cover
next page >
Cover
title: author: publisher: isbn10 | asin: print isbn13: ebook isbn13: language: subject public...
38 downloads
646 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
cover
next page >
Cover
title: author: publisher: isbn10 | asin: print isbn13: ebook isbn13: language: subject publication date: lcc: ddc: subject:
Vernacular Universals and Language Contacts : Evidence from Varieties of English and Beyond Routledge Studies in Germanic Linguistics ; 14.14 Filppula, Markku.; Klemola, Juhani.; Paulasto, Heli. Taylor & Francis Routledge 0415992397 9780415992398 9780203883426 English Universals (Linguistics) , English language--Variation, Languages in contact, Linguistic change. 2009 P204.V47 2009eb 425 Universals (Linguistics) , English language--Variation, Languages in contact, Linguistic change.
cover
next page >
< previous page
page_i
next page >
Page i Vernacular Universals and Language Contacts
< previous page
page_i
next page >
< previous page
page_ii
next page >
Page ii Routledge Studies in Germanic Linguistics SERIES EDITORS: EKKEHARD KÖNIG, Free University Berlin, Germany JOHAN VAN DER AUWERA, Antwerp University, Belgium 1. Negative Contexts Collocation, Polarity and Multiple Negation Ton van der Wouden 2. When-Clauses and Temporal Structure Renaat Declerck 3. On the Meaning of Topic and Focus The 59th Street Bridge Accent Daniel Büring 4. Aspectual Grammar and Past-Time Reference Laura A. Michaelis 5. The Grammar of Irish English Language in Hibernian Style Markku Filppula 6. Intensifiers in English and German Peter Siemund 7. Stretched Verb Constructions in English David Allerton 8. Negation in Non-Standard British English Gaps, Regularizations and Asymmetries Lieselotte Anderwald 9. Language Processing in Discourse A Key to Felicitous Translation Monika Doherty 10. Pronominal Gender in English Peter Siemund 11. The Grammar of Identity Intensifiers and Reflexives in Germanic Languages Volker Gast 12. Dislocated Elements in Discourse Syntactic, Semantic, and Pragmatic Perspectives Edited by Benjamin Shaer; Philippa Cook; Werner Frey; Claudia Maienborn 13. English and Celtic in Contact Markku Filppula, Juhani Klemola, and Heli Paulasto 14. Vernacular Universals and Language Contacts Evidence from Varieties of English and Beyond Edited by Markku Filppula, Juhani Klemola, and Heli Paulasto
< previous page
page_ii
next page >
< previous page
page_iii
next page >
Page iii Vernacular Universals and Language Contacts Evidence from Varieties of English and Beyond Edited by Markku Filppula, Juhani Klemola & Heli Paulasto New York London
< previous page
page_iii
next page >
< previous page
page_iv
next page >
Page iv First published 2009 by Routledge 270 Madison Ave, New York, NY 10016 Simultaneously published in the UK by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business This edition published in the Taylor & Francis e-Library, 2008. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk. © 2009 Taylor & Francis All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging in Publication Data Vernacular universals and language contacts : evidence from varieties of English and beyond / edited by Markku Filppula, Juhani Klemola, and Heli Paulasto. p. cm. -- (Routledge studies in Germanic linguistics ; 14) “This Overview and all the other articles in this volume are based on papers presented at an international symposium on World Englishes: Vernacular Universals vs. Contact-Induced Change, held at the University of Joensuu Research Station at Mekrijärvi, Finland, 1–3 September, 2006.” Includes bibliographical references and index. 1. Universals (Linguistics) 2. English language--Variation. 3. Languages in contact. 4. Linguistic change. I. Filppula, Markku. II. Klemola, Juhani. III. Paulasto, Heli. P204.V47 2009 425--dc22 2008031878 ISBN 0-203-88342-X Master e-book ISBN ISBN10: 0-415-99239-7 (hbk) ISBN10: 0-203-88342-X (ebk) ISBN13: 978-0-415-99239-8 (hbk) ISBN13: 978-0-203-88342-6 (ebk)
< previous page
page_iv
next page >
< previous page
page_v
next page >
Page v Contents Vernacular Universals and Language Contacts: An Overview MARKKU FILPPULA, JUHANI KLEMOLA, AND HELI PAULASTO
1
I The Theory of Vernacular Universals 1 Cognition and the Linguistic Continuum from Vernacular to Standard J.K. CHAMBERS 2 Vernacular Universals and Angloversals in a Typological Perspective BENEDIKT SZMRECSANYI AND BERND KORTMANN
19 33
II Consonant Cluster Reduction and Default Singulars: Prototypical Vernacular Universals? 3 How Diagnostic Are English Universals? DANIEL SCHREIER 4 Number Agreement in Existential Constructions: A Sociolinguistic Study of EighteenthCentury English TERTTU NEVALAINEN 5 There Was Universals; Then There Weren’t : A Comparative Sociolinguistic Perspective on ‘Default Singulars’ SALI A. TAGLIAMONTE
57 80 103
III Universals and Contact in Varieties of English 6 Irish Daughters of Northern British Relatives: Internal and External Constraints on the System of Relativisation in South Armagh English (SArE) KAREN P. CORRIGAN
< previous page
page_v
133
next page >
< previous page
page_vi
next page >
Page vi 7 The Case of Bungi: Evidence for Vernacular Universals ELAINE GOLD 8 The Regularisation of the Hiatus Resolution System in British English: A ContactInduced ‘Vernacular Universal’? DAVID BRITAIN AND SUE FOX 9 The Interplay of ‘Universals’ and Contact-Induced Change in the Emergence of New Englishes DONALD WINFORD 10 Digging for Roots: Universals and Contact in Regional Varieties of English MARKKU FILPPULA, JUHANI KLEMOLA, AND HELI PAULASTO
163 177 206 231
IV Methodological and Theoretical Perspectives 11 Methods and Inferences in the Study of Substrate Influence TERENCE ODLIN 12 Some Offspring of Colonial English Are Creole SALIKOKO S. MUFWENE 13 Vernacular Universals and the Sociolinguistic Typology of English Dialects PETER TRUDGILL 14 Linguistic Universals and Vernacular Data PETER SIEMUND 15 Why Universals VERSUS Contact-Induced Change? SARAH G. THOMASON
Contributors Name Index Subject Index
< previous page
265 280 304 323 349 365 373 379
page_vi
next page >
< previous page
page_1
next page >
Page 1 Vernacular Universals and Language Contacts An Overview Markku Filppula, Juhani Klemola, and Heli Paulasto 1. THE RISE OF THE CONCEPT OF ‘VERNACULAR UNIVERSALS’ OR ‘VERNACULAR ROOTS’ The concept of ‘vernacular universals’ (henceforth abbreviated as VUs) has attracted a great deal of attention in recent years, thanks mainly to the pioneering work by the Canadian sociolinguist Jack Chambers (see, esp. Chambers 2003, 2004). In Chambers (2004), he describes VUs—or ‘vernacular roots’, as he calls them—as follows: [A] small number of phonological and grammatical processes recur in vernaculars wherever they are spoken…. [T]hese features occur not only in working-class and rural vernaculars but also in child language, pidgins, creoles and interlanguage varieties. Therefore, they appear to be natural outgrowths, so to speak, of the language faculty, that is, the species-specific bioprogram that allows (indeed, requires) normal human beings to become homo loquens…. They cannot be merely English. (Chambers 2004:128 f.) Chambers’s oft-cited examples of possible VUs are the following (Chambers 2004:129): • (ng) or alveolar substitution in final unstressed –ing (walkin’, talkin’, runnin’) • (CC) or morpheme-final consonant cluster simplification (pos’ office, han’ful ); • final obstruent devoicing (hundret for hundred , cubbert for cupboard ); • conjugation regularization, or levelling of irregular verb forms (Yesterday John seen the eclipse ; Mary heared the good news); • default singulars, or subject-verb nonconcord (They was the last ones);
< previous page
page_1
next page >
< previous page
page_2
next page >
Page 2 • multiple negation, or negative concord (He didn’t see nothing); • copula absence, or copula deletion (She smart ; We going as soon as possible). Although here picked out from English, VUs are meant to denote, as Chambers points out in the previous quotation, features that are found (more or less) universally across all kinds of (nonstandard) varieties of different languages. Indeed, VUs have since become an object of study for linguists working on a wide variety of languages, although it is true to say that, as yet, VUs have mainly preoccupied the minds of those working on English and its many regional and other varieties. The terminology, however, may differ from one researcher to another, and the same general idea is familiar from many works long before the concept of VUs came to the fore. To give just one example, in his description of Welsh English Alan Thomas (see Thomas 1994) speaks of ‘general vernacular features’, by which he means those features of Welsh English dialects that are shared by many, if not most, other nonstandard varieties of English and, in the Welsh case, are therefore not due to substratum influence from Welsh. Although Thomas—like many others writing on the same type of phenomena—does not claim any universal status for his ‘general vernacular features’, more recent research does not shy away from the term ‘universal’, as can be seen from most of the articles contained in this volume and many other previous works. As is typical of linguistic studies, a host of new terms have emerged for the same general notion, reflecting different points of view and even different ‘degrees of universality’ of a given feature. A good example of a ‘weaker’ type of universal is the notion of ‘angloversals’, launched by Christian Mair (see Mair 2003). By this term, he refers to features that are characteristic of so-called New Englishes, but not necessarily of other varieties, let alone other languages: … Tendenzen, die bei der Herausbildung postkolonialer Standards regelmässig beobachtet werden, ohne dass historisch-genetische Zusammenhänge nachgewiesen werden können…. [M]öglicherweise das Ergebnis von Lernstrategien von Nichtmuttersprachlern.2 (Mair 2003:83) In this volume, Benedikt Szmrecsanyi and Bernd Kortmann propose several different types of ‘-versals’ (as they call them) depending on the researcher’s point of view: the primary division is between those that are of interest to typologists and those that are of interest to sociolinguists. Under the former label, there are ‘genuine universals’, ‘typoversals’, ‘phyloversals’, and ‘areoversals’, whereas the latter group comprises ‘vernacular universals’, ‘angloversals’, ‘francoversals’, and so on, and ‘varioversals’ (see their contribution below for further description). For our present purposes, it is interesting to note that, for Szmrecsanyi and Kortmann, ‘vernacular
< previous page
page_2
next page >
< previous page
page_3
next page >
Page 3 universals’ are just one category among many other ‘-versals’ and are limited to those features that are common to spoken vernaculars (such as double negation). Even this brief description of VUs and other kinds of universals makes it clear that there are different views on the nature and development of putative universals and the criteria by which they can be identified. In sections 2 and 3 of this Overview, we will look at two main approaches to these questions on the basis of previous literature and the contributions to this volume. The first sees VUs as (primarily) ‘natural developments’ in languages and varieties of languages, whereas the second considers them ‘unmarked’ features in some sense or another. As will be seen, these two standpoints need not exclude each other, and, in fact, they often overlap. In section 4, we will move on to discuss the relationship between VUs and language or dialect contacts, which could be considered an alternative, or at least additional, source for putative VUs, and which also have a connection with processes of second-language acquisition (SLA). Section 5, then, is devoted to some other relevant historical and sociolinguistic issues, such as the possible roles played by language change in general, prescriptivism, and language awareness. In section 6, we will return to the vexed question of where exactly we are to find VUs: in nonstandard or standard varieties, or in both. 2. VUs AS ‘NATURAL DEVELOPMENTS’ Regardless of the theoretical framework, underpinning most research on VUs as natural developments is the idea that their proper locus is nonstandard varieties, which are said to represent more ‘natural’ developments than (more or less strictly) codified standard varieties. Nonstandard varieties are therefore better amenable to the study of possible universal features in syntax or phonology. In the earlier quotation, Chambers (2004:128) speaks of VUs as ‘natural outgrowths of the language faculty’, deriving ultimately from the species-specific ‘bioprogram’. As such, they can be contrasted with standard varieties which differ from others in that they resist “natural tendencies in the grammar and phonology”, as he points out in an earlier connection (Chambers 2003:254). In a similar vein, although not committing himself to the notion of a bioprogram, Kortmann (2004) writes that [i]n many domains of grammar, regional and social non-standard varieties conform to cross-linguistic tendencies where the relevant standard varieties do not. (Kortmann 2004:1) In an effort to narrow down criteria for ‘naturalness’, some scholars have defined it in terms of (relative) productive and perceptual ease. This
< previous page
page_3
next page >
< previous page
page_4
next page >
Page 4 approach has been found to be useful, for example, in studies of language contacts and language shift. In the latter type of context, in particular, those features that are the easiest to produce and perceive are the most likely to be transferred to an emerging contact variety (see, e.g., Thomason and Kaufman 1988:26). Not surprisingly, naturalness has also featured centrally in studies on pidgins and creoles. For example, Mühlhäusler (1986:61f.) defines naturalness in terms of several properties of what he terms ‘languageinternal natural rules’. They are: more resistant to change; more frequent in terms of frequencies of tokens; more frequent across languages; more likely to be the basis of neutralization; and, finally, more likely to provide the model for analogical change than ‘abnatural’ categories. For Mühlhäusler, the best evidence of the operation of natural rules is to be found in pidgins, which typically have the following features in common (1986:61f.): • they are characterised by the absence of highly marked sounds, such as rounded front vowels and clicks; • voiced sibilants are replaced by voiceless ones; • tonal distinctions are lost; • there is no passive; • the present infinitive is used for verbs in almost all tenses; • the masculine is used for all genders; • the singular form of nouns caters for the plural meaning as well; and • there is a general preference for continuous constituents. Pidgins are thus “maximally natural languages”, and it is only when they become more exposed to influences from the superstrate language that “natural solutions” gradually give way to abnatural ones. At that stage, pidgins become ‘cultural artefacts’, to use Mühlhäusler’s words (1986:62f.). It is but one sign of the complexities involved in the notion of naturalness that Mühlhäusler’s account is in sharp contrast with Bickerton’s well-known ‘bioprogram’ theory (see Bickerton 1981), according to which the proper locus for the operation of natural rules is first-generation creoles and not pidgins. Writers in this volume also approach the question of naturalness from various points of view. To begin with, Jack Chambers discusses some cognitive aspects of naturalness, focusing on ‘look-back’ vs. ‘look-ahead’ mechanisms (for further discussion, see Chambers, Chapter 1). Naturalness is also mentioned as an explanatory factor by Karen Corrigan, who investigates possible cross-linguistic constraints on relativisation phenomena using South Armagh English as her principal database. Writing on factors governing the so-called Consonant Cluster Reduction in varieties of English, Daniel Schreier makes use of the generalisation put forward by Wolfram, Childs, and Torbert (2000:18), according to whom consonant clusters followed by another consonant are more likely to be reduced than clusters followed
< previous page
page_4
next page >
< previous page
page_5
next page >
Page 5 by a vowel because the latter represent a more ‘natural’ sequence. Salikoko Mufwene is yet another writer to rely on the concept of naturalness, albeit from a slightly different perspective: he argues that creoles should be treated as dialects of their lexifier languages, hence as results of perfectly ‘natural’ processes of language evolution, instead of placing too much emphasis on the role of contact in their development. Finally, Donald Winford opens up an SLA dimension to naturalness, arguing that ‘indigenised’ varieties of English, such as Irish English, Singapore Colloquial English, and Barbadian creole, all arise as the result of ‘natural’ or ‘untutored’ SLA. Therefore, as Winford writes, any unified explanation of the origins of contact varieties must look to SLA studies for theoretical support. 3. VUs AS ‘UNMARKED’ FEATURES ‘Markedness’ is another concept which comes up in some way or another in discussions on VUs. In general linguistic theory, it is often associated with generative grammar and especially with the notion of Universal Grammar (UG). Of course, there is no uniform theory of UG even today, and it has been formulated in many different ways over the years. In the most general terms, UG is said to constitute the main contents of the genetically inherited ‘language acquisition device’, which is activated in conditions of ‘restricted input’ from the environment, i.e., in the early stages of first-language acquisition. Particularly influential within the UG tradition, and also relevant to our discussion of markedness here, has been Chomsky’s (1981) ‘principles and parameters’ approach, which contains as a central component a general theory of markedness. This theory determines the hierarchical order of preference among the parameters of UG: the unmarked parameter values are selected unless there is evidence to the contrary, and they therefore constitute the ‘default parameter settings’. The marked values apply only if the evidence forces deviation from the default parameter settings (Chomsky 1981, Chapter 1). It is noteworthy that, regardless of the exact formulation of the theory of UG, it does not allow for markedness to be defined in (the measurable) terms of productive and perceptual ease; furthermore, what is specified as ‘unmarked’ by UG need not necessarily be more frequent cross-linguistically. Besides firstlanguage acquisition, which formed the basis for the Chomskyan approach, the theory of UG has been used to explain pidgin and creole genesis, which is similarly characterised by the ‘poverty of the stimulus’ from the immediate learning environment (see, e.g., Bickerton 1981). Other approaches, not (necessarily or explicitly) tied with the UG framework, usually rely on naturalness, ‘complexity’, or some other similar concept as (one of) their criteria for markedness. Haspelmath’s (2006) survey of 12 different senses of markedness provides a good backdrop to the discussions in this volume. He groups them into four major classes:
