the carrots’ or even i bi-n es huus am baue ‘I am about to build a house’ for the Swiss dialects. Cf. for southwestern Germany map III/1.403 of the South West German dialect atlas (SSA).

Non-standard evidence in syntactic typology

89

References Abraham, Werner 2002 Characteristics of spoken vernaculars: parsing strategies? The case of German. In: Jan Berns and Jaap van Marle (eds.), Present-day Dialectology, 1–34. Berlin/New York: Mouton de Gruyter. Andersen, Henning 1988 Center and periphery: adoption, diffusion and spread. In: Jaþek Fisiak (ed.), Historical Dialectology, 39–83. Berlin/New York: Mouton de Gruyter. Auer, Peter 1993 Zur Verbspitzenstellung im gesprochenen Deutsch. Deutsche Sprache 3: 193–222. Bernhardt, J. 1903 Zur Syntax der gesprochenen Sprache (Ein Versuch). Niederdeutsches Jahrbuch 29: 1–25. Bhatt, Christa and Claudia M. Schmidt 1993 Die am + Infinitiv-Konstruktion im Kölnischen und im umgangssprachlichen Standarddeutschen als Aspekt-Phrasen. In: Werner Abraham and Josef Bayer (eds.), Dialektsyntax, 71–98. Opladen: Westdeutscher Verlag. Bucheli Berger, Claudia MS Copredicative agreement and the development of a copredicative marker in Swiss German dialects. MS, University of Zurich. Cornips, Leonie and Karen Corrigan forthc. Convergence and divergence in grammar. In: Peter Auer, Frans Hinskens and Paul Kerswill (eds.), Dialect Change. Cambridge: CUP. Ebert, Karen 2000 Progressive markers in Germanic languages. In: Östen Dahl (ed.), Tense and Aspect in the Languages of Europe, 605–653. Berlin/New York: Mouton de Gruyter. Eichhoff, Jürgen 2000 Wortatlas der deutschen Umgangssprachen. Vol. 4. Bern/München: K.G. Saur. Eroms, Hans-Werner 1984 Indikativische periphrastische Formen mit doa im Bairischen als Beispiel für latente und virulente Regeln. In: Peter Wiesinger (ed.), Beiträge zur bairischen und ostfränkischen Dialektologie, 123–135. Göppingen: Kümmerle.

90

Peter Auer

Fleischer, Jürg MS Typology of Relative Clauses in German Dialects. Paper presented at the Methods IX Conference, Joensuu 2002. Girnth, Heiko 2000 Untersuchungen zur Theorie der Grammatikalisierung am Beispiel des Westmitteldeutschen. Tübingen: Niemeyer. Glaser, Elvira 1993 Syntaktische Strategien zum Ausdruck von Indefinitheit und Partititivät im Deutschen (Standardsprache und Dialekt). In: Werner Abraham and Josef Bayer (eds.), Dialektsyntax, 99–116. Opladen: Westdeutscher Verlag. 1995 Die syntaktische Nullstelle – eine Kennform des Alemannischen? In: Heinrich Löffler (ed.), Alemannische Dialektforschung, 65–79. Tübingen/Basel: Francke. Günther, Johannes 1967 Die städtische Umgangssprache von Freiburg im Breisgau – eine sprachsoziologische Untersuchung. Diss. Freiburg. Henn-Memmesheimer, Beate 1986 Nonstandardmuster. Tübingen: Niemeyer. 1997 The patterning of nonstandard syntax in German. In: Jenny Cheshire and Dieter Stein (eds.), Taming the Vernacular, 232–249. London: Longman. König, Werner 2001 dtv-Atlas Deutsche Sprache. München: Deutscher TaschenbuchVerlag. Kortmann, Bernd 1999 Typology and dialectology. In: Bernard Caron (ed.), Proceedings of the 16th International Congress of Linguists, Parix 1997. CD-Rom. Amsterdam: Elsevier Science. 2002 New prospects for the study of English dialect syntax: Impetus from syntactic theory and language typology. In: Sjef Barbiers, Leonie Cornips and Susanne van der Kleij (eds.), Syntactic Microvariation, 185–213. Amsterdam: SAND. (http://www. meertens.nl/books/synmic/) Langer, Nils 2001 Linguistic Purism in Action. Berlin/New York: Mouton de Gruyter. Löffler, Heinrich and Werner Besch 1977 Alemannisch. Düsseldorf: Pädagogischer Verlag Schwann. Lötscher, Andreas 1983 Schweizerdeutsch. Frauenfeld: Huber.

Non-standard evidence in syntactic typology 1993

91

Zur Genese der Verbverdopplung bei gaa, choo, laa, aafaa (‘gehen’, ‘kommen’, ‘lassen’, ‘anfangen’) im Schweizerdeutschen. In: Werner Abraham and Josef Bayer (eds.), Dialektsyntax, 180–200. Opladen: Westdeutscher Verlag. 1997 ‘Guet, sind Sie doo’ – Verbstellungsprobleme bei Ergänzungssätzen im Schweizerdeutschen. In: Arno Ruoff and Peter Löffelad (eds.), Syntax und Stilistik der Alltagssprache: Beiträge der 12. Arbeitstagung zur alemannischen Dialektologie; 25.–29. September 1996 in Ellwangen and Jagst, 85–95. Tübingen: Niemeyer. Merkle, Ludwig 1975 Bairische Grammatik. München: Heimeran. Rohdenburg, Günter 2002 Die Umschreibung finiter Verbformen mit DOON ‘TUN’ und die Frikativierung stammauslautender Plosive in nordniederdeutschen Mundarten. Nowele 40: 85–104. Russ, Charles V.J. (ed.) 1990 The Dialects of Modern German: A Linguistic Survey. London: Routledge. Scheutz, Hannes 1992 Apokoinukonstruktionen. Gegenwartssprachliche Erscheinungsformen und Aspekte ihrer historischen Entwicklung. In: Andreas Weiss (ed.), Dialekte im Wandel, 242–264. Göppingen: Kümmerle. Schmidt, Christa 2000 Die Verbverdopplung im Zürichdeutschen. Unpublished M.A.thesis, University of Freiburg, Deutsches Seminar I. Seiler, Guido 2002 Prepositional dative marking in Upper German: A case of syntactic microvariation. In: Sjef Barbiers, Leonie Cornips and Susanne van der Kleij (eds.), Syntactic Microvariation, 243–279. Amsterdam: SAND. Selting, Margret 1995 Prosodie im Gespräch. Tübingen: Niemeyer. Stellmacher, Dieter 2000 Niederdeutsche Sprache. 2nd edition. Berlin: Weidlich. Stein, Dieter 1997 Syntax & Varieties. In: Jenny Cheshire and Dieter Stein (eds.). Taming the Vernacular, 35–50. London: Longman. van Marle, Jaap 1997 Dialect versus standard language: nature versus culture. In: Jenny Cheshire and Dieter Stein (eds.), Taming the Vernacular, 13–34. London: Longman.

92

Peter Auer

von Polenz, Peter 1994 Deutsche Sprachgeschichte vom Spätmittelalter bis zur Gegenwart. Vol 2: 17. und 18. Jahrhundert. Berlin/New York: Mouton de Gruyter. Weber, Albert 1948 Zürichdeutsche Grammatik. Zürich: Schweizer Spiegel Verlag. Werlen, Iwar 1994 Neuere Fragestellungen in der Erforschung der Syntax der deutschen Dialekte. In: Klaus Mattheier and Peter Wiesinger (eds.), Dialektologie des Deutschen, 49–75. Tübingen: Niemeyer. Wunderlich, Hermann 1894 Unsere Umgangssprache in der Eigenart ihrer Satzfügung. Weimar: Felber. Zehetner, Ludwig 1985 Das bairische Dialektbuch. München: Beck.

The typology of motion and posture verbs: A variationist account Raphael Berthele Abstract According to work by Talmy, Slobin and others, satellite-framed (S)-languages tend to express the path of motion in a verb particle and the manner of motion in the verb (jump, run, etc.), whereas verb-framed (V)-languages express the path in the verb (enter, exit) and the manner in an adverbial – if at all. In this paper, this typological distinction will be applied to data from Swiss German (Muotathal dialect), Standard German and French. As expected, French narratives use more path verbs than German narratives. However, there are surprising differences within the S-framed German varieties: Muotathal dialect rarely expresses manner, in some cases even less frequently than French. It is argued that two typological categories are not sufficient once dialects are taken into account. The categories not only need to be refined. It also needs to be recognized that there are additional dimensions (e.g. orality/literacy) which are important factors affecting typological affiliations. 1. 2. 3. 4. 5. 6.

Introduction Typological approaches to motion verbs Data The verb slot: Motion and posture verbs The verb-sister slot: Path and place arguments Conclusions 1

1. Introduction

In the present contribution, issues from the typological research on motion verbs are carried into the domain of variationist/dialectological linguistics. Three sets of parallel data are analyzed and compared here: Swiss German (SG) dialect data (from the Muotathal area in central Switzerland, MU) and Standard High German (SHG) data are compared with standard French (FR) data. The main aim is twofold: first, it will be shown that the assimilation of theoretical frameworks in typology (and – since they are

94

Raphael Berthele

tightly connected – with frameworks in psycholinguistics and cognitive linguistics) can lead to new dialectological insights. As will be shown, the former can in particular shed some new light on ‘hidden’ or at least nonsalient differences between dialects and their corresponding standard languages. Second, this contribution will argue that certain categories and distinctions that have sometimes been taken for granted in typology and cognitive linguistics have to be fundamentally revised if one takes into account dialectological and variationist data. In section 2, some of the main issues of ‘motion verb research’ in typological and cognitive linguistics will be addressed. Section 3 presents the data sample used in this contribution, as well as some necessary notes on the sociolinguistics of Swiss German. Section 4 then singles out the verb slot and shows different usage patterns of the ‘manner’ and ‘displacement’ verbs in the varieties used in this paper. In section 5, the path and place arguments in the verb-sister position are compared, followed by a demonstration of the sometimes striking differences in the linguistic overtness of spatial information. These analyses, finally, are taken as the basis of a detailed discussion of the typological assumptions and what, in the case of this particular domain, dialectologists and typologists can learn from each other (section 6). 2. Typological approaches to motion verbs As Tesnière (1959: 307–308) has pointed out, verbs used in constructions referring to moving objects in space can be categorized into two groups, which he calls movement and displacement. The former are verbs such as French marcher, courir, trotter, galoper, sauter, sautiller, ramper, voler, nager; i.e. verbs which, according to Tesnière, express the somatic conditions of the moving object. Displacement verbs, on the other hand, express concepts tied to the spatial ground on which the motion takes place (examples: monter, descendre, aller, venir, entrer, sortir). It has long been observed that the languages of the world do not all pay the same amount of attention to these two types of verbs. Among the scholars working on this phenomenon, Talmy (1985, 2000b: 213–288) has probably been the most influential. Slobin in particular (1991; 1996; forthcoming]; Slobin and Bocaz 1988) has used Talmy’s framework for the crosslinguistic comparison of spatial expressions (mainly motion verbs) and their acquisition. Given the considerable resonance of Talmy’s and Slobin’s

The typology of motion and posture verbs

95

work, this paper will use Talmy’s conceptual and terminological framework in order to analyze dialectal and standard language data. It is thus necessary to provide a short introduction into the basic categories and distinctions made by Talmy and the scholars who build upon his ideas. According to Talmy, the conceptualization of a motion event (and, as we will see in a moment, of static spatial settings as well) can be analyzed as integrating a series of different conceptual components. In Figure 1, two instances of locating expressions and their conceptual components are presented. [

figure

activating process

association function

transition fixity the cup the linguist

is is running

ground ]

core schema on into

the table the library

Figure 1. Conceptual structure of a locating expression (adapted from Talmy 2000b: 221)

The entity on which the attention in discourse is predominantly centered is called the figure. This is the object or being which is going to be located or traced in space. The figure is typically expressed in the subject NP of a spatial expression, as illustrated in Figure 1. The second conceptual part of a motion event is the activating process. It can have two distinct parameter settings, either transition or fixity. Despite the fact that Talmy himself uses the term ‘motion events’ (2000b: 226) for all instances covered by the conceptual content rendered in Figure 1, I consider it to be more appropriate to use a more generic and less ‘dynamic’ term, namely spatial localization event. The activating process is generally mapped onto the verb, if the conceptual content is linguistically expressed. The third conceptual entity in Talmy’s framework is the association function, which relates the figure to its spatial environment. In the examples in Figure 1, the association function is expressed in the preposition (P) of the linguistic expression. The last entity, the ground, is (part of) the spatial environment in respect to which the figure is either statically located or moving. Since the conceptual core of any spatial localization event is by

96

Raphael Berthele

definition the localization of a figure with respect to the ground, the association function together with the ground can be seen as the core schema of the whole conceptualization. In some cases, however, the association function alone can be seen as the core schema, particularly in utterances which do not openly specify the ground (e.g. The boy fell down [from the tree].) As we have seen in Figure 1, the verb carries the concept of transition/fixity. In the second example in Figure 1, however, the verb not only encodes motion, but it adds information about the nature of the motion. The verb run thus is an instance of Tesnière’s mouvement. Talmy analyses these verbs as integrating two components, namely [motion] and [manner (of motion)]. This is one instance of the more general phenomenon of integrating more than one event into a single clause (Talmy 2000b: 214– 216). The other type of event integration which is important for spatial expressions is the integration of motion and path, i.e. the co-integration of the association function into the verb. Figure 2 depicts the enhanced mapping of conceptual content into syntactic constituents.

[ figure

activating association ground process function (path)

the linguist is running into

[ figure

Åsupport relation

framing event

[event]

co-event

(manner of motion)

the library

activating association ground process function

the linguist is entering

]

]

the library

Figure 2. Co-event (=manner) integration and path verbs

In the first example in Figure 2, information about the manner of motion (an example of what Talmy calls a co-event) is mapped onto the verb. In the second example, the association function (=the path) is mapped onto the verb. According to Talmy, the languages of the world show different patterns of conceptual mappings in the verb slot. For the present purposes, it is sufficient to distinguish two groups: either the languages tend to integrate a co-event in the verb slot, or they tend to integrate the association

The typology of motion and posture verbs

97

function into the verb slot (Talmy 2000b: 223). If the association function is not mapped to the verb slot, it is expressed in a constituent which stands in a sister position to the verb: either a satellite or a PP. A satellite is a particle such as the German separable and inseparable verb-prefixes (“Präfixe und Halbpräfixe”, cf. Eisenberg et al. 1998: 435). If, in a VP, a satellite occurs alone, there is no explicit ground element in the surface structure. The ground then can easily be inferred from the (textual) context. Depending on this difference of lexicalization patterns, Talmy proposes to distinguish between satellite-framed (S-) languages and verb-framed (V-) languages: Languages that characteristically map the core schema into the verb will be said to have a framing verb and to be verb-framed languages. Included among such languages are Romance, Semitic, [...]. On the other hand, languages that characteristically map the core schema onto the satellite will be said to have a framing satellite and to be satellite-framed languages. (Talmy 2000b: 222)

According to Talmy and Slobin (1991, 1996), there are important differences between S- and V-languages in the realm of motion events. Slanguages (e.g. English, German) express the core schema of the spatial relation, generally the path of a moving entity, in a particle, whereas Vlanguages (e.g. Spanish, French) map the core schema onto the verb. Since the verb slot in S-languages is available for additional information, they tend to integrate co-events in their expressions, mainly the manner of motion of the figure. As Slobin (1996: 218) shows convincingly, these lexicalization patterns contribute to typical “typologies of rhetoric”, i.e. the speakers of a particular language tend to tell different stories, depending on whether their language is of the S- or V-type. As an example, Slobin (1996) has been able to show that English displays a higher degree of mannersaliency than Spanish, a V-language: Spanish narratives use many fewer manner-specifying verbs when describing moving figures in space; the verb inventory as a whole is smaller than the one used by English speakers. The English and Spanish examples (1) and (2) represent this difference; both describe exactly the same picture from the Frog Story (Mayer 1969; examples are taken from Slobin, forthcoming): (1) (2)

the boy tumbles down from the branch y el niño se cae

98

Raphael Berthele

As Berman and Slobin (1994) point out, this difference seems to apply to S- and V-languages in general: The satellite-framed languages in our sample also tend towards greater specification of manner, probably because the lexicon provides a large collection of verbs that conflate manner with change of location (crawl, swoop, tumble, etc.), often conflating cause as well (dump, hurl, shove, etc.). In verb-framed languages, such elaboration is more of a “luxury”, since path and manner are elaborated in separate expressions, which are generally optional, and which are less compact in form [e.g., ‘exit flying (from the hole)’ vs. ‘fly out (of the hole)’]. As a consequence of these differences, it seems – at least in our data – that English and German narrations are characterized by a great deal of dynamic path and manner description, while Spanish, Hebrew, and Turkish narrations are less elaborated in this regard, but are often more elaborated in description of locations of protagonists and objects and of endstates of motion. (Berman and Slobin 1994: 118–119)

There is no doubt that, rather than applying a rigid binary distinction, the languages of the world can be positioned on a cline of manner saliency, with (proto)typical V-languages on one end and typical S-languages on the other end of the scale (Slobin, forthcoming). The list of languages which are analyzed using the ‘Talmy-Slobin’ paradigm is considerable (cf. the extensive list of Frog Story studies in the Appendix of Strömqvist and Verhoeven, forthcoming). However, I am not aware of systematic research on the relationship between dialects and their corresponding standard languages. And this, of course, is what the analysis in the present contribution aims at. Before we can proceed to the analysis of the French and German data, one additional theoretical approach has to be discussed briefly. In his (2001) paper, Wälchli proposes a more fine-grained typology of motion verbs; more precisely, he focuses on a typology of displacement, in the sense introduced by Tesnière (see above). First, we can distinguish three different ‘loci’ of the expression of displacement: “(V) Verbal encoding (i.e. by the verb stem), (AN) adnominal encoding (i.e. by prepositions, postpositions or case marking), and (AV) adverbal encoding (i.e. by verb affixes or verb particles).” (Wälchli 2001: 301) Wälchli’s verbal encoding is thus the predominant lexicalization pattern in typical V-languages, and the two other categories (AN, AV) are the loci of the core schema in S-languages. An innovative aspect of Wälchli’s proposal is his subcategorization of the different types of path. He proposes

The typology of motion and posture verbs

99

6 “cardinal kinds of displacement”, which he labels with the help of Latin prepositions: AD, IN, SUPER, AB, EX, DE (cf. Table 1). Based on a corpus of parallel texts (the Gospel according to Mark), 40 languages are analyzed, and the coding of path in what Wälchli calls “basic intransitive motion” (2001: 300), i.e. motion “of humans (not animals or objects) moving without special haste and without vehicles”, is analyzed. One main result of this study is a cross-tabulation of the languages in Wälchli’s sample, based on the number of path distinctions made in the adnominal and in the verbal slot. Typical S-languages are those which show up to 6 adnominal distinctions (e.g. Western Apache, Russian), whereas typical Vlanguages offer up to 6 verbal distinctions but only very few adnominal distinctions (e.g. Tagalog, Kiswahili). This cross-tabulation allows for a two-dimensional representation of the path expression, and in particular it shows that some languages employ both verbal and adnominal encoding of the path (Udmurt, Erzya Mordvin). Whereas this representation allows us to draw a more adequate picture of the locus of expression of path, it is – and Wälchli is, of course, perfectly aware of that – based on a problematic text corpus and focuses on only one particular type of displacement. Wälchli's list of cardinal paths proves to be helpful not only for the typological comparison of a great number of genetically different languages, but also for our present purposes. However, there are two aspects in Wälchli's list of cardinal paths which need some modification in order to match the needs of the present analysis. Firstly, in Wächli's list, displacement along the vertical axis is conflated with a specified movement away or towards a ground element. E.g., as Table 1 shows, downward movement coincides with movement away from the source. This association is an unnecessary constraint for the analysis of the data here, since a motion verb coding a downward path can open a slot for either the source (3) or the goal (4) or even both source and ground objects (5), as the French examples below show clearly: (3)

le garçon tombe de l’arbre ‘the boy falls from the tree’

(4)

le garçon tombe sur le sol ‘the boy falls onto the ground’

(5)

le garçon tombe de l’arbre sur le sol ‘the boy falls from the tree onto the ground’

100

Raphael Berthele

Thus, I propose to categorize the cardinal paths in a more general manner, not specifying for source or goal in the vertical dimension. Secondly, Wälchli associates deixis and displacement, as becomes manifest in the use of the verbs 'come' and 'go' (cf. Table 1). I propose to disregard deixis as well, since the deictic center taken depends on the deictic stance taken in a particular narrative. For example, it is perfectly possible to have a path such as “F come into G”. With these two modifications, the list of cardinal paths can be set up as in Table 1. Table 1. Cardinal kinds of displacement AD IN SUPER AB EX DE

Wälchli F go to G F go into G F go onto G F come from G F come out of G F come down from G

modified (RB) F displace to G F displace into G F displace up F displace away from G F displace out of G F displace down

Wälchli’s idea, together with the core notions from the theoretical and empirical work by Talmy and Slobin, provide the framework within which the dialectological data in the remainder of this contribution will be analyzed. After a short account of the data used here, we will first discuss the lexical patterns found in the verb slot (Wälchli’s verbal coding of displacement), then those in the adverbal and the adnominal slot. 3. Data The data have been collected using two well-tested tools. The Frog Story (Mayer 1969) was used in order to elicit reference to motion events. Mayer’s picture book consists of 24 panels which show the trajectory of a little boy and his dog searching for their pet frog. The story is well suited for eliciting all kinds of motion in space, since both the boy and the dog run across the forest, fall into a ditch, climb trees, etc. The informants (cf. Table 2) are asked to study the picture book briefly, in order to get acquainted with the plot of the story. Subsequently, they are asked to retell the story, backed up by the picture book. This retelling is recorded and later transcribed and coded for the phenomena discussed here. Note that Talmy’s framework does not only apply to motion events. Expressions referring to

The typology of motion and posture verbs

101

static spatial relations underlie very similar constraints, and they are thus included where it is appropriate. Static spatial expressions are elicited by using another picture book which has been developed by the Max Planck Institute for Psycholinguistics in Nijmegen (cf. Bowerman 1996). This consists of 71 drawings, each of which has an object (the figure) marked by an arrow. The informants are asked to describe where this figure is located with respect to its physical environment (ground). This task is relatively monotonous, and it is sometimes hard both for the informant as well as for the dialectologist to go through the whole picture book, particularly with older informants who are often unable to see the point of this kind of task. Table 2. Number of informants per variety for the two data series Variety Standard High German (SHG) Muotathal Dialect (MU) French (FR)

Frog Story 10 8 10

Static Relations 5 6 5

The speakers of Standard High German in the sample are all Germans, from different areas of the country. Some of them have been living in Switzerland for some time. Since in Germany and especially in Switzerland the dialect–standard situation must be taken into account, all SHG informants have been chosen on the basis of two criteria: first, they must not have a German dialect as their first language (only a very slight accent was allowed), and second – for those living in Switzerland – they must not 2 speak a SHG which is markedly mixed with Swiss German features. The Swiss German dialects are represented here by the Muotathal (MU) dialect. The Muotathal is located in the central part of Switzerland, in the canton of Schwyz. The valley is not situated particularly high up in the mountains (altitude: 610 m. above sea level), nor is it particularly inaccessible, but nevertheless it is a place with a high degree of linguistic and cultural idiosyncrasy. The population is characterized by relatively dense, close-knit and multiplex social networks, which can be easily inferred from the small set of last names borne by a large portion of the native population. The natives are very proud of their area, and they show great loyalty to their valley. The Muotathal dialect has been chosen here because of its relatively traditional language-ecology conditions. Since I am interested in factors such as the conceptional orality vs. literacy of a variety (cf. section 6), this very traditional area of German-speaking Switzerland

102

Raphael Berthele

with relatively low levels of migration, higher education, and valleyexternal ties is most likely to offer a contrast to the Standard language. Some remarks on the Swiss German language situation are necessary to understand the categories in Table 2 (see also Haas 2000: 57–138 and Rash 1998 for a general characterization of the SG situation). The most important point here is that the Alemannic dialects spoken in Switzerland are the first language of virtually all natives, and that there is no ‘standard’ Swiss German dialect. With rare exceptions (basically internal migrants who mix SG dialects), every Swiss German speaker speaks some clearly regionally marked variety. Although a certain leveling of the small-scale dialect features can be observed, there is no tendency towards a Swiss German koiné, as the study by Christen (1998) has shown clearly (assertions to the contrary, namely in German sociolinguistic publications [cf. e.g. Dittmar 1997: 80], are based on traditional misunderstandings and projections of the German situation onto the Swiss context). Although the term diglossia certainly has some problematic aspects, the Swiss German situation still resembles very much the picture given in Ferguson (1959). The Swiss case of diglossia is characterized by functional differentiation – SG is used orally and SHG in all formal written contexts. The main domains of oral use of SHG in German-speaking Switzerland are as the language of instruction (although, once again, not categorically) and the language of some programs in the electronic media (news, movies). This medially constrained diglossia is an important aspect in the attempt to explain the variation patterns shown in sections 4 and 5 below. Since one of the points of the present contribution is to show that Slanguages (such as SG) do not always behave the way typologists would expect them to, French has been added to the data sample in order to compare the two S-language varieties of German with a V-language. The speakers of French in the sample are all native speakers of French. They are all natives of the French-speaking part of Switzerland, although some of them live or have lived abroad (in France, in the US). Since most of the features which differentiate Swiss French from Parisian French are lexical and prosodic, they do not affect our data. Pleonastic constructions such as monter en-haut ‘climb up’ instead of the simple use of the path verb monter, which can be observed in colloquial French in Switzerland and 3 elsewhere, do not occur at all in our data. Since the data under investigation here are part of an ongoing project on spatial language in German and its neighboring Romance varieties, the number of informants is not the same for all varieties. However, given the

The typology of motion and posture verbs

103

tradition both in typology and dialectology of taking one single text or informant to represent a dialect or language, the sample here is nevertheless more ‘representative’ than what has been the rule so far. One of the aims of the aforementioned larger-scale project on spatial language will be to investigate social stratification of variation phenomena such as those discussed here. For the sake of brevity, however, sociological variables such as gender, age, and education have to be disregarded in the present analysis. 4. The verb slot: Motion and posture verbs As discussed in section 2, one of the major typological issues in the realm of motion verbs is the difference in the use of path verbs and manner verbs. In this section, the verb slot in the Frog Story retellings will be analyzed. The three varieties/languages used in this contribution will be juxtaposed and compared, using the categories established both by Talmy and by Wälchli. In the second part of the section, the verb slot of the static data will be discussed and compared in a similar way. 4.1. Motion verbs The Frog Story data have been analyzed with respect to the way the verb slot of a motion event is filled. Whenever a figure changes its location in the retelling, a more or less complex construction referring to movement (=activating process), path (=association function) and – optionally – ground (either the source of the trajectory, a milestone on the trajectory, or the goal) and maybe manner of motion is used. Figure 3 shows the proportion of path verbs relative to the total number of motion verbs, grouped by the three languages/varieties. The non-shaded parts of the columns represent the percentage of verbs which do not code any path information.