< previous page
page_5
next page >
< previous page
page_6
next page >
Page 6 (i) markedness as complexity, (ii) markedness as difficulty, (iii) markedness as abnormality, and (iv) markedness as a multidimensional correlation. By complexity, Haspelmath means, e.g., semantic or formal markedness, which can be expressed by the following type of statements: “In the English opposition dog/bitch , dog is the unmarked member because it can refer to male dogs or to dogs in general”. Or: “In English, the past tense is marked (by -ed) and the present tense is unmarked” (Haspelmath 2006:26). Difficulty can be either phonetic, morphological, or cognitive. Examples are: “On the scale b>d>g>G, the consonants to the right are increasingly more marked”; “A singular/plural pair like book/books is less marked than sheep/sheep because the latter is not iconic”; “The plural category is marked because it requires more mental effort and processing time than the singular” (op.cit., 26). Abnormality, then, can be defined in terms of, e.g., textual, situational, typological, or distributional markedness (rarity). This is exemplified by statements like “For direct objects, coreference with the subject is marked and disjoint reference is unmarked”; “For marked situations, languages typically use complex expressions”; “The syllable coda position is marked in contrast to the onset position”; and “Object-verb word order is the marked case: it occurs only with negation” (op.cit., 26). Finally, markedness as a multidimensional correlation refers to cases in which structures or categories that are cross-linguistically comparable share the same markedness values for the different markedness dimensions. For example, some categories (e.g., the future tense) are semantically complex and overtly coded, occur infrequently in texts, and are found only in some languages, whereas other categories (such as the present tense) are their exact opposites. Haspelmath (op.cit., 38) quotes here Croft’s (2003:110) generalisation, which states that “[i]f tokens of a typologically marked value of a category occur at a certain frequency in a given text sample, the tokens of the unmarked value will occur at least as frequently in the text sample”. Having surveyed all these different senses of markedness, Haspelmath raises the question of whether this concept is needed at all. His own answer is a definite and somewhat anticlimactic ‘No’ because structural asymmetries can, according to him, be explained just as well by frequency asymmetries, phonetic difficulty, or pragmatic inferences. Haspelmath elaborates on his verdict as follows: [I]t seems that the intuitive shared sense of ‘marked/unmarked’ is not distinguishable from the sense of everyday words like uncommon/common, abnormal/normal, unusual/usual, unexpected/expected. Apart from the larger class of markedness as abnormality (section 2.3), we also find markedness as complexity (section 2.1) and as difficulty
< previous page
page_6
next page >
< previous page
page_7
next page >
Page 7 (section 2.2), but since complexity and difficulty typically lead to lower frequency, abnormality is in effect what all markedness senses share. But we do not need a technical linguistic term for abnormality / uncommonness / unusualness / unexpectedness. Simple everyday concepts should be expressed by simple everyday words. (Haspelmath 2006:63) Of the writers in this volume, an approach similar to that of Haspelmath is adopted by Benedikt Szmrecsanyi and Bernd Kortmann, who establish the degree of ‘complexity’ and ‘analyticity’ as the most significant factor dividing different Englishes. Complexity for them means a smaller amount of features that arguably simplify syntactic rules, whereas analyticity leads to a greater number of ‘autonomous’, i.e., invariable and periphrastic, features. Applying Principal Components Analysis to their database of 46 different varieties of English, Szmrecsanyi and Kortmann are able to demonstrate which varieties of English come on top for each of these parameters. Jack Chambers’ chapter, in contrast, comes closest to the UG type of approach especially insofar as he sees VUs as features that are determined by the genetically inherited bioprogram and are ‘natural outgrowths’ of the innate language faculty. He does, however, make use of the concept of ‘difficulty’ because relative cognitive difficulty or ease form the background to his discussion on ‘look-back’ and ‘look-ahead’ mechanisms. Sali Tagliamonte, writing on ‘default singulars’ in this volume, is yet another to adhere to the concept of markedness when explaining potential VUs on the basis of data from varieties of English spoken in North America, the Caribbean, and the United Kingdom. She sees a link between the ongoing typological change of English from a synthetic to an analytic language and the tendency towards regularisation and loss of inflections. Under this scenario, typologically marked features such as third singular -s come under particular pressure and may disappear. Other writers relying on one or the other of the notions discussed so far include David Britain and Sue Fox, who use markedness as the key concept in their account of the rise of the glottal stop as a hiatus breaking strategy in some present-day English dialects. They rely here on Lombardi (2002), who, after examining different realisations of epenthesis in a wide range of languages, argues that glottal stops are the least marked epenthetic consonants because they have the least marked place of articulation for a consonant. Because there is also evidence that the glottal stop is the default option in children’s speech, the use of glottal stops to break hiatus in especially some London dialects of English can, according to Britain and Fox, be seen as an example of the ‘emergence of the unmarked’. Daniel Schreier, whose view on the naturalness of certain kinds of consonant clusters was already briefly mentioned previously, elaborates on the ‘typologically marked’ nature of the phonotactic system of English in his
< previous page
page_7
next page >
< previous page
page_8
next page >
Page 8 contribution by noting that English “belongs to the minority of the world’s languages whose phonotactic systems admit groups of consonants (C) in a single syllable structure”. According to Schreier, such clusters are typologically marked, by which he means that they are cross-linguistically less frequent than other structures, historically unstable, and, hence, liable to change in conditions of language contact, and acquired relatively late. Sarah Thomason starts off by referring to the usual definition of ‘markedness’ in terms of the range of occurrence of a feature in the world’s languages and age of learning by first-language acquirers. For her, too, these criteria suffice as “rough indicators of marked (harder to learn) vs. unmarked (easier to learn) linguistic structures”. As is evident from her formulation, markedness for her essentially equals ease of learning. She mentions the English interdental fricatives as a typical example of universally marked segments because they are rare in the world’s languages and are learned later than other fricatives in L1 acquisition of English. However, [n] can be considered to be a universally unmarked segment because of its ubiquity in the world’s languages and its early acquisition by children. 4. LANGUAGE CONTACTS AND UNIVERSALS OF SLA A comparative cross-linguistic approach to vernacular universals necessarily entails study of the possible role of language contacts as an alternative or at least additional explanation for the emergence of VUs, whether conceived of as ‘natural outgrowths’ in the UG sense or as something else. To begin with, it is a well-known fact that language contacts have been downplayed especially by those seeking to prove that language change is primarily conditioned by language-internal factors or follows from some language-independent universal principles. This is, however, alien to those scholars who have worked on vernacular data or contact languages. Indeed, the contributions to this volume are strong testimony to the central role that language contacts play in any discussion of putative universals or language change in general. Assuming that language contacts are a factor to be reckoned with when dealing with VUs, the question is: what exactly is the relationship between language contact phenomena and vernacular universals, and to what extent can we distinguish them from each other? In her contribution to this volume, Sarah Thomason argues that a strict dichotomy between vernacular universals and contact-induced change is not possible because many linguistic changes involve both various processes of contact-induced change and universal tendencies of various kinds. According to her, such a dichotomy embodies presuppositions that are empirically untenable—namely, that similar or identical changes in languages or varieties in contact arise from the same cause; that linguistic change generally has a single motivation; and that processes underlying putative universal structural tendencies are necessarily different from those underlying
< previous page
page_8
next page >
< previous page
page_9
next page >
Page 9 contact-induced changes. On this last point, she argues that ease of learning features in both universal tendencies and contact-induced change. In a similar vein, although approaching the matter from a slightly different perspective, Peter Trudgill claims that contacts versus VUs is not the relevant issue; the relevant opposition is between ‘low-contact’ and ‘high-contact’ languages or varieties, and a low level of contact or ‘language isolation’ is just as relevant to the discussion on vernacular universals as is a high level of contact. By way of documentation, Trudgill discusses some conservative varieties of English which are relatively isolated, yet have certain linguistic characteristics in common that are not shared with ‘high-contact’ varieties such as colonial Englishes. He further asks whether the characteristics of isolated varieties represent in some sense the reverse of those found in high-contact languages. Salikoko Mufwene, in turn, tackles the issue of contacts versus universals from the perspective of creole formation. He, too, questions the meaningfulness of trying to distinguish between vernacular features and contact-derived features and asserts that contacts are the principal trigger of change in both so-called ‘native’ English, on the one hand, and ‘indigenized’ or ‘non-native’ Englishes, as well as English creoles, on the other hand. In his view, English creoles have undergone the same restructuring processes that we find in the evolution of English everywhere, and language and/or dialect contacts always have a central role in these processes. This also means that we should critically reexamine the traditional distinction between internally and externally motivated change, which in turn has been used to distinguish creoles and indigenized varieties from the native varieties. As he points out, the latter have also arisen as a result of contacts between (dialects of) English and other Western European languages in what he describes as “conditions of relative absorption of the then non-English-speaking populations by the native-speaking populations”. David Britain and Sue Fox also present evidence to demonstrate how language contacts are a factor behind certain types of simplification and regularisation processes. Their example is the different strategies that speakers of two English dialects use to resolve hiatus, e.g., with the indefinite and definite articles in prevocalic and nonprevocalic positions. While the traditional dialect of the rural Fens, which has an overwhelmingly White population, maintains the traditional, complex hiatus resolution system, the speech of multicultural London is undergoing a process of simplification and regularisation, leading to a system that uses glottal stops to block hiatus. The main explanatory factor behind this change, the authors argue, is the extent to which speakers in London are exposed to language and sociocultural contact. They conclude that, rather than contact being a factor blocking the rise of vernacular universals, it turns out to be crucial to their emergence. In their own research article, the editors of this volume (see Filppula, Klemola, and Paulasto) seek to find out what impact language contacts and
< previous page
page_9
next page >
< previous page
page_10
next page >
Page 10 universal factors have had upon the behaviour and rates of occurrence of three syntactic features in regional or national varieties of English spoken in the British Isles, on the one hand, and in several colonial contexts, on the other hand. The features investigated are absence of plural marking with nouns of measurement, certain nonstandard uses of the definite article, and the use of the progressive form with stative verbs. These can be assumed to represent different levels of universality and language/dialect contact backgrounds, whereas the varieties of English were chosen so as to represent different positions on the vernacular—standard continuum, as well as different sociohistorical backgrounds either as so-called ‘L1 varieties’ or as ‘contact’ or ‘L2 varieties’. Through this kind of comparative setup, the authors set out to find answers to the question of whether there are vernacular universals in the sense of Chambers (2004) and, if so, what kinds of varieties of English constitute their proper locus, and furthermore, what is the crosslinguistic validity of the results. Their findings suggest that, of the three features, absence of plural marking with measurement nouns emerges as a good candidate for a vernacular universal in the demanding sense, being widely attested in nonstandard varieties of English all over the world and also in a large number of other languages or varieties of them. As for the two others, they have a more limited distribution and appear to be more in evidence in varieties with an L2 (and, hence, a language contact) background than in traditional vernacular dialects of England (which are an example of ‘low-contact’ varieties). The colonial contact dimension is very much to the fore in Elaine Gold’s contribution, where she discusses the rare ‘case’ of the so-called Bungi dialect. This is an English vernacular that arose in the mid-eighteenth century along the Hudson Bay trade routes in Canada, in a region which is today part of the province of Manitoba. Bungi is a contact variety which resulted from contacts between Scottish English-speaking Hudson Bay Company traders and the indigenous Cree. For most of its lifespan, extending into the twentieth century, this variety was both socially and geographically relatively isolated from mainstream or standard Canadian English. As such, it offers an interesting point of comparison to St. Helenian English, discussed by Daniel Schreier elsewhere in this volume. Among the linguistic features discussed by Gold are, first, four syntactic and phonological features which appear on Chambers’s list of VUs, viz. alveolar realization of final unstressed -ing, final consonant cluster simplification, subject-verb nonconcord, and final obstruent devoicing. In addition to finding support for these in Bungi, Gold proposes three other candidates for VU status: lack of distinction between he and she, extended use of progressive constructions, and use of the auxiliary be in perfect constructions in place of standard English have . To sum up so far, it seems evident in the light of the foregoing that VUs cannot be adequately dealt with without assuming some role for language contacts. Having established this, the next question would be the relationship
< previous page
page_10
next page >
< previous page
page_11
next page >
Page 11 between contact-induced change and the concepts of ‘naturalness’ and ‘markedness’, defined in one or the other of the ways described above. It is a common observation that universally marked features are less likely to be transferred in language contact than those that are unmarked. Thomason and Kaufman (1988) note that this holds of situations of language shift, in particular, where the usual direction of change is thus from more marked to less marked. In situations where the contacting languages are both maintained, that is, in conditions of relatively stable bilingualism, marked features have a better chance of being transferred than in language shift situations (Thomason and Kaufman 1988:51ff.). In her contribution to this volume, Sarah Thomason reasserts the same generalisation, associating the concept of markedness with ease of learning: the easier a given feature is for a shifting speaker to acquire, the more likely it is to establish itself as part of the resultant contactlanguage grammar. Whether the result is an overall simplification of the contact-language grammar, as is often claimed, appears to be a much more controversial issue. Relying on examples from Lithuanian and a number of other languages, Thomason emphasises that contacts can lead to both simplification and complication. The connection between markedness and ease of learning brings us to the language learning or acquisition perspective on VUs and contact-induced change. Besides Thomason, this perspective is explored in this volume by Terence Odlin, according to whom recent research in SLA vindicates the role of substrate influence, hence language contact, in the formation of learners’ ‘interlanguage’ and also casts doubt on Bickerton’s (1981) wellknown claims about the decisive role of the ‘bioprogram’ in the genesis of creoles, as opposed to pidgins. Odlin does, however, leave a certain role for ‘something like a bioprogram’ when it comes to choices between types of restructuring processes in creoles, but underlines the need to consider the inputs from substrate languages in the case of creoles, as well as in that of other contact varieties. Donald Winford is another contributor to this volume to point out the importance of substratal contact influences and SLA universals for the emergence of creoles and other kinds of contact varieties such as Irish English. Despite shared processes such as simplification and regularisation, which characterise both contact languages and early stages of SLA in general, and could therefore be described as being vernacular universals, Winford detects such major differences across contact languages, e.g., in their tense and aspect systems, as are best explained in terms of influence from relevant substrate languages or of diffusion. As examples of these, he discusses some differences in the ‘sociohistorical ecologies’ among Irish English, Barbadian Creole (Bajan), and Singapore English, which have resulted in both similarities and differences in the linguistic outcomes of the contacts. However, like most of the other writers quoted above, Winford wants to see universals and substrate influence not as opposing factors in the genesis of
< previous page
page_11
next page >
< previous page
page_12
next page >
Page 12 contact languages, but as complementary, “with universal principles acting as constraints on the role of substrate influence and other factors”. In his aforementioned article on Consonant Cluster Reduction (CCR), Daniel Schreier also pays attention to the SLA dimension of CCR and the so-called ‘voicing constraint’ commonly associated with it. He first notes that the overall rate of use of CCR increases in contact conditions, suggesting that language contacts play a role in determining ‘reducible’ clusters and their extent of use. This, then, explains why the voicing constraint on CCR is categorical in BrE and AmE, both of which can be considered to be relatively ‘low-contact’ varieties, but almost or entirely absent in a variety like Indian English or Hispanic English, which have been subject to extensive contacts with other languages. This leads to greater amounts of CCR among speakers of the latter varieties, as well as of other non-native varieties of English. Schreier attributes this to the wellknown SLA process of ‘rule generalisation’. In this case, it means adoption of a rule or process present in the target (superstrate) variety, but with subsequent generalisation to a larger number of contexts than are allowed in the source language. 5. HISTORICAL, SOCIOLINGUISTIC, AND FUNCTIONAL-TYPOLOGICAL PERSPECTIVES ON VUs The issues of naturalness, markedness, and language contacts discussed earlier do not by any means exhaust the range of problems associated with VUs. Equally important are the possible roles played by language change and ‘drift’ in general, prescriptive norms guiding language usage at different periods and speakers’ awareness of them, interaction between structural and sociolinguistic factors, as well as functionaltypological issues. All of these topics are addressed in some way or another in the contributions to this volume. The role of language change and the interaction between structural and sociolinguistic factors receive thorough treatments in the articles on ‘default singulars’, one of the prime candidates for VU status in Chambers’ list, by Terttu Nevalainen, Daniel Schreier, and Sali Tagliamonte. Using a corpus of eighteenthcentury correspondence (the so-called ‘CEECE corpus’) as her database, Nevalainen’s study focuses on the agreement patterns with the copula be with plural existential NP subjects. Her findings give a clear indication of the impact of social factors such as gender and level of education on the extent of use of singular concord in eighteenth-century English. She further argues that prescriptive grammarians of the time played a role in stigmatising subject–verb nonagreement, which is shown by the decreasing frequencies of use of default singulars towards the end of the period. Among the linguistic factors considered by Nevalainen are the semantic and
< previous page
page_12
next page >
< previous page
page_13
next page >
Page 13 syntactic types of the subject and present–past tense variation. All of these are found to be influential in determining agreement or nonagreement. Default singulars are the second point of focus in Daniel Schreier’s chapter, alongside CCR already mentioned. He discusses the findings on default singulars on the basis of previous works and, more specifically, of his own research on the hitherto little-known variety of English spoken on the remote South Atlantic island of Tristan da Cunha. Being one of the ‘New Englishes’ and because of the isolation of its speakers from those of other varieties, Tristan da Cunha English (TdCE) opens a particularly interesting window to the issue of VUs. Indeed, Schreier’s results show that default singulars are, or rather were, almost categorical in this variety, but a change has taken place in recent decades, most probably due to the gradual opening up of TdCE to influences from other varieties, including more standardised ones. Gender has also played a role in this process, with even younger men adhering more generally to the traditional default singulars than women. While also focusing on default singulars, Sali Tagliamonte’s chapter adopts a cross-dialectal comparative approach. Her data come from 13 different present-day dialects of English spoken in widely differing settings in North America, in the Caribbean (Samaná, spoken in the Dominican Republic), and in the United Kingdom. By means of sophisticated multivariate analyses, she traces the interaction among social and linguistic factors such as grammatical person, the constraint known as ‘The Northern Subject Rule’, existentiality, and negation. Her conclusion is that, although it is apparent that universal principles underlie variation in this domain of English usage, the observed variation is subject to constraints which may themselves reflect universal tendencies. Also writing on the interplay of social and linguistic, but arguing from a different type of data, Karen Corrigan investigates possible universal constraints, such as the well-known Case Accessibility Hierarchy (CAH), on relativisation strategies in South Armagh English (SArE), spoken in the north of Ireland. Briefly, the CAH predicts that the subject position, being the most prominent one semantically, is the most accessible to relativisation, whereas other positions involving a greater degree of perceptual/processing difficulty are less accessible. What adds to the linguistic interest of SArE in this connection is its sociohistorical background: originally settled by speakers of northern British English dialects, this dialect preserves many conservative features of these dialects, but has also been exposed to influences from the Irish language, which was spoken in South Armagh until the early part of the twentieth century. Having assessed the possible convergence between SArE relativisation strategies and those of Irish, Corrigan concludes that, instead of the substratum hypothesis, the evidence in this case rather supports an explanation in terms of internal constraints on relativisation. They, in turn, are universal in nature in the sense that many aspects of relative postmodification in SArE are predicted by the CAH and are also in evidence in both
< previous page
page_13
next page >
< previous page
page_14
next page >
Page 14 the historical development of English and Scots as well as in contemporary contact vernaculars and in the process of language acquisition in general. Peter Siemund’s approach differs from, and interestingly complements, the others in that he brings into the present discussion on universals the theoretical framework developed within the ‘school’ of functional typology. From this vantage point, he discusses three grammatical domains of English, which are pronominal gender, reflexives and reflexive marking, and tense and aspect. Using data drawn from a number of vernacular Englishes, he argues that the putative universal features in these domains can be best interpreted in terms of ‘semantic maps’ and ‘conceptual space’, as developed within functional typology, and that the observed variation across different varieties of English with respect to these features tallies well with semantic maps resulting from cross-linguistic typological comparisons. 6. CONCLUSION: WHERE TO LOOK FOR VERNACULAR UNIVERSALS? Finally, there remains the question of the ‘locus’ of VUs: where exactly are we to find them, and are we looking in the right place if we focus on nonstandard varieties alone? As already mentioned, Jack Chambers, in setting the agenda for research on VUs in his seminal works (see Chambers 2003, 2004), sees their proper locus in nonstandard varieties. Although this view has served as a starting point for most of the contributions in this volume, too, there are clear differences of opinion between the writers, which makes this question all the more intriguing. Perhaps the most critical stand on VUs as features of primarily nonstandard varieties is adopted by Peter Trudgill, who argues that standard varieties are just as relevant to the study of VUs as are nonstandard ones. For him, the important distinction with respect to universal features is not between standard and nonstandard varieties, but between high-contact (both standard and nonstandard) and low-contact varieties, such as Traditional Dialects of English. The latter are, according to him, a hitherto little researched area, yet they could reveal some ‘genuine’ examples of vernacular universals that cannot be attributed to diffusion from other varieties to the same extent as features of high-contact varieties. Along the same lines as Trudgill, Peter Siemund argues that “there is no reason to believe in a set of universals valid for vernacular data alone, at least not for English”. According to him, linguistic universals of any type are always domain-specific and may be found in both standard and nonstandard varieties; thus, universals of language change, language contact, language acquisition, and the structure of discourse may all be different and have no necessary connection with each other. What also complicates the issue with regard to English, in particular, is its differentiation into numerous varieties worldwide, such as native varieties, shift varieties, L2 varieties,
< previous page
page_14
next page >
< previous page
page_15
next page >
Page 15 creoles, and pidgins. Therefore, Siemund concludes, assuming a uniform set of universal properties shared by all of these and lumping them under the single heading of ‘vernacular universals’ runs the risk of oversimplification. Another writer critical of Chambers’ original notion of VUs and its underlying Bickertonian bioprogram hypothesis is Donald Winford, who agrees with Siemund as regards the nature of VUs: it is the language universals in general, including those of language architecture, language acquisition, processing, and development, that are behind the grammars of contact varieties, such as creoles and indigenized Englishes. However, in many of the chapters in this volume, the concept of VUs has been found useful and worthy of further exploration. As said previously, most writers use it at least as a starting point or heuristic device even if they do not fully subscribe to it or argue for a different approach. Thus, Szmrecsanyi and Kortmann explicitly adopt the concept of vernacular universals as their starting point in their account of the different types of universals shared by varieties of English worldwide. Daniel Schreier, in turn, finds empirical support for two of the candidates for VUs in Chambers’ original list, viz. CCR and the levelling of past be to was , in his analysis of these phenomenona in Indian and New Zealand English, St Helenian, and West Virginian Appalachian English: both features occur in all the varieties analysed and can, according to Schreier, be considered ‘true universals’ of at least English, in the sense that they are found in varieties that have no sociohistorical connections with each other. Similar support for vernacular universals emerges from Elaine Gold’s aforementioned description of the Bungi dialect, which makes frequent use of four syntactic or phonological features that appear on Chambers’ list of putative VUs. Gold also points out that the occurrence of the same features in such distant vernaculars as Newfoundland English and African Nova Scotian English provides further support for the existence of VUs. Finally, one can mention that the notion of VUs is also backed by some of the findings reported in the chapter co-authored by the editors (see Filppula, Klemola, and Paulasto, Chapter 10, this volume). They discuss evidence which suggests that such syntactic features as lack of number agreement with nouns of measurement can be considered vernacular universals, appearing as it does not only in many different nonstandard varieties of English both in the British Isles and other parts of the world, but also in a wide range of genetically unrelated languages. In conclusion, the contributions to this volume show that the notion of VUs has sparked off a great deal of original new research and produced exciting results. However, they reveal that debates on the nature of VUs and their relationships with other kinds of universals and explanatory factors such as language and dialect contacts will continue. Much further work is no doubt needed, but already at this stage it is clear that research on vernacular universals has opened up important new perspectives on features shared by varieties of English and of other languages—perspectives that may well have been overlooked had there not been such a conceptual tool available.
< previous page
page_15
next page >
< previous page
page_16
next page >
Page 16 NOTES 1. This Overview and all the other articles in this volume are based on papers presented at an international symposium on World Englishes: Vernacular Universals vs. Contact-Induced Change , held at the University of Joensuu Research Station at Mekrijärvi, Finland, 1–3 September, 2006. The Symposium was organised by the research project on “Vernacular Universals vs. Contact-Induced Language Change” (2005–2008), funded by the Humanities and Social Sciences Section of the Academy of Finland (Grant No. 210702). We gratefully acknowledge the support from the contributors to this book and from Erica Wetter, Research Editor at Routledge/Taylor & Francis, in making this publication possible. Our thanks are also due to Minna Korhonen for editing and preparing all the texts and other materials for publication. 2. “… tendencies that are regularly observed in the genesis of postcolonial standards, and for which it is not possible to ascertain historical or genetic connections…. Possibly the result of the learning strategies of nonnative speakers.” [Our translation] REFERENCES Bickerton, D. 1981. Roots of Language. Ann Arbor: Karoma Publishers. Chambers, J.K. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Implications . Oxford, UK and Cambridge, USA: Blackwell. ———. 2004. Dynamic typology and vernacular universals. In Kortmann, ed. 2004. 127–45. Chomsky, N. 1981. Lectures on Government and Binding . Dordrecht: Foris Publications. Croft, W. 2003. Typology and Universals. 2nd edition. Cambridge: Cambridge University Press. Haspelmath, M. 2006. Against markedness (and what to replace it with). Journal of Linguistics 42: 25–70. Kortmann, B. 2004. Introduction. In Kortmann, ed. 2004. 1–10. ———, ed. 2004. Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective . Berlin and New York: Mouton de Gruyter. Lombardi, L. 2002. Coronal epenthesis and markedness. Phonology 19: 219–51. Mair, C. 2003. Kreolismen und verbales Identitätsmanagement im geschriebenen jamaikanischen Englisch. In Zwischen Ausgrenzung und Hybridisierung: Zur Konstruktion von Identitäten aus kulturwissenschaftlicher Perspektive. Identitäten und Alteritäten 14, edited by E. Vogel et al., 79–96. Würzburg: Ergon. Mühlhäusler, P. 1986. Pidgin & Creole Linguistics. Oxford: Basil Blackwell. Thomas, A. 1994. English in Wales. In English in Britain and Overseas: Origins and Development, Vol. V of The Cambridge History of the English Language, edited by R. W. Burchfield, 94–147. Cambridge: Cambridge University Press. Thomason, S.G., and T. Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley/Los Angeles/London: University of California Press. Wolfram, W., B. Childs, and B. Torbert. 2000. Tracing language history through consonant cluster reduction: Comparative evidence from isolated dialects. Southern Journal of Linguistics 24: 17–40.