104

Raphael Berthele 100

80

no path

60

% of total motion verb use

non-cardinal paths 40

SUPER EX

20

DE AD AB

0 SHG

MU

FR

Figure 3. Cardinal kinds of displacement in the three linguistic varieties (see Table 1)

As typologists expect and predict, speakers of French, on the whole, use more path verbs than speakers of German. The difference is considerable both when comparing French with the Muotathal data as well as with 4 SHG. Additionally, the figure not only shows global differences of path verb usage, but also displays the exact locus of this difference: whereas French has a path verb expressing EX paths (sortir), German does not. French also has a path verb for IN paths (entrer), however, due to the particular features of the plot, entrer does not show up at all in the data. Another important difference is the very frequent use in French of DE verbs (mainly the verb tomber), whereas German fallen and Muotathal dialect kchiie are used less frequently. The French data also display more usage of SUPER verbs (monter); interestingly, this category of path verbs is almost negligible in the Muotathal data. If we compare the path verbs of the Muotathal dialect and SHG, we discover a very frequent use of AD verbs in the former, which is basically due to the high frequency of the verb cho (SHG kommen [=‘come’]), as in 5 examples (6) and (7).

The typology of motion and posture verbs

(6)

dä chömmeds änefür wider a öppis then come-they yonder again at something ‘then again they come to something’

(7)

und obe isch ejä de Uhu cho and above is yes the owl come ‘and above the owl appeared’

105

aane towards

It is at least questionable whether the verb cho in all instances represents a path verb. Whereas in some examples it clearly has the semantics of a path leading towards a ground object (as in example 6), in others (example 7) the verb simply means ‘to appear’ or ‘to move’. This non-path usage of cho~kommen is not impossible in SHG, but it is not nearly as frequent as in the Muotathal dialect. Since the line between path and non-path usage of cho is not easily drawn, all instances of cho have nevertheless been tagged as path verbs. It thus could be argued that for the Muotathal dialect the proportion of the AD path verbs in Figure 3 could shrink to a value close to the SHG value. This would even increase the difference between SHG and the MU in terms of overall path verb usage. Whereas Figure 3 is an account of the motion verb usage in the data, Figure 4 renders the degree of co-event integration in the verb slot. As we have seen in section 2 above, the ‘standard’ co-event in the realm of motion verbs is the manner of motion, i.e. additional information about the modality of displacement. Examples (8) to (10) contain instances of manner verbs from all three languages/varieties. (8)

schlich er sich leise heimlich aus seinem Terrarium sneaked he himself silently secretly out-of his+DAT terrarium ‘he sneaked silently out of his terrarium’

(9)

dass de gugger da so umeschwirrt that the cuckoo there [] around-whirrs ‘that the cuckoo whirrs around there’

(10)

il saute sur la ruche ‘he jumps onto the beehive’

Again, just as in the case of the motion verbs, there is no doubt that all varieties do have manner-of-motion verbs. The question here is how

106

Raphael Berthele

frequent they use them. Figure 4 displays the frequencies of manner verb usage, again grouped by the three varieties. 60

50

40

% of manner verbs

30

20

10

0 N=

10

8

10

SHG

MU

FR

Figure 4. Proportion of manner-of-motion verb usage in SHG, MU, and FR

As Figure 4 shows clearly, the usage of manner verbs varies considerably. Boxes represent the middle 50% of the values, the median (dark line) as the best representative of the distribution, and the vertical lines indicate the scattering of the whole sample. The box plot was chosen as the best chart type since it shows medians and distributions at the same time. Every speaker of SHG, MU, and FR is treated as an individual case with his/her score for manner verb use. The median, as the best representative of a distribution of ordinal-scale values, shows clearly that FR has the lowest values for manner verb usage, confirming the typologists’ expectation. For SHG, on the other hand, the plot reflects a considerably higher use of manner verbs, with very little overlap with the FR data. The speaker of SHG with the lowest use of manner verbs uses about as many manner verb tokens as the speaker of FR with the highest scores. MU, in the middle of the chart, shows a peculiar distribution. First, the scattering is very wide, with the highest values close to the highest value of SHG and the lowest value even lower than the lowest FR values. Second, the median is much

The typology of motion and posture verbs

107

lower than that of SHG, and the MU sample as a whole is statistically significantly different from the SHG sample (independent sample t-test, p [µ]. In these dialects, we see alternations such as the following (Kondosopoulos 1988; Mansfield and Trudgill 1994): (3)

kali kale

[kali] [kale]

but kala kalo kalous

[kaµa] [kaµo] [kaµus]

In fact, of course we are rather used to changes which involve switching between /l/ and /r/. However, it is most unusual for the /r/ to be of this type. Languages which have a retroflex approximant are rare enough. Only 15 of

The impact of language contact and social structure

445

the 317 languages cited in Maddieson (1984: 245) have such an articulation, i.e. 4.7%. And languages which have it as an allophone of /l/ are presumably even rarer. 5.2. Sound systems So far we have looked at unusual sound changes, but I would suggest that the same will also hold true for what we can perhaps refer to as unusual sound systems. Although relatively little work seems to have been done on the vowel systems of Greek dialects, no doubt because many of them have the apparently uninteresting five-vowel /i, e, a, o, u/ system, we can observe something strange about the vowel system of the Cretan dialect of Sfakia. Typically, 5-vowel systems are very stable and very common: 31% of the world’s languages have such systems (Maddieson 1984: 127). They also appear to make maximum usage of available vowel space, and one is not surprised when such a system, in order to maximise distinctiveness and naturalness, consists, as the Standard Greek system does, of a close front unrounded vowel, a close back rounded vowel, and open central vowel, and, in between and equidistant from these, a mid-front unrounded vowel and a mid-back rounded vowel. The Sfakian dialect is not at all like this. The high vowels /i/ and /u/ are indeed where we would expect them to be. The low vowel /a/, however, is very back [$]. And the mid vowels are actually much closer than mid, approximately [e, o]. Large areas of phonetic space in the vowel trapezium thus go unused, as it were. Most remarkable, however, is something which has happened in Cyprus, and which, I again venture to suggest, is typical of the sorts of changes which tend to occur in peripheral communities. As I have already mentioned, Cypriot Greek and other southeastern dialects have preserved Ancient Greek geminates, as in /filla/ ‘leaves’. However, this is only half the story. As is well-known, following the work of Brian Newton (1968), modern Cypriot Greek has also in the intervening period acquired geminates from other sources. Firstly, there are geminates which result from borrowings from Italian and Turkish, which also have geminates, in items such as /kappellos/ from Italian cappello. Then there are geminates which result from assimilation, such as /niffi/ ‘bride’ from earlier /nimfi/. Then, famously and somewhat more mysteriously, are the cases which Newton labels “spontaneous gemination”, where for reasons which are not

446

Peter Trudgill

entirely understood, although Newton goes some way towards explaining what has happened, single consonants have turned into geminates, as in /otti/ ‘whatever’ from earlier /oti/. Newton points out that spontaneous gemination has occurred in other languages as well. However, the most remarkable thing about Cypriot Greek is that, unlike Ancient Greek, it also has word-initial geminates: in word-initial position, single and geminate consonants are in opposition. Word-initial geminates in Cypriot Greek may be the result historically of spontaneous gemination, as in /nne/ ‘yes’, from earlier /ne/; or of assimilation, as in /66illos/ ‘dog’ from earlier /skilos/. They also occur in loans from Turkish, as in /ppullin/ ‘stamp’ from Turkish pul – which is also rather mysterious since Turkish does not have word-initial geminates. And they also occur in loans from English, as in tennis, which is /ttenis/ in Cypriot Greek. (This is presumably because the aspiration of word-initial voiceless plosives in English is interpreted as a sign of gemination – see below.) Note, however, that since English and Turkish do not have wordinitial geminates, we can be sure that these loans must have post-dated the development of such geminates in Cypriot, in order for the borrowings to have taken this form. We cannot say that word-initial geminates occur in Cypriot Greek as a result of borrowings from English and Turkish. Whatever the source of these geminates, however, the fact remains that word-initial single and geminate consonants contrast, as in /ppefti/ ‘he falls’ versus /pefti/ ‘Thursday’. Of these geminates, particularly remarkable are the word-initial geminate stops of Cypriot Greek. Geminates are rather rare in the world’s languages. Maddieson (1984) lists 19 of the 317 sample languages in his data base as having long consonants, i.e. only 6%. Phonotactics are not dealt with in Maddieson’s book, but I believe that it is reasonable to suppose that only a small minority of this 6% of languages with geminates will have them in word-initial position. And an even smaller number will have word-initial geminate stops. It is not difficult to see why this would be. Only when a long consonant is potentially heterosyllabic, and when the length distinction occurs intramorphemically, can such a consonant be classed as a geminate (see Catford 1977: 210– 211), but the main phonetic reason for the rarity of word-initial geminates will be the problem of signalling a length difference on initial consonants, especially stops, and most especially voiceless stops. According to Abramson (1987), for example, in Pattani Malay the “length” distinction between word-initial voiceless stops is actually not a length distinction at all but is maintained by differences in the relative amplitude of the

The impact of language contact and social structure

447

following vowel. So it is in fact not at all surprising that the phonetics and phonology literature cites only a small number of languages where genuine word-initial geminate stops are known to occur (see Hume, Muller and van Engelenhoven 1997; Davis 1999). One phonetically well-established case is that of LuGanda, a Bantu language of Uganda which has a true single/geminate contrast (see Butcher forthcoming). Voiced geminates in this language occur in syllable-initial position mainly as the result of the historical loss of an intervening high vowel. Butcher’s data show that the initial geminates are about 60% longer than the singles. His conclusion is that the difference between the two is achieved through differences in both peak pressure and stricture duration. As far as Cypriot Greek is concerned, Arvaniti has shown that word-initial geminate stops are distinguished from single consonants partly by aspiration but that they are also, crucially, genuinely and substantially longer than singletons (Arvaniti 1999; Arvaniti and Tserdanelis, forthcoming; Tserdanelis and Arvaniti, forthcoming). We are thus on sure ground when we maintain that the situation of Cypriot Greek as regards initial geminates is very unusual universally, and extraordinarily unusual amongst European languages. The proportion of the world’s languages with a contrast between word-initial geminate and single stops in monomorphemic words must be infinitesimally small. Even if there are, say, 30 such languages, this will be less than 1%. My suggestion is that it may not be a coincidence that the variety of Greek which has this unusual feature is spoken on an island at the geographical periphery of the Greekspeaking world. 6. Conclusion There is evidence from linguistic varieties around the world, including Greek dialects, to suggest that the distribution of structural characteristics over the world’s languages may not be entirely random from a sociolinguistic point of view. For example, we have seen indications that long-term contact involving child bilingualism may lead to increased complexity, including redundancy. Conversely, contact involving adult second language acquisition may lead to increased simplification. Furthermore, communities with dense, tightly-knit social networks may be more likely to demonstrate fast-speech phenomena and the consequences of this, and more likely to experience unusual sound changes. I would like to suggest, moreover, that insights of this type can complement research in

448

Peter Trudgill

linguistic typology by giving an explanatory edge to the findings of this discipline. And I would also suggest that these insights should give some sense of urgency to typological research: if it is true that certain sorts of linguistic structure are to be found more frequently, or possible only, in dialects spoken in smaller and more isolated communities, then we had better research these types of community as rapidly as we can while they still exist. Acknowledgements I am very grateful for comments and information I have received on issues discussed in this paper to Amalia Arvaniti, Andy Butcher, Eric Hamp, Mark Janse, Brian Joseph, Paul Kiparsky and Jennifer Muller.

References Abramson, Arthur S. 1987 Word-initial consonant length in Pattani Malay. Haskins Laboratories Status Report on Speech Research SR-91: 143–147. Andersen, Henning 1988 Center and periphery: Adoption, diffusion, and spread. In: Jacek Fisiak (ed.), Historical Dialectology: Regional and Social, 39–84. Berlin/New York: Mouton de Gruyter. Arvaniti, Amalia 1999 Effects of speaking rate on the timing of single and geminate sonorants. In: Proceedings of the XIVth International Congress of Phonetic Sciences, San Francisco: 599–602. forthc. Comparing the phonetics of single and geminate consonants in Cypriot and Standard Greek. In: Proceedings of the Fourth International Conference on Greek Linguistics, Nicosia, 17–19 September 1999. Arvaniti, Amalia and Georgios Tserdanelis forthc. On the phonetics of geminates: Evidence from Cypriot Greek. In: Proceedings of International Conference on Spoken Language Processing (ICSLP2000), Beijing, 17–20 October 2000. Butcher, Andrew R. forthc. Fortis/lenis revisited one more time: The aerodynamics of some oral stop contrasts in Europe, Australia, and in between. Clinical Linguistics and Phonetics.

The impact of language contact and social structure

449

Catford, John C. 1977 Fundamental Problems in Phonetics. Edinburgh: Edinburgh University Press. Chambers, Jack K. 1995 Sociolinguistic Theory. Oxford: Blackwell. Davis, Stuart 1999 On the representation of initial geminates. Phonology 16: 93–104. Dawkins, Richard M. 1916 Modern Greek in Asia Minor: A Study of the Dialects of Silli, Cappadocia and Pharasa with Grammar, Texts, Translations and Glossary. Cambridge: Cambridge University Press. Dressler, Wolfgang U. 1984 Explaining natural phonology. Phonology Yearbook 1: 29–51. Drettas, Georges 1997 Aspects Pontiques. Paris: Association de Recherches Pluridisciplinaires. Garman, Michael 1990 Psycholinguistics. Cambridge: CUP. Grace, George 1990 The ‘aberrant’ (vs. ‘exemplary’) Melanesian languages. In: Philip Baldi (ed.), Linguistic Change and Reconstruction Methodology, 155–173. Berlin/New York: Mouton de Gruyter. Harris, John 1990 Segmental complexity and phonological government. Phonology 7.2: 255–300. Holton, David, Peter Mackridge and Irene Philipaki-Warburton 1997 Greek: A Comprehensive Grammar of the Modern Language. London: Routledge. Horrocks, Geoffrey 1997 Greek: A History of the Language and Its Speakers. London: Longman. Hume, Elizabeth, Jennifer Muller and Aone van Engelenhoven 1997 Non-moraic geminates in Letti. Phonology 14: 371–402. Janse, Mark 1998 Grammaticalization and typological change: The clitic cline in Inner Asia Minor Greek. In: Mark Janse and An Verlinden (eds.), Productivity and Creativity: Studies in General and Descriptive Linguistics in Honor of E.M. Uhlenbeck, 521–545. Berlin/New York: Mouton de Gruyter. Joseph, Brian 1983 The Synchrony and Diachrony of the Balkan Infinitive. Cambridge: Cambridge University Press.

450

Peter Trudgill

Kaye, Jonathan, Jean Lowenstamm and Jean-Roger Vergnaud 1990 Constituent structure and government in phonology. Phonology 7: 193–231. Kondosopoulos, Nikolaos 1988 Glossikos Atlas tis Kritis. Iraklio: Panepistimiakes Ekdosis Kritis. Lenneberg, Eric 1967 Biological foundations of Language. New York: Wiley. Maddieson, Ian 1984 Patterns of Sounds. Cambridge University Press. Mansfield, Peter and Peter Trudgill 1994 A sex-specific linguistic feature in a European dialect. Multilingua 13.4: 181–186. Martinet, André 1962 A Functional View of Language. Oxford: Oxford University Press. Milroy, James and Lesley Milroy 1985 Linguistic change, social network and speaker innovation. Journal of Linguistics 21: 339–384. Newton, Brian 1967 The phonology of Cypriot Greek. Lingua 18: 382–411. 1968 Spontaneous gemination in Cypriot Greek. Lingua 20: 15–57. 1972 The Generative Interpretation of Dialect. Cambridge: Cambridge University Press. Nichols, Johanna 1992 Linguistic Diversity in Space and Time. Chicago: University of Chicago Press. Trudgill, Peter 1989a Language contact and simplification. Nordlyd 15: 113–121. 1989b Contact and isolation in linguistic change. In: Leiv Egil Breivik and Ernst H. Jahr (eds.), Language Change: Contributions to the Study of Its Causes, 227–237. Berlin/New York: Mouton de Gruyter. 1989c Interlanguage, interdialect and typological change. In: Susan Gass, Carolyn Madden, Dennis Preston and Larry Selinker (eds.), Variation in Second Language Acquisition: Psycholinguistic Issues, 243–253. Clevedon: Multilingual Matters. 1992 Dialect typology and social structure. In: Ernst H. Jahr (ed.), Language Contact and Language Change, 195–212. Berlin/New York: Mouton de Gruyter. 1995 Grammaticalisation and social structure: Nonstandard conjunctionformation in East Anglian English. In: Frank R. Palmer (ed.), Grammar and Meaning: Papers in Honour of Sir John Lyons, 136– 147. Cambridge: Cambridge University Press.

The impact of language contact and social structure 1996

451

Dialect typology: Isolation, social network and phonological structure. In: G. Guy et al. (eds.), Towards a Social Science of Language, vol. 1: 3–21. Amsterdam/Philadelphia: Benjamins. 1998 Typology and sociolinguistics: linguistic structure, social structure and explanatory comparative dialectology. Folia Linguistica 31: 349–360. 1999 Language contact and the function of linguistic gender. Poznan̗ Studies in Contemporary Linguistics 35: 9–28. 2001 Linguistic and social typology. In: Jack K. Chambers, Natalie Schilling-Estes and Peter Trudgill (eds.), Handbook of Linguistic Variation and Change, 707–728. Oxford: Blackwell. Tserdanelis, Georgios and Amalia Arvaniti forthc. The acoustic characteristics of geminate consonants in Cypriot Greek. In: Proceedings of the Fourth International Conference on Greek Linguistics, Nicosia, 17–19 September 1999.

Jespersen’s cycle and the interaction of predicate and quantifier negation in Flemish Johan van der Auwera and Annemie Neuckermans Abstract The paper focuses on the construction en niemand niet ‘no one’ (literally ‘not no one not’) in East Flemish dialects, and finds it to be unaccounted for in current typological work about negation. We present a hypothesis on how predicate negators like not and quantifier negators like no one interact and develop. Crucial reference is made to what is known as “Jespersen’s cycle”. Within this hypothesis, en niemand niet will be analyzed as an intermediate stage. From the point of view of dialect geography, en niemand niet is also an intermediate form. The paper furthermore discusses an iemand niet construction which has the same ‘no one’ meaning (literally ‘someone not’), and a sentence-final predicate negator niet. 1. 2. 3. 4.

Introduction Triple negation Jespersen’s cycle An integrated scenario

5. 6. 7.

The dialect geography of (en) niemand (niet) Brabantic iemand niet Conclusion

1. Introduction This paper discusses a type of triple negation in which one semantic negation is expressed with two so-called “predicate negators” (like not) and one “quantifier negator” (like no one). In section 2 this type will be further characterized and distinguished from similar constructions. In section 3 we briefly discuss what has been called “Jespersen’s cycle”, the observation by Jespersen (1917) that predicate negators may replace one another in a peculiar way. We also pay attention to the relevance of Jespersen’s cycle for quantifier negators and we conclude that the present typological literature cannot provide an account for the triple negation discussed in section 1. In section 4 we supply a unified account of the cyclicity of predicate and 1 quantifier negators. We describe various “Jespersen cycles” and integrate them into one scenario. In this hypothesis, the triple negation that we started with is analyzed as an intermediate stage. In section 5 we turn to the

454

Johan van der Auwera and Annemie Neuckermans

geographical spread of the triple negation and point out that the dialect geography of the construction conforms to this point of view. In section 6 we turn to another construction, “discovered” during the field work. In keeping with the title of the book, the paper shows that dialectology and typology are of high mutual relevance. The paper is primarily based on data elicited in the SAND project. SAND stands for Synctactische Atlas van de Nederlandse Dialecten 2 ‘Syntactic Atlas of the Dutch Dialects’. Exceptions are some of the sentences used in sections 2 and 5. The SAND data and the maps shown in this paper are preliminary in more than one sense. First, some of the data derive from sampling points that will not be shown in the atlas envisaged by the project. Second, we do not have access yet to all the materials of the sampling points for the eventual atlas. Third, for the purpose of this paper, we only represent the Belgian dialects of Dutch, for it is in these dialects that the relevant type of triple negation is attested. The sampling points that the present paper is based on are shown on Map 1. The map also includes the names “West Flemish”, “East Flemish”, “Brabantic”, and “Limburgish”. The adjective “Flemish” and the noun “Flanders” are commonly used in two senses: (i) to refer to the Dutch-speaking area of Belgium, and (ii) to refer to the Southwestern part of the Dutch-speaking area, covering the Belgian provinces of East Flanders, West Flanders, but also French Flanders and the southern part of Dutch Zealand. In the second sense, we can speak about “West Flemish” and “East Flemish” dialects. Together with Brabantic and Limburgish they are dialects of Flemish (first sense).

West Flemish

East Flemish

Map 1. Sampling points

Brabantic Limburgish

Jespersen’s cycle and the interaction of predicate and quantifier negation

455

2. Triple negation The type of triple negation studied in this paper is illustrated in (1): (1)

East Flanders (SAND) ‘k en ben niemand I not am nobody ‘I have not met anyone.’

ni not

tegengekomen met

First of all, it is important to point out that (1) is not an isolated occurrence. The exact spread of the phenomenon will be discussed in section 3. Secondly, semantically, there is only one negation. (1) says that it is false that the speaker has met someone. (1) is the negation of (2). (2)

‘k ben iemand I am someone ‘I have met someone.’

tegengekomen met

Formally, however, sentence (1) contains three negative markers: (i) en is an old negative marker; as the sole marker of negation, it seems to have disappeared in all of Germanic, except in some Flemish dialects, and then only in some frozen expressions. (ii) ni is a relatively new negative marker; through the phenomenon often called ‘Jespersen’s cycle’, which will be discussed later on, it was first used in combination with the old en marker, and then replaced it; in all of Germanic the replacement stage has been reached, but some of the Flemish dialects retain the en ... ni stage; Map 2 is a SAND map that shows the sampling points where en ... ni has surfaced, either in the subordinate clause or in both the main and the subordinate clause. (iii) niemand is the negative indefinite pronoun meaning ‘no one’; morphologically it is composed of the positive indefinite pronoun iemand ‘someone’ and the negative prefix ne, an allomorph of the negative marker en under (i).

456

Johan van der Auwera and Annemie Neuckermans

Ŷ Informant uses en … ni in both the main and the subordinate clause Ƈ Informant uses en … ni in the subordinate clause Ɣ Informant does not use en … ni Map 2. The use of en ... ni in main and subordinate clauses

We will call the en and the ni of (1) “predicate negators”, and niemand a “quantifier negator”. For the meaning, we will speak about “predicate” and 3 “quantifier negation”. We can thus state that the kind of triple negation that we focus on in this paper involves the expression of one semantic negation with the help of two predicate negators and one quantifier negator. Just like the survival of en in Flemish dialects is probably unique within present-day Germanic, this construction too is unique, and, as we will see in detail later on, it is not simply a consequence of the survival of en, for en niemand ni appears in only part of the en ... ni zone. The triple negation focussed on in this paper is of course just one of several types of triple negation. There are two parameters: (i) how many negations are there (semantically)? (ii) how many of the negators are predicate negators? how many quantifier negators? Different parameter settings give different types of triple negation. Some types are illustrated in (3–5).