< previous page
page_16
next page >
< previous page
page_17
next page >
page_17
next page >
Page 17 I The Theory of Vernacular Universals
< previous page
< previous page
page_18
next page >
page_18
next page >
Page 18 This page intentionally left blank.
< previous page
< previous page
page_19
next page >
Page 19 1 Cognition and the Linguistic Continuum from Vernacular to Standard J.K. Chambers 1. INTRODUCTION The theory of Vernacular Roots begins with the hitherto unexploited observation that dialects become more complex as they become more standard or literary. The continuum from vernacular to standard is marked by definable layers of increasing structure-dependent grammatical devices and articulated phonological contrasts. One of the fundamental tenets is that structure-dependent grammatical devices and articulated phonological contrasts carry cognitive cost. They are excrescences upon more natural systems, or, in the terms I have used elsewhere, they are learned rather than primitive processes (Chambers 2003:271–3). For example, the principle driving conjugation regularization—“Make past tense and past participle the same”—reduces irregular forms, and irregular forms have to be memorized rather than generated, potentially creating cognitive overload (Chambers 2003:260). Pinker and Prince (1999:236–7) explicitly relate regular and irregular verbs to human conceptualization. Regular verbs form classical (Aristotelian) categories defined by necessary and sufficient criteria. Irregular verbs form prototype categories sometimes marked by family resemblances. “The classical category consisting of regular verbs,” they say, “is defined completely and explicitly by the nature of a rule in the context of a formal system, in this case, a rule within English grammar that applies to any word bearing the part-of-speech symbol ‘verb’ (unless it has an irregular root)” (Pinker and Prince 1999:236). “Family resemblance categories, in contrast,” in their terms (Pinker and Prince 1999:236), “are generalization of patterns of property correlations within a set of memorized exemplars.” Viewed from a cognitive perspective, they say, “[T]he properties of the regular and irregular classes of verbs in English show that classical categories and family resemblance categories …. are products of two different kinds of psychological processes: a formal rule system and a memorized, partially structured list of exemplars” (Pinker and Prince 1999:237). It follows that formal rule systems carry less cognitive cost than retrieval of memorized exemplars. From this perspective, certain grammatical
< previous page
page_19
next page >
< previous page
page_20
next page >
Page 20 tendencies take on cognitive motivation. For instance, the age-old tendency toward conjugation regularization in English grammar is driven by the need to limit the cognitive cost of retrieving memorized exemplars from the lexicon. From the earliest historical records of the English language, there has been a tendency to reduce unique forms in verb paradigms by making them the same as other forms in the same paradigm. For example, Modern English helped is both past tense and past participle, whereas Old English had healp (past tense singular), hulpon (past tense plural), and holpen (past participle). The tendency continues to this day with, for instance, contemporary regularization of past participles proved and mowed , identical to the past-tense forms, replacing obsolescent proven and mown. Vernacular dialects are ahead of standard dialects in conjugation regularization, as I have shown elsewhere (Chambers 2003:260–5). Vernaculars appear to be cognitively more efficacious than standard dialects, or perhaps it is more precise to view the continuum from the other end. Standard dialects may codify grammatical and phonological complexities partly as markers of social rank. As Kroch (1978:18) says, “Dominant social groups tend to mark themselves off symbolically as distinct from the groups they dominate and to interpret their symbols of distinctiveness as evidence of superior moral and intellectual qualities.” For Kroch and for Pinker and Prince (and for me), the idea that cognitive cost is involved in some kinds of linguistic complexity, however cogently based on common sense, remains an inference. In what follows, I demonstrate that it has empirical content by looking at the typological limits of one form of structuredependent grammatical processing. 2. STRUCTURAL LAYERS OF SUBJECT–VERB AGREEMENT In an earlier study in sociolinguistic typology, I assembled evidence from several studies showing that dialects add variable number agreement hierarchically according to a universal (English) progression based on person of subjects (Chambers 2004). By way of illustration, here is an AfricanAmerican Vernacular sentence from a blues recording made in Hollywood in 1947, written and sung by Chas Q. Price: (1) The things you done for me, baby, is too good to ever be told … By comparison to standard dialects, the sentence holds a couple of basilectal features, including is with plural subject (default singular) and done as past tense (conjugation regularization), not to mention the split infinitive. However, it is worth noting in passing that the subgenre is urban blues, not country blues, and other lines of the lyric are linguistically sophisticated. For
< previous page
page_20
next page >
< previous page
page_21
next page >
Page 21 instance, the resolution of this verse (the third line) is stated as a modal past perfect passive VP— (2) The things you done for me, baby, is too good to ever be told. The things you done for me, baby, is too good to ever be told. You must have been sent from heaven to satisfy my soul. For my purposes, the main interest is the subject–verb nonagreement in the first two lines. Subject is NPpl (the things you done for me ), but the copula appears to be ‘singular’. Actually, in this vernacular variety, the copula is invariant in form, and the default form is the one that is called the singular in the standard dialect. The universal English progression from basilect to standard in terms of subject–verb agreement is graphically illustrated in Figure 1.1. The grammatical subjects are shown on the abscissa. NPpl is near the left end, where agreement occurs early in the progression toward acrolects, represented here by the urban dialects of York, England, and Sydney, Australia.
Figure 1.1 Percentage of standard were with grammatical subjects in five communities (Chambers 2004:138, based on Schreier 2002: Table 3, 85).
< previous page
page_21
next page >
< previous page
page_22
next page >
Page 22 3. A DISCREPANCY IN STANDARD AGREEMENT One observation in Figure 1.1 is that even in acrolects subject–verb agreement never attains categoricity because of variable agreement with expletive there . Highly regulated though standard varieties are, speakers find it impossible to maintain subject–verb agreement with there and occasionally use nonagreeing variants like the (b) sentences: (3) a. There’re two men at the door. b.There’s two men at the door. (4) a. We are well aware that there are often unforeseeable discoveries coming out of pure research. b.We are well aware that there is often unforeseeable discoveries coming out of pure research. The probable cause of this grammatical lapse is that subject–verb agreement with expletive there requires a look-ahead mechanism. There is featureless (discussed in Chambers 2004:140 and in countless other places). It occupies the subject NP slot, but it is not logically (or semantically) the subject. Instead, the logical subject follows the copula. The verb is expected to agree with the displaced logical subject. In order for this agreement to take place, speakers must code the verb with a marker that agrees with a constituent that has not yet been uttered. Cognitively, the language faculty must look ahead at a constituent that is not yet explicit in the linear order of the utterance. Verb marking depends on registering grammatical properties (number in English, person and gender as well as number in some languages) of a constituent that is not yet in short-term memory. The terminology I use throughout the discussion is illustrated in the labels on the following sentence and defined under it: (5) The expletive fills the subject slot, but is a featureless nominal unmarked for person, number, gender, or other attributes. The displaced logical subject is the NP that would fill the subject slot if the expletive were not in it; expletive sentences can always be paraphrased with the displaced subject in the subject NP slot, with no loss of meaning but altered focus, as in Representatives of eleven nations are at this conference. Look-ahead mechanism is the anticipatory coding on the verb of attributes of an NP that follows it in the linear sequence of the utterance.
< previous page
page_22
next page >
< previous page
page_23
next page >
Page 23 4. COGNITIVE COST OF LOOK-AHEAD MECHANISMS The failure of agreement in sentences with there expletives causes some tension in normative situations. It is considered an error and subject to correction, but sociolinguistic research shows that nonagreement is highly frequent in natural conversations in this context. Its persistence in the face of active suppression, and especially its frequency, suggests that it might be a kind of grammatical breaking point. The look-ahead mechanism required by subject–verb agreement with expletives appears to tax human cognition to its limits. If so, it might provide a window into the language faculty. For if it is a limitation of human cognition, then it cannot merely be a fact about English. Look-ahead mechanisms must cause similar effects in all languages wherever they occur. I show that it is indeed a universal feature of human cognition based on evidence from three sources. First, languages in which the expletive subject is homophonous with a nonexpletive pronoun always avoid lookahead mechanisms by marking agreement as if the expletive were the pronoun. Second, languages that have agreement with postverbal subjects make grammatical compromises by agreeing with left conjunct only in conjoined subjects (which I illustrate from English, Arabic, Spanish, and Biblical Hebrew). Third, VS languages, in which subject–verb agreement would necessarily require a look-ahead mechanism, show a greater than chance tendency to have no subject–verb agreement rules (Keenan’s principle, which I illustrate from Celtic). The overriding generalization is that variability is expected, perhaps inevitable, with look-ahead processes. 5. THIRD-PERSON SINGULAR PRONOUNS AND EXPLETIVES First, languages in which the expletive subject is homophonous with a paradigmatic pronoun always avoid look-ahead mechanisms by marking agreement as if the expletive were the pronoun. In French, the expletive il is homophonous with the third-person singular masculine pronoun. Agreement, in effect, is with the homophone whether the displaced or logical subject is singular (a) or plural (b, c; sentences from O’Grady 2005:97): (6) a. Il est arrivé un homme. there be.Sg arrived a man ‘There arrived a man.’ b. Il est arrivé deux hommes there be.Sg arrived two men ‘There arrived two men.’
< previous page
page_23
next page >
< previous page
page_24
next page >
Page 24 c. Il *sont arrivé deux hommes …. be.Pl … Nonagreement holds in English in exactly the same way when it is used as a dummy subject,1 as in cleft sentences like these: (7) a. It was the three penalties in the first period that cost us the game. b. It *were the three penalties … The mismatch in (b) between it as singular and were as plural seemingly rules out agreement. Moreover, eliminating the mismatch between subject and verb as in (c) rules out the dummy subject reading: c. *They were the three penalties … Subject–verb agreement never occurs in languages with expletives that are identical to third-person singular pronouns. Superficially, it might look as if the form of the pronoun dictates in some sense the third-person singular form of the verb because the default singular form of the verb happens to be the same as third-person singular masculine form. But it and il in these examples do not express third-person singularity. They have no content at all unless it is the content of the logical subject. If we go looking for an explanation for the synonymy of the forms, rather than the pronoun determining the verb form, it is probably the opposite. The difficulty of look-ahead agreement and the need to avoid it ostensibly favours the default singular form of the verb. In this view, it is the compatibility of that verb form with the third-person singular pronoun that leads languages to use it as an expletive. Some English vernacular dialects prefer it to there as the expletive. AAVE uses it about 80 percent of the time (Labov 1972:270). In AAVE, nonagreement is categorical with either it or there : (8) a. It’s a little boy at the door = (uncommonly) There’s a little boy at the door. b. It’s two little boys at the door = (uncommonly) There’s two little boys at the door. Typologically, expletives can either be featureless pronominals (like there ) or forms identical to third-person singular masculine pronouns. The universal tendency to avoid look-ahead mechanisms in grammatical processing means that most languages use default singular verb forms after expletive subjects. Default singulars are usually similar or identical to third-person singular masculine verb forms, giving the appearance that the expletive is somehow equivalent to the third-person singular masculine pronoun regardless of the
< previous page
page_24
next page >
< previous page
page_25
next page >
Page 25 properties of the logical subject. This superficial compatibility apparently explains why many languages choose that pronoun as the expletive. 6. LIMITING THE SCOPE OF LOOK-AHEAD MECHANISMS Second, dialects that require look-ahead agreement tend to make grammatical compromises that lighten the cognitive load. For instance, when the logical subject is compound, most grammars expediently limit the scope of looking ahead. The obvious compromise is for verbal agreement to ‘see’ only the left conjunct in conjoined subjects. Standard English again provides a case in point. (9) a. In hockey there is a goalie and two defencemen. b. … *are … c. In hockey there is a goalie, a centre, two wingers and two defencemen. d. In hockey there are two defencemen, a goalie, a centre and two wingers. Notice, in passing, that (d) is also possible with default singular: In hockey there is two defencemen, a goalie, a centre and two wingers . The variability does not affect the argument developed in this section. Left-conjunct agreement lightens the cognitive load by requiring the look-ahead mechanism to process only the most proximate constituent instead of the entire constituent. In instances where the most proximate constituent is singular and the logical subject constituent is plural, agreement marking is strictly speaking incongruous: a. The motive for left-conjunct agreement is to limit the scope of the look-ahead mechanism in order to make it manageable. Again, if limiting the scope of look-ahead mechanisms belongs to human cognition, it must show up in grammars of other languages besides English. Left-conjunct agreement occurs, for instance, in Spanish and Arabic with VS clauses (Doron 2000:76–7; Doron also provides examples from Modern Irish, Greek, Biblical Hebrew, and Modern Hebrew): (10) Spanish: Estaba abierta la tienda y el mercado was.3s open.3fs the shop.f and the market.m ‘The shop and the market were open.’
< previous page
page_25
next page >
< previous page
page_26
next page >
Page 26 (11) Standard Arabic: la ibat maryam wa- zayd fi-l-bayt played3FS MariamF and Zaydm in-the-house ‘Mariam and Zayd played in the house.’ (Rana Fahoum, p.c.) In both sentences, the displaced logical subject NP is a plural compound made up of feminine singular and masculine singular NPs. The verb, however, is coded for agreement with feminine singular, that is, with the attributes of the first NP in the compound. The result is strictly speaking ungrammatical—if the subject NP preceded the verb in these sentences, instead of following it, the agreement marking on the verb would be third person plural. However, the result is not ungrammatical because agreement is always marked for the left conjunct only in these structures:
The motive appears to be to constrain the scope of the look-ahead mechanism. Biblical Hebrew provides telling examples of the cognitive cost of look-ahead mechanisms or, more precisely, the need to avoid them. Biblical Hebrew allows both SV and VS constructions. In SV constructions, with the subject NP preceding the verb, whether it is conjunct or not, the verb takes on agreement marking with the whole NP, as expected (Doron 2000:79): (12) u-mo:šε: aharo:n w- u:r a:lu: ro:š haggib a: and-Moses Aaron and-Hur climbed3mp head the.hill ‘And Moses, Aaron and Hur went up to the top of the hill.’ (Exodus 17:10) However, when the subject NP follows the verb, agreement is with the left conjunct only, as in the Spanish and Arabic examples discussed previously (Doron 2000:80): (13) wat-tašar dbo:ra: u:-ba:ra:q bεn abi:no: am and-sang3FS Deborah and-Barak son Abinoam ‘Then sang Deborah and Barak the son of Abinoam.’ (Judges 5:1) There are, then, two distinct agreement principles in Biblical Hebrew (as there are in Spanish and other languages) depending on whether the constituent order is SV or VS. Incontrovertible evidence that the difference in the agreement patterns is the result of the cognitive cost of looking-ahead rather than from some fine-tuned grammatical nuance comes from another of Doron’s sentences (2000:80):
< previous page
page_26
next page >
< previous page
page_27
next page >
Page 27 (14) wat-ta:qa:m ribqa: w-na aro:tεyha: wat-tirkabna: al and-rose3fs Rebecca and-maids.her and-rode3fp on haggəmalli:m the.camels ‘And Rebeka arose, and her damsels, and they rode upon the camels.’ (Genesis 24:61) literally : And arose-she Rebecca and her maids, and rode-they on he camels. The sentence is compound, with two clauses linked by the conjunction wat- . Constituent order in both conjuncts is verb first with compound subject NPs following the verb. In the first clause, as expected, the verb ‘arose’ agrees only with the left conjunct, ‘Rebecca’, in the subject NP ‘Rebecca and her maids’, that is, the agreement on the verb is third person feminine singular. However, in the second clause, the verb ‘rode’ agrees with the entire subject NP, ‘Rebecca and her damsels’, that is, the agreement on the verb is third person feminine plural. Although the subject NP is elided in the second clause, its presence in the first clause, where it precedes the verb, triggers subject–verb agreement that is characteristic of SV structures. The agreement mechanism functions like this: Agreement is marked differently in the two clauses. But the two clauses do not differ structurally. The first clause marks agreement with the left conjunct, in the normal manner for VS clauses in the language. The second clause marks agreement ‘incorrectly’ for a VS clause. Instead, it marks agreement as if it were an SV clause. In fact, it agrees with the subject NP as it occurs in the preceding clause. No agreement rule in any language stipulates subject–verb agreement with subjects in preceding clauses. The second clause marks agreement as if it were SV instead of VS, but the change is accidental, not structural. The occurrence of the full subject NP preceding the verb of the second clause triggers agreement with the full NP. Agreement in the second clause proceeds as if it were a look-back mechanism. Languages mark agreement with the full NP when it does not require looking ahead, but they settle for compromises like left-conjunct agreement if it requires looking ahead. 7. IGNORING OBLIGATORY LOOK-AHEAD MECHANISMS Another strategy for coping with the cost of look-ahead mechanisms in language would obviously be to ignore them. As we will see, this strategy is also common.
< previous page
page_27
next page >
< previous page
page_28
next page >
Page 28 When we look at VSO languages, in which subject–verb agreement obligatorily requires a look-ahead mechanism, we become aware of further compelling evidence that grammars avoid look-ahead mechanisms. VS languages show a strong tendency to have no subject–verb agreement rules at all (Payne 1990:15, citing Edward Keenan’s principle). This seems to hold almost without exception for languages in which the pragmatically neutral clause order has the subject immediately after the verb, without other constituents between, that is, strict VSO languages, which include Celtic, Eastern Nilotic, Polynesian, and Jacaltec (Payne 1990:11).2 For Celtic, Hendrick (2000:29–31) points out that when the subject is a pronoun, the verb form is inflected with a suffix encoding person and number. For instance, these Welsh sentences consist of a verb stem and the pronominal suffix expressing the person and number of the subject: (15) a. Canodd singPASTsingular ‘He sang’ b. Canon singPASTplural ‘They sang’ However, when the subject NP is overt, the syntax takes on what Hendrick (2000:29) calls “an analytic pattern (homophonous to the third singular form) which remains invariant despite the grammatical features of the subject.” In other words, there is nonagreement. In the following Welsh sentences, the subject is Sion ‘Sean’, third person singular, in the first, and y plant ‘the children’, third person plural, in the second. In both sentences, the verb form is canodd, the default form for ‘sing’ in the past tense: (16) Welsh a. Canodd Sion. sing past Sion ‘Sion sang.’ a. Canodd y plant. singpast the children ‘The children sang.’ Similarly, in these two Irish sentences, the verb is invariably chuirfeadh notwithstanding the number difference in the subject NPs, which are sé ‘he’, third person singular, in the first sentence and siad ‘they’, plural, in the second: (17) Irish a. Chuirfeadh sé isteach ar an phost sin. putcond he in on the job that ‘He would apply for that job.’
< previous page
page_28
next page >
< previous page
page_29
next page >
Page 29 b. Chuirfeadh siad isteach ar an phost sin putcond they in on the job that ‘They would apply for that job.’ As Hendrick (2000:29) says, “sentences (a) and (b) show that the verb does not change its morphological shape as the overt subject changes in number.” Obviously, VSO languages require a look-ahead mechanism in order to mark subject agreement on the verb. Typologically, they avoid it. VSO languages, wherever they occur in the world, regardless of genetic affiliation and Sprachbund, tend to lack verb–subject agreement as grammatical processes. Languages with other word orders often include in their grammars subject–verb agreement in acrolectal varieties. In those word orders, subject– verb agreement is a look-back mechanism, and the subject constituent with which verb is marked for agreement is present in short-term memory in the linear sequence of the utterance. 8. LOOK-BACK MECHANISMS CARRY COGNITIVE COST Evidence from these three sources all point to the conclusion that look-ahead mechanisms tax human cognition. First, languages that mark agreement with postverbal subject NPs limit the scope of looking ahead by agreeing with the closest conjunct, the leftmost one, in postverbal subject NPs. Second, languages that use expletives (featureless place holders) for displaced subject NPs tend to avoid the look-ahead mechanism that would be required for verb–subject agreement by using the default singular verb form. As a corollary, the third-person singular pronoun often occurs as the expletive apparently because it is congruous with the default singular verb form. Third, strict VS languages generally avoid what would be obligatory look-ahead mechanisms by having no subject–verb agreement. One immediate conclusion that follows from the problems posed by look-ahead mechanisms is to account for the discrepancy in subject–verb agreement patterns in standard varieties of English. As we saw previously, English dialects show a hierarchical tendency to increase domains of subject–verb agreement as one goes up the continuum from vernaculars to standard dialects. In the basilect, represented in Figure 1.1 by Tristan da Cunha in 1850–1945, there is no subject–verb agreement. In the acrolects, represented in the figure by urban dialects of York and Sydney, subject–verb agreement is virtually categorical with all subject NPs but one. In between, mesolects usually have variable subject–verb agreement on an implicational order shown on the abscissa of Figure 1.1. The discrepancy in subject–verb agreement in the acrolects comes with there subjects. Subject–verb agreement after there drops off precipitously,
< previous page
page_29
next page >
< previous page
page_30
next page >
Page 30 settling at around 50 percent or less. In the light of what we now know about look-ahead mechanisms in grammar, this result is no longer surprising. Instead, it is expected. This conclusion will undoubtedly prove disquieting to some speakers of standard dialects. Fastidious speakers, in fact, decry sentences with default singulars (as in There’s too many cooks in the world). Sociolinguistic evidence has shown that such sentences are highly frequent in the spoken language for all social classes, not excluding fastidious speakers. The evidence I have presented about the cognitive cost of look-ahead mechanisms suggests that they are inevitable. More generally, we can see that look-ahead mechanisms will be a source of dialect variation whenever they are required in grammars. Many grammatical rules attain invariability or categoricity in the acrolect. This is true even of rules with numerous exceptions that require memorization, as in the irregular verbs discussed earlier. Adult speakers of the standard dialect almost never falter on even the most irregular paradigms, say, go-went-gone or do-does-did-done. However, they do occasionally falter on subject–verb agreement with expletives. The difference shows that speakers find it easier to retrieve irregular lexemes as they arise in the stream of speech than to process grammatical markers determined by lexemes not yet spoken. The difference again emphasizes the cognitive challenge posed by look-ahead mechanisms. By the same reasoning, it seems clear that vernacular varieties will eschew any grammatical devices that require look-ahead mechanisms. Indeed, as we have seen, acrolectal varieties of VSO languages such as Celtic themselves eschew verb–subject agreement that inevitably requires looking ahead. 9. IMPLICATIONS FOR LANGUAGE SPECIFICITY AND STRUCTURE DEPENDENCE Chomsky’s position on the language faculty maintains that it is made up of language-specific rules and representations. In the strongest version of his theory, such rules and representations are held to be independent of other general cognitive principles. They are moreover structure dependent, meaning that they apply over domains that are not necessarily linear in the temporal stream of an utterance, but also structural in the sense that they operate on constituents that can be multilayered and discontinuous. The problems posed by look-ahead mechanisms for the language faculty, it seems to me, call into question both language specificity and structure dependence, at least in the strongest interpretation. Rules of grammatical production are, we have seen, impeded or compromised when look-ahead mechanisms come into play. Languages like English that include in their grammars subject–verb agreement in pragmatically neutral word orders have variable agreement in verb–subject constructions. Languages like Spanish and Biblical Hebrew that have full-subject
< previous page
page_30
next page >
< previous page
page_31
next page >
Page 31 agreement in subject-first constructions relax the restrictions when the verb precedes the subject and require only left-conjunct agreement. The difference in the two situations is that, in the categorical agreement contexts, the NP that will be coded inflectionally for number, person, gender, and the like on the verb has already been uttered when the verb stem that takes the inflection is uttered. The constituent is already explicit in the linear order of the utterance. It is already registered in short-term memory. Explicitness and memory are cognitive adjuncts that are applied in all kinds of cognitive tasks. When they are missing, as in verb–subject constructions, their absence appears to impinge on grammatical processing. Similarly, processing in terms of constituency appears to get compromised when look-ahead mechanisms come into play. Languages that otherwise include full-subject verbal agreement, as we have seen, settle for agreement with only the closest constituent even though that constituent may be, as in conjoined subject NPs, a mismatch for the inflection that would appear otherwise. The likely cause is the failure of access to the full structure of the construction as it is being uttered. Some aspects of structure are perhaps incomplete or incipient during speech acts, particularly those structures that are multilayered or compounded. Here, then, are instances where grammatical processing becomes incomplete or irregular or nonexistent not for any grammatical principle, but because memory and explicitness are challenged. The search for an explanation as to the failure of English acrolects to impose subject–verb agreement with there expletives leads ultimately to evidence of cognitive limits on the language faculty. Structure dependency breaks down with look-ahead mechanisms, and grammars seemingly violate their own principles by recognizing partial and incomplete constituents. Variation becomes an inevitable aspect of grammar. NOTES 1. Dummy subjects are expletives in the sense that they are featureless pronominals, but they fail the paraphrase test noted earlier. They cannot be paraphrased by replacing the expletive with the logical subject: *The three penalties in the first period was that cost us the game. 2. These are language families that are strictly VSO. I did not look at other verb-first orders like VOSObl (Fijian, Toba Batak), VOOblS (Malagasy, Tzeltal), or Vfree-NP-order (Tagalog) (Payne 1990:11). REFERENCES Carnie, A., and E. Guilfoyle, eds. 2000. The Syntax of Verb Initial Languages. Oxford, UK: Oxford University Press. Chambers, J.K. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Significance . 2nd edition. Oxford, UK, and Malden, US: Blackwell.
< previous page
page_31
next page >
< previous page
page_32
next page >
Page 32 ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective , edited by B. Kortmann, 127–45. Berlin and New York: Mouton de Gruyter. Doron, E. 2000. VSO and left-conjunct agreement: Biblical Hebrew vs. Modern Hebrew. In Carnie and Guilfoyle, eds. 2000. 75–95. Hendrick, R. 2000. Celtic initials. In Carnie and Guilfoyle, eds. 2000. 13–37. Kroch, A. 1978. Toward a theory of social dialect variation. Language in Society 7: 17–36. Labov, W. 1972. Language in the Inner City . Oxford: Basil Blackwell. O’Grady, W. 2005. Syntactic Carpentry: An Emergentist Approach to Syntax . Mahwah, NJ, and London: Lawrence Erlbaum Associates. Payne, D.L. 1990. The Pragmatics of Word Order: Typological Dimensions of Verb Initial Languages. Berlin and New York: Mouton de Gruyter. Pinker, S., and A. Prince. 1999. The nature of human concepts: evidence from an unusual source. In Language, Logic and Concepts: Essays in Memory of John Macnamara, edited by R. Jackendoff, P. Bloom, and K. Wynn, 221–62. Cambridge, US, and London, UK: MIT Press. Schreier, D. 2002. Past be in Tristan da Cunha: The rise and fall of categoricality in language change. American Speech 77: 70–99.