Jespersen’s cycle and the interaction of predicate and quantifier negation

457

(3)

West Flanders (Haegeman 1995: 132) da Valère an niemand niets gezeid en-oat that Valère to none nothing said not-had ‘that Valère had not said anything to anyone’ Semantically: 1 negation Formally: 1 predicate negator and 2 quantifier negators

(4)

West Flanders (Haegeman 1995: 132) da Valère nooit an geen mens niets gezeid oat that Valère never to no person nothing said had ‘that Valère had never told anyone anything’ Semantically: 1 negation Formally: 3 quantifier negators

(5)

West Flanders (Haeberli and Haegeman 1999: 108) da Valère nie en-durft niets zeggen that Valère not not-dares nothing say ‘that Valère does not dare not to say anything’ Semantically: 2 negations Formally: 2 predicate negators and 1 quantifier negator

Like (1), the sentences in (3) and (4) express one semantic negation with three negators, but instead of the two predicate negators and the one quantifier negator of (1), sentence (3) employs one predicate negator and two quantifier negators, and sentence (4) employs three quantifier negators. Like (1), sentence (5) employs two predicate negators and one quantifier negator, but the semantic effect is that of two negations. Of course, the “triplicity” of the negation can also be seen as a parameter. There are also quadruple and even quintuple negations. Subtypes are illustrated in (6) and (7), but they will also remain outside the scope of the paper. (6)

West Flanders (Haegeman 1995: 132) da Valère nooit an geen mens niets gezeid en-oat that Valère never to no person nothing said not-had ‘that Valère had never told anyone anything’ Quadruple negation Semantically: 1 negation Formally: 1 predicate negator and 3 quantifier negators

458

(7)

Johan van der Auwera and Annemie Neuckermans

West Flanders (Haegeman 1995: 133) da Valère nooit an geen mens nie niets gezeid en-oat that Valère never to no person not nothing said not-had ‘that Valère had never said nothing to anyone’ Quintuple negation Semantically: 2 negations Formally: 2 predicate negators and 3 quantifier negators

3. Jespersen’s cycle It was Dahl (1979: 88) who coined the felicitous term “Jespersen’s cycle” for the scenario, described by Jespersen (1917), in which, for reasons of heightened expressivity, a negative marker may be accompanied by another word, which later takes over the role of marking the negation to the extent that the original marker disappears. A three-step representation for French is given in (8). (8)

1 2 3

jeo ne dis je ne dis pas je dis pas ‘I do not say’

Old French Modern Standard French Modern Colloquial French

(9) shows a five-step representation, in which the bracketing indicates the optional addition and the optional omission of a marker. (9)

1 2 3 4 5

jeo ne dis je ne dis (pas) je ne dis pas je (ne) dis pas je dis pas

The representation of the three or five stages of the cycle is one thing, the interpretation is something else. There are at the moment two dominant perspectives. The first is a grammaticalization account, involving the usual dynamics of strengthening (and weakening) and bleaching.

Jespersen’s cycle and the interaction of predicate and quantifier negation

(10)

459

Perspective I

1

jeo ne dis

2

je ne dis (pas)

the negator is strengthened

3

je ne dis pas

the strengthener bleaches and becomes part of the negator

4

je (ne) dis pas

the original negator loses ground

5

je dis pas

One finds this account both in functionalist and in formalist linguistics, e.g. both in Haspelmath (1997: 204–205) and in Van Kemenade (2000), with the former taking grammaticalization to be driven by semantics, and the latter stressing its morphosyntactic features. The second type of perspective includes the first one, but it furthermore stresses the role of a word order change. Perspective I (11)

Perspective II = Perspective I + the negator is preverbal

1

jeo ne dis

2

je ne dis (pas)

the negator is strengthened

3

je ne dis pas

the strengthener bleaches and becomes part of the negator

4

je (ne) dis pas

the original negator loses ground

5

je dis pas

the discontinuous negator embraces the verb

the negator is postverbal

What is at least as essential as the grammaticalization, according to the advocates of the second approach, is the development of a preverbal to a postverbal negation, via an intermediate stage in which the negation is

460

Johan van der Auwera and Annemie Neuckermans 4

discontinuously both preverbal and postverbal (“embracing”). Again, versions of this type of account are found in functionalist and in formalist writings, e.g. in Van der Horst and Van der Wal (1979: 24–26) or Horn (1989: XXI) and in Beukema (1999:10). Some connect the word order change with a change in the general word order profile of the language, an idea which surfaced with Vennemann (1974). With respect to this word order change, proponents of the first perspective hold that it is not part of the Jespersen cycle itself, only a possible consequence. With respect to this controversy, we support version I. We have two reasons. The first is that languages can be said to evidence the typical Jespersonian process of negative renewal without any effect on word order. For Latin it would be reasonable to posit the scenario – in a three-step format – shown in (12). (12)

1 2 3

ne V ne oenum V ‘not one’ non V

We start with a preverbal negator, we end with one, and the midstage negator is preverbal, too. The second reason for preferring perspective I is that even for languages for which the word order version of the cycle has been proposed, the application is not without problems. Dutch is such a language. For main clauses, the cycle indeed involves the expected word order change, but in subordinate clauses the negators remain preverbal throughout the cycle. (13)

Declarative main clauses 1 en V 2 en V niet 3 V niet

Subordinate clauses 1 en V 2 niet en V 3 niet V

Given that we support version I, we have the option to discuss word order or to leave it out. For reasons of space, we will leave it out. Jespersen’s cycle is a generalization about the development of predicate negators. But there is nevertheless a double link with quantifier negators. First of all, the element that strengthens the original negator and that will grow into the new negator may be a quantifier negator. There is no necessity, though. French pas ‘step’ or Latin oenum ‘one’ are not quantifier negators. Dutch niet, on the other hand, or English not are. Their original

Jespersen’s cycle and the interaction of predicate and quantifier negation

461

meaning was ‘nothing’. Secondly, quantifier negators have been analyzed with a Jespersen cycle too, albeit a shorter one, with only the last two stages of the three-stage scenario. Stages (8.2–3) for French predicate negators are nicely paralleled for quantifier negators (Haspelmath 1997: 204; cf. also Burridge 1983: 31–33). (8)

1 2 3

jeo ne dis je ne dis pas je dis pas ‘I do not say’

(14)

1 2 3

-je ne dis rien je dis rien ‘I say nothing’

The effect of the quantifier negator mini-cycle is that it creates a universally undesirable pattern, that of expressing the negation of a sentence containing a quantifier negator only once, viz. in a negativized quantifier. The universal default, as argued most persuasively by Haspelmath (1997: 201–205), building on earlier work by Bernini and Ramat (1992) and Kahrel (1996), is to express the quantifier negator in the company of a predicate negator. This universal preference, Haspelmath argues, can be withstood in two kinds of circumstances, one of which involves Jespersen’s cycle. But, as Haspelmath (1997: 205) further shows, the preference is so strong that it can undo the effect of Jespersen’s cycle. For Québec French, for instance, there is a (14.4) stage (the data are from Muller 1991: 262; further support in Daoust-Blais 1976: 1127). (14)

1 2 3 4

-je ne dis rien je dis rien je dis pas rien ‘I say nothing.’

This now is the point to turn to the type of triple negation that is central to this paper. Can the existence of (1) be explained with any hypothesis mentioned so far?

462

(1)

Johan van der Auwera and Annemie Neuckermans

East Flanders (SAND) ‘k en ben niemand I not am nobody ‘I have not met anyone.’

ni not

tegengekomen. met

The answer is negative. Given the parallelism between the predicate and quantifier cycles (8) and (14), valid also for Dutch, one would expect (15). (15)

‘k ben niemand I am nobody ‘I have not met anyone.’

tegengekomen met

(15) is indeed available: it is simply the standard Dutch strategy. From the point of view of the clash between Jespersen’s cycle and the universal preference for a quantifier negator to be accompanied by a predicate negator, and the fact that the universal default can get the upper hand, sketched with Québec French (14.4), one would expect (16). (16)

‘k ben niemand I am nobody ‘I have not met anyone.’

ni not

tegengekomen met

And indeed, (16) also occurs, and as we will see in section 5, it occurs in a wide variety of the Flemish dialects (and elsewhere in West Germanic). But nothing, so far, can explain why one could get both the old predicate negator en, the new predicate negator ni and the quantifier negator. This is what we will do in the next section. 4. An integrated scenario In what follows, we sketch the development of some of the Flemish predicate and quantifier negator constructions. The focus is not on the exact shape of the markers nor on the exact chronology, only on the general processes. We also do not discuss the various proposals about factors that can speed up the cycle, such as the difference between main clause and subordinate clause, or the position of the negator. For data we rely on Van der Horst and Van der Wal (1979), De Meersman (1980), Burridge (1983) and Beheydt (1998). It will be useful to work in terms of the five-stage

Jespersen’s cycle and the interaction of predicate and quantifier negation

463

model. The generalization that will be sketched will apply only to Dutch. To capture the negation facts from French or Latin, for example, the cycle 5 will have to provide more scenarios. In the oldest stage of Dutch, predicate negation could employ the single predicate negator en. This is stage 1 of the Jespersen cycle. On the side of quantifier negation, it seems that the language behaved in conformity with the universal preference for marking negation with both a predicate negator and a quantifier negator. In Figure 1 we note en niemand ‘not no one’, which we will follow up to the present, and en niet ‘not nothing’, which we will only follow up to the point that niet is marshaled into predicate negator 6 service. Stage 1

Predicate negator en

Quantifier negator en niet en niemand

Figure 1. Dutch negation: Stage 1 of the Jespersen cycle

Obviously, niet and niemand are complex forms, consisting of a negation and a positive quantifier iet ‘something’ and iemand ‘someone’. We assume that it is the ordinary predicate negator en that combines with the positive quantifier (cf. Stoett 1923: 160–161, and Bernini and Ramat 1996 on Old Saxon and Gothic). This, we further assume, leads to a pattern that goes against the universal preference. This time it is not Jespersen’s cycle that creates the problem but the second of the two processes that Haspelmath (1997: 203) recognizes as the disturbing factors, viz. what he 7 calls “negative absorption”, the fusion of negator and quantifier. The effect is then undone by reintroducing the predicate negator. Stage

1

Predicate negator

en

Quantifier negator en iet ѭѮ niet Ѩ (en) niet Ѩ en niet

en iemand ѭѮ niemand Ѩ (en) niemand Ѩ en niemand

Figure 2. Dutch negation: Developments for quantifier negation within stage 1 of the Jespersen cycle

464

Johan van der Auwera and Annemie Neuckermans

Stage 1 is attested in the oldest documents of Dutch, but these also show that the language has entered stage 2. The quantifier negator niet is reanalyzed as a strengthener of the predicate negator. Maybe in that period, a formal split or divergence occurs. The ‘nothing’ quantifier is formally differentiated from the predicate negator strengthener: ‘nothing’ becomes 8 niets, the strengthener remains niet. Stage

Predicate negator

1

en

2

Ѩ en (niet)

Ѯ

Quantifier negator en iet ѭѮ niet Ѩ (en) niet Ѩ en niet Ѩ en niets

en iemand ѭѮ niemand Ѩ (en) niemand Ѩ en niemand Ѩ en niemand

Figure 3. Dutch negation: Stages 1 and 2 of the Jespersen cycle

Between stages 2 and 3 the niet strengthener becomes progressively more and more obligatory. In the course of this obligatorification, one can imagine it spreading to quantifier negator constructions. In Figure 4, where we henceforth only mark the fate of en niemand, we show two scenarios. In one scenario niet is added as a strengthener and in the other it is not. Stage

Predicate negator

1

en

2

Ѩ en (niet)

Ѯ

Quantifier negator ... ... Ѩ Ѩ en niets en niemand Ѩ Ѯ ѭ en niets en niemand (niet) en niemand

Figure 4. Dutch negation: The optional spread of niet into quantifier negation

Both scenarios make sense, and even for the same reason. If niet is optionally added in a quantifier negation context, it might be through analogy with its optional presence in a predicate negation context. The absence of niet in the quantifier negation context can also be linked with the forces of analogy. With predicate negation en (niet), we are working towards double negative marking, a model already set by quantifier

Jespersen’s cycle and the interaction of predicate and quantifier negation

465

negation. Because of this analogy, there is no analogical pull towards triple 9 marking in the case of quantifier negation. There is also some empirical support. On Map 3, the two triangles roughly indicate the present-day main clause en niet zone. In the left triangle en niet occurs together with en niemand – one of the two scenarios hypothesized in Figure 4. In the right triangle it occurs together with en niemand (niet) – the other scenario.

en niet en niemand (niet)

Ì Informant uses en niemand niet z Informant does not use en niemand niet Map 3. en niet, en niemand, en niemand (niet)

The data on Map 3 show us a synchronic, present-day dialectological picture. As evidence for a diachronic scenario, a synchronic picture cannot be sufficient, but at least it shows that the predicate negator strategy en niet can coexist as easily with the quantifier negator strategy en niemand as 10 with the quantifier negator strategy en niemand (niet). In stage 3 of the Jespersen cycle obligatorification has succeeded, in the predicate negation context, of course, but also in one of the two scenarios for quantifier negation. Furthermore, the analogical pressure to add niet to en niemand must be even stronger than in stage 2, for now niet has become obligatory for predicate negation.

466

Johan van der Auwera and Annemie Neuckermans

Stage 2

3

Predicate negator ... Ѩ Ѯ en (niet) Ѩ en niet

Ѩ en niets

Quantifier negator ... Ѯ ѭ en niemand (niet) en niemand Ѩ en niemand Ѩ Ѯ en niemand (niet) Ѯ en niemand niet

Figure 5. Dutch negation: Stages 2 and 3 of the Jespersen cycle

In stage 4 the old en negator is disappearing, for predicate negation as well as for quantifier negation, and in both scenarios, and in stage 5 en is lost completely. Once again there is analogical pressure for the niet-less quantifier negator strategy to get enriched by niet, and at stage 5 pressure is particularly strong, as the niemand stage also goes against the universal 11 default. Stage ...

3

4

5

Predicate negator ... Ѩ en niet

Ѩ (en) niet

Ѩ niet

Quantifier negator ... Ѩ en niemand Ѩ Ѯ en niemand (niet) Ѯ en niemand niet Ѩ (en) niemand Ѩ Ѯ (en) niemand (niet) Ѯ (en) niemand niet Ѩ niemand Ѩ Ѯ niemand (niet) Ѯ niemand niet

Figure 6. Dutch negation: Stages 3 to 5 of the Jespersen cycle

Jespersen’s cycle and the interaction of predicate and quantifier negation

467

Figure 6 contains the parallelism noted when we paired (8) with (14), marked in Figure 7 with the uninterrupted lines. But it also shows three Jespersen cycles going from en niemand to niemand niet, marked with the interrupted lines. They differ with respect to the stage at which niet is allowed. Two have a fully classical en niemand niet intermediate stage, one has (en) niemand (niet). In this way, we could say that the full scenario integrates at least five more or less classical Jespersen cycles. Stage

Predicate negator

1

en

2

Ѩ Ѯ en (niet)

3

Quantifier negator en iets ѭ Ѯ niets Ѩ (en) niets Ѩ en niets Ѩ en niets

Ѩ en niet

en iemand ѭ Ѯ niemand Ѩ (en) niemand Ѩ en niemand Ѯ en niemand (niet)

4

5

Ѩ (en) niet

Ѩ niet

Ѩ en niemand

Ѩ

en niemand niet

en niemand (niet) Ѯ

Ѯ

Ѩ

Ѩ (en) niemand

Ѩ

Ѩ niemand

(en) niemand (niet) Ѯ (en) niemand niet

niemand niet

Figure 7. Five Jespersen cycles

ѭ en niemand

niemand (niet) Ѯ

Ѯ

468

Johan van der Auwera and Annemie Neuckermans

5. The dialect geography of (en) niemand (niet) We have seen that en niemand niet is a stage that is intermediate between en niemand and niemand niet. We will now show that this intermediacy is reflected geographically as well. Map 4, first of all, gives an idea of the wide geographical spread of the construction niemand niet. Informants were asked whether sentence (17) is acceptable in their dialect. (17)

er wil niemand niet there wants no one not ‘There is no one that wants to dance.’

dansen dance

Ì Informant uses niemand niet z Informant does not use niemand niet Map 4. niemand niet

Map 4 shows that niemand niet is widely accepted. There is small resistance zone in Brabant, which we will comment on in section five, but most resistance has been found in West Flanders. In West Flanders, especially in the southern part, informants translate niemand niet as en niemand. This is shown in Map 5. Some informants spontaneously translated sentence (17) using en niemand niet. Map 6 shows that en niemand niet occurs in southern East 12 Flanders.

Jespersen’s cycle and the interaction of predicate and quantifier negation

469

Ì Informant uses en niemand z Informant does not use en niemand Map 5. en niemand

ÌInformant uses en niemand niet z Informant does not use en niemand niet Map 6. en niemand niet

470

Johan van der Auwera and Annemie Neuckermans

If we compare Maps 4 to 6, we see that en niemand niet can easily be considered a zone of transition between niemand niet and en niemand. But there is more. Informants that volunteered en niemand niet also gave alternative translations. Some went with southwestern en niemand, marked on Map 7 with an arrow pointing in that direction (ҥ). Some gave the central-eastern niemand niet variant; we marked them with ҧ.

ҧ

ҧ ҥ ҧ ҥ ҧ ҧ ҥ ҧ ҧ ҥ ҧ

Ì Informant uses en niemand niet z Informant does not use en niemand niet Ÿ Informant also uses niemand niet  Informant also uses en niemand Map 7. en niemand niet between niemand niet and en niemand

We conclude that the present-day dialect geography is in conformity with the cyclicity hypothesis presented in section 3. It is important, however, to stress that we are not claiming that the present-day geographical facts prove the correctness of the cyclicity hypothesis. In the cyclicity hypothesis the intermediate nature of en niemand niet has a system-internal explanation: it is because of analogy with their predicate negator system, because of the Jespersonian succession of weakening and strengthening or because of a conformity with the universal default that speakers of en niemand niet have added niet. But a system-external explanation is possible, too. In this case speakers adapt their system in analogy with another system. En niemand niet can thus be seen as a mix between en niemand and niemand niet. It concerns conservative speakers who adjust their stage 3 en niemand or their stage 4 (en) niemand and insert a niet out of the progressive system of their neighbors, more precisely out of these speakers’ fifth stage niemand niet

Jespersen’s cycle and the interaction of predicate and quantifier negation

471

strategy. Interestingly, the actual steps in such a development are already part of the cycle. In Figure 8 we highlight both scenarios. Within this system-external account en niemand niet is actually intermediate also: it is a conservative construction with a progressive feature of someone else’s system. Stage ...

Predicate negator ... Ѩ

3 en niet

Ѩ 4 (en) niet ...

...

Quantifier negator ... Ѩ en niemand Ѩ Ѯ en niemand (niet) Ѯ en niemand niet Ѩ (en) niemand Ѩ Ѯ (en) niemand (niet) Ѯ (en) niemand niet ...

Figure 8. The development of Dutch predicate and quantifier negators – stages 3 and 4

6. Brabantic iemand niet Let us now return to the Map 4, repeated here with an expanded legend as Map 8. In a small Brabantic area, around the city of Aarschot, informants do not come up with the expected niemand niet, but rather with iemand niet ‘someone not’. This was noted already by Pauwels (1958: 457-460). He argues that iemand niet must have developed out of niemand niet, and that negative niemand can only be replaced by positive iemand when niet follows. In other words, it is not the case that iemand has itself become a negative quantifier. We have entered the change from niemand niet to iemand niet in a preliminary way onto the general scenario in Figure 9.

472

Johan van der Auwera and Annemie Neuckermans

Ì Informant uses iemand niet ‫ ڏ‬Informant uses niemand niet ‫ ڗ‬Informant uses neither iemand niet nor niemand niet Map 8. iemand niet Stage ...

5

Predicate negator ...

Ѩ niet

Quantifier negator ... Ѩ

niemand (niet) Ѯ niemand niet Ѩ iemand niet

Ѩ niemand Ѯ

Figure 9. The development of Dutch predicate and quantifier negators – stage 5

Of course, previous stages of the Aarschot dialect – or any other dialect, for that matter – might have had iemand niet too, for the niet has been around since stage 2, and there is no reason why the presence of en couldn’t have allowed the change, too. The point is that there is so far no evidence for constructions such as en iemand niet. Positive iemand is, of course, found in the first stage, before the univerbation. Pauwels (1958: 458) explicitly refers to the ancient univerbation of en iemand into niemand, and he says that the new Aarschot construction can be seen as a development in the

Jespersen’s cycle and the interaction of predicate and quantifier negation

473

reverse sense. This is true, but it is no less true that the development should not be considered as a simple process that can run in two directions. niemand (en) niemand en ѧ nie- ѧ (en) ѧ en iemand nieniemand mand mand

Ѭ

ѭ

ѧ nie-

mand (niet) ѭ

Ѭ

ѧ (en) niemand (niet)

ѭ

nie- ѧ iemand mand niet niet

(en) Ѭ niemand niet

en ѧ en nie- Ѭ niemand mand niet (niet)

Figure 10. From en iemand to iemand niet

Interestingly, in the dialectogial literature on Dutch negation, Pauwels (1958) is famous for his description of a special clause-final negator, a predicate negator that is optionally added to the predicate negator (en) niet, to the triple negation strategy studied in this paper, and even to the iemand niet construction. (18)

Brabant (Pauwels 1958: 438) Hij zal niet komen nie. he will not come not ‘He won’t come.’

(19)

Brabant (Pauwels 1958: 454) Dat is iets dat ik niet that is something that I not ‘That is something that I don’t want.’

(20)

Brabant (Pauwels 1958: 443) Ik geloof dat er niemand I believe that there no one ‘I believe that no one comes.’

en wil not want

niet not

nie. not

en komt not comes

nie. not

474

(21)

Johan van der Auwera and Annemie Neuckermans

Brabant (Pauwels 1958: 457) Ik heb iemand niet gezien I have some not seen ‘I haven’t seen anyone.’

nie. not

Pauwels (1958) is not the only source. Clause-final nie(t) was probably first described by Blancquaert (1923), and it has been discussed as a possible source of a similar negator strategy in Afrikaans. In our SAND work, however, we have so far found no trace of it. So it may have disappeared. This Brabantic, clause-final ni makes sense as yet another round of the cycle, with the clause-final negator strenghtening the earlier one(s) – cp. the analysis of a similar structure in Brasilian Portuguese in Muller (1991: 215). If this is correct, then the above scenarios are still too simple. Even the Jespersen cycle for predicate negator, which has hitherto been kept constant, has to be adjusted. Instead of (22) we need something like (23). (22)

en ѧ en (niet) ѧ en niet ѧ (en) niet ѧ niet

(23)

en ѧ en (niet) ѧ en niet ѧ (en) niet ѧ niet ѭ (en) niet (niet) ѧ niet (niet) Ѭ

Once again we see that intra-linguistic variation keeps pace with crosslinguistic variation. 7. Conclusion In this paper we have offered an hypothesis for the development and the interaction of predicate negators and quantifier negators. It draws on the typological literature, but it expands it, for the task was to provide for a construction consisting of an old and a new predicate negator in combination with a quantifier negator, a construction which had hitherto not been discussed in the typological literature on negation. This triple negator pattern is argued to be intermediate in the usual Jespersen-cycle sense, an intermediacy which is reflected in the attestations in present-day Flemish dialects. The integrated scenario also provides for a construction consisting of the new predicate negator in combination with a positive quantifier negator, which developed out of a negative one, and we briefly discuss a clause-final predicate negator. Both testify to the fact that intra-

Jespersen’s cycle and the interaction of predicate and quantifier negation

475

linguistic variation can be as challenging as cross-linguistic variation. A theory of variation has to deal with both. Acknowledgments Special thanks are due to Andreas Ammann, Lieselotte Anderwald, Bernd Kortmann, Ceyhan Temürcü, and Reinhild Vandekerckhove for their comments on earlier versions of this paper. Notes 1. 2.

3.

Since there is more than one, we will use “Jespersen cycle”, without ’s and ensuing definiteness. The project runs from 2000 to 2003 and is funded by the Vlaams-Nederlands Comité voor Nederlandse taal en cultuur ‘Flemish-Dutch Commission for Dutch language and culture’ (G.2214.00). On the Flemish side data have been gathered by the second author, as well as by Gunther De Vogelaer (Ghent) and Vicky Van Den Heede (Antwerp). Help came from the whole binational SAND team, but especially from Magda Devos (Ghent), Guido Vanden Wyngaerd (Brussels) and Sjef Barbiers (Amsterdam). This article extends the analysis offered in van der Auwera and Neuckermans (in press). There are many other terms in the literature, and some are used with more than one sense. Table 1 lists some of the more common ones and shows how they correspond to our usage of “predicate negation” and “quantifier negation”. Table 1. Terms for the negative markers of (1) en, ni, niemand negation sentence negation

en, ni

niemand

predicate negation standard negation sentence negation bipartite negation

quantifier negation quantifier negation constituent negation negative concord

negative concord

this paper Payne 1985 Van Kemenade 1999 Haegeman 1995 Van der Wouden 1997

negative concord

Haspelmath 1997

476

Johan van der Auwera and Annemie Neuckermans

4.

In these discussions, “verbal” refers to a finite verb, whether lexical or auxiliary. 5. French is different at least because neither the strengthener pas nor the negative quantifier personne are originally negative. Latin is different because the strengthener fuses with the old negator and it is this univerbation that will become the new one. Also, the old negator does not disappear. 6. We have dialect data only for niemand. We therefore do not claim that the story of niemand generalizes to all the other quantifier negators. 7. Interestingly, judging on the basis of Haspelmath’s (1997: 209) claims on English, he would not accept any absorption analysis for niemand and niet. 8. In some West Flemish dialects niet can mean ‘nothing’. The split would not seem to have occurred there. nieten is another ‘nothing’ word. It makes sense to analyze it as iet with both a proclitic and an enclitic en. Both observations are due to Reinhild Vandekerckhove. 9. There is no systematic diachronic study of triple negations, but it is clear that they are not a recent phenomenon (see e.g. Beheydt 1998: 88). 10. English seems to have had a strong preference for the second scenario, a preference deemed worthy of the name “Jack’s Law” by Laing (2002: 299) – the reference is to Jack (1978a, b). 11. The idea that the predicate negator may make its entry in the quantifier negation system at stage 5 of the predicate negator cycle was also expressed by Blancquaert (1923: 61). 12. This is not to say that en niemand niet does not occur outside of this area. Thus Haegeman (1995: 116) accepts it for West Flemish.