< previous page
page_32
next page >
< previous page
page_33
next page >
Page 33 2 Vernacular Universals and Angloversals in a Typological Perspective Benedikt Szmrecsanyi and Bernd Kortmann 1. INTRODUCTION In this study, we endeavor to take a typological perspective on language-internal variation in English. Our inquiry will be based on the largest comparative survey to date of entire grammatical subsystems of varieties of English worldwide (Kortmann and Szmrecsanyi 2004). A catalogue of 76 morphosyntactic features taken from the 11 core areas of English morpho-syntax will be investigated for 46 (groups of) nonstandard varieties of English around the world. Our analytical point of departure is the notion that there are different reasons that languages, or varieties of a given language, should exhibit the same linguistic features. Such features may fall into any one of the following categories: (i) genuine universals (e.g. all languages have vowels); (ii) typoversals, i.e. features that are common to languages of a specific typological type (e.g. SOV languages tend to have postpositions ); (iii) phyloversals, i.e. features that are shared by a family of genetically related languages (e.g. languages belonging to the Indo-European language family distinguish between masculine and feminine gender); (iv) areoversals, i.e. features common to languages which are in geographical proximity to each other (e.g. languages belonging to the Balkan Sprachbund have finite complement clauses); (v) vernacular universals, i.e. features that are common to spoken vernaculars (e.g. spoken vernaculars tend to have double negation); (vi) features that tend to recur in vernacular varieties of a specific language: angloversals, francoversals, etc. (e.g. in English vernaculars, adverbs tend to have the same morphological form as adjectives); (vii) varioversals, i.e. features recurrent in language varieties with a similar socio-history, historical depth, and mode of acquisition (e.g. L2 varieties of English tend to use resumptive pronouns in relative clauses). With areoversals constituting the level of generalization below which orthodox linguistic typology normally does not stray, the present chapter
< previous page
page_33
next page >
< previous page
page_34
next page >
Page 34 focuses on (iv)–(vii) in varieties of English worldwide. Taking the notion of ‘vernacular universals’ (cf., for instance, Chambers 2004) as its starting point, the present study embarks on an in-depth exploration of candidate (uni)versals and some implicational tendencies in different types and areal groups of the world’s Englishes. We also utilize principal component analysis to explore large-scale parameters of variance, which are possibly recurrent cross-linguistically, in our global survey of varieties of English. On a methodological note, then, this study is in keeping with the Freiburg program of combining functional Greenbergian typology with dialectology, which argues, in a nutshell, that the observable patterns of cross-dialectal and crossvarietal variation can be best analyzed and interpreted in terms of the same methodological and interpretational apparatus familiar from the typological study of large-scale cross-linguistic variation (see Kortmann 2004b for a collection of papers in this spirit). 2. DATA: THE HANDBOOK OF VARIETIES OF ENGLISH This study rests empirically on the Handbook of Varieties of English (Kortmann et al. 2004), the most comprehensive reference work to date on the phonology (Volume 1) and morphosyntax (Volume 2) of varieties of English around the globe. The Handbook contains survey articles by about 100 specialists, covering some 60 almost exclusively nonstandard varieties or groups of varieties: native vernaculars (henceforth: L1 varieties), which include all main national standard varieties, distinctive ethnic, regional, and social varieties, e.g. African American Vernacular English; English-based pidgins and creoles, such as Tok Pisin; and the major English as a Second Language (henceforth: L2 varieties) varieties, e.g. Malaysian English. Crucially, all of the varieties surveyed are spoken ones. What will take center stage in the present study is the morphosyntactic survey of the multimedia reference tool, available on CD-ROM and online (http://www.mouton-online.com), accompanying the Handbook (cf. Kortmann and Szmrecsanyi 2004). This first-ever comprehensive survey of nonstandard English morphosyntax worked in a very simple way: we compiled a catalogue of 76 features—essentially, the ‘usual suspects’ in previous dialectological, variationist, and creolist research—and sent out this catalogue to the authors of the chapters in the morphosyntax volume of the Handbook. For each of these 76 features, the contributors were asked to specify into which of the following three categories the relevant feature falls: A pervasive (possibly obligatory) or at least very frequent B exists but a (possibly receding) feature used only rarely, at least not frequently C does not exist or is not documented
< previous page
page_34
next page >
< previous page
page_35
next page >
Page 35
Figure 2.1 Varieties sampled in the morphosyntactic database of the Handbook. Kortmann and Szmrecsanyi (2004:1142–5) discuss the survey procedure, the varieties of English sampled, as well as the advantages and drawbacks of the method in considerable detail. Suffice it to say here that the 40 Handbook authors provided us with data on 46 nonstandard varieties of English. Figure 2.1 visualizes the geographical distribution of varieties of English in the survey. As can be seen, all seven anglophone world regions (British Isles, America, Caribbean, Australia, Pacific, Asia, and Africa), as well as a fair mix of L1 varieties, L2 varieties, and pidgins/creoles are represented in the survey. The features are numbered from 1 to 76 (see the appendix for the feature catalogue in its entirety) and include all major phenomena discussed in previous survey articles on grammatical properties of (individual groups of) nonstandard varieties of English, with a slight bias towards features observed in L1 varieties. They cover 11 broad areas of morphosyntax: pronouns, the noun phrase, tense and aspect, modal verbs, verb morphology, adverbs, negation, agreement, relativization, complementation, and discourse organization and word order. 3. VERNACULAR UNIVERSALS OR VERNACULAR ANGLOVERSALS? We have drawn on the Handbook’s database elsewhere to offer an extended empirical discussion of (i) unrestricted angloversals, i.e. the most frequent features in varieties of English across the board (Kortmann and Szmrecsanyi 2004:1153–60); (ii) unrestricted areoversals (Kortmann and Szmrecsanyi 2004:1160–84); and (iii) unrestricted varioversals, i.e. top features in New
< previous page
page_35
next page >
< previous page
page_36
next page >
Page 36 Englishes (Kortmann and Szmrecsanyi 2004:1184–94). Here, we seek to explore in more depth the sociodialectological notion of ‘vernacular universals’ or ‘vernacular roots’. According to Jack Chambers, these universals comprise “a small number of phonological and grammatical processes [that] recur in vernaculars wherever they are spoken […] not only in working class and rural vernaculars, but also in […] pidgins, creoles and interlanguage varieties” (2004:128). Crucially, the putative ubiquity of such features is argued to be unlikely to be due to sociolinguistic diffusion, which is why they must be “primitive features of vernacular dialects” (Chambers 2003:243)—that is, unlearned and thus innate. Chambers (2004:129) lists the following four morphosyntactic candidates for vernacular universals (indicated in square brackets are the features in our 76-features catalogue which correspond most closely to the four morphosyntactic processes named by Chambers): (1) conjugation regularization, or levelling of irregular verb forms: John seen the eclipse, Mary heared the good news [36–39]; (2) default singulars, or subject–verb nonconcord: They was the last ones [55, 59; marginally 53 and 54]; (3) multiple negation, or negative concord [44]; (4) copula absence, or copula deletion: She smart , We going as soon as possible [57; possibly 58, 73]. So, how frequent are these features in vernacular varieties of English sampled in our database? Figure 2.2, which plots the percentages of varieties where they are attested (as either pervasive or existing, but rare), provides details. Certainly, these features are rather frequent: auxiliary deletion in wh- questions [73] recurs in 78% of the varieties in the survey, multiple negation [44] is attested in 76%, and regularization of irregular verb paradigms [36] in 70%. Crucially, Figure 2.2 also suggests that in Englishes spoken in the
Figure 2.2 Frequencies of some candidates for vernacular angloversals in the Handbook’s database.
< previous page
page_36
next page >
< previous page
page_37
next page >
Page 37 Americas, these features are even more widespread; indeed, it is here that they are truly universal. Yet on a global scale, none of the features is attested in more than 80% of varieties of English worldwide; thus, frequent they are, but universal they are not. What is more, there are features in our survey which are considerably more pervasive globally (although note that no feature in our 76-feature survey is actually attested in each and every variety sampled). This is why we would like to suggest that these features— conjugation regularization, multiple negation, and copula/auxiliary absence—be more appropriately considered areoversals precisely because of their preponderance in vernacular Englishes spoken in America. Along these lines, we believe that some terminological caution might be warranted in the use of the term ‘universal’. Notice here that this would be true even if we found a nonstandard feature attested in each and every variety of English in the world: according to most definitions, vernacular universals should have counterparts in the vernaculars of other languages as they are “the outgrowths of […] rules and representations in the bioprogram” (Chambers 2004:129). Yet, of the four candidates given, multiple negation is, at least at first glance, the only convincing one in a truly cross-linguistic perspective. What about copula absence? It turns out that in a cross-linguistic perspective, zero copulas are simply not that frequent. Figure 2.3 reproduces Map 120 (Stassen 2005) of the World Atlas of Language Structures (Haspelmath et al. 2005) which plots languages where zero copula for predicate nominals (of the type He __ a sailor) is possible against languages where it is impossible. What emerges is that, in a significant majority (211:175) of languages sampled by Stassen, prenominal zero copulas are impossible. If zero copulas were somehow part of the language faculty,
Figure 2.3 Zero copula for predicate nominals—the cross-linguistic perspective (Stassen 2005). Black triangles indicate languages where the phenomenon is impossible, and gray dots indicate languages where the phenomenon is possible. Sample size: 386 languages.
< previous page
page_37
next page >
< previous page
page_38
next page >
Page 38 there should be more languages that have them—and this means that the phenomenon is, as a matter fact, not such a promising candidate for a vernacular universal.1 As for subject–verb nonconcord (or default singulars), we wish to highlight that a language needs structural means to display subject–verb concord (at least historically) to exhibit subject–verb nonconcord in any nonvacuous way. Yet many languages—for instance, highly isolating languages such as Vietnamese—do not show agreement at all; so vernacular Vietnamese showing nonconcord is an utterly unremarkable fact. Conversely, one encounters the problem that there are many languages (e.g. Hungarian) that have a lot more agreement than English; this is unexpected if the ‘bioprogram’ has a proclivity toward nonconcord. Along similar lines, a language must have irregular verb forms (again, at least historically) to exhibit conjugation regularization. Yet Turkish, for instance, is commonly considered not to have irregular verbs— how could it, then, display conjugation regularization? It might be contended that Turkish does not, in fact, refute the argument because Turkish can be seen as a maximally consistent implementation of the ‘conjugation regularization’ principle. But then again, if total consistency is theoretically possible, every single language with a good deal of irregular verbs (e.g. Italian) would play havoc with the bioprogram explanation. In short, what we mean to suggest here is that features such as default singulars or conjugation regularization are things which are rather typical of languages that, like English, have some inflectional morphology but are in the process of getting rid of what remains. But what is happening in nonstandard varieties of English and, possibly, languages belonging to the same morphological type as English does not necessarily apply to vernaculars of inflectional or agglutinating languages such as Finnish, Hungarian (which have a great deal of grammatical agreement), or Turkish. A feature such as default singulars is maybe better referred to as a vernacular typoversal—a feature, in other words, which is typical of vernaculars of inflecting languages. It is unlikely that the ‘language faculty’ would provide for special rules and representations applying to English-like vernaculars; default singulars and conjugation regularization are just too conditioned on the typological type of English to be cross-linguistically ‘universal’. At the same time, notice that it is not only loss of agreement or loss of redundancy that we can observe in vernaculars; individual vernaculars have and can indeed be shown to currently develop a more elaborate inflectional morphology or, for example, an agreement system than the standard variety has (cf. several studies in Barbiers et al. 2002; Kortmann 2004b). Given these observations, we believe that a candidate feature for a vernacular universal should, at a minimum, fulfill the following criteria: • The candidate feature should be attested in a vast majority of a given language’s vernacular varieties.
< previous page
page_38
next page >
< previous page
page_39
next page >
Page 39 • The candidate feature should not be patterned geographically or according to variety type (in the case of English: L1, L2, or pidgin/ creole). • For the sake of cross-linguistic validity, the candidate feature should not be tied to a given language’s typological make-up (inflectional, agglutinating, etc.).2 • The candidate feature should be cross-linguistically attested in a significant number of the world’s languages (especially among the world’s many languages without a literary/philosophical tradition). 4. VARIETIES OF ENGLISH: IMPLICATIONAL TENDENCIES We now turn to a number of correlational tendencies and implications between individual nonstandard features in the Handbook’s survey. Analysts of cross-linguistic variation make a distinction between two types of correlational tendencies: ‘biconditional implications’, also known as ‘equivalences’ (e.g. ‘if in a language the genitive follows the noun, then the complement follows the adposition, and vice versa’; Greenberg 1963), and ‘one-way implications’, also known as ‘preferences’ (e.g. ‘if a language has a marked singular, it has also a marked plural, but not necessarily vice versa’; Greenberg 1966). Crucially, most analysts consider that the correlations need not be absolute or perfect, but can have exceptions. 4.1 Biconditional Implications Let us begin by discussing some biconditional implications (or equivalences) exhibited in the Handbook’s survey. To reiterate, biconditional implications are nonhierarchical correlations, such that varieties will attest just one of two features, but not the other. Out of 76 × 75/2 = 2,850 potential candidate pairings (thanks to the survey’s coverage of 76 nonstandard features), there are no biconditional implications which hold, without exceptions, for all varieties in our sample. Admitting a limited number of counterexamples as acceptable, however, the following biconditional implications emerge as the front runners for biconditionalimplicational angloversals: • [45] ain’t as the negated form of be—[46] ain’t as the negated form of have (this biconditional implication holds in 94% of all varieties) • [12] noncoordinated subject pronoun forms in object function—[13] noncoordinated object pronoun forms in subject function (89%) • [23] habitual do—[27] do as a tense and aspect marker (89%) • [63] relative particle as—[64] relative particle at (89%) So, 94% of the varieties sampled either have both ain’t as the negated form of be and ain’t as the negated form of have or they have neither of
< previous page
page_39
next page >
< previous page
page_40
next page >
Page 40 these things. This relationship—which is consonant with many previous dialect descriptions (cf. Anderwald 2003:149–50)—can be neatly visualized in a tetrachoric table where unpredicted cells are shaded in gray:
Australian Vernacular English is a typical variety in that it exhibits both ain’t as the negated form of b e and ain’t as the negated form of have . Scottish English also conforms with the biconditional implication, in that it has neither of these features. New Zealand English, Irish English, and Welsh English, by contrast, are the three exceptions in our database to the statistical tendency in that only one but not the other of the uses of ain’t is attested in them. In a similar vein, noncoordinated subject pronoun forms in object function [12] and noncoordinated object pronoun forms in subject function [13] also tend to go together, as do habitual do [23] and do as a tense and aspect marker [27].3 Likewise, if a variety exhibits the relative particle at [64], it will have the relative particle as [63] and vice versa. In short, pairings such as these are best seen as feature bundles instead of pairs of independent features. What is to be said about biconditional-implicational varioversals? Distinguishing among L1 varieties, L2 varieties, and pidgins and creoles, Table 2.1 provides a list of some perfect—i.e. exceptionless-biconditional implications. Thus, among L1 varieties (Table 2.1a), we find a correlation such that when in a variety past forms of irregular verbs can replace participle forms [38], unmarked forms [37] are also possible and vice versa. Further, if a L1 vernacular displays noncoordinated object pronoun forms in subject function [13], noncoordinated subject pronoun forms in object function [12] are also possible and vice versa. No L1 vernacular in our sample, then, displays just one of these things. Among both L2 varieties and English-based pidgins and creoles (Table 2.1b–c), we observe that if a variety has one use of ain’t (either for be, have , or as a generic negator), it will also have the other two uses of the form. This versatility of ain’t , then, can be considered a biconditional-implicational universal of New Englishes. In L2 varieties, we also note, for instance, that if a variety has no as a preverbal negator [50], it will also display the past tense/anterior marker been [29] and vice versa. In English-based pidgins and creoles (Table 2.1c), lack of inversion in main clause yes/no questions [74] and lack of inversion/lack of auxiliaries in wh- questions [73] always occur in tandem, which is another way of saying that English-based pidgins and creoles either invert in questions or they don’t.