References Beheydt, Griet 1998 Het gebruik en de vorm van de negatie in het Zuidelijke Nederlands in een diachronisch perspectief (15de–20ste eeuw.) [The Use and the Form of Negation in Southern Dutch from a Diachronic Perspective (15th–20th c.)] Licentiaatsthesis, Katholieke Universiteit Leuven. Bernini, Giuliano and Paolo Ramat 1996 Negative Sentences in the Languages of Europe. Berlin/New York: Mouton de Gruyter.

Jespersen’s cycle and the interaction of predicate and quantifier negation

477

Beukema, Frits 1999 Five ways of saying no: the development of sentential negation in a Government and Binding perspective. In: Ingrid Tieken-Boon van Ostade, Gunnel Tottie and Wim van der Wurff (eds.), Negation in the History of English, 9–27. Berlin/New York: Mouton de Gruyter. Blancquaert, E[dgard] 1923 Over de dubbele ontkenning en nog wat. Handelingen van het Vlaamse filologencongres 6: 60–69. Burridge, Kate 1983 On the development of Dutch negation. In: Hans Bennis and W.U.S. van Lessen Kloeke (eds.), Linguistics in the Netherlands 1983, 31– 40. Dordrecht: Foris. Dahl, Östen 1979 Typology of sentence negation. Linguistics 17: 79–106. Daoust-Blais, Denise 1976 Étude de quelques constructions syntaxiques du parler français de Montréal: quantificateurs et négation. In: Marcel Boudreault and Frankwalt Möhren (eds.), Actes du XIIIe congrès international de linguistique et philologie romanes, Volume II: 1119–1132. Québec: Les presses de l’Université de Laval. De Meersman, Alfons 1980 Ontkenningen in ouder Nederlands. Studia Germanica Gandensia 21: 5–14. Haeberli, Eric and Liliane Hageman 1999 Negative concord and verb projection raising in Old English and West Flemish. In: Ingrid Tieken-Boon van Ostade, Gunnel Tottie and Wim van der Wurff (eds.), Negation in the History of English, 101–119. Berlin/New York: Mouton de Gruyter. Haegeman, Liliane 1995 The Syntax of Negation. Cambridge: Cambridge University Press. Haspelmath, Martin 1997 Indefinite Pronouns. Oxford: Clarendon Press. Horn, Laurence R. 1989 A Natural History of Negation. Chicago: The University of Chicago. Press. Jack, George B. 1978a Negative adverbs in early Middle English. English Studies 59: 295– 309. 1978b Negation in later Middle English prose. Archivum Linguisticum n.s. 9: 58–72.

478

Johan van der Auwera and Annemie Neuckermans

Jespersen, Otto 1917 Negation in English and Other Languages. Copenhagen: A.F. Høst. Kahrel, Peter 1996 Aspects of Negation. Ph.D. dissertation, University of Amsterdam. Laing, Margaret 2002 Corpus-provoked questions about negation in early Middle English. Language Sciences 24: 297–321. Muller, Claude 1991 La négation en français: syntaxe, sémantique et éléments de comparaison avec les autres langues romanes. Genève: Droz. Pauwels, J[an] L. 1958 Het dialect van Aarschot en omstreken. Brussel: Belgisch Interuniversitair Centrum voor Neerlandistiek. Payne, John R. 1985 Negation. In: Timothy Shopen (ed.), Language Typology and Syntactic Description, Volume 1: Clause Structure, 197–242. Cambridge: Cambridge University Press. Stoett, F[rederic] A. Middelnederlandsche spraakkunst, syntaxis. Den Haag: Nijhoff. 19233 van der Auwera, Johan and Annemie Neuckermans forthc. Een Oost-Vlaamse driedubbele negatie. Taal en Tongval Van der Horst, J[ohannes] M. and M[aria] J. van der Wal 1979 Negatieverschijnselen en woordvolgorde in de geschiedenis van het Nederlands. Tijdschrift voor Nederlandse Taal-en Letterkunde 95: 6–37. Van der Wouden, Ton 1997 Negative Contexts. Collocation, Polarity and Multiple Negation. London: Routledge. Van Kemenade, Ans 1999 Sentential negation and clause structure in Old English, In: Ingrid Tieken-Boon van Ostade, Gunnel Tottie and Wim van der Wurff (eds.), Negation in the History of English, 147–165. Berlin/New York: Mouton de Gruyter. 2000 Jespersen’s cycle revisited: formal properties of grammaticalization. In: Susan Pintzuk, George Tsoulas and Anthony Warner (eds.), Diachronic Syntax: Models and Mechanisms, 51–74. Oxford: Oxford University Press. Vennemann, Theo 1974 Topics, subjects and word order: From SXV to SVX via TVX. In: John M. Anderson and Charles Jones (eds.), International Conference on Historical Linguistics I, 339–376. Amsterdam: North-Holland.

“Gendered” pronouns in English dialects – A typological perspective Susanne Wagner Abstract Gender as a grammatical category manifests itself in modern English primarily in the form of pronominal gender agreement (he, she vs. it) based on “natural” (biological) criteria. A number of non-standard varieties, however, show gender assignment rules based on a mass-count distinction in nouns, where count nouns can be referred to anaphorically by “gendered” pronouns (i.e. he, him; she, her), whereas mass nouns only employ neuter it. With the demise of traditional dialects, these particular assignment rules are slowly giving way to the system of Standard English. Earlier studies by Ossi Ihalainen point to interesting paths these changes follow. Ihalainen suggested that standard forms first occur in less accessible positions along the Noun Phrase Accessibility Hierarchy (Keenan and Comrie 1977). When tested against corpus material from Southwest England and Newfoundland, this hypothesis proves true. In all of the tested texts, there is a strong tendency for Standard English forms to be more frequent in less accessible positions on the Accessibility Hierarchy (AH). 1. 2. 3. 4. 5. 6.

Introduction Gender in English “Gendered” pronouns Ihalainen’s hypothesis Corpus study Conclusion and outlook

1. Introduction The topic of this paper cuts across various approaches to linguistics. At its core, it will be concerned with a phenomenon found in non-standard varieties of English, more specifically traditional West Country and 1 Newfoundland dialects. The specific study presented here, however, is of a typological nature, concerning the distribution of traditional non-standard forms in comparison with their modern standard counterparts. At the same

480

Susanne Wagner

time, the analysis touches upon aspects of language change, particularly the steps from dialect to standard, as can be witnessed by the distribution of “gendered” pronouns in the two varieties discussed here. Traditionally, typologists only deal with standard varieties of a language, and the results from such studies often contrast with rather than support the typological predictions. For example, the predictions made by the Accessibility Hierarchy (cf. [1]) concerning relativisation are problematic in Standard English (StE). While the hierarchy predicts that gapping or zero relativisation (like any relativisation strategy) should be possible in all positions to the left of any position where it occurs, StE does not allow gapping in subject position (e.g. *The man came in is my husband). In English dialects, however, the predictions are borne out (cf. Kortmann 1999, 2002; similarly Fleischer, this volume, on relativisation in German dialects). (1)

subject > direct object > non-direct/indirect object > possessor/oblique (cf. Keenan and Comrie 1977; Comrie 1989: 155–164; Croft 1990: 108–111)

The inclusion of non-standard varieties in typological studies promises new insights into many problematic issues. In addition, a combination of corpus-linguistic, dialectological and typological methods may point to explanations of paths in language change that have been unaccountable or seemed unsystematic before. 2. Gender in English It is common knowledge that gender as a grammatical category is of only marginal importance in Present-Day English (cf. e.g. Wales 1996: 134). While earlier stages of English showed gender agreement patterns that are familiar from, for example, modern French or German, Present-DayEnglish’s only contribution in terms of gender-differentiated forms can be found in the third person singular of the personal pronoun paradigm. From a typological point of view, such a system can be called a “pronominal gender system” (Corbett 1991: 5). In addition, some gender-marked noun endings such as -or vs. -ess also exist, but are by no means as omnipresent as the personal pronouns he, she vs. it.

“Gendered” pronouns in English dialects

481

Gender systems such as that found in Present-Day English are often described as “natural” (cf. e.g. Corbett 1991: 9). Criteria for such systems are widespread in the world’s languages; often, the general division is one between human and non-human, and humans are divided into male and female in turn (cf. Corbett 1991: 11). Sometimes the dividing line is animate – inanimate instead of human – non-human. English might be an example of such a division, as animals (particularly domestic animals) are usually masculine or feminine according to sex. However, the factors that may influence pronoun choice are manifold and complex, even in StE – a boat can be she, fill ‘er up is regularly heard when talking about refuelling a car, and in emotional situations, he can refer to all types of things (cf. e.g. Morris 1991 or Mathiot and Roberts 1979). 3. “Gendered” pronouns When looking not only at standard varieties, but also at non-standard varieties of English, one is confronted with an even wider array of gender assignment rules than are familiar from StE. A peculiar system of gender assignment is described in much of the 19th century dialectological literature on Southwest English (“West Country”) dialects. One of the most influential accounts can be found in a publication by Frederic Elworthy. His three major works on West Country dialect (Elworthy 1875, 1877, 1886) focus mainly on West Somerset English, which, judging from various other studies on traditional (19th century) West Country dialect, can be taken as representative of both the area and time. The Dorset poet William Barnes, for example, uses the phrase “two classes of things” (cf. Barnes 1886: 17) when commenting on a system that Elworthy describes as follows: (2)

a.

Every class or definite noun, i.e. the name of a thing or object which has a shape of its own, whether alive or dead, is either masculine or feminine, but nearly always the former; indeed, the feminine pronouns may be taken as used only with respect to persons. (Elworthy 1877: 32; my italics) Examples: pitcher = he; “tool, book, house, coat, cat, letter etc., are all spoken of as he”. (Elworthy 1877: 33)

482

Susanne Wagner

b.

It is simply an impersonal or abstract pronoun, used to express either an action or a noun of the undefined sort, as cloth in the quantity, water, snow, air, etc. (Elworthy 1877: 33; my italics) Examples: weather, hay, beer = it (Elworthy 1877: 33)

In modern terms, the distinction that is made by these authors between different types of nouns is that of mass versus count nouns. Gender distinctions are based on that division, so that only count nouns use the forms we know as masculine and feminine, while mass nouns use neuter it 2 exclusively. This two-fold paradigm is basically that of ProtoIndoeuropean, where the basis for noun classification was that of animate (i.e. “masculine” with subcategory “feminine” for female humans and animals) vs. inanimate (i.e. “neuter”) instead of count vs. non-count (cf. e.g. Szemerényi 19904: 164–165). Judging from the modern literature and comments by experts, gender systems based on a mass-count distinction are very rare, although not unknown, in the world’s languages. Occasional examples are, however, only mentioned for non-standard varieties of Germanic languages rather than the respective standard ones, for example Danish dialects in Jutland (Siemund 2002); Low German dialects (Rohdenburg 2002); Tasmanian Vernacular English (Pawley 2002). These “gendered” pronouns can be found in Southwest England and Newfoundland English to the present day, as both traditional (20th century) material such as the SED Basic Material as well as the unpublished incidental material of the SED fieldworker notebooks (cf. Orton 1962) and comparatively modern data (e.g. from oral history projects of the late 20th century) show. By “gendered” pronouns I understand instances of personal pronouns which are marked for gender (i.e. either masculine or feminine) but which refer to inanimate (count) nouns. Some illustrative examples 3 follow in (3) to (6). (3)

ªMXÛJcGªE¥,«|ªS8GªEcG«¥SWXۍQQªVWLÛPQ (when) you’ve got a boil, put a bottle up to it and steam it (SED 36 Co 3, book VI)

(4)

'DWGNLÛSªªLÛGD8Q that keeps it [cart] down (SED 31 So Montacute, book I)

“Gendered” pronouns in English dialects

(5)

'4,ªGŠH,QG,,P,QW'4WªEŠ8N they drained it [field] into that brook (SED 15 He Lyonshall, book IV)

(6)

DÛYª«oÛVQ|LÛªYDÛ«GÌ`8ÛP,ªScNW I’ve lost it [knife]; it fell through my pocket (SED 36 Co 2, book IX)

483

“Gendered” pronouns are of typological interest for two reasons in particular: First, as already mentioned above, the mass-count basis of the system of gender assignment seems to be typologically rare. Moreover, where it does occur, it seems to be restricted to non-standard varieties. Second, the modern distribution of “gendered” pronouns shows some interesting patterns, which have first been observed and presented by the Finnish dialectologist Ossi Ihalainen in one of his papers on West Country English. After discussing the relevant background information, “Ihalainen’s hypothesis” will be applied to corpus material from Southwest England and Newfoundland. 4. Ihalainen’s hypothesis In his publications on West Country English in general and Somerset English in particular (cf. Ihalainen 1985a, 1985b, 1991b), Ossi Ihalainen stresses his interest in two details of pronominal distribution: – What is the relationship between standard it and non-standard he-forms? – What is the distribution of subject and object forms in this context? Ihalainen’s studies of interest here are based on the analysis of various recordings from the county Somerset in Southwest England. According to Ihalainen, he used five random samples of 30 minutes each for this investigation (1985b). For another study, he states that “some 7000 words” were used (Ihalainen 1985a: 69), while no details at all are offered for the third study (Ihalainen 1991b, based on a 1983 paper). It is to be assumed that the corpora were largely identical for all of these studies; at least the figures for “gendered” pronouns are identical for the (1985a) and (1991b) study. The figures presented in the articles show interesting distributions whose proportional relation should hold for other data sets as well. As to

484

Susanne Wagner

the first of the two questions posed above, the reader is presented with the following table which, according to the author, lists the occurrences of all pronouns corresponding to StE it. They were then divided into the respective categories of “thing” and “mass”, or count and non-count, as has been proposed by earlier researchers. Table 1. Distribution of personal forms vs. it for “thing” and “mass” referents (from Ihalainen 1985b: 157) thing mass

personal form (he, him) 69 79% 0 0%

it 18 95

21% 100%

From Table 1 one can deduce that dialect forms have lost some territory to their standard equivalents over the past century. With 21% of it on traditional he-turf, the traditional system seems to be losing ground. However, one should be careful in passing premature judgement, as none of the 19th-century dialectologists presented figures to support their claims. Although gender assignment is always presented as a clear-cut system in the traditional dialect literature, authors (who at that time only rarely analysed actual data) may have been idealizing the situation. Consequently, it is by no means impossible that Ihalainen’s figures would have held at Elworthy’s times, too. Ihalainen draws on the perhaps most widely known typological hierarchy to describe the pattern underlying the results he observed in his corpus. When looking for “gendered” pronominal forms, he noticed that their distribution across grammatical relations, i.e. basically subject and 4 (direct) object, was by no means equal. Ihalainen’s figures and ratios are reproduced in Table 2. Table 2. Distribution of personal forms vs. it for “thing” referents according to syntactic context

subject object

ratio personal forms : it 1991b/ 1985a 1985b 7.3 : 1 5:1 2.4 : 1 3:1

As can be seen from Table 2 (based on Ihalainen 1991b: 115, 1985a: 69 and 1985b: 161), the difference in distribution is even more pronounced for

“Gendered” pronouns in English dialects

485

the (1991b) data than for the (1985b) corpus. Based on these figures, the author suggests that standard forms establish themselves in the dialect system in less accessible positions in the Noun Phrase Accessibility Hierarchy (recall [1]) and spread from there. (1)

subject > direct object > non-direct/indirect object > possessor/oblique (cf. Keenan and Comrie 1977; Comrie 1989: 155–164; Croft 1990: 108–111)

Thus, it is more frequent in object positions than one might expect, at least in contexts where he is also possible in dialect (cf. Ihalainen 1985a: 69–70). Or, in Ihalainen’s own words: “[T]he most prominent linguistic contexts are also the most favourable to dialectal forms. Interpreted diachronically, this means that the changes concerned arose in non-prominent contexts and are spreading to more salient ones.” (Ihalainen 1991b: 105) Ihalainen’s frequency-based line of arguing is well-known from such cases as irregular verb paradigms (particularly be) or the personal pronoun system itself, where it can be assumed that the only thing preventing gender from disappearing from the paradigm altogether is the high frequency of third person pronominal forms: In these systems, a frequent irregular form can survive much more easily than an infrequent one, which will generally be regularized to fit the rest of the paradigm. The link between frequency and accessibility is established, for example, in Keenan (1987: 49): “The frequency with which people relativise in discourse conforms to the [AH], subjects being the most frequent, then direct objects, etc.” However, this should not lead us to the conclusion that (a) new forms (either standard or non-standard) always enter paradigms from less frequent positions and spread to more frequent ones and (b) traditional/ conservative forms generally only survive in grammatically frequent relations. Such a link between conservatism and frequency definitely exists for a number of paradigms, and may indeed be the “normal” path for morphological change (cf. e.g. Krug 2003: 18). In addition to the paradigm investigated here, Herrmann (forthcoming) makes similar observations for the demise of the relative particle as, which only survives in subject position, but has been superseded in lower positions by StE relative pronouns which entered the paradigm from less frequent positions and are spreading upwards on the AH.

486

Susanne Wagner

On the other hand, Herrmann also found that the non-standard relative particle what entered the paradigm via the subject – i.e. the most frequent – position and is now spreading down the AH, taking the reverse path. For the present investigation, standard it forms seem to have “invaded” the territory of personal forms in object position first, later spreading to the more prominent subject contexts (cf. also Ihalainen 1991a: 114–115). In a wider context, Ihalainen uses the AH as reference point since aspects of frequency of grammatical functions across different categories cumulate in this widely known and accepted hierarchy. Moreover, it also serves as an illustrative device to establish a link between frequency observations and language change: The more frequent the forms, the less likely they are to change (here: from dialect to standard) – at least in morphology (cf. Krug 2003: 18). This is a hypothesis that seems worth testing against a considerably larger corpus. 5.

Corpus study

5.1. The corpora The material used in the current investigation stems from two regions, namely the Southwest of England (particularly Somerset) and Newfoundland. Both sub-corpora are comparable in so far as they consist 5 of oral history material. For Newfoundland, the archive at Memorial University (Memorial University of Newfoundland Folklore and Language Archive – MUNFLA) stores a large number of interviews with natives that were conducted for a range of different studies, both individual (generally students’ papers and theses) and project-oriented (e.g. a collection of folktales; cf. Halpert and Widdowson 1996). As most interviews are conducted by a member of the informant’s (extended) family, the interviewer’s language has generally no interferencing effects on the informant’s language. A further bonus has to be seen in the fact that MUNFLA employs professional transcribers who produce non-standardized orthographic transcripts. This is not the place to discuss the problems that emerge when trying to render dialect in a still legible orthography. Although a phonetic transcript would certainly be preferable, temporal factors and issues of legibility and practicality are in favour of an orthographic version. For this study, 31 interviews, amounting to 132,000 words altogether, recorded between the late 1960s and the early 1980s, constitute the Newfoundland component.

“Gendered” pronouns in English dialects

487

The material for the Southwest of England largely stems from the Oral Archive of the Somerset Rural Life Museum, where conditions similar to those at MUNFLA apply. Despite the fact that no professional transcribers were used, the transcripts are surprisingly accurate in rendering dialect features. 31 interviews, recorded between 1968 and 1996, totalling ca. 174,500 words, were selected for the present investigation. 5.2. Method In order to be able to undertake an analysis of forms in certain syntactic contexts, a number of previous steps are necessary. It was decided to tag the corpus data for the three most frequently occurring grammatical relations, namely subject, direct object and prepositional object. As mentioned above, Ihalainen’s analysis is restricted to the former two levels, but as the corpora contain enough data for prepositional objects, this context was included as well. If the study is to produce statistically relevant results, not all texts are equally suitable for the analysis. Thus, a sub-corpus made up of texts meeting the following criteria was compiled: (a) (b)

total of exceptional masculine pronouns • 10 total for each grammatical relation • 1

In addition, there had to be enough standard forms, i.e. its, in the texts to be able to come up with statistically significant results at all. For that reason, such an analysis is not possible for a truly traditional dialect text, as such a text usually does not, or at least should not, contain any instances of it 6 referring to countable nouns. In the end, only six texts from the Newfoundland material (31,035 words) and four from the Somerset material (32,700 words) proved suitable for testing Ihalainen’s hypothesis. The results, though not always 7 statistically significant, make the point nevertheless. There is a strong tendency for standard it to be more frequent in less accessible positions on the AH. In all texts we can observe a slight, steady or even steep increase in percentages or likelihood in favour of Ihalainen’s hypothesis. Figures 1(a,b,c) to 2(a,b) illustrate this for the most significant texts. Background information on these figures can be found in Tables 3 to 7.

488

Susanne Wagner

Figure 1(a). Newfoundland text 1

Figure 1(b). Newfoundland text 2

Figure 1(c). Newfoundland text 3 Figure 1.

Distribution of “gendered” pronouns vs. it in selected corpus texts – Newfoundland

Table 3. Distribution of “gendered” forms across grammatical relations (C626) C626 subject object prep. obj.

# of gendered forms 12 17 7

%

# of it

%

0.706 0.386 0.368

5 27 12

0.294 0.614 0.632

“Gendered” pronouns in English dialects

489

Table 4. Distribution of “gendered” forms across grammatical relations (C631) C631 subject object prep. obj.

# of gendered forms 23 18 1

%

# of it

%

0.719 0.409 0.056

9 26 17

0.281 0.591 0.944

Table 5. Distribution of “gendered” forms across grammatical relations (C631) C2914 subject object prep. obj.

# of gendered forms 50 11 5

%

# of it

%

0.806 0.500 0.417

9 26 17

0.194 0.500 0.583

Figure 2(a). Somerset text 1 Figure 2.

Figure 2(b). Somerset text 2

Distribution of “gendered” pronouns vs. it in selected corpus texts – Southwest England

Table 6. Distribution of “gendered” forms across grammatical relations (SRLM 23) SRLM 23 subject object prep. obj.

# of gendered forms 21 21 1

%

# of it

%

0.600 0.500 0.091

14 21 10

0.400 0.500 0.909

490

Susanne Wagner

Table 7. Distribution of “gendered” forms across grammatical relations (SRLM 62) SRLM 62 subject object prep. obj.

# of gendered forms 15 1 0

%

# of it

%

0.484 0.048 0

16 20 2

0.516 0.952 1.000

Further evidence supporting Ihalainen’s hypothesis can be found in isolated examples in almost all of the other 52 interviews. Speakers are much more likely to use a formulation as in (8a) than one as in (8b): (8)

a. b.

He (the river) got trout in it. It got trout in un. (Ihalainen 1991b: 114)

The examples in (9) and (10) provide illustrative material of the side-byside existence of traditional (masculine) forms and modern (neuter) forms within a single utterance, generally referring to the same entity. Ihalainen’s “accessibility hypothesis” seems to be a very good explanation to account for the presence of both forms. While the subject form he still defends its territory, positions further down (to the right) on the AH are losing ground. Thus, even though masculine forms are still very frequent in comparison with other traditional dialect phenomena, we have to assume that the dialect system is slowly giving way to the system of StE. (9)

a. b. c. d.

So, therefore, if they had to use a stone jar they used to get the basket maker to put some wicker around it, so if he had a bump he wouldn’t break. [SRLM (055)] ... when I saw the plough nobody valued it, if he had been kept dry he would have been good now. [SRLM (003a)] [talking about the right time to pick apples] Well, like if you picked one an’ cooked it early he isn’t same as when he’s been picked and kept, is he? [SRLM (317)] ... so anyway Father come home, looked down, “What’s that?” – “Sewing machine.” – “You’d better take he back where you had him from then.” [...] Ooh I’ve still got the machine down home today; what can I do with it? [SRLM (102)]

“Gendered” pronouns in English dialects

(10)

491

e.

... and he’d tie up a bundle, oh a decent bundle with two strings round it, and the thing, when he was such a weight he had the tipper there, he’d go round and throw out the sheaves. And then there was a man had to pick up that bundle and take him back wherever the man was building a stack, a straw stack. [SRLM (108)]

a.

Yea, well he was made, you could have it for a bank one time and dey turned into a work box. He was made for a money bank, ya know perhaps dey shoffed deir coppers down it. [MUNFLA 70-003: C0627] [Int.: A bane (i.e. bean) jar, and what would you use that for? Oh dey have he to put banes into it. [MUNFLA 70-003: C0627] ... you had a flap roll around it, ‘round the lock, keep he dry, perhaps you had a muskrat’s fur, rabbit’s fur ‘round it for the water, hold the water ... [MUNFLA 75-164: C2829] ... when the great wave used to come rolling, we’d run as fast as we could to get up out of it before he’d get us. [MUNFLA 78-006: C3321] And he’s [song] all there but there’s different words in it you know. [MUNFLA 70-037: C751]

b. c. d. e.

As is to be expected, the change from one system to the other is in progress for many speakers, resulting in seemingly contradictory uses of pronouns. Very often, we still find the occasional un in (direct) object position, although the speaker will generally use standard it in those positions, while masculine he is still dominant in subject position (cf. 11). (11)

a.

b.

He said dat was, dat was true, he said, their camp’d be down every, made no difference how dey put en up, how strong dey put it up he’d be down flat when dey come back. So dey, his uncle said we considered dere must be something, dere must be, must be a grave dey was sot on you know, so dey shift and he never come down no more. [MUNFLA 70-003: C0633] Oh yes he [certificate] had to go in wid it [application for old age pension] see, he had to go in wid en [application?], now he won’t be, come back no more ‘fore I gets 70, if I lives till dat. Dey’ll send it back den, dey don’t want it. Dey sent back

492

Susanne Wagner

mi mother’s, my mother’s, they sent back he when she got in her 70’s, and said, we don’t need en no longer, you know. [MUNFLA 70-003: C0632] 6. Conclusion and outlook The present study has tried to provide explanations for observed paths of language change in dialects. It could be shown that in the specific case discussed here, the change from dialect to standard language follows predictions that were made in other (typological, frequency-related) contexts. Despite the clearly observable tendencies outlined here, a number of questions remain unanswered. For one, material from only two varieties, West Country English and Newfoundland English, has been analysed. The lack of a sufficient amount of suitable data from the other varieties which supposedly have similar gender assignment systems prevents us from generalising our observations. Although it might be sufficient from a typological point of view to state that other such systems exist, it would be desirable to be able to back these claims with more data. Moreover, the analysis of the distribution of “gendered” vs. standard pronouns has outlined a path that language change may take. It was shown that the traditional dialect forms are most resistant to standard influence in their most frequently occurring positions. But despite a certain systematicity in the distribution of forms, we are still only talking about tendencies – actual language use is by no means as systematic as some studies may suggest. Frequency of occurrence has proved to be highly significant in predicting paths of language change. However, frequency criteria do not allow generalisations across categories in general. In the actual case investigated here, it was shown that in some – particularly morphological – paradigms (here: “gendered” pronouns) traditional forms will be preserved (only) in the most frequent environments (here: subject positions). The exact opposite, however, can be found in other paradigms (e.g. phonology; cf. Krug 2003: 18). It is thus neither possible to generally equate conservatism with frequency nor innovativeness with frequency. Many of these patterns will only become visible when analysing non-standard as well as standard varieties, particularly in a language like Present-Day English, where the standard variety is “the odd one out” more often than not.

“Gendered” pronouns in English dialects

493

Notes 1. 2.

3. 4.

5. 6.

7.

The research presented here is based in part on corpora collected for the project “Vergleichende Dialektsyntax am Beispiel der Britischen Inseln”, funded by the Deutsche Forschungsgemeinschaft (Ko 1181/1-1,2). In reality, the system is much more complex, but it would lead too far to attempt explanations here. The factors influencing gender assignment in a number of varieties of English, including StE, are discussed in detail in Wagner (forthcoming). The examples from the fieldworker notebooks are identified by the respective county number + county code + location number + book number. Indirect objects are frequently realised as prepositional objects in English (give me it vs. give it to me), basically eliminating the possibility of analysing indirect objects. Even when realised, both indirect and prepositional objects are comparatively rare in spoken language. For details on the purpose and make-up of oral history interviews, compare the homepage of the Oral History Society at www.oralhistory.org.uk. This is one of the reasons why the SED material cannot be used here, although it contains many relevant examples. Neither the make-up of the Basic Material nor the fieldworker notebooks admit an analysis according to grammatical relations. Chi-square values range between truly insignificant (at 25%) via marginally significant to highly significant at the 0,003%-level. Details for the individual texts look as follows: C626: relevant at 5,7%; C627: 1,6%; C 628: 6,3%; C631: 0,003%; C1187/8: 25%; C2914: 0,28%; SRLM 23: 1,2%; SRLM 62: 0,2%; SRLM 224: 99%; SRLM 83: 0,05%.

References Barnes, William 1886² A glossary of the Dorset Dialect with a Grammar. Guernsey/St. [1970] Peter Port: Toucan Press. Comrie, Bernard 1989² Language Typology and Linguistic Universals. Chicago: University of Chicago Press. Corbett, Greville 1991 Gender. Cambridge: Cambridge University Press. Croft, William 1990 Typology and Universals. Cambridge: Cambridge University Press. Du Pree, Victor 1969 MUNFLA Tapes 70-003/C626-633.

494

Susanne Wagner

Elworthy, Frederic Thomas 1877 An Outline of the Grammar of the Dialect of West Somerset. [1965] London: Trübner and Co. [Vaduz; Kraus Reprint Ltd.] 1886 The West Somerset Word-book. London: Trübner and Co. [Vaduz; Kraus Reprint Ltd.] [1965] Fleischer, Jürg this vol. A typology of relative clauses in German dialects. Fudge, Jesse 1969 MUNFLA Tape 70-037/C751. Halpert, Herbert and John D.A. Widdowson 1996 Folktales of Newfoundland – The Resilience of the Oral Tradition. 2 vols. St. John’s: Breakwater. Herrmann, Tanja forthc. Relative clauses in English dialects. In: Bernd Kortmann, Tanja Herrmann, Lukas Pietsch and Susanne Wagner, A Comparative English Dialect Grammar: Agreement, Gender, Relative Clauses. Berlin/New York: Mouton de Gruyter. Ihalainen, Ossi 1985a Synchronic variation and linguistic change: evidence from British English dialects. In: Roger Eaton et. al. (eds.), Papers from the 4th International Conference on English Historical Linguistics, Amsterdam, 10–13 April 1985, 61–72. Amsterdam: Benjamins. 1985b He took the bottle and put ‘n in his pocket: The object pronoun it in present-day Somerset. In: Wolfgang Viereck (ed.), Focus on: England and Wales (VEAW, G4), 153–161. Amsterdam: Benjamins. 1991a The grammatical subject in educated and dialectal English: comparing the London-Lund Corpus and the Helsinki Corpus of Modern English Dialects. In: Stig Johansson and Anna-Brita Stenström (eds.), English Computer Corpora. Selected Papers and Research Guide, 201–214. Berlin: Mouton de Gruyter. 1991b On grammatical diffusion in Somerset folk speech. In: Peter Trudgill and Jack Chambers (eds.), Dialects of English. Studies in Grammatical Variation, 104–119. London/New York: Longman. Keenan, Edward L. 1987 Variation in Universal Grammar. In: Edward Keenan, Universal Grammar: 15 Essays, 46–59. London: Croom Helm.

“Gendered” pronouns in English dialects

495

Keenan, Edward L. and Bernard Comrie 1977 Noun phrase accessibility and Universal Grammar. Linguistic Inquiry 8: 63–99. Kortmann, Bernd 1999 Iconicity, typology and cognition. In: Max Nänny and Olga Fischer (eds.), Form Miming Meaning: Iconicity in Language and Literature, 375–392. Amsterdam/Philadelphia: Benjamins. 2002 New Prospects for the Study of Dialect Syntax: Impetus from Syntactic Theory and Language Typology. In: Hanns Bennis and Sjef Barbiers (eds.), Syntactic Microvariation, 185-213. Amsterdam. (http://www.meertens.nl/books/synmic/) Krug, Manfred 2003 Frequency as a determinant in grammatical variation and change. In: Günther Rohdenburg and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 7–67. Berlin/New York: Mouton de Gruyter. Lehr, Genevieve 1977 MUNFLA transcript of Tape 78-006/C3321. Mathiot, Madeleine and Marjorie Roberts 1979 Sex roles as revealed through referential gender in American English. In: Madeleine Mathiot (ed.), Ethnolinguistics: Boas, Sapir and Whorf Revisited, 1–47. The Hague: Mouton. Morris, Lori 1991 Gender in modern English: The system and its uses. PhD dissertation, Département de langues et linguistique, Université Laval (Quebec). Orton, Harold 1962 Survey of English Dialects – Introduction. Leeds: Edward Arnold. Pawley, Andrew 2002 Using he and she for inanimate nouns in English: questions of grammar and worldview. In: Nick Enfield (ed.), Ethnosyntax: Explorations in Grammar and Culture, 110–137. Oxford: Oxford University Press. Rohdenburg, Günther 2002 Grammatical parallels between English and Low German dialects and their typological implications. In: Dieter Kastovsky, Gunther Kaltenbröck and Susanne Reichl (eds.), Anglistentag 2001 Wien – Proceedings, 35–56. Trier: Wissenschaftlicher Verlag.

496

Susanne Wagner

Siemund, Peter 2002 Animate pronouns for inanimate objects: Pronominal gender in English regional varieties. In: Dieter Kastovsky, Gunther Kaltenbröck and Susanne Reichl (eds.), Anglistentag 2001 Wien – Proceedings, 19–34. Trier: Wissenschaftlicher Verlag. Szemerényi, Oswald Einführung in die vergleichende Sprachwissenschaft. Darmstadt: 19904 Wissenschaftliche Buchgesellschaft. Simms, Garfield 1975 MUNFLA Tapes 75-164/C2829-2831. Wagner, Susanne forthc. Gender in English pronouns – myth and reality. In: Bernd Kortmann, Tanja Herrmann, Lukas Pietsch and Susanne Wagner, A Comparative English Dialect Grammar: Agreement, Gender, Relative Clauses. Berlin/New York: Mouton de Gruyter. Wales, Katie 1996 Personal Pronouns in Present-Day English. Cambridge: Cambridge University Press.

Population linguistics on a micro-scale. Lessons to be learnt from Baltic and Slavic dialects in contact Björn Wiemer Abstract This article aims at demonstrating what kinds of problems specialists in areal and contact linguistics have to cope with if they want to do justice to structural variation inside small contact regions, but with a comparative view towards larger areas and typological research. Nichols’ (1992) view that typology should be viewed as ‘population linguistics’ is applied here to small areas with genetically closely affiliated language varieties and within comparatively short time spans. The varieties examined all belong to North Slavic and East Baltic, the features analysed are taken from the gender and case systems, number marking in verbs, functions of reflexive markers and prepositions. It is argued that typological markedness is often not a sensible way of explaining what has been going on in contact varieties of a small area, since the relevant features are often area-specific and/or language-specific and the types of functional oppositions require very subtle distinctions. Furthermore, the cognitive and communicative basis for typological unmarkedness can be overridden by factors like social prestige and success. “Ideological” factors have been claimed to be decisive, for example by Chambers (2000). It is for this reason that his reservations against contact-induced changes must themselves be met with reservation. 1. Markedness and socio-geographical framework 2. Case studies 3. Lessons to be learnt from small-scale contact areas 4. Conclusions

1. Markedness and socio-geographical framework The premise of this volume centers around the general observation that regional varieties which differ from their standards often behave “more normally” from a typological perspective than the latter. From a typological viewpoint, ‘markedness’ is defined as “a universal property of a conceptual category, not a language-particular property of a language-particular

498

Björn Wiemer

grammatical category” (Croft 22003: 88; original emphasis). However, it seems arguable whether it always is appropriate to judge on the type and degree of markedness on a world-wide scale, since every language (both the standard variety and its regional dialects) can be seen as interacting with other varieties on some larger, albeit not on a global area. In this respect, it appears to be more realistic to establish the peculiarities of the structural features of a given variety relative to its dialectal and diastratic environment. ‘Markedness’ would then simply refer to ways of grammatical encoding which are encountered less frequently and which are more restricted in structural and behavioral terms (on these properties cf. Croft 22003: 91–99) than the unmarked values of the respective category in the given variety and those varieties which are close to it in areal terms. Such an approach seems justified since areas themselves may vary as to their typological biases. This fact has led people to postulate the existence of so-called linguistic convergence areas (Sprachbünde), or similar agglomerates of areally affiliated languages with a set of shared features on all structural levels. Furthermore, some very general typological features like, for instance, basic word order, alignment or head vs. dependent marking also have been shown to be biased if measured on a continentwide scale. Suffice it to mention Nichols (1992), who used the results of her investigations to advocate the view that linguistics should better be understood as a “population science”. From this perspective, emphasis is to be placed not on structural properties as universal features of languages isolated from contact, but on structural properties which arise due to diverse contact conditions of speakers (populations). Thus, languages are treated as communicative means among populations of speakers which, by contact with populations using other languages (dialects), are subject to multilingualism, and this plays a decisive role in structural changes of the respective languages (dialects), regardless of their “genetic” affiliation: “diversity itself (...) emerges as a stable and unmarked property of sets of languages” (Nichols 1992: 2). Contrary to Nichols, I do not intend to deal with reconstruction over great time depths, and I will be concerned only with a tiny geographic region (at least in comparison to Nichols’ literally global approach), namely, the Baltic-Slavic contact area (BSA); see Map 1 at the end of the paper. Nonetheless, I want to look at language varieties in that area under the auspices of “population linguistics”. Inter alia this means that we should approach the shared peculiarities of different varieties of the same or genetically closely-related languages with respect to their immediate

Population linguistics on a micro-scale

499

environment in geographic, as well as in sociolinguistic terms (i.e. vernaculars vs. standards). As will, I hope, become clear below, such an approach on a “micro-areal scale” needs some adjustments. In general, the problem with applying global typological methodology to small areas is that we have to find appropriate taxonomies of structural properties which allow us to assess what is really distinctive of the varieties of the area chosen for description. In each instance, we are forced to establish a gauge which permits us to specify in which areas and varieties a given feature is marked on a dialectal and/or diastratic background. For instance, we already know that certain grammatical features may be rare from a world-wide viewpoint, but wide-spread in Europe: e.g. the have-perfect and, as its presupposed feature, the existence of a general transitive verb like to have, an external possessor marked by the dative, particles used for comparison, or the existence of relative pronouns. This leads to notions like “Europeanisms” or Standard Average European (SAE); cf. Haspelmath (1998; 2001), where the respective features are discussed in detail. The next step “down” on a geographical scale (which would also in part be comprised by the notion of SAE) would be observations that some features, again, may be common in certain areas of Europe, so that they might be judged as “marked” if compared to a general European picture, but quite common from a world-wide perspective. Here we could mention Bulgarian and Macedonian, which have a definite, but no really grammaticalized indefinite article, whereas more central SAElanguages have both definite and indefinite articles. The latter, however, is a very rare phenomenon on a world-wide scale, thus Balkan Slavic (as representing more peripheral SAE-languages) is closer to the “world-wide trend” than typical SAE-languages. Another larger region of interesting convergences is the so-called Circum-Baltic Area, which has been made the subject of a double volume edited by Dahl and Koptjevskaja-Tamm (2001). It partly intersects with SAE, and its eastern part can be regarded as the immediate background against which structural peculiarities found in the BSA have to be interpreted (see below). In fact, the typical features of just one area in Europe can serve as a background for peculiarities of a few varieties inside that particular area. On the one hand, some features may be “typologically weird” and/or “weird” in comparison to the standard variety, but “normal” with respect to the area where the variety in question is used. On the other hand, if a variety shows “typologically normal” behavior, but is peculiar in

500

Björn Wiemer

comparison to its immediately neighboring languages (varieties), this needs some extra explanation, which cannot just be based on reference to typological unmarkedness. For it is hardly conceivable how the speakers of a particular local variety in some small area of the world should be “inclined” to make their dialect conform with the “world-wide average”. In other words: typological unmarkedness needs some motivation in cognitive and/or communicative terms. Typologically frequent coding devices cannot be an explanatory principle by themselves, but must derive from some deeper principles, which might also be decisive in the case of the marked behavior of a variety with respect to the area it belongs to. Such principles have been claimed to be at work from a sociolinguistic perspective by Chambers (2000). On the basis of world-wide studies of English vernaculars, he suggests that certain recurrent linguistic features of them are near universal and constitute a closed set (cf. also Chambers 2 2003: 258–266 and in this volume). For instance, he gives a list of six features (Chambers 2000: 12), three of which are phonological 1 (pronunciation of unstressed final morphemes like Eng. {ing} , final consonant cluster simplification, final obstruent devoicing), one morphological (“default singulars”, i.e. the use of the standard singular gram instead of the plural gram with plural subjects) and two syntactic 2 (multiple negation, lack of the copula ). My aim here is to show that the qualifying judgement ‘marked’ vs. ‘unmarked’ depends on the sociolinguistic and areal background chosen for comparison. This fact becomes the clearer, the more “fine-grained” analyses and criteria are required to clarify the respective properties – and the less they can be backed up by rough statistical analyses as those which were conducted, e.g. by Nichols (1992). These are the modifications we have to keep in mind when doing population linguistics on a geographical “micro-scale”. In addition, I want to demonstrate the primary importance of language contact for explaining structural convergence especially in small areas. Language contact may be even more important than “vernacular universals”. In this sense I will take issue with Chambers’ reservations against the role of language contact in structural change. Besides this, it is often difficult to distinguish the role of contact from the role of such cognitively motivated universals. For the present purposes we get the following gradation of areal comparison:

Population linguistics on a micro-scale global (e.g. Nichols 1992) Baltic Area (intersect) > Belarusian borderlands.

501

> (Northern) Eurasia > SAE/CircumBaltic-Slavic contact area (BSA) > Lithuanian-

Figure 1. The successive inclusion of areas from the viewpoint of linguistic variation

SAE and the Circum-Baltic Area can be opposed to each other (as well as each of them to, e.g. the Balkan Sprachbund), but we have to keep in mind that they intersect to some extent (Swedish, Danish, Low German, standard Polish). The BSA practically comprises the south-eastern part of the Circum-Baltic Area. Its major contact varieties all belong to East Slavic, Polish (a West Slavic language) and Baltic. As for East Slavic, beside Belarusian dialects, the archaic dialect of the Russian-speaking Old Believers should be named here; as for Polish, we will be concerned with the northern variety of the so-called polszczyzna kresowa (PolKres), especially its rural (sub)varieties which were built up on a Belarusian and 3 Lithuanian substratum during the last 150 years. On the northern edges of the BSA, Balto-Finnic languages have also played a considerable role in contacts (cf. the relevant articles in Dahl and Koptjevskaja-Tamm 2001). Finally, the Lithuanian-Belarusian borderlands (with Latgalia in the northeast) can be regarded as the central part of the BSA; see Map 1 at the end of the paper. The main part of this paper, section 2, addresses the problem of how to find an adequate scale and set of criteria for an assessment of linguistic properties in a small area, namely, the Lithuanian-Belarusian borderlands. I will illustrate this point on the basis of some morphosyntactic features chosen from local varieties of that region. Here East Slavic and Polish dialects exhibit features which are remarkable from the point of view of standard Russian and Polish, whereas they are rather usual in that area – or even expected as convergences based on multilingualism, language shift, etc. in the eastern part of the Circum-Baltic Area. 2.

Case studies

2.1. The functional range of the reflexive marker All Slavic and Baltic languages of the area have reflexive markers that show a wide range of functions, and this range appears to be higher than the typological average; it is also higher than in other European languages

502

Björn Wiemer

with polyfunctional reflexive markers (e.g. German, Scandinavian or French and other Romance languages). The varieties of the BSA differ in the expression format of the reflexive marker: East Slavic and Latvian have an agglutinative postfix (ESlav. {s’a}, Latv. {s}), Polish (including PolKres) has an enclitic (siĊ), Lithuanian has the morpheme {si} which (at least in the standard and most of its dialects) is intermediate between clitic and agglutinative status in that it moves between the postfix position and the position immediately previous to the stem, depending on whether the stem has at least one prefix (including the negation prefix {ne}). However, these morphotactic differences are irrelevant for the comparison of the functional ranges of these markers. The point I want to make is the following: although from a typological perspective all the varieties we are analyzing here are characterized by highly polyfunctional reflexive markers, Lithuanian and the Slavic varieties each show some remarkable resistance to certain kinds of functions which a spreading reflexive marker otherwise acquires easily. Lithuanian is very productive in the reflexive-benefactive function (“transitive reflexives”, e.g. nusipirkti ledǐ ‘to buy oneself some ice-cream’; cf. Geniušienơ 1987: 126–133, 290–296), but does not use the reflexive marker for the passive, although we can encounter the so-called ‘modal passive’ (e.g. Batai gerai nešiojasi ‘The shoes wear well’; cf. Geniušienơ 1987: 109–118, 261–266). Conversely, East Slavic uses its reflexive marker for a passive, but it does not use it as reflexive-benefactive marker. Standard Polish abolished “reflexive passives” somewhere in the 19th century (in PolKres the reflexive marker is still, though rarely, used as a passive marker, sustained by East Slavic). One might suggest that in the Slavic varieties at stake here the reflexive marker is probably not suitable as a marker of reflexivebenefactive, because this function is fulfilled by an unbound pronoun in the dative (Russ. sebe, Pol. sobie), whereas in East Baltic a morphological distinction between accusative-genitive and dative forms does not exist. However, morphological case distinctions do not constitute an ultimate obstacle for this function to be adopted via calquing; we will see this immediately below. As far as the functional distribution of the reflexive marker in the varieties of the BSA is concerned, we arrive at a clear-cut area-internal distinction. Now, given appropriate contact conditions the respective “barriers“ can be overcome. On the one hand, some Lithuanian dialects whose speakers were in contact with speakers of Belarusian for the last few

Population linguistics on a micro-scale

503

centuries show an extension of their reflexive marker to the function of a passive; cf. (1)

Javai jau veža-si Ƴ ten already transport.PRS.3.RM in(to) there corn.NOM ‘The corn is being transported there already’ (Garliava) (quoted from Jakulienơ 1968: 215).

(2)

toki maišai audza-si sacks.NOM sew.PRS.3.RM such.NOM ‘such sacks are sewn’ (quoted from Kardelytơ 1975: 79).

On the other hand, some Belarusian and/or PolKres speakers who have been in intense contact with Lithuanian speakers (or who have recently changed from Lithuanian to Belarusian) apply “their” reflexive marker in benefactive function; cf., for instance, Br. prynjasusja xleba ‘I will bring myself some bread’, which is a morphological calque from Lith. atsinešiu duonos ‘id.’. Cf. Sudnik (1974), who adds that she came across analogous examples in PolKres. Such an example can be quoted from Maslennikova 4 (1973b: 77): przyniosáam s’e mleka ‘I (have) brought myself some milk’. Although such instances are very rare, they demonstrate that features or functions which are marked (i.e. unusual or even regarded as grammatically impossible) in some particular varieties of a small contact area, can nonetheless be encountered. The transitive reflexives of Belarusian and Polish varieties are marked if compared to the broader Slavic-speaking background. Conversely, occasional examples of a reflexive passive in the south-east Lithuanian varieties are marked with regard to the whole rest of Lithuanian varieties. Whether all these converging properties in Lithuanian and Slavic varieties are marked on a European or even world-wide background, is another question, which still requires special investigation. Whatever might be the answer to this question, for the characterization of the particular contact situation and its structural repercussions in the involved varieties it is of no big significance. Notice, furthermore, that only within a small area and with the aid of a detailed taxonomy of reflexivemarker-functions can these contact-induced differences between (sub)varieties of that area be noticed and described. The following two cases are likewise meant to illustrate analogous changes via intense contact (bi- or multilingualism). In both cases the trigger for the change is thus the same as with the functional expansion of

504

Björn Wiemer

reflexive markers discussed above. However, the next case (2.2) illustrates not an expansion of functions, but a reduction of a morphological distinction. Besides this, the next two cases differ from one another in the typological and/or cognitive markedness of the result, the first “confirming” this markedness (2.2), the second running counter to it (2.3). 2.2. Loss of number distinction in the 3rd person of verbs The SG:PL-distinction in the 3rd person of finite verbs has been loosened in the same Belarusian variety in which transitive reflexives can be encountered; i.e. the form of the singular may be used “instead of” the plural (cf. Sudnik 1974). This reminds us of Chambers’ “default singulars”, which he included into his list of vernacular universals (see section 1). Superficially, the Baltic and Belarusian cases look very much alike. However, Chambers implies that vernacular universals manifest themselves spontaneously, and he argues against language contact (the “diffusionist hypothesis”) as a factor general enough to explain the ubiquity of properties of English vernaculars (and of vernaculars in general), inter alia of “default singulars” (Chambers 22003: 266–269; this volume). Convincing as Chambers’ argument may seem at first sight, the decisive role of language contact cannot be dismissed in some particular cases. And it is in this respect where the partly converging BSA-varieties differ from each other and from the English vernaculars discussed by Chambers. On the one hand, the dialectal Belarusian “default singular” is to be considered as the consequence of a clear Lithuanian substratum, since it is only Baltic which has never had a number distinction in the 3rd person during the thousands of years for which we can speak of contacts between Baltic and Slavic. It is thus a very recent innovation, which arose under intense contact conditions, and it has the status of no more than a facultative option. On the other hand, the lack of a SG:PL-distinction in Baltic is of prehistoric age, regardless of whether we consider it to be an innovation with respect to some Indo-European precursor or an archaism. These two cases with basically identical “output” belong to totally different periods, and we can only speculate whether the East Baltic reduction of the paradigm or the retention of an archaism was itself the result of language 5 contact, too. At any rate, their paths of development are most probably not comparable to “default singulars” in English vernaculars.