< previous page
page_40
next page >
< previous page
page_41
next page >
Page 41 Table 2.1 Some Perfect (i.e. Exceptionless) Biconditional-Implicational Varioversals Exhibited in the Database* Feature 1 Feature 2 a. L1 Varieties [38] levelling of preterite/ppt verb forms: past [37] levelling of preterite/ppt verb forms: unmarked forms replacing the participle [13] noncoordinated object pronoun forms in [12] noncoordinated subject pronoun forms in object subject function function b. L2 Varieties [46] ain’t as the negated form of have [45/47] ain’t as the negated form of be/ain’t as generic negator before a main verb [50] no as preverbal negator [29] past tense/anterior marker been c. English-Based Pidgins and Creoles [46] ain’t as the negated form of have [45/47] ain’t as the negated form of be/Ain’t as generic negator before a main verb [74] lack of inversion in main clause yes/no [73] lack of inversion / lack of auxiliaries in wh- questions questions *The occurrence/non-occurrence of feature 1 is conditioned on the occurrence/non-occurrence of feature 2, and vice versa. The designation of pair parts as ‘feature 1’ and ‘feature 2’, respectively, is arbitrary. 4.2 One-Way Implications We now turn our attention to one-way implications of the type ‘if A then B but not necessarily vice versa’. Overall, the varieties sampled in our survey exhibit the surprisingly large number of 247 perfect (i.e. 100% felicitous) bidirectional statistical correlations, out of 76 × 75 = 5,700 potential candidate pairings; at the 85% felicity level, there are no less than 721 such correlations. Needless to say, a good deal of these prima facie implications turn out, upon closer inspection, to be meaningless due to lack of statistical significance (cf. Cysouw 2003). Still, the database appears to contain a healthy number of substantial one-way implications. In an inevitably eclectic fashion, Table 2.2 lists some perfect and statistically significant oneway implications, indicating statistical robustness in the rightmost column.4 In Table 2.2a, we find a number of one-way implicational angloversals, i.e. one-way implications that hold for each and every variety of English in the sample. Thus, any variety that has would in if-clauses [31] also displays loosening of the sequence of tense rule [30], but not necessarily vice versa. The following tetrachoric table illustrates this (as before, unpredicted cells are shaded in gray):
< previous page
page_41
next page >
< previous page
page_42
next page >
Page 42 Table 2.2 Some Perfect (i.e. Exceptionless) One-Way Implicational Angloversals (a.) and Areoversals (b.–d.) Exhibited in the Database* Feature 1 Feature 2 Fisher’s Exact Test a. All Varieties [31] would in if clauses [30] loosening of sequence of tense rule p < .001 [47] ain’t as generic negator before a main [45/46] ain’t as the negated form of be/have p < .005 verb [63] relative particle as [61] relative particle what / [66] gapping or zero- p < .05 relativization in subject position b. L1 Varieties [27] do as a tense and aspect marker [23] habitual do p < .005 [45] ain’t as the negated form of be [44] multiple negation / negative concord p < .05 [22] habitual be [26] be as perfect auxiliary p < .05 [73] lack of inversion / lack of auxiliaries in [74] lack of inversion in main clause yes/no p < .005 wh-questions questions c. L2 Varieties [75] like as a focussing device [76] like as a quotative particle p < .05 [4] regularized reflexives-paradigm [8] generic he/his for all genders p < .05 [63] relative particle as [61] relative particle what / [66] gapping or zero- p < .05 relativization in subject position d. English-Based Pidgins and Creoles [38] levelling of preterite/ppt verb forms: [39] levelling of preterite/ppt verb forms: participle p < .01 past replacing the participle replacing the past form [48] invariant don’t for all persons in the [44] multiple negation / negative concord p = .05 present tense *If a variety has feature 1, it also has feature 2, but not necessarily vice versa. Rightmost column provides results of one-tailed Fisher’s Exact tests on corresponding 2 × 2 tetrachoric tables.
< previous page
page_42
next page >
< previous page
page_43
next page >
Page 43
Newfoundland English has both [30] and [31], Colloquial American English has neither, and East Anglia has [30] but not [31]. All of these configurations are predicted by the one-way implication. What is not predicted is a variety that has [31] (would in if-clauses) but not [30] (loosening of the sequence of tense rule); indeed, such a variety is not attested in our database. A possible interpretation of this particular one-way implication seems to be that loosening of the sequence of tense rule triggers, or is a necessary precondition to, would in if- clauses; the phenomenon, in short, appears to be a specific manifestation of the loosening-ofthe-sequence-of-tense-rule phenomenon. In cross-linguistic typology, one-way implications are often interpreted in terms of historical evolution. This interpretation appears to be appropriate for the implications obtaining between ain’t as generic negator before a main verb and ain’t as the negated form of be/have . We had seen before that in New Englishes there is an equivalence among the three uses of ain’t . We can now refine this analysis by stating that if L1 vernaculars are added to the picture, we obtain a preference such that we will not find ain’t as generic negator before a main verb [47] unless we also see the two more restricted uses of ain’t [45/46]. This may well have a diachronic explanation such that a L1 vernacular that displays ain’t as a generic negator must have developed the more specific uses of ain’t first. This line of argument squares nicely with Anderwald (2003:48), who, in her book-length treatment of nonstandard negation in the British Isles, argues that generic ain’t (more specifically, ain’t for do) is an extension of ain’t for be and have . The second bundle of implications in Table 2.2a falls into the domain of relativization strategies. It turns out that any given variety will not have the relative particle as [63] unless that variety also has (i) the relative particle what [61] and (ii) gapping or zero relativization in subject position [66]. This implication, too, is not entirely unexpected given what we know about the typology of relativization strategies in dialects of English: zero relativization, for instance, is known to be a precondition to having as or what as relative markers (Herrmann 2003:91). This line of analyzing one-way implications also presents an opportunity to test a hypothesis to be found in the literature on vernacular universals predicting that if a variety exhibits the was–weren’t split [51], it will also have was/were generalization [59], but not vice versa (cf. Britain 2002:19; Chambers 2004:132). This should be the case because was/were generalization is the more basic phenomenon that may—but need not —develop into the was–weren’t split. Indeed, we do find evidence for this preference in our
< previous page
page_43
next page >
< previous page
page_44
next page >
Page 44 survey: although the one-way implication between [51] and [59] is not perfect and has an exception— CollAusE has [51] but not [59]—the resulting tetrachoric table is statistically significant at p = .050. Table 2.2b reports a number of one-way implications that hold true for English L1 vernaculars only. For instance, a L1 variety needs habitual do [23] in its inventory in order for do to be attested as a tense and aspect marker [27]. Once again, this is in line with what we know from the dialectological literature (e.g. Kortmann 2004a). In like manner, ain’t as the negated form of be [45] always implies multiple negation [44]. Be as perfect auxiliary [26] is a necessary (but not sufficient!) condition for habitual be [22], and lack of inversion in main clause yes/no questions [74] can trigger lack of inversion/lack of auxiliaries in whquestions [73], but not the other way around. This is an interesting preference because we had seen in Table 2.1c that this specific pairing is actually an equivalence in English-based pidgins and creoles. In English L1 vernaculars, by contrast, the pairing is a one-way street. Turning to some selected one-way implications in L2 varieties of English (Table 2.2c), we note, for one thing, that like as a quotative particle [76] implies like as a focusing device [75]. Furthermore, the possibility of regularized reflexives paradigms [4] appears to necessitate generic he/his for all genders [8]. Last, Table 2.2d presents a couple of one-way implications that are significant in English-based pidgins and creoles: for one thing, for a pidgin or creole to exhibit past forms of irregular verbs replacing participle forms [38], it must also exhibit participle forms replacing the past form [39]. Second, not unlike L1 varieties requiring multiple negation to have ain’t for be, English-based pidgins and creoles exhibit invariant don’t [48] only if they also exhibit multiple negation [44]. 5. PARAMETERS OF VARIATION IN WORLD ENGLISHES: A BIRD’S EYE PERSPECTIVE In this section, we adopt a bird’s eye perspective to explore what generalizations can be made about underlying dimensions of morphosyntactic variance in World Englishes, drawing on principal component analysis (cf. Kim and Mueller 1978) of our database. Principal component analysis is a statistical technique that reduces a number of independent variables to a smaller number of hypothetical constructs, or dimensions, known as ‘components’ or ‘factors’, which can be assigned meaningful interpretations by the analyst. What counts in principal component analysis, then, is the arrangement of data along—ideally— meaningful dimensions (or components). In linguistics, principal component analysis and similar techniques have been utilized, for instance, in studies of register variation (cf. Biber 1988, who uses factor analysis , a close relative to principal component analysis). Our aim here is to probe where principal component analysis can take us in terms of the Handbook’s survey: which bundles of morphosyntactic features account, in
< previous page
page_44
next page >
< previous page
page_45
next page >
Page 45 concert, for the most variance between varieties of English? Treating varieties as variables and features as cases, we used principal component analysis (with Varimax rotation and Kaiser normalization) to extract two components. The first component accounts for 23.2% of the observable variance in the database, whereas component 2 explains 15.2% of the variance. This means that component 1 is statistically more robust than component 2 and that the two components cumulatively capture c. 38% of the variance5 in our survey. Figure 2.4 visualizes the factor scores for both of our principal components in a two-dimensional coordinate plane. As for the linguistic interpretation of the two components, we suggest that component 1 displays increased levels of ‘morphosyntactic complexity’. Although we are fully aware that ‘complexity’ is a notoriously controversial and ill-defined notion, we tentatively define morphosyntactic complexity as follows: a given variety X is morphosyntactically more complex than a given variety X′ if variety X exhibits (i) fewer features that arguably simplify syntactic rules, (ii) a smaller amount of features that can be said to aid processing, and (iii) more features that are indicative of ‘distinctions beyond communicative necessity’ (notice that this latter part of our definition heavily draws on McWhorter 2001:125). Further, we believe component 2 to indicate a given variety’s degree of ‘analyticity’, a notion which we operationally define as bringing about a greater number of features that are rather autonomous—that is, invariable and periphrastic—in nature, much along the lines of Vincent’s concise definition: Construction C is relatively more analytic than another construction C ′ having approximately the same grammatical content as C to the extent that the constituent elements of C show greater morphosyntactic and phonological autonomy than do those in C ′. (Vincent 1997:99) What is the evidence for this specific interpretation of the axes in Figure 2.4? We offer that those features characteristic of varieties towards the right pole of the x-axis (i.e. component 1s) are, more often than not, indicative of increased levels of morphosyntactic complexity, and that many of those features that are characteristic of varieties scoring high on the y-axis (i.e. component 2) are rather analytic in nature. Thus, as one moves upwards on the vertical axes, one increasingly finds varieties attesting analytical and invariant features such as zero past-tense forms [40], invariant present-tense forms [53], and never as a preverbal past-tense negator. As one moves rightwards on the horizontal axis, one increasingly encounters varieties that attest morphosyntactically more complex features such as double comparatives and superlatives [19], inverted word order in indirect
< previous page
page_45
next page >
< previous page
page_46
next page >
Page 46
Figure 2.4 Visualization of principal components of variance in the 76 × 46 database. Squares represent L1 varieties, triangles represent L2 varieties, circles represent English-based pidgins and creoles, and dotted boxes indicate group memberships.
< previous page
page_46
next page >
< previous page
page_47
next page >
Page 47 questions [69], and multiple negation [44] (for a detailed discussion, qualitative and statistical, see Szmrecsanyi and Kortmann, to appear).What is important here is that visual inspection of Figure 2.4 reveals four more or less distinct groupings, which are indicated by dotted boxes in Figure 2.4: English-based pidgins (box 1), English-based creoles (box 2), L2 varieties of English (box 3), and L1 English vernaculars (box 4). This pattern is neat and faces two outliers only: Norfolk, classified as an L1 variety in our survey, is located in the pidgin/creole box; and Bahamian English (BahE), classified as a pidgin/creole variety, ends up in the L1 box (although only marginally so). Consider, now, that the classification of Norfolk and BahE is actually not as clear-cut as the survey’s classification (which was the product of lively discussions among the Handbook’s editors) might suggest. In his survey article on Norfolk, Mühlhäusler (2004:789) notes that this variety “shares the characteristic of many creoles, koinés and mixed languages of not having a great deal of inflectional and derivational morphology”, whereas Burridge (2004:1116) classifies Norfolk as a contact variety. Reaser and Torbert (2004:392–3), in contrast, devote a whole section of their survey article on BahE to the question as to why the linguistic status of the varieties that the label refers to is a problematic one. At any rate, it is well known that the settlement history of the Bahamas is rather unique. These two outliers notwithstanding, with regard to morphosyntactic complexity (component 1), the hierarchy in (1) emerges: (1) English-based pidgins < English-based creoles < L2 varieties of English < English L1 vernaculars Is this hierarchy meaningful from a theoretical perspective? We would like to argue that it is. As for (morpho-)syntactic complexity, we would like to draw attention to an ongoing debate in the literature about what (if anything) it is that sets apart creoles, as a synchronic class, from other languages. According to McWhorter, if all of the world’s languages could be ranked on a scale of complexity, there would be a delineable subset beginning at the ‘simplicity’ end and continuing towards the ‘complexity’ one all of which were creoles. (McWhorter 2001:162) McWhorter argues that this is because [c]reole languages are unique in having emerged under conditions which occasioned the especial circumstance of stripping away virtually all of a language’s complexity … such that the complexity emerging in
< previous page
page_47
next page >
< previous page
page_48
next page >
Page 48 a creole is arising essentially from ground zero, rather than alongside the results of tens of thousands of years of other accretions. As such, creoles tend strongly to encompass a lesser degree of complexity than any older grammar. (McWhorter 2001:155) Although these statements are not uncontroversial, especially not among the creolist community, our modest empirical inquiry can be seen to suggest that McWhorter’s view may not be entirely implausible (at least as far as varieties of English and English-based creoles are concerned). Conflating the English-based pidgins in our survey (box 1) with the English-based creoles (box 2), one finds that in Figure 2.4 all of the leftmost varieties are English-based pidgins and creoles, with Bislama, Tok Pisin, and so on constituting the most extreme cases. Moving rightwards, one then finds overlap areas between pidgins, creoles, and L2 varieties of English: Nigerian Pidgin English, for example, scores as high on the complexity (i.e. horizontal) dimension as Belizean Creole and Butler English (according to Hosali (2004:1032), it is hard to say whether ButlE should be considered a pidgin or an early fossilized interlanguage). Where boxes 2 and 3 overlap, we find creoles, such as Aboriginal English, that are as complex as some L2 varieties, such as Ghanaian English. Last, one finds a number of L2 varieties—Chicano English, Indian South African English, and so on—that are as complex as the average L1 vernacular, although there also appear to be L1 vernaculars (e.g. Orkney and Shetland English) that appear to be less complex morphosyntactically than some L2 varieties. What can be said about the other dimension in Figure 2.4? Analyticity (component 2) yields the following hierarchy: (2) English-based pidgins / English-based creoles / L2 varieties of English > English L1 vernaculars This is another way of saying that, in terms of the vertical axis (and, thus, in terms of analyticity), there is not much variance among English-based pidgins, English-based creoles, and L2 varieties. Crucially, however, these three groups score (the usual exceptions, e.g. BahE, apply) higher on the analyticity dimension than any L1 vernacular in the sample. It is well known in this context that creoles in general tend to be highly analytic if not isolating languages (cf. Hagège 1985; McWhorter 1998; Mufwene 1990; Schuchardt 1979). Similarly, the adult L2 acquisition process is known to be hostile toward inflectional morphology (cf. Klein and Perdue 1997), a fact which might ultimately be responsible for the analytic/isolating nature of creole languages (cf. DeGraff 1997). At any rate, given the literature, it is reasonable that among the varieties sampled in our database, L1 vernaculars should be the least analytic ones.