Population linguistics on a micro-scale

505

Now, from a typological viewpoint it is plausible to assume that the singular is the unmarked member in the 3rd person of verbs; this assumption is basically confirmed by the data, although there are some counter-examples (cf. Corbett 2000: 185–186). We may further suppose that the unmarked status of the singular in verbs mirrors basicness in cognitive terms. This seems to be also Chambers’ assumption (see above). Given such premises, the loosening of the SG:PL-distinction in bilingual Belarusian-Lithuanian populations would demonstrate that the Belarusian variety in which marking of the plural in the 3rd person has become optional conforms to a cognitively based (and hence typologically expectable) default of the singular. Nonetheless, it does not seem plausible to assume that the Belarusian dialect referred to by Sudnik (1974) was on its way towards losing its 3rd person plural mainly because of some cognitive predispositions and typological unmarkedness of the singular, for lack of a SG:PL-distinction in the 3rd person of verbs is a marked feature on the background of both the Circum-Baltic Area and the SAE, and it is a highly marked feature for Slavic languages. So why should speakers of just one Slavic variety give up a categorial distinction which belongs to central features in its morphology and which is ubiquitous in the vernaculars of the investigated area, except for Lithuanian and Latvian? Obviously, the decisive factor is not cognitive and typological markedness relations as such, but thorough contact with Lithuanian (i.e. bilingualism of speakers). Only the latter supplies the necessary trigger for this change. 2.3. ‘behind’ as preposition in comparative constructions The next example illustrates the same point against Chambers’ reservations concerning language contact as an important factor in the spread of innovations in vernaculars. But this time the example stands for a typologically highly marked feature, which has nonetheless spread widely from Lithuanian (or remained as its substratum) into Slavic contact varieties. According to Haspelmath (2001: 1499), one of the features of SAElanguages is the use of a particle in comparative constructions (such as Engl. than, Germ. als). In Eurasia another frequent way of coding the base for comparison is the “separative” technique, i.e. the use of the ablative case or a preposition meaning ‘from’ (plus appropriate case). Both techniques can be encountered in standard Polish and Russian, compare:

506

Björn Wiemer

Pol. Brat jest wiĊk-sz-y niĪ siostr-a ‘The brother is larger than the sister’ (with particle), Brat jest wiĊk-sz-y od siostr-y ‘id.’ (“separative” technique) and Russ. Brat bol’-še, þem sestr-a ‘id.’ (with particle), Brat bol’-še sestr-y 6 ‘id.’ (“separative” technique with the bare genitive ). Contrary to these, the use of a preposition meaning ‘behind’ is typologically rare, but it is a 7 frequent option in Lithuanian ; compare Lith. Brolis yra did-esn-is už seserƳ.ACC ‘id.’ (lit. ‘The brother is bigger behind the sister’). In Latvian such a function has become obsolete, but aiz ‘behind’ used for comparison is attested in Latvian folk songs at least until the beginning of the 20th century (cf. Ơndzelin 1905: 19). It is this technique which has been calqued from Lithuanian in Belarusian and Polish dialects; compare here an example from local Polish (PolKres): (3)

u on máód-sz-y za mnie a behind me.ACC and he young.COMP.NOM.SG.M ‘and he is younger than me [lit. ‘behind me’]’ (Buivydžiai, Lithuania, 1987; quoted from: Grek-Pabisowa and Maryniakowa 1999: 136).

za + ACC in comparative constructions is encountered also in Russian dialect regions which are contiguous with the BSA, namely, in dialects around Pskov (cf. SRNG 1972/IX: 236) and around Smolensk (cf. SSmolG 1985/IV: 32). It is attested even in some Ukrainian dialects, which may be supposed to have calqued this coding technique from Lithuanian indirectly (cf. ýekmonas 2001: 119–120; Koptjevskaja-Tamm and Wälchli 2001: 684). This is a clear case of areal convergence which (for whatever reasons) runs counter to markedness patterns at least on a northern Eurasian scale. We see that a highly marked feature can be taken over under conditions of intense language contact (bilingualism). Since a marked feature is adopted, this case is opposite to the loss of number distinctions (see 2.2), but it occurs in contact varieties of roughly the same area. Again, the “odd ones out” in terms of the BSA are standard Russian and Polish. But this time these standard languages conform more closely to the pattern not only in SAE, but also in Eurasia; actually they show parallel use of the unmarked patterns of both areas.

Population linguistics on a micro-scale

507

2.4. Nominal inflection (gender and object marking) Since the existence of nominal inflection is by itself a phenomenon with very uneven distribution in macroareal terms, it does not make very much sense to consider peculiarities of noun inflection in varieties of the BalticSlavic region against a world-wide background, but it does make sense to consider noun inflection in a European (and broader, a Eurasian) and, even more so, in an inner-Slavic context. I will pick out two phenomena which are typical for the Slavic varieties of the BSA. They differ not only with respect to their role in the grammar, but also with respect to the degree with which they have resisted influence from surrounding contact varieties. 2.4.1. Accusative ending = genitive ending as a marker of animacy and the category of ‘virility’ Animacy is to be understood broadly here, as it includes also more specific distinctions like those based on [r human] or [r virile]; if necessary I will specify the range of this feature. Animacy effects manifest themselves in object marking and in plural marking, either by the choice among two (or more) alternative explicit markers or by the (optional or obligatory) use of an explicit marker (vs. its lack). Cf. standard Russian examples for the marking of (in)animate object NPs: ja vižu starš-ego brat-a (ACC = GEN) ‘I see the/my older brother’ vs. ja vižu star-yj dom-‡ (ACC = NOM) ‘I see a/the old house’ (the nouns brat ‘brother’ and dom ‘house’ both are masculine), and standard Polish examples for the marking of (non)virile plural NPs: radoĞn-i studenc-i przysz-l-i ‘The joyful students have come’ (virile) vs. radosn-e list-y, dziewczyn-y, wnuk-i przysz-á-y ‘The joyful letters, girls, grandchildren have come’ (non-virile). Animacy reflected in object and in plural marking can be considered as logically independent phenomena, though both may be combined or may interact. This fact is demonstrated by the history of all northern Slavic languages and the present-day situation in West Slavic. I am unaware of whether this combination is unique on a world-wide scale; relevant typological studies do not seem to exist (M. Haspelmath, p.c.). At any rate, this combination can be regarded as a (if not the) distinctive areal feature of north (= West + East) Slavic against a European background. From a typological perspective, what is peculiar for Slavic as a whole is that it is 8 the genitive ending which replaces the accusative ending. The roots of this

508

Björn Wiemer

inner-Slavic innovation have to be searched for in the prehistoric, commonSlavic period (i.e. before the 9th c. AD), when the beginnings of the ACC=GEN-rule for masculine singular nouns emerged. Later on Slavic languages developed different patterns in the plural, with the most outstanding peculiarities in standard Polish and Czech. The Polish and the East Slavic systems will be considered more closely below. In general, differential object marking is rarely based on animacy alone, but interrelated with or even dominated by definiteness. According to the systematic survey given by Bossong (1998), the contemporary north Slavic languages are virtually the only language group in Europe in which 9 differential object marking is entirely determined by animacy, its diachronic connection with definiteness (to be observed still in Old Church Slavonic) has become obscure altogether. North Slavic languages display the highest degree of grammaticalization of differential object marking also in terms of obligatoriness, determined by membership in certain classes on the animacy hierarchy (cf. Bossong 1998: 211). Bossong (1998: 254–255) comes to the conclusion that differential object marking is an areal phenomenon, with the Slavic languages inside the broad zone in which this marking has spread, and the Baltic languages (as well as the Germanic and Celtic ones) outside of it. The “demarcation line” runs thus through the midst of the BSA. Baltic languages do not share with Slavic its innovation of the ACC=GEN-pattern, nor do they have any other grammatical device of distinguishing animate, personal or virile nouns from the rest. In sum, whereas among the languages of the world differential object marking is a very common phenomenon (Bossong 1998: 207), nothing definite can be said about the world-wide distribution (and possible areal or genetic biases) of animacy effects alone, since more often than not they appear to cooccur with (or be overshadowed by) effects arising from definiteness. Furthermore, it is by no means always the case that such a distinction goes beyond declensional paradigms and manifests itself in 10 clausal syntax. Against this background, Slavic languages are remarkable for their technique of marking animacy in direct object position: the ending of the genitive (of the respective declensional paradigm) has replaced the original accusative ending. Slavic languages, however, differ considerably as to whether they show the ACC=GEN-pattern only in the singular or in the plural, too, as well as to the gender and declensional classes involved, and whether there is a further distinction relevant to the gender system. Without going into details of the often complex distribution of this pattern in the 11 Slavic languages and of its even more complex history, I want to restrict

Population linguistics on a micro-scale

509

myself to a brief comparison of the systems of standard Polish and standard Russian and have a look at what has been going on in non-standard varieties of Polish and East Slavic in the Baltic-Slavic region. Let me add here that the Baltic languages do not share with Slavic its innovation of the ACC=GEN-pattern, nor do they have any other grammatical device of distinguishing animate, personal or virile nouns from the rest. Russian (both standard and almost all of the dialects) and other East Slavic dialects which are geographically close to the Russian-speaking territory have the ACC=GEN-pattern also in the plural, regardless of the gender and declensional class to which the noun belongs (cf. Avanesov/Orlova 21965: 97; Pen’kovskij 1975: 153–154); for instance, deti ljubjat svo-ix roditel-ej, mater-ej ‘the children love their parents, mothers’ vs. uþeniki ljubjat letn-ie kanikul-y ‘the pupils like summer holidays’. The situation in Polish is more complicated. Standard Polish has an opposition of ‘virile : non-virile’ nouns (Pol. kategoria mĊskoosobowa): somewhat simplifying the picture – largely with regard to morphonological alternations at the end of the stem (caused by the vowel following it) – the nominative plural of nouns denoting adult male persons (‘virile’) is marked by the ending {i} (with softening of the stem) or {ov’e}, whereas the rest of nouns takes {y}, {e} (for masc. and fem.) or {a} (for neuter nouns). The same nouns which are marked this way as ‘virile’ show the ACC=GENpattern also in the plural, while non-virile nouns have ACC-endings which are identical with the nominative (ACC=NOM); for instance, dzieci kochają swo-ich rodzic-ów (vs. swo-je matk-i) ‘the childen love their parents’ (vs. ‘their mothers’). In the standard language, virility is marked not only on the noun, but on all attributes that agree with it (adjectives, demonstratives), on anaphoric pronouns (virile oni vs. non-virile one ‘they’) and on past tense 12 forms of the verb (virile {li} vs. non-virile {áy}). See the examples at the beginning of this subsection. From the diachronic point of view we can say that, by and large, the contemporary Russian system represents the most advanced stage in terms of lexical expansion of the ACC=GEN-pattern in both numbers (first in the singular, then in the plural), whereas in standard Polish an expansion to the plural has not lasted. The evolution of these gender distinctions is intimately connected to a restructuring process of the inherited system of nominal declensions, which I will not dwell on here. Following Corbett (1991: 161–168) we should classify these distinctions as ‘subgenders’. As for Belarusian, only its north-eastern dialects show the ACC=GENpattern in the plural more or less consistently also with respect to nouns

510

Björn Wiemer

denoting animals (= Russian pattern); cf. Avanesauң et al. (1969: Map 20, 46). The (south)western dialect group shows patterns which are either totally consistent with standard Polish or are intermediate between the Russian and the standard Polish pattern (e.g. by marking ACC=GEN in nouns denoting women, but not animals, and lacking a declensional distinction in NOM.PL). In turn, investigators of PolKres have often noticed a lack of the ‘virile : non-virile’-opposition, manifested both by the unification of the nominative plural ending {i, y} and the ACC=GEN-pattern for all genders (cf., for instance, Grek-Pabisowa and Maryniakowa 1999: 13 34–35; KaraĞ 2001: 176–178; Masojü 2001: 42–46). This patterning must be understood as a result of an East Slavic substratum, whereas another part of the variation can be explained by the orientation towards standard Polish, which for many (if not most of) the PolKres speakers constitutes a particularly prestigious variety. If we now compare the coding patterns in the standard varieties and the Slavic vernaculars of the Baltic-Slavic contact region, Russia and Poland, we by and large notice a geographical continuum: the farther to the east (and north), the more the coding pattern confirms to standard Russian; the pattern totally corresponding to standard Polish can be encountered only in dialects of western and southern Poland (Silesia and Wielkopolska, large parts of Maáopolska). However, in the prevailing number of Polish dialects the endings of the nominative plural and the correlated agreement patterns of nouns, adjectives, demonstratives and past tense forms of verbs do not allow for any (sub)gender distinctions in the plural (cf. again Dejna 21993: 232–234 and Map 70 therein). In other words: these Polish dialects as well 14 as the predominant number of East Slavic vernaculars have levelled out the role of grammatical number in their animacy distinctions; in the standard Polish (as well as in the Czech) animacy distinction the role of number has not been neutralized, although it has remained under domination of the role of the direct object case (accusative). We may now summarize: The variety which is the odd one out happens to be standard Polish (together with dialects in western and southern Poland), not only with regard to its East Slavic neighbors (and “sisters”), but also in a broader European context. Among the languages relevant for the BSA it is the only one with a clear-cut virility distinction. It shares this feature with Czech and Upper Sorbian. It could therefore be called a “specialty” of the western edges of North Slavic (excluding Slovak and the eastern part of Poland). At least in Europe this situation seems to be unique, whereas the East Slavic

Population linguistics on a micro-scale

511

varieties and the Polish varieties bordering on East Slavic territory or spoken “inside” of it (i.e. PolKres) just behave more normally in broader areal terms. Whether they do so also in global typological (macro-areal) terms remains to be investigated. With respect to North Slavic and the BSA, the distribution of the patterns allowing for subgenders displays an areal continuum from the (south)west to the (north)east. The predominant pattern is ACC=GEN also in the plural regardless of the gender (and largely also declensional class) of the noun (= standard Russian pattern). It is also the prevailing pattern for Slavic varieties in the BSA. To this we should add the following. The category of animacy has not in the slightest been affected by contacts with Baltic (Lithuanian, Latvian). Nor have, in turn, Lithuanian or Latvian varieties in tight contact with (East) Slavic developed any kind of differential object markers (based on animacy or definiteness). We may thus conclude that marking of animacy (virility etc.) is a phenomenon that has remained entirely restricted to innerSlavic mutual influences; beyond these it turns out to be remarkably resistant to language contact. 2.4.2. Disappearance or weakening of neuter gender Contrary to the animacy and virility distinctions – i.e. the rise of subgenders – discussed in the preceding subsection, the loss of the neuter gender (or the tendency to lose it) gives an example of a change in nominal inflection (and of agreement patterns connected with it) which is due to contacts crossing the Slavic-Baltic borderline. The reduction of a system originally composed of three genders (understood as agreement classes) is known for Scandinavian dialects as well. It might thus be considered as a (weak) areal feature concerning different varieties of most of the IndoEuropean languages in the Circum-Baltic region. Anyway, the reasons for this reduction are different and concrete results are as a rule not directly comparable; in particular, the Scandinavian cases cannot really be equalled to what we can observe in Slavic varieties (cf. Koptjevskaja-Tamm and Wälchli 2001: 694–695). Latvian lost its neuter gender a long time ago and most varieties of Lithuanian have retained some traits of it only in demonstrative pronouns (e.g. Lith. tai, tatai vs. Lith. tas.NOM.SG.M, ta.NOM.SG.F ‘this, that’),

512

Björn Wiemer

adjectives (forms ending in unstressed {a} or in stressed short {ù}) and certain participles (all ending in unstressed {a}). With respect to the latter two word classes we can speak of a target gender without a controller gender (in the sense of Corbett 1991), since “neuter” adjectives and participles are practically exclusively used as predicates of subjectless clauses or of clauses with tai ‘this’ (see above) as a formal subject. Slavic varieties in the vicinity of Lithuanian vernaculars show a considerable degree of decay of the neuter as a controller gender, too. This has become manifest even in the codified variety of Belarusian, which has {i, y} as allomorphs for the NOM.PL of neuter nouns (cf. Podlužnyj 1997: 194), where standard Russian and Polish (and most of their regional varieties) have the “inherited” ending {a, ’a}. This corresponds to Belarusian vernaculars, the Russian of Old Believers in the region under discussion here (for Old Believers’ speech cf. ýekmonas 2001: 125), to PolKres (cf., inter alia, Maslennikova 1973a: 86) as well as to the Russian prostoreþie, an agglomerate of urban dialect mixing which I will return to below. Two reasons for the decay of the neuter in Belarusian and PolKres have been discussed in the first place: (i) substratum (or simply interference) from Lithuanian, which has only a two-gender system (see above), (ii) akan’e, i.e. the pronunciation of phonological /o/ and /e/ as [a] in unstressed positions. The substratum hypothesis is difficult to dismiss, while the “phonetic” explanation cannot be the only one. This can already be seen from the fact that the Belarusian nominative plural of neuter nouns ends in {i, y}, i.e. in vowels which are by no means the result of akan’e, but must have appeared by analogy to the endings of the other two genders, which are represented by a much larger number of nouns. There is, of course, the possibility of a spontaneous process, not triggered by contact with Lithuanian showing a system of two controller genders. In favour of such a possibility it might be mentioned that various Russian dialects not contiguous to the BSA show tendencies of weakening the paradigm of neuter nouns, too: in such cases not only the nouns, but even many attributive adjectives take on feminine endings. As a rule, this occurs only in the nominative and the accusative, whereas in the oblique cases agreement normally conforms to the usual paradigm of neuters; but these in turn coincide with the masculine paradigm. This is true in particular for south Russian dialects, quite far away from the BSA. The mechanism seems to be the same as in PolKres (see below), namely, akan’e plus analogy, analogy being relevant for those cases where the ending {o} in neuter nouns is stressed and akan’e does not apply. We thus get not only

Population linguistics on a micro-scale

513

cases like bol’šája stáda ‘large flock’ (standard bol’šóe stádo, pronounced [stáda]), but also bol’šája seló ‘large village’ (instead of bol’šóe seló); cf. Avanesov and Orlova (21965: 100). The weakening of neuter nouns as a separate agreement class is to be observed also in urban vernaculars, which Russian scholars usually subsume under the term prostoreþie (cf. Zemskaja and Kitajgorodskaja 1984: 72–73). By and large, prostoreþie arose as the result of dialect mixing in towns, when speakers of different dialects settled there and their non-standard dialects began to converge. It has no codified norm, but in some respects a more regular morphology than the standard language. Among other things, declensional endings are more unified, there are fewer exclusions (e.g. indeclinables), and the case-number markings of nouns are in many cases not distinguished according to their gender and declensional class. Basically the same process can be observed in PolKres, which arose on a Belarusian and/or Lithuanian substratum. Anyway, in this case the process seems to have proceeded further than in Belarusian and Russian dialects, as it shows more symptoms of morphologization. Investigators both of rural PolKres and of a PolKres variety spoken in Vilnius (which stands a little closer to standard Polish) give examples in which a feminine third person pronoun (ona) is used with reference to a noun corresponding to a neuter in standard Polish; cf. (4a–b) for a rural variety (in the late 15 1960s), (5a–b) for Polish spoken in Vilnius (recorded in the early 1990s): (4)

a.

wáokna ln’ana – of_flax.NOM.SG.F fibre.NOM.SG ona n’e papsuji s’e 3.NOM.SG.F NEG decay.FUT.3.SG RM ‘The flax fibre – it [lit. she] will not decay’.

b.

zrob’ona okna made.NOM.SG.F window.NOM.SG i zasufka w n’ej and bolt.NOM.SG.F in 3.LOC.SG.F ‘(There is) a window made, and a bolt in it [lit. her]’ (both examples from Maslennikova 1973b: 90).

514

(5)

Björn Wiemer

a.

Miasto odrzuciáo nasze propozycje (...) proposal.ACC.PL city.NOM.SG.N reject.PAST.3.SG.N our ona chce jedenastu tysiĊcy (hektarów) it [lit. she] want.PRS.3.SG eleven.GEN thousand.GEN.PL ‘The city has rejected our proposals (...) it [lit. she] wants eleven thousand (hectars).’

b.

kaĪda apteka powinna each.NOM.SG.F pharmacy.NOM.SG.F obliged.NOM.SG.F posiadaü lekarstwo, bynajmniej jedną one.ACC.SG.F possess.INF medicine.ACC.SG.N at_least z grupy odnoĞnie do choroby illness from group relative.ADV to ‘Each pharmacy is obliged to have (the proper) medicine, one (sort) from (each) group concerning the illness’ (both examples from Masojü 2001: 40).

The only difference between the rural and the Vilnius variety is that in the rural variety the respective nouns show akan’e more consistently, whereas in the Vilnius variety respective nouns often are pronounced according to the Polish standard norm (i.e. without akan’e). The crucial fact is that nonetheless even in the latter case anaphoric pronouns and quantifiers are inflected like feminines, not like neuters. Whereas it might still be argued that ona in (5a) can be analysed as the neuter pronoun óno with akan’e (rendered as [óna]), such an interpretation is altogether excluded for cases like the quantifier jedną (phonetically [LHGQR˙]) ‘one (of)’ in the ACC.SG; if akan’e were a factor, we should expect jédna (< jédno). Thus jedną is doubtlessly a feminine form of the quantifier. Likewise the locative pronoun (w) niej in (4b) cannot be the result of akan’e of a neuter pronoun, because the locative of ono would be (w) nim. Consequently, akan’e in neuter nouns has led to a partial restructuring of the gender system in PolKres. Forms of the (standard Polish) neuter paradigm exist alongside forms of the feminine paradigm even in oblique cases, first of all the genitive; cf. ta wiadra.NOM.SG ‘that bucket’ (instead of to wiadro) with the GEN.SG tej wiadry (o FEM) // tego wiadra (o MASC or NEUT) or the GEN.SG form Īadnej sáowy ‘no word’ (instead of Īadnego sáowa) and the DAT.SG form kaĪdej oknie ‘every window’ (instead of kaĪdemu oknu), 16 attested in different subvarieties of PolKres. We have to admit that the whole paradigm of neuter nouns has not vanished as such and that there has

Population linguistics on a micro-scale

515

always been a great deal of (virtually free) variation of forms in virtually all grammatical cases. Nonetheless the paradigmatic status of neuter nouns is severely delimited by the morphological consequences of akan’e, and in many subvarieties of PolKres no morphological markers have remained which would be unique to the neuter: the endings of nouns and attributive words conform either to the masculine or to the feminine paradigm (for more details cf., first of all, Maslennikova 1973b: 85, 88, 94). From a broader areal perspective this process is by no means unique: we can observe it even in other Slavic varieties, both diastratically and dialectally. In Slavic varieties of the BSA it is evidently sustained and magnified by Lithuanian influence (substratum), even though the stronger assertion that Lithuanian influence has triggered the decay of the neuter cannot ultimately be proven. The Lithuanian-Belarusian contact region (with PolKres largely “in-between”) itself forms a kind of transitional zone 17 between the genderless Balto-Finnic languages (to the north) and West Slavic with their more or less stable three-gender systems. 2.5. Subsuming the peculiarities in nominal declension The facts from subsections 2.2–2.4 are summarized in the following table. The results from the discussion on the functional range of reflexive markers (see 2.1) are not very suitable for being presented in the form of a table; for this reason I have not included them here. Table 1. Peculiarities of varieties in the BSA (compared with standard Polish and Russian varieties outside the BSA) distinction of SG vs. PL in 3rd person of verbs

‘behind’ used ACC=GEN in comparative (with semantic constructions distinction implied) (optional)

tendency to lose neuter gender

standard Polish

+

–

+ [r virile]

–

standard Russian

+

–

+ [r animate]

–

PolKres

+

+

+ [r animate]

+

516

Björn Wiemer distinction of SG vs. PL in 3rd person of verbs

‘behind’ used ACC=GEN in comparative (with semantic constructions distinction implied) (optional)

tendency to lose neuter gender

Old Believers’ Russian

+

+

+ [r animate]

+

Russian vernaculars outside the BSA

+

(+) (only in neighboring regions)

+ [r animate]

(+), not in all regions

standardized Belarusian

+

–

+ [r animate]

+

northern Belarusian dialects

– (only in one variety)

+

+ + (different ranges from [r animate] to [r virile])

Lithuanian

–

+

–

Latvian

–

–

–

+ (total loss) + (total loss without traits)

3. Lessons to be learnt from small-scale contact areas For many of the surveyed features it would hardly be adequate to even try to make judgments concerning their markedness on a pan-European, let alone a world-wide background. One reason is that some of these features are rather specific to language groups in either genetic terms (e.g. animacy expressed by a ACC=GEN-pattern) or in terms of morphological typology (e.g. weakening and loss of the neuter gender, which presupposes gender as a grammatical category, or so-called syncretisms of cases, presupposing case paradigms). Thus, some coding techniques which show an interesting variation in Europe and the particular subarea chosen here, may altogether be absent if we look for some more global tendencies. This is a reason why the method developed by Nichols (1992) has its limits if applied to small areas and, first and foremost, to structural properties on a much more fine-grained level than the properties analyzed

Population linguistics on a micro-scale

517

in her book. How can one measure features which are not present in a larger amount of languages and which to a large extent depend on the morphological type of the language (variety), but which are nonetheless crucial indicators of structural variation among language varieties of a common morphological type in a smaller area? In fact, Nichols does account for morphological complexity. But on a global scale this parameter must remain very coarse, since it cannot even include prepositional phrases, for many languages simply lack this syntactic category (Nichols 1992: 64). Even more important is the following point: a comparison of varieties (whether genetically associated or not) of a particular contact area in which morphosyntactic properties already converge to a certain degree does not require a measurement of (morphological) complexity as such (provided an adequate measurement of such a kind can be found at all!). Rather what is needed is a quite subtle functional analysis of the particular means (of probably equal complexity) which are exploited for the same grammatical and, in the last resort, communicative purpose. Recall, for instance, prepositions meaning ‘behind’, which are used in varieties of the BSA in comparative constructions. The usage of these prepositions is rather “exotic” if looked at from a broader Eurasian perspective. But this formation is by no means more or less complex than the “particle technique” of typical SAE-languages or the “separative technique” of languages in northern Eurasia. Consider furthermore the issue made above (2.1) with reflexive markers: what is pertinent in this case is the functional extension of derived diatheses; this extension can be established only on the basis of a distributional analysis which has far more facets and subtleties than the simple ‘causative : inchoative’ distinction (cf. the taxonomic approach in Geniušienơ 1987). We can count the number of types of derived diatheses and idiosyncrasies of verbs with a reflexive marker in the respective varieties, but the measurement we achieve thereby aims at a complex functional variation, not just a complexity on the morphosyntactic surface. A similar remark may be made with respect to processes of expansion and reduction of gender patterns or number and case markers discussed in section 2. It is exactly such functional differences which should be at stake if we are interested in the description and explanation of the internal variation within contact zones and comparisons with broader areas which they are part of. Finally, what about Chambers’ reservation against contact as a factor responsible for the rise and spread of vernacular innovations (the so-called “diffusionist hypothesis”)? The discussion of some of the properties in

518

Björn Wiemer

section 2 should have demonstrated that Chambers’ negative stand, wellmotivated as it appears in particular cases, has to be relativized. Two types of cases suggest that he seems to have thrown out the baby with the bathwater: (i) Features that are highly marked typologically and against the background of continent-sized areas may nonetheless spread into a large number of contact varieties and thereby become the unmarked case for the pertinent small area; e.g. the use of the preposition meaning ‘behind’ in comparative constructions. (ii) Features which are highly marked in areal and genetic terms, but correspond to “vernacular universals”, are adopted by speakers of contact varieties only under conditions of intensive bi- or multilingualism; for instance, “default singulars” in the BSA supply a very good illustration of this point. These observations strongly suggest that internal, structural reasons of linguistic change and external, contactinduced motives for the spread of linguistic innovations should be counterbalanced against one another in each particular case. Contact-induced reasons for change cannot be dismissed in a “wholesale” manner, only because the diffusionist hypothesis happens to be ridiculous if applied to the small number of cases considered by Chambers. Since I cannot continue this discussion here, let me stress that this should not mean to deny (or neglect) the cognitive and communicative basis of many of the structural features we encounter in vernaculars (as well as in first- and second-language acquisition and in the formation of pidgins and creoles). Chambers, however, seems to have forgotten a crucial communicative tenet propagated by himself: the roots of dialectal differences, and thus also the spread of innovations and other changes, are not so much of a structural nature, but are rather to be sought for in social attitudes, prestige and other factors underlying people’s everyday ideological orientation (cf., inter alia, Chambers 22003: 254 and passim). With this in mind, we should assume that in order for vernacular universals to become manifest, they must first overcome (or not run counter to) the fetters of attitudes and differences in social prestige. Since ideological stands (in the largest possible sense) and the varieties of their representatives are bound to historical and societal conditions, we cannot simply “skip” or neglect them in our theorizing about language change and the relationship between standard and vernacular varieties (in particular if we abide by Chambers’ own tenets).

Population linguistics on a micro-scale

519

4. Conclusions The aim of this article was to show what kinds of problems specialists in areal and contact linguistics have to cope with if they want to do justice to structural variation inside small contact regions, but with a comparative view towards larger areas and typological research. First of all, the (un)markedness of a particular feature is of a relative nature, since (un)markedness on a typological or broad areal background may occur to be a rather inadequate criterion for “measurement” in a small area (and populations in complex multilingual/multidialectal settings, respectively). If we want to do ‘population linguistics’ on a micro-scale and comprehend the crucial facts of (and reasons for) structural variation among the varieties involved, we, among other things, need a theory of diversity which allows for very fine-grained functional oppositions and criteria of functional expansion and reduction which are beyond (or “beneath”) the scope of normal typological research, but which are as a rule also neglected (or ignored) by dialectologists and specialists in contact linguistics, who typically deal with a very limited number of varieties. Another important issue which needs to be incorporated into a comprehensive theory of contact linguistics is the diastratic opposition ‘vernacular : standard’. But here we have to be careful, too. Neither can typological (un)markedness be a criterion in and of itself, because it results from a certain cognitive “equipment” and/or communicative requirements (in this respect it can be regarded to be epiphenomenal). Nor can structural properties of vernaculars, which probably derive from the same requirements and therefore tend to be universal, be claimed to supersede contact-induced changes in any case. The reason most probably is that speakers are guided not only by ease in communication and cognition, but they orient themselves also towards ways of speaking which they – for some reason or other – regard as prestigious and worth practicing in order to be socially successful. This orientation may not fit natural tendencies of vernacular speech and thus easily blur the power of the cognitive underpinnings of these tendencies. Together with this we have to account for the unifying effects of codification, school education, and writing traditions (not treated in this article). In the case of the BSA these factors are “distributed” unevenly among the languages involved. If we take all these dimensions together, we see how difficult a task it is to counter-balance them against each other. An enormous research program would be necessary to investigate them even in such a small (but hitherto

520

Björn Wiemer

neglected) contact area as the BSA, all the more if research is to be extended beyond the few cases discussed in section 2.

Map 1. Sociolinguistic situation and boundaries of Lithuania (Grumadienơ and Stundžia 1997: 1914)

Population linguistics on a micro-scale

521

Abbreviations of language varieties Br. ESlav. Lith. Latv. Pol. PolKres Russ.

Belarusian East Slavic Lithuanian Latvian Polish (standard) polszczyzna kresowa (northern variant) Russian (standard)

Acknowledgments I would like to thank Walter Breu and Martin Haspelmath, whose remarks on animacy effects allowed me to evade some shortcomings which I otherwise might have been fallen victim to in section 2.4.1. I also want to thank Axel Holvoet for his helpful comments on Latvian prepositions (2.3). I am especially indebted to Bernhard Wälchli for his invaluable remarks and discussions, which indirectly have entered into the present paper – although he might not notice it. Thanks also to Bridget Drinka, Ranko Matasoviü and Bernd Kortmann for their comments on a previous, longer version of this paper. Notes 1. 2. 3. 4. 5.

6.

Curly brackets will be used for units on the morphemic level, square brackets for phonetic, and slashes for phonological units. Somewhat misleadingly, Chambers calls this last feature “verbal adjectives”. For a comprehensive overview of the contact varieties relevant here, and their history, cf. Wiemer (2003). It is not clear whether s’e is etymologically to be regarded as the dative (< *sebČ) or the accusative form (< *sĊ). For the functional extension (via calquing or not) this is irrelevant. Hardly anybody has considered whether this common trait of the Baltic languages can be explained as the result of language contact, first of all with Balto-Finnic. To my knowledge, only Trost (1981: 83–84) has dwelt upon this question. This derives etymologically from the Indo-European ablative.

522 7. 8.

9. 10.

11. 12. 13. 14. 15. 16. 17.

Björn Wiemer In Lithuanian (at least the standard language, Lith. bendrinơ kalba) the use of particles (negu, nei) is possible, too. Bossong (1998: 244–245) rightly stresses that the morphological coincidence of the inherited endings of the accusative and the genitive in Balto-Finnic is the result of a merely phonetic process; it thus has nothing to do with the specifically Slavic type of animacy-based object-marking. In two other cases, namely, Ossetic (Iranian, spoken in the Causasus) and Mordvin (Ugro-Finnic, NE of the European part of Russia), differential object marking is based on a ACC=GEN-rule, which probably arose by Russian influence. However, the crucial factor in both cases is not animacy, but definiteness (Bossong 1998: 231–232, 242–243). A notable further exception is Romani, where it also is animacy alone which decides about the type of object marking (Bossong 1998: 234). Although Bossong does not mention this, we may assume Slavic contact-influence here. This is to exclude such (petrified) animacy effects as the distinction between weak and strong declensions in the history of German, whereby nouns denoting men or animals have kept their inherited distinction between nominative and accusative (e.g. der Fürst.NOM vs. den Fürst-en.ACC ‘duke’) and nouns denoting women or inanimates did not (e.g. der Haken.NOM vs. den Haken.ACC ‘hook’); cf. Breu (1988: 51). Similar distinctions of declensions and genders can be traced back to very ancient Indo-European oppositions of [r animacy] (Bossong 1998: 207–208; Breu 1988: 54). If we attempted to account for these facts as well, animacy would turn out to be a ubiquitous phenomenon in all subsystems of the grammar, hard to handle in a typological classification. For a comprehensive survey on the contemporary languages and Old Church Slavonic cf. Bossong (1998: 209–218). These forms in general show gender-number agreement with the subject in all Slavic languages. There is, of course, considerable variation (as it seems, free variation) between different case endings and NP-internal agreement patterns, which we need not go into here. Ukrainian differs from Russian only rather insignificantly, in that plural direct object nouns denoting animals may be marked alternatively by the ACC=GENending or by the ACC=NOM-ending (Bossong 1998: 216). The notation here and in the following has partially been adapted to conventional Polish orthography. For the data see, among others, Grek-Pabisowa and Maryniakowa (1999: 32– 34), KaraĞ (2001: 130–133), Maslennikova (1973b), Masojü (2001: 39–40). On the loss of gender altogether in Low Latvian dialects (situated outside the BSA) as an effect of Finnic substratum cf. Koptjevskaja-Tamm and Wälchli (2001: 695–697).

Population linguistics on a micro-scale

523

References Avanesauң, Ruben Ivanoviþ, Kandrat K. Atraxoviþ and Juzơfa F. Mackeviþ 1969 Linhvistyþnaja heahrafija i hrupou̧ka belaruskix havorak [Linguistic geography and location of Belarusian dialects]. Minsk: Navuka i tơxnika. Avanesov, Ruben Ivanoviþ and Varvara G. Orlova 2 1965 Russkaja dialektologija [Russian dialectology]. Moskva: Nauka. Bossong, Georg 1998 Le marquage différentiel de l’objet dans les langues d’Europe. In: Jack Feuillet (ed.), Actance et Valence dans les Langues de l’Europe [Actancy and valence in the languages of Europe], 193–258. Berlin/New York: Mouton de Gruyter. Breu, Walter 1988 Die grammatische Belebtheit als Genusgrammem. In: Vsevolod Setschkareff, Peter Rehder and Herta Schmid (eds.), Ars Philologica Slavica: Festschrift für Heinrich Kunstmann, 43–55. (Sagners slavistische Sammlung 15.) München: Sagner. Chambers, Jack K. 2000 Universal sources of the vernacular. Sociolinguistica 14: 11–15. 2 2003 Sociolinguistic Theory (Linguistic variation and its social significance). Oxford/Malden, MA: Blackwell. this vol. Dynamic typology and vernacular universals. Corbett, Greville G. 1991 Gender. Cambridge: Cambridge University Press. 2000 Number. Cambridge: Cambridge University Press. Croft, William 2 2003 Typology and Universals. Cambridge: Cambridge University Press. ýekmonas, Valerijus 2001 Russian varieties in the southeastern Baltic area: Rural dialects. In: Östen Dahl and Maria Koptjevskaja-Tamm (eds.), Circum-Baltic Languages, vol. 1: Past and Present; vol. 2: Grammar and Typology, 101–136. (Studies in Language Companion Series 54– 55.) Amsterdam/Philadelphia: Benjamins. Dahl, Östen and Maria Koptjevskaja-Tamm (eds.) 2001 Circum-Baltic Languages, vol. 1: Past and Present; vol. 2: Grammar and Typology. (Studies in Language Companion Series 54–55.) Amsterdam/Philadelphia: Benjamins. Dejna, Karol 2 1993 Dialekty polskie [Polish dialects]. Wrocáaw etc.: Ossolineum.

524

Björn Wiemer

Ơndzelin, Jan 1905 Latyšskie predlogi, I þast’ [Latvian prepositions, part one]. Jur’ev: K. Mattisen. Geniušienơ, Emma 1987 The Typology of Reflexives. (Empirical Approaches to Language Typology 2.) Berlin/New York: Mouton de Gruyter. Grek-Pabisowa, Iryda and Irena Maryniakowa 1999 Wspóáczesne gwary polskie na dawnych kresach póánocnowschodnich [Contemporary Polish dialects in the former northeastern border region]. Warszawa: Slawistyczny OĞrodek Wydawniczy. Grumadienơ, Laima and Stundžia, Bonifacas 1997 Lithuania. In: Hans Goebl, Armin Burkhardt, Gerold Ungeheuer, Herbert Ernst Wiegand, Hugo Steger and Klaus Brinker (eds), Kontaktlinguistik (Ein internationales Handbuch zeitgenössischer Forschung), 2. Halbband, 1912–1919. Berlin/New York: Mouton de Gruyter. Haspelmath, Martin 1998 How young is Standard Average European? In: Paolo Ramat (ed.), Areal Typology, 271–287. Oxford: Pergamon. 2001 The European linguistic area: Standard Average European. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals (An International Handbook), vol. 2, 1492–1510. Berlin/New York: Mouton de Gruyter. in prep. Occurrence of nominal plurality. To appear in: Matthew S. Dryer, Martin Haspelmath, David Gil and Bernard Comrie (eds.), World Atlas of Language Structures. Oxford: Oxford University Press. Jakulienơ, Audrone 1968 Lietuviǐ kalbos pasyvo formavimasis ir sangrąžiniai veiksmažodžiai [The rise of the passive in Lithuanian and reflexive verbs]. Baltistica IV: 211–220. KaraĞ, Halina (ed.) 2001 JĊzyk polski na KowieĔszczyĨnie (Historia. Sytuacja socjolingwistyczna. Cechy jĊzykowe. Teksty) [Polish in the region of Kaunas (History. Sociolinguistic situation. Linguistic features. Texts)]. Warszawa/Wilno: Elipsa. Kardelytơ, Jadvyga 1975 Gervơþiǐ tarmơ (Fonetika ir morfologija) [The dialect of Gervơþiai (Phonetics and morphology)]. Vilnius: Mintis.

Population linguistics on a micro-scale

525

Koptjevskaja-Tamm, Maria and Bernhard Wälchli 2001 The Circum-Baltic languages: An areal-typological approach. In: Östen Dahl and Maria Koptjevskaja-Tamm (eds.), Circum-Baltic Languages, vol. 1: Past and Present; vol. 2: Grammar and Typology, 615–750.(Studies in Language Companion Series 54–55.) Amsterdam/Philadelphia: Benjamins. Maslennikova, Ljudmila Ivanovna 1973a O kategorii oduševlennosti-neoduševlennosti v pol’skom govore derevni Ornjany Litovskoj SSR [On the animacy category in the Polish dialect of the village Ornjany in the Lithuanian SSR]. In: Viktor V. Martynov (ed.), Pol’skie govory v SSSR, þ. 1: Issledovanija i materialy 1967–1969 gg., 66–106. Minsk: Nauka i texnika. 1973b Nekotorye osobennosti kategorii srednego roda v pol’skom govore derevni Ornjany Litovskoj SSR [Some peculiarities of the category of neuter in the Polish dialect of the village Ornjany in the Lithunian SSR]. In: Viktor V. Martynov (ed.), Pol’skie govory v SSSR, þ. 2: Issledovanija i materialy 1969–1971 gg., 73–96. Minsk: Nauka i texnika. Masojü, Irena 2001 Regionalne cechy systemu gramatycznego wspóáczesnej polszczyzny kulturalnej na WileĔszczyĨnie [Regional features of the grammatical system of the contemporary cultural Polish dialect in the Vilnius region]. Warszawa: Elipsa. Nichols, Johanna 1992 Linguistic Diversity in Space and Time. Chicago/London: The University of Chicago Press. Pen’kovskij, A.B. 1975 Zametki o kategorii oduševlennosti v russkix govorax [Remarks on the animacy category in Russian dialects]. In: Elena Vasil’evna Nemþenko (ed.), Russkie govory (K izuþeniju fonetiki, grammatiki, leksiki), 152–163. Moskva: Nauka. Podlužnyj, Aleksandr Iosifoviþ 1997 Belorusskij jazyk [The Belarusian language]. In: Viktoria Nikolaevna Jarceva, Viktor Alekseeviþ Vinogradov and E.A. Poceluevskaja (eds.), Jazyki Rossijskoj federacii i sosednix gosudarstv I: A–I, 189–201. Moskva: Nauka. SRNG (1972/IX): Filin, Fedot Petroviþ and Fedor Pavloviþ Sorokoletov (eds.), Slovar’ russkix narodnyx govorov, vypusk IX [Dictionary of Russian rural dialects, vol. 9]: Erepenja-Zaglazet’sja. Leningrad: Nauka.

526

Björn Wiemer

SSmolG (1985/IV): Ivanova, Anastasija I. and L.Z. Bojarinova (eds.), Slovar’ smolenskix govorov, vypusk 4 [Dictionary of the Smolensk dialects, vol. 4]: E–I. Smolensk: Smolenskij gos-nyj ped. institut im. K. Marksa. Sudnik, Tamara Mixajlovna 1974 Iz morfologiþeskix nabljudenij nad govorami litovsko-slavjanskogo pograniþ’ja [Observations on the morphology of dialects in the Lithuanian-Slavic border region]. Balto-slavjanskie issledovanija 1974: 215–219. Moskva: Nauka. Trost, Pavel 1981 Kategorie þísla v baltském slovese [The category of number in the Baltic verb]. Sbornik prací 400. výroþí založení univerzity ve Vilniusu, 81–85. Praha/Vilnius: Univerzita Karlova. Wiemer, Björn 2003 Dialect and language contacts on the territory of the Grand Duchy of Lithuania from the 15th century until 1939. To appear in: Kurt Braunmüller and Gisella Ferraresi (eds.), Aspects of Multilingualism in European Language History. Amsterdam/Philadelphia: Benjamins. Zemskaja, Elena Andreevna and Margarita V. Kitajgorodskaja 1984 Nabljudenija nad prostoreþnoj morfologiej [Observations on the morphology in prostoreþie]. In: Elena Andreevna Zemskaja and Dmitrij Nikolaeviþ Šmelev (eds.), Gorodskoe prostoreþie (Problemy izuþenija), 66–102. Moskva: Nauka.

Addresses of authors Lieselotte Anderwald Englisches Seminar Albert-Ludwigs-Universität Freiburg Rempartstr. 15 D-79085 Freiburg Germany Peter Auer Deutsches Seminar Albert-Ludwigs-Universität Freiburg D-79085 Freiburg Germany

Östen Dahl Stockholms Universitet Institutionen for Lingvistik S-109 91 Stockholm Sweden Gunther De Vogelaer Vakgroep Nederlandse taalkunde Universiteit Gent Blandijnberg 2 9000 Gent Belgium

Raphael Berthele Departement für Germanistik Linguistik des Deutschen Universität Freiburg/Fribourg Miséricorde CH-1700 Freiburg/Fribourg Switzerland

Jürg Fleischer Phonogrammarchiv der Universität Zürich Freiestr. 36 CH-8032 Zürich Switzerland

Walter Bisang Institut für Allgemeine und Vergleichende Sprachwissenschaft Universität Mainz Jakob Welder-Weg 18 D-55099 Mainz Germany

Bernd Kortmann Englisches Seminar Albert-Ludwigs-Universität Freiburg Rempartstr. 15 D-79085 Freiburg Germany

J.K. Chambers Department of Linguistics University of Toronto Toronto Ontario M5S 3H1 Canada

Yaron Matras Department of Linguistics University of Manchester Oxford Road Manchester M13 9PL United Kingdom

528

Addresses of authors

Jim Miller Department of Applied Language Studies and Linguistics University of Auckland PB 90219 Auckland New Zealand

Peter Trudgill English Linguistics Fribourg University Miséricorde 1700 Fribourg Switzerland

Annemie Neuckermans Linguïstiek (GER) Universiteit Antwerpen (UIA) Universiteitsplein, 1 B-2610 Antwerpen Belgium

Johan van der Auwera Linguïstiek (GER) Universiteit Antwerpen (UIA) Universiteitsplein, 1 B-2610 Antwerpen Belgium

Günter Rohdenburg Fakultät für Kulturwissenschaften Anglistik/Sprachwissenschaft Universität Paderborn Warburger Str. 100 D-33098 Paderborn Germany

Susanne Wagner Englisches Seminar Albert-Ludwigs-Universität Freiburg Rempartstr. 15 D-79085 Freiburg Germany

Guido Seiler Universität Zürich Deutsches Seminar Schönberggasse 9 CH-8001 Zürich Switzerland

Björn Wiemer Universität Konstanz FB Sprachwissenschaft / Slavistik Postfach 55 60, D 179 D-78457 Konstanz Germany

Peter Siemund Universität Hamburg Institut für Anglistik und Amerikanistik Von-Melle-Park 6 D-20146 Hamburg Germany

Language and dialect index A African American Vernacular English (AAVE) 61, 253–254, 257, 268 Ainu 264 Alabama English 26 Alemannic (German dialect) 73, 76–78, 85–88, 102, 121–123, 217, 224, 369–370 High 74, 78–80, 85, 88, 262 Low 217, 224–225, 230–231 American English 64, 315, 316 Amharic 151, 429 Appalachian English 253–254 Arabic 61, 149, 173 Modern Standard 149 Armenian 309 B Baltic 4–5, 28, 150, 176, 497–501, 504–511, 521–522 East 497, 502–504 Balto-Finnic 501, 515, 521–522 Bari 264 Basque 263 Bavarian (German dialect) 70, 73, 76–80, 86–88, 217–220, 223, 227, 232–235, 238, 262 Bayso 37–39 Belarusian 501–506, 509, 512–516, 521–522 Belizean Creole 246, 249, 252 Breton 263 Brabantic (Flemish dialect) 187– 190, 194, 453–454, 468, 471–474 Bulgarian 289, 308–309, 312, 437, 499

C Cape Flats English 256, 260, 273 Celtic 263–264, 412–413, 420, 425, 508 Brythonic 263 Irish Gaelic 7, 54, 61, 238–239, 246, 250, 255, 258, 266, 314, 401–413, 420–430 Scottish Gaelic 54, 253, 314, 320 Welsh 49, 237, 246, 250, 263, 295, 301, 320, 336, 412, 425– 426, 430 Chinese 13, 418–419 Cushitic 29, 37–40 Cypriot Greek 445–447 Czech 237, 301, 508–510 D Dalecarlian (Swedish dialect) 161– 163, 167–172 Danish 147–148, 152–161, 167– 170, 179, 226, 234, 238–239, 308, 501 dialects 482 Standard 178 Dutch 70, 181–191, 197–207, 216, 229, 236–237, 245–247, 252, 260–261, 266, 269, 337, 341, 360, 454, 460–466, 471–476 Middle 261 Standard 70, 181–184, 187–188, 201–203 dialects 3–7, 181–191, 196, 200– 204, 206 Flemish 3–4, 8, 187–191, 194– 199, 202, 205, 208, 453–456, 462, 468, 471–476

530

Language and dialect index

Frisian 189, 196, 204–206, 216, 229, 235–237, 308 Groningen 186, 196, 203 Limburgish 454 Zeeland (group of dialects) 188– 189, 204 E East Franconian (German dialect) 88, 217–220, 227 East Pomeranian (German dialect) 217–218, 233–235 English 2–8, 18–20, 27, 50–51, 59– 65, 70, 97–98, 110–112, 116– 118, 127–135, 141–144, 148, 152–155, 161, 171–174, 202, 216, 220–221, 224–226, 234– 235, 245–268, 305–320, 323– 331, 336–338, 340–349, 352– 356; 401–403, 406–407, 410– 413, 420–430, 437–440, 446, 460, 476, 479–481, 492–493, 500, 504 Old 311, 316, 337, 343, 410 Middle 261, 316, 410 Early Modern 249, 252, 336, 358, 407–411, 422 Late Modern 259, 356–357 non-standard 18, 258, 260, 305, 311, 315–316, 320, 335, 350, 358–359, 426 spoken 57, 252, 266, 306, 311– 313, 320, 337 Standard 47, 50–54, 57, 70, 182– 184, 226, 247–248, 251, 254– 255, 268, 313, 316, 319, 320– 322, 336–337, 343, 351–352, 355, 401, 415, 479–480 individual varieties of English: African American Vernacular English (AAVE) 61, 253– 254, 257, 268

Alabama English 26 American English 64, 315, 316 Appalachian English 253–254 Belizean Creole 246, 249, 252 Cape Flats English 256, 260 Gullah 246, 249–253, 257 Hebridean English 412 Irish English 7, 238–239, 246, 250, 256–258, 266, 401, 404– 411, 420–421, 426 Jamaican Creole (English) 60, 252 Lumbee 131, 146 New World Englishes 257 Newfoundland English 482, 486–492 Ocracoke 128, 131 Scottish English 238, 252, 306– 308, 313–315, 318–322, 412 Somerset dialect (SW England) 337, 343, 345–348, 351 Southern (British) English 238, 318 Southwest English 481–492 Sranan (Creole English) 60 Tasmanian Vernacular English 482 Welsh English 246, 250, 320, 412, 425–426, 430 White Southern US vernacular 253 Ethio-Semitic 29 F Faroese 159, 176 Finnish 61, 161, 172, 236, 300, 429, 483 Flemish 3–4, 8, 187–190, 194–199, 205, 453–456, 462, 474–476 Brabantic 187–190, 194, 453– 454, 468, 471–474 West 188, 198, 454, 476