< previous page
page_48
next page >
< previous page
page_49
next page >
Page 49 6. SUMMARY AND CONCLUSIONS Although we are fully aware that the quantitative approach adopted in this study can never replace, but only complement, an in-depth qualitative dialectological inquiry, we hope that this study has succeeded in highlighting the many ways in which methods of typological analysis can be brought to bear on a database mapping language-internal variation. For one thing, we sought to demonstrate that an analytical distinction between different types of versals does more than just proliferate terminology—rather, these are distinctions that, we believe, are indispensable for achieving descriptive and explanatory accuracy. Recall here that in terms of unrestricted versals in particular, we have shown that not everything that is widespread in one particular world region or in one particular variety type is necessarily widespread in other varieties of English, and not everything that is recurrent in varieties of English must necessarily be universal or even pervasive cross-linguistically. In this spirit, we have argued that ‘vernacular universal’, for instance, is a rather strong term that should be used with caution. We also aimed to demonstrate that instructive generalizations can follow not only from considering absolute frequencies, but also from co-occurrence patterns. Carrying the distinction between angloversals and varioversals over to the study of implicational tendencies, we have suggested that different bundles of features, or biconditional implications, characterize different groups of varieties, and that the set of observable one-way implications appears to be conditioned on the variety type under analysis. Such differences may well reflect different pathways of historical development and grammaticalization, an issue which—we believe—is certainly worth paying attention to in future study. Finally, drawing on a principal component analysis of our survey, we argued that varieties of English can be considered to vary along two fundamental dimensions: morphosyntactic complexity and analyticity. These dimensions yield hierarchies which arrange varieties of English according to variety type, rather than according to other factors. Crucially, these hierarchies can, we believe, be considered angloversals, or ‘angloversal continua’, in their own right. Should similar patterns emerge in the study of vernacular varieties of other languages, we are likely to be dealing with cross-linguistically robust vernacular continua . APPENDIX: THE FEATURE CATALOGUE Note: For a version of the feature catalogue annotated with linguistic examples, see Kortmann and Szmrecsanyi (2004:1146–8). Pronouns, Pronoun Exchange, and Pronominal Gender 1. them instead of demonstrative those 2. me instead of possessive my
< previous page
page_49
next page >
< previous page
page_50
next page >
Page 50 3. special forms or phrases for the second person plural pronoun 4. regularized reflexives paradigm 5. object pronoun forms serving as base for reflexives 6. lack of number distinction in reflexives 7. she/her used for inanimate referents 8. generic he/his for all genders 9. myself/meself in a nonreflexive function 10. me instead of I in coordinate subjects 11. nonstandard use of us 12. noncoordinated subject pronoun forms in object function 13. noncoordinated object pronoun forms in subject function Noun Phrase 14. absence of plural marking after measure nouns 15. group plurals 16. group genitives 17. irregular use of articles 18. postnominal for-phrases to express possession 19. double comparatives and superlatives 20. regularized comparison strategies Verb Phrase: Tense and Aspect 21. wider range of uses of the progressive 22. habitual be 23. habitual do 24. nonstandard habitual markers other than do 25. levelling of difference between present perfect and simple past 26. be as perfect auxiliary 27. do as a tense and aspect marker 28. completive/perfect done 29. past tense/anterior marker been 30. loosening of sequence of tense rule 31. would in if-clauses 32. was sat/stood with progressive meaning 33. after -perfect Verb Phrase: Modal Verbs 34. double modals 35. epistemic mustn’t Verb Phrase: Verb Morphology 36. levelling of preterite and past participle verb forms: regularization of irregular verb paradigms
< previous page
page_50
next page >
< previous page
page_51
next page >
Page 51 37. levelling of preterite and past participle verb forms: unmarked forms 38. levelling of preterite and past participle verb forms: past form replacing the participle 39. levelling of preterite and past participle verb forms: participle replacing the past form 40. zero past-tense forms of regular verbs 41. a- prefixing on ing-forms Adverbs 42. adverbs (other than degree modifiers) have same form as adjectives 43. degree modifier adverbs lack -ly Negation 44. multiple negation/negative concord 45. ain’t as the negated form of be 46. ain’t as the negated form of have 47. ain’t as generic negator before a main verb 48. invariant don’t for all persons in the present tense 49. never as preverbal past-tense negator 50. no as preverbal negator 51. was–weren’t split 52. invariant nonconcord tags Agreement 53. invariant present-tense forms due to zero marking for the third person singular 54. invariant present-tense forms due to generalization of third person -s to all persons 55. existential/presentational there’s , there is, and there was with plural subjects 56. variant forms of dummy subjects in existential clauses 57. deletion of be 58. deletion of auxiliary have 59. was/were generalization 60. Northern Subject Rule Relativization 61. relative particle what 62. relative particle that or what in nonrestrictive contexts 63. relative particle as 64. relative particle at 65. use of analytic that his/that’s, what his/what’s, at’s , as’ instead of whose
< previous page
page_51
next page >
< previous page
page_52
next page >
Page 52 66. gapping or zero relativization in subject position 67. resumptive/shadow pronouns Complementation 68. say -based complementizers 69. inverted word order in indirect questions 70. unsplit for to in infinitival purpose clauses 71. as what/than what in comparative clauses 72. serial verbs Discourse Organization and Word Order 73. lack of inversion/lack of auxiliaries in wh-questions 74. lack of inversion in main clause yes/no questions 75. like as a focussing device 76. like as a quotative particle NOTES 1. Some readers might contend that copula absence before NPs is quite rare even in varieties of English (cf. Labov 1969:731 on constraints on copula deletion in AAVE). In this connection, it should be noted that in varieties of English such as AAVE, copula deletion before NPs may be rare but possible. 2. If it actually has such a tie, ‘vernacular typoversal’ might be a better term for a candidate feature. 3. Having said that, we shall see in section 4.2 that in L1 vernaculars, a one-way implication holds between habitual do and do as a tense and aspect marker. 4. The specific statistical measure used is Fisher’s Exact test (cf. Cysouw 2003:91–2). 5. Statistically speaking, this value is fairly good—by way of comparison, the first two factors in Biber (1988:83) account for 26.8% and 8.1%, respectively, of the shared variance. REFERENCES Anderwald, L. 2003. Negation in Non-Standard British English: Gaps, Regularizations and Asymmetries (Routledge Studies in Germanic Linguistics). London/ New York: Routledge. Barbiers, S., L. Cornips, and S. van der Kleij, eds. 2002. Syntactic Microvariation . Amsterdam: SAND. Biber, D. 1988. Variation Across Speech and Writing . Cambridge: Cambridge University Press. Britain, D. 2002. Diffusion, leveling, simplification and reallocation in the past tense of BE in the British Fens. In Investigating Change and Variation Through Dialect Contact , Special Issue of Journal of Sociolinguistics 6, edited by L. Milroy, 16–43. Burridge, K. 2004. Synopsis: Morphological and syntactic variation in the Pacific and Australasia. In Kortmann et al., eds. 2004. Vol 1. 116–1131.
< previous page
page_52
next page >
< previous page
page_53
next page >
Page 53 Chambers, J. K. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Implications . Oxford Malden: Blackwell. ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology, edited by B. Kortmann, 127–45. Berlin and New York: Mouton de Gruyter. Cysouw, M. 2003. Against implicational universals. Linguistic Typology 7: 89–101. DeGraff, M. 1997. Verb syntax in, and beyond, creolization. In The New Comparative Syntax , edited by L. Haegeman, 64–94. London: Longman. Greenberg, J. H. 1963. The Languages of Africa. Bloomington: Indiana University. ———. 1966. Language Universals, With Special Reference to Feature Hierarchies. The Hague: Mouton. Hagège, C. 1985. L’homme de paroles. Paris: Fayard. Haspelmath, M., M. Dryer, D. Gil, and B. Comrie, eds. 2005. The World Atlas of Language Structures. Oxford: Oxford University Press. Herrmann, T. 2003. Relative clauses in dialects of English: a typological approach. PhD thesis, University of Freiburg. Available fromhttp://www.freidok.uni-freiburg.de/volltexte/830/ (accessed 14 November 2006). Hosali, P. 2004. Butler English: Morphology and syntax. In Kortmann et al., eds. 2004. Vol. 2. 1031–44. Kim, J. O., and C. W. Mueller. 1978. Introduction to Factor Analysis: What It Is and How To Do It (Quantitative Applications in the Social Sciences). Newbury Park/ London/New Delhi: Sage Publications. Klein, W., and C. Perdue. 1997. The basic variety (or: Couldn’t natural languages be much simpler?). Second Language Research 13: 301–47. Kortmann, B. 2004a. Do as a tense and aspect marker in varieties of English. In Kortmann, ed. 2004b. 245– 75. ———, ed. 2004b. Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective . Berlin and New York: Mouton de Gruyter. ———, E. Schneider, K. Burridge, R. Mesthrie, and C. Upton, eds. 2004. A Handbook of Varieties of English, 2 vols. Berlin and New York: Mouton de Gruyter. ———and B. Szmrecsanyi. 2004. Global synopsis: Morphological and syntactic variation in English. In Kortmann et al., eds. 2004. Vol. 2. 1142–202. Labov, W. 1969. Contraction, deletion and inherent variability of the English copula. Language 45: 715–62. McWhorter, J. 1998. Identifying the creole prototype: Vindicating a typological class. Language 74: 788–818. ———. 2001. The world’s simplest grammars are creole grammars. Linguistic Typology 6: 125–66. Mühlhäusler, P. 2004. Norfolk Island-Pitcairn English (Pitkern Norfolk): Morphology and syntax. In Kortmann et al., eds. 2004. Vol. 2. 789–801. Mufwene, S. 1990. Transfer and the substrate hypothesis in creolistics. Studies in Second Language Acquisition 12: 1–23. Reaser, J., and B. Torbert. 2004. Bahamian English: Morphology and syntax. In Kortmann et al., eds. 2004. Vol. 2. 391–406. Schuchardt, H. 1979. The Ethnography of Variation: Selected Writings on Pidgins and Creoles. Ann Arbor, MI: Karoma. Stassen, L. 2005. Zero copula for predicate nominals. In Haspelmath et al., eds. 2005. 486–9. Szmrecsanyi, B., and B. Kortmann. To appear. The morphosyntax of varieties of English worldwide: A quantitative perspective. Lingua. Vincent, N. 1997. Synthetic and analytic structures. In The Dialects of Italy, edited by M. Maiden, 99–105. London: Routledge.
< previous page
page_53
next page >
< previous page
page_54
next page >
page_54
next page >
Page 54 This page intentionally left blank.
< previous page
< previous page
page_55
next page >
Page 55 II Consonant Cluster Reduction and Default Singulars Prototypical Vernacular Universals?
< previous page
page_55
next page >
< previous page
page_56
next page >
page_56
next page >
Page 56 This page intentionally left blank.
< previous page
< previous page
page_57
next page >
Page 57 3 How Diagnostic Are English Universals? Daniel Schreier Another universal that is becoming established [in sociolinguistic theory] after more than thirty years of empirical research is that certain variables appear to be primitives of vernacular dialects in the sense that they recur ubiquitously all over the world. (Chambers 1995:242) There is nothing so important as trifles. Never trust to general impressions, but concentrate yourself upon details. (Sherlock Holmes; in Doyle 1890) 1. INTRODUCTION In his groundbreaking book Sociolinguistic Theory , Jack Chambers (1995) identified and discussed a number of features shared by all varieties of English around the world: 1) (ng ), an alveolar nasal /n/ instead of a velar one in - ing suffixes (e.g. ‘singing’); 2) consonant cluster reduction (I met him at the pos’ office; she lef’ her umbrella on the bus this morning); 3) what he labelled “default singulars”, i.e. past be levelling with pivot form was (in we was good friends then, the cows was out all night, etc.); 4) regularisation of pasttense forms of irregular verbs (he come around last night); and 5) multiple negation (‘I didn’t do no queuejumping’). Chambers argued that English universals of this kind were of immense interest to sociolinguists simply on account of the fact that they made an appearance in sociohistorically unrelated varieties (e.g. Falkland Islands, Newfoundland, and Hong Kong English). They are thus to some extent independent of the varieties’ sociohistorical backgrounds and of contact-induced change mechanisms in various (post-)colonial settings (i.e. pidginisation and/or creolisation, koinéisation, nativisation, etc.; cf. Schneider 2003). The manifestation of English universals does carry much research potential for a number of linguistic disciplines indeed. To formulate but some
< previous page
page_57
next page >
< previous page
page_58
next page >
Page 58 of the most imminent questions: what reasons can we offer for the presence of ‘primitives of vernacular dialects’ throughout the English-speaking world? Are