Language and dialect index East 191, 195–196, 202, 453– 454 Franconian (German dialect) East Franconian 88, 217–220, 227 Middle Franconian 79 Moselle Franconian 78, 217– 219, 224 French 93–94, 97–104, 108–112, 117–118, 122, 171–174, 232, 311–312, 417, 422, 429, 454, 458–463, 476, 480, 502 Standard 117, 182, 458 Guyanais Creole French 60 Haitian Creole French 60 Frisian 189, 196, 203–205, 308 North 216, 229, 235–237, 308 West 308 G Gammalsvenska (dialect of Gammalsvenskby/Zmejevka, Ukraine) 160, 179 Gbe 23–24 German 4–5, 69–87, 93, 97–98, 101–104, 107–112, 116–120, 151–153, 182–184, 189, 206, 211–222, 225–226, 229–239, 245–247, 260–262, 266–268, 292, 295, 308, 312, 337, 341– 347, 353, 360, 369, 372, 394, 415–422, 429, 480, 502, 522 Old High 85, 222, 232, 239 Old Low 232 Middle 74–77, 85 Middle High 232, 261 Middle Low 232–233 Early New High 75 Low 26, 73–74, 80, 87–88, 215– 217, 225, 229, 232, 260, 268, 335, 338–360, 482, 501 non-standard 83

531

spoken 71, 75, 81–84, 260–261 Standard High 93, 101, 117 Upper 75–76, 88, 216–217 individual non-standard varieties of German: Alemannic 73–80, 85–88, 102, 121–123, 217, 224–225 230– 231, 262, 369–370 Bavarian 70, 73, 76–80, 86–88, 217–220, 223, 227, 232–235, 238, 262 Franconian 78–79, 88, 217–220, 224, 227 East Pomeranian 217–218, 233– 235 Leibitz/Lubica (German linguistic island in eastern Slovakia) 217, 223–224, 227, 234–235 North Saxon 217–219, 222–226, 233–235, 239, 260 Ripuarian (group of dialects) 79, 83–84, 87–88 Swiss German 3, 70, 88–89, 93– 94, 101–105, 120–123, 212, 216, 229, 236–237, 367–373, 381–384, 394–396 Upper Saxon 86, 217–219, 222– 225, 231–235 Wendland dialect 344–347, 350, 360 Westphalian 217–222, 227, 231– 233 Germanic 4–5, 22, 75, 119, 151, 184, 216–217, 229–238, 246, 263–266, 269, 344, 413–416, 421–423, 427–428, 455–456, 482, 508 North 4 West 3–5, 8, 147, 182, 245–247, 261, 264–266, 423, 462 Greek 4, 234, 278–280, 289, 292, 300, 309, 435–447

532

Language and dialect index

Ancient 445–446 Asia Minor 440 Asia Minor Greek dialects 440 Cypriot Greek 445–447 dialects 435–437, 443–447 Modern 149, 237, 435, 438–439 Pontic (group of dialects) 441 Sfakian 445 Groningen (Dutch dialect) 186, 196, 204 Guinea-Bissau Creole Portuguese 60 Gullah 246, 249–253, 257 Guyanais Creole French 60 H Haitian Creole French 60 Hebridean English 412 Hindi 13, 264 Hungarian 236, 289 I Icelandic 159, 176 Old 150 Indo-Aryan 22, 278–280 Indo-European 22, 61, 236, 246– 247, 264, 311, 429, 504, 511, 521–522 Indo-Iranian 22, 263 Irish (Gaelic) 7, 54, 61, 238–239, 246, 250, 255, 258, 266, 314, 401–413, 420–430 Irish English 7, 238–239, 246, 250, 256–258, 266, 401, 404–411, 420–421, 426 Italian 3, 134, 191, 200, 308, 312, 417, 422, 445 J Jamaican Creole (English) 60, 252 Japanese 418

Javanese 418 Jiwarli 23–24 K Karelian 36, 236 Kobon (language spoken in New Guinea) 264 Kwa 24 L Latin 99, 150, 232, 236, 460, 463, 476 Latvian 150–152, 300, 502, 505– 506, 511, 516, 521–522 Leibitz/Lubica (German linguistic island in eastern Slovakia) 217, 223–224, 227, 234–235 Limburgish (Dutch dialect) 454 Lithuanian 284–285, 300, 501–506, 511–516, 521–522 Lumbee (English dialect) 131 M Macedonian 289, 308, 437, 499 Maltese 308 Mandinka (African language) 61 Mano 264 Middle Franconian (German dialect) 79 Mordvin 99, 522 Moselle Franconian (German dialect) 78, 217–219, 224 Muotathal (Swiss German dialect) 93, 101, 104–105, 122 N Navajo 264 New World Englishes 257 Newfoundland English 482, 486– 492 Nilo-Saharan 29, 37

Language and dialect index North Saxon (German dialect) 217– 219, 222–226, 233–235, 239, 260 Norwegian 147–148, 152–158, 167–170, 178–179 Bokmål 155–156, 179 Nynorsk 155–156, 179 O Occitan 308 Ocracoke (English dialect) 128, 131 Omotic 29 Orokaiva 264 Ossetic 263, 522 P pidgins and creoles 5, 8, 22, 36, 40, 60–61, 128–129, 135, 139, 245– 254, 257–259, 264–267, 402– 403, 438, 518 Atlantic 36, 250–254, 257–258, 266 Caribbean 249–250, 256–257, 266 Polish 217, 501–515, 522 polszczyzna kresowa 501, 521 Pontic (group of Greek dialects) 441 Portuguese 60, 422, 474 Guinea-Bissau Creole Portuguese 60 Proto-Uralic 236 Q Quechua 36 R Ripuarian (group of German dialects) 79, 83–84, 87–88 Romance 22, 97, 102, 122, 232, 312, 417, 422, 502

533

Romani 3–4, 7, 277–286, 289–295, 299–301, 522 Proto- 278 Russian 36, 49, 61, 99, 305–311, 324–331, 410, 417–418, 429, 501, 505–506, 509–516, 521–522 spoken 306, 309, 327 standard 324, 328–330, 501, 506–512, 515 S Samaná 131, 142 Scandinavian 147, 152, 155, 158– 159, 167–171, 174–179, 202, 234, 422, 502, 511 Continental 147, 153–154, 176, 184, 202–204 Scottish English 238, 252, 306–308, 313–315, 318–322, 412 Sfakian (Greek dialect) 445 Sinitic 13, 41 Sko 264 Slavic/Slavonic 4–5, 118–119, 150, 202, 235, 312, 324–325, 344, 439, 497–498, 501–512, 515, 522 Old 152 Balkan Slavic 499, 501, 507 East 501–502, 508–510, 521 North 497, 508–511 West 501, 507, 515 Slovak 235, 289–293, 300–301, 510 Somerset dialect (SW England) 337 Sorbian 308, 510 Southwest English 481–492 Spanish 36, 44, 97–98, 118, 182– 183, 195, 201–202, 417–422, 429 spoken 338–339 Sranan (Creole English) 60 Sudanese 418 Swedish 7, 147–148, 152–163, 167, 174–179, 216, 308, 501 dialects 7, 159–161, 170

534

Language and dialect index

Gammalsvenska (dialect of Gammalsvenskby/Zmejevka, Ukraine) 160, 179 Standard 157, 160–161, 165– 171, 176 Swiss German 3, 70, 88–89, 93–94, 101–102, 120, 123, 367–369, 372, 381, 384, 395–396 Muotathal 93, 101, 104–105, 122 Thusis (Grisons district) 370– 373 Zurich German 212, 216, 229, 236–237, 371, 382, 394–395 T Tasmanian Vernacular English 482 Thai 418 Thusis (Grisons district, Swiss German dialect) 370–373 Tigre 37–38, 45 Tucano 264 Turkish 98, 237, 309, 360, 439– 441, 445–446 U Ukrainian 506, 522 Upper Saxon (German dialect) 86, 217–219, 222–225, 231–235 Usarufa 264

V Vepsian 236 W Warlpiri 23–24, 42 Welsh 49, 237, 246, 250, 263, 295, 301, 320, 336, 412, 425–426, 430 Welsh English 246, 250, 320, 412, 425–426, 430 Wendland (Low German dialect) 345–347, 350, 360 Westphalian (German dialect) 217– 222, 227, 231–233 White Southern US vernacular 253 Y Yessan-Mayo 264 Yiddish 118–119, 149, 211–212, 216–217, 220–224, 227–236, 239, 281 Z Zeeland (group of Dutch dialects) 188–189, 204 Zurich German (Swiss German dialect) 212, 216, 229, 236–237, 371, 382, 394–395

Subject index A Accessibility Hierarchy 7, 211–215, 218–221, 228, 230, 233–234, 479, 480, 485 accommodation 135 acrolect 130, 133, 135, 141 adjectives 37, 60, 70, 85, 147–152, 155, 164–165, 345, 354–360, 509–512, 521 agreement 6–7, 20, 26, 85, 137, 184, 205, 263, 286–287, 359, 421, 429, 479–480, 510–513, 522 agreement markers 141, 182– 185, 187–198, 200–205, 441 complementizer agreement 189, 196–199, 206 grammatical agreement 26, 85, 141, 181–206 nonagreement 18–19 analyticity 7, 70, 266 Animacy Hierarchy 7, 238 anterior 251, 257–259, 267, 310, 315 area(s) convergence 5, 7, 498 dialect 8, 16, 57–58, 64, 77, 160–161, 165–167, 170, 174, 176, 194, 249, 345, 444 linguistic 1, 14–16, 23, 27–30, 39, 71–72, 75–76, 184, 190, 211, 234–235, 437 Circum-Baltic 28, 499–501, 505, 511 Ethiopia 29, 39 Southeast Asia 23, 28 see also Sprachbund, Standard Average European

areal continuum 368, 380, 383–384, 511 articles 37, 76, 151, 152, 172–174, 179 definite 76, 85, 147–178, 280, 300, 343, 499 indefinite 6, 77, 174, 499 aspect 6–8, 80, 83, 88, 245–254, 257–261, 264, 267–268, 305– 309, 324–325, 329, 331, 401– 404, 413, 417, 425, 428 completive 254 habitual 245–250, 256, 261, 264, 266, 309, 325, 405, 414–415, 425 perfective 245, 325 progressive 80, 88, 245, 259, 261, 404, 413, 425 atlas 12, 17, 75, 88, 206, 263, 454 dialect 29, 88 language 16 see also SAND, World Atlas of Language Structures auxiliary 6, 24, 47, 50–53, 57–65, 78, 107, 121, 123, 248, 263, 348– 352, 370–374, 377, 381–382, 386, 393, 405–410, 418, 421– 422, 429, 476 vs lexical/full verb 6, 348–353, 371–381, 386, 393–394 B basilect 130–135, 139 bilingualism 403, 437, 447, 505– 506 bioprogram 36, 128–129, 139

536

Subject index

British National Corpus (BNC) 57– 59, 65, 339–340, 349, 353, 358, 361 C case systems 278, 497 Cat Corpus (Swedish dialects) 160– 169, 172–174 child bilingualism 437, 447 cliticization 50 clitics 6, 24, 50, 132, 181, 188–192, 195–197, 436, 441, 502 clitic pronouns 6, 188–191, 195, 197, 436, 441 codification 87, 519 cognitive linguistics 94, 122 comparative constructions 505–506, 515–518 complementation 6, 7, 277–278, 282, 285–289, 293, 295–299 complementizers 78, 181, 186, 189, 195–202, 206 complexity 138, 230, 338–341, 437–438, 442, 447, 517 principle 335–339, 342, 359 conceptional literacy 118–120 conceptional orality 101, 120 conceptual space 7, 277, 298–299 conditional clauses 256, 259, 262, 268 configurationality 23–24 conformity 435, 442, 463, 470 consistency 7, 69–71 constraints 52, 83–84, 120, 127, 132–133, 139–143, 154, 158, 187, 191, 205, 214, 355, 383–397 ranked 386 tied 385, 390–392 unranked 388–390 contact 5–11, 15, 21, 27–29, 34–39, 118–119, 127, 135, 234–235, 279–281, 289–299, 401–406,

412, 413, 421–427, 430, 435– 439, 444, 447, 497–507, 510– 512, 515–522 between dialects 234 between languages 5, 7–11, 15, 21, 35–37, 119, 235, 281, 289–299, 402, 406, 435–439, 500, 504–506, 511, 521 linguistics 11, 39, 497, 519 continuum 81, 118, 132–135, 140, 159, 176, 257–258, 281–282, 287–289, 294–295, 299, 345, 368, 380, 383–384, 510–511 contraction auxiliary 47, 50–54, 57–60, 63– 64 negative 47, 50–54, 57–60, 63– 64 control 282, 287, 291–292, 326, 442 copula 36, 60–62, 83, 127, 286 absence 60–62, 65, 129, 500 deletion 61, 65, 129 corpus 3–4, 56–58, 99, 119, 160– 169, 172–174, 216, 256, 268, 284, 312–313, 316–321, 337, 349, 353, 357, 361, 428, 479– 480, 483–489, 493 see also British National Corpus, Cat Corpus, Freiburg English Dialect Corpus, Northern Ireland Transcribed Corpus of Speech count nouns 172–174, 344–347, 479, 482 creoles 5, 36, 60, 65, 128, 135, 139, 245, 248–259, 264–267, 403, 518 D database 281–284, 289–293, 299, 316 default singulars 18, 127–143, 500, 504, 518

Subject index depictives 85 dialect contact 234 definition of 11 rural 307 urban 512 dialectologists 2, 5–9, 12, 16, 20, 94, 101, 111, 118–121, 331, 436, 483–484, 519 dialectology 1–4, 7–16, 27–30, 33– 36, 39, 47, 70, 103, 118, 121– 123, 127, 130, 211–212, 237, 248, 277–280, 299, 367–368, 404, 427, 454 traditional 36 dialectometry 12, 17 diffusion 12, 14, 16, 30–34, 38–39, 132, 281, 437 discourse 12, 21, 31, 34–35, 95, 121, 181–184, 254, 261, 269, 306, 326, 337–338, 341, 440, 485 dislocation 335, 340, 342, 360 left 75, 335–343, 360 do/tun-periphrasis 6, 8, 73–74, 80, 83, 88, 260–263 double determination 7, 148, 152, 155, 170, 178 E event integration 96, 105, 109, 282, 296 existentials 132–133, 238, 317–319, 407–409, 414–420, 423, 428 F factuality 289, 294–296 fast-speech phenomena 440–442, 447 fieldwork(er) 4, 15–16, 20, 283, 299, 482, 493

537

figures 71–72, 95–96, 104, 106, 109, 113, 115, 137, 139, 258, 295–296, 299, 345, 424–425, 463–467, 471–473, 488–489, 501 ground 95–97, 101–104, 108– 109, 114–117, 121 Founder Principle 257–258, 266 Freiburg English Dialect Corpus (FRED) 56–57, 64, 250, 255– 256, 268 Frog Story 97–103, 113, 117 functionalism 2, 3, 30, 459–460 integrative 2, 11–15, 30–31 G geminate consonants 436, 446 gender 16, 26, 103, 134, 140, 218– 220–223, 280, 329, 343–347, 429, 479–485, 492–493, 497, 507–517, 522 agreement 479–480 systems 6, 335, 343–346, 359, 360, 480–482, 508, 512–515 geography 15, 27, 367, 372, 453, 454, 468–470 grammar 1–6, 13, 21, 28–31, 36, 53, 64, 70, 73–75, 128–131, 136, 139, 245, 250, 258, 281, 326, 336, 360, 367–369, 372, 381– 397, 403–405, 422–424, 507, 522 generative 24, 28, 64, 128, 205, 336, 385, 422–424 universal 28, 128, 424 grammaticalization 2, 7, 147, 152, 167, 176–179, 233–246, 252, 257, 264–267, 416, 420, 424, 429, 458–459, 508 paths 148, 178, 245–246, 262– 265 ground 94–99, 101–105, 111, 116, 206, 257, 341, 403, 447, 459, 484, 490

538

Subject index

H head/modifier pattern 369–370 head-final 370–371, 377–378, 384– 386, 393–394 languages 370–373 orderings 6, 371, 375, 378 head-initial 370–371, 377, 384, 394 languages 370–373 orderings 6, 371 hierarchy 1, 7, 14, 19, 25–27, 32, 48–52, 58–60, 63–64, 130–133, 135–138, 141–143, 211–221, 228, 230, 233–234, 238, 277, 294–299, 386, 479–480, 484– 486, 508 of individuation 7, 335, 343– 345, 359 I iconicity 8, 30, 291, 351 imperfective 264, 305–306, 310, 324–325, 413, 417, 428 vs perfective 257, 266, 305–306, 310, 324–325, 417, 428 implicational hierarchies 7, 19, 26, 293 implicational universals 14, 19, 25, 370 intensifiers 335, 354–359 invisible hand (theory) 11 L language contact 5–11, 15, 21, 35–37, 119, 235, 281, 289, 299, 402, 406, 435–500, 504–506, 511, 521 sampling 8, 21, 38 spoken 50, 53, 57, 69–72, 80, 120, 159, 179, 267, 305–307, 331, 493

written 70–71, 120, 155, 159, 179, 232, 236, 306, 327, 354, 429 M manner-salient languages 118 map(s) 16–17, 27–30, 55–57, 75, 87–88, 97, 110–111, 160–163, 176–177, 295, 299, 309, 372– 379, 386, 396, 454–456, 465, 468–472, 498, 501, 510, 520 semantic 295, 299 markedness 8, 36, 47–50, 64, 278, 299, 497, 504–506, 516, 519 local 8, 47–50, 60–64 mass nouns 77, 172, 344–347, 479, 482 mass-count distinction 479, 482 methodology 8, 211, 216, 277–281, 424, 427, 499 microparametric syntax 3 modal verb 296, 371–375, 379–382, 385–386, 393–395 mood 245–247, 260 subjunctive 80, 247, 259–262, 266, 286 morphology 1, 47, 85, 278, 281, 486, 505, 513 motion verbs 93–94, 98–99, 103– 109, 119–121 multidimensional scaling 15–17, 35 multilingualism 498, 501–503, 518 N naturalness 7, 140–141, 445 negation 6, 48, 247–248, 263, 300, 453–466, 473–475, 502 double 73–74 multiple 7, 129, 140, 500 predicate 463–466, 475

Subject index quantifier 453, 456, 463–466, 475–476 triple 8, 453–456, 461, 473, 476 Northern Ireland Transcribed Corpus of Speech (NITCS) 56, 404–406, 411, 421–422, 428 Northern Subject Rule 133 number marking 497, 513 O Optimality Theory 4, 52, 367–369, 384–386, 392–395 Stochastic OT 4, 52, 392 P partitive constructions 77, 172 passive 78, 286, 305–307, 317–318, 325–330, 502–503 past tense forms 6, 248–250, 254, 258, 335, 349–352, 359, 509–510 Perfect 6, 249–254, 264–266, 305– 317, 321–330, 345, 401–410, 416–423, 427–429 after- 314, 406, 411–412, 423 be- 268, 319, 406, 410–411, 421–422 in Russian 305–306, 309–310, 324–331 with ever, yet 321 with time adverbs 322–325 Perfect Passive 306 perfective 257, 266, 324, 417, 428 person marking 7, 24, 181–188, 190–194, 199, 201–204, 206 phonology 47, 139–140, 278, 307, 335–354, 359, 440, 443, 447, 492 pidgins 5, 8, 36, 60, 128–129, 135, 139, 245, 249–254, 257–259, 264–267, 438, 518 population linguistics 8, 497–500, 519

539

possessives 74– 5, 85–87, 175, 216, 305, 308–311, 317, 324, 328– 330, 343, 410, 421 posture verbs 6, 93, 103, 108–109 pragmatics 81, 261, 297 prepositions 61, 78, 83, 88, 95, 98– 99, 119–121, 215–216, 224–229, 232, 286, 292, 370, 412, 497, 505, 517–518, 521 principle of rhythmic alternation 335–336, 354–359 pro-drop 185, 189–190, 195, 202– 204 non-Pro-Drop 181, 182, 202, 203 progressive 80, 83–84, 88, 245– 249, 257–261, 263–266, 313, 404, 413, 425, 428, 470 pronominal gender 479–480 pronouns 8, 19, 26–27, 61, 76, 81, 88, 133, 147, 150, 153, 158–163, 181–193, 195–206, 211, 218– 224, 227–233, 238, 280, 335– 338, 341–343, 436, 439–441, 455, 479–484, 487–492, 502, 509, 511, 513, 514 personal 191, 223–229, 235, 238, 341, 344–345, 480–482, 485 possessive 216 reflexive 6 relational 153 relative 186, 219, 222, 228–229, 232–233, 236, 239, 485, 499 resumptive 6, 228, 229, 231, 235, 335–342, 359, 360 weak 188, 192, 196, 203, 206 prosody 164 prototype theory 62 psycholinguistics 94, 122, 141 purpose clauses 285–287, 292, 295– 298

540

Subject index

Q quantity principle 336, 351–353, 359 questionnaires 15, 19–20, 283–284, 301, 328, 372, 382, 391, 396 R Radical Construction Grammar 28, 39 raising 75–76, 444 redundancy 7, 112–116, 120, 140, 183, 336, 437–439, 447 regularities 49, 69–71, 127, 130, 226 regularization 129, 135, 140 relative clauses 6, 26, 37, 76, 152, 211–221, 224–232, 234, 236, 239, 318, 339–342, 370 relative markers 218, 231, 234 resultative 251, 265, 268, 305–311, 317–320, 324–330, 409–410, 414–415, 419–420, 423, 429 S SAND (Syntactic Atlas of the Dutch Dialects) 3, 191, 199, 205–206, 454–455, 462, 474–475 satellite-framed languages (Slanguages) 97–98, 102 Sprachbund 5, 27, 437, 501 semantics 19, 47–49, 62, 81, 105, 112, 117, 153, 261, 287, 459 Simple Past 249–251, 254–255, 305–306, 310–316, 322–324 simplicity 149 social attitudes 7–9, 518 social networks 9–11, 29, 32, 101, 135, 435–439, 447 society 8, 65, 134–135, 439, 442 structure 9, 435–436 types of 435–436

sociolinguistics 8, 94, 128–130, 435–436 sound change 443–447 spatial language 102, 108, 121–123 spontaneous spoken language 305– 307, 331 Standard Average European 6, 181, 236–237, 499 standard language vs. vernacular speech 69–71, 94, 98, 120–121, 135, 147–148, 152, 165, 168, 175, 184, 217, 233, 236–237, 308, 327, 401, 492, 509, 513, 519, 522 subject doubling 190–195, 201, 206 subject tripling 190–191, 195, 206 subject-verb concord 137 subordination 79, 233 substrate/substratum 36, 239, 267, 401–408, 411–413, 421–429, 501, 504–505, 510–515, 522 superlatives 153–155, 165 superstrate 135, 248–250, 258–259, 267, 401–413, 421–429 Survey of English Dialects (SED) 12, 55–57, 482–483, 493 syntax 1–5, 48, 69–72, 79–87, 128, 142, 178, 212, 277–278, 283, 326, 372, 396, 508 microparametric 3 spoken 70–72 written 70–72 T tense 6–8, 48, 52–53, 57–62, 64–65, 75, 80, 123, 245–263, 266–269, 278, 286, 305–309, 316–319, 325, 329–331, 335, 345, 348– 352, 355, 358–359, 401–403, 405–409, 413–423, 426–429, 509–510 tense-carrier 261–263

Subject index theory 4, 52, 62, 139–140, 367, 369, 384–386, 392, 443, 475, 519 syntactic 228, 268 typologists 1–6, 9–14, 20–21, 25, 79, 94, 102–106, 110, 117–118, 121, 130, 206, 267, 307, 309, 324, 331, 437, 480 typology 1–3, 6–16, 19, 21, 25, 31, 34–39, 47, 69–70, 75, 79, 84, 86, 93, 98, 103, 117–118, 122–123, 183, 205, 211–212, 218, 228, 231, 237, 277, 279, 281, 299, 305–308, 329, 367–369, 414, 424, 427, 435, 448, 454, 497, 516 areal 1, 6, 8, 27, 308, 437 dynamic 2, 127–130 functional 1, 60, 64, 335, 359 variationist 8, 143 U universals 7, 12, 31–33, 37–38, 47, 299, 401–404, 423–425, 500 implicational 14, 19, 25–26, 370 vernacular 7, 61, 127–130, 248, 500, 504, 518 V variable output grammar 385, 392 variation conditioned 382–384, 394 covariation 12–15, 18, 25, 35–38 cross-dialectal 1, 370 cross-linguistic 1–2, 14, 130, 369, 372, 474–475 diatopic 381–382, 394 free 222, 227, 367, 382–383, 394, 515, 522

541

dialect 1, 16, 119, 178, 185, 205, 277–279, 367–370, 380–381, 394 grammatical 3, 335–336, 354 internal 161, 381–382, 384–385, 391, 517 intralinguistic 5, 127, 130, 474, 475 linguistic 1–2, 11–14, 35, 71, 130, 369, 372, 474–475, 501 structural 1, 11–13, 23, 36, 85, 497, 517, 519 syntactic 5, 20, 69, 72, 74, 83, 85 verb clusters 6, 367–374, 377–387, 390, 394–395 verb duplication 78 verb-framed languages (Vlanguages) 97–98 vernacular 7, 61, 128–132, 135– 136, 139–143, 161, 250–254, 258, 268, 482, 499–500, 504– 505, 510–513, 516–519 processes 143 universals 7, 61, 127–130, 248, 500, 504, 518 W word order 19, 185–192, 197, 202– 206, 226, 369, 410, 420, 459– 460, 498 verb-first 79–82 World Atlas of Language Structures (WALS) 30, 263 Z zero relatives 211, 218, 226, 230, 234, 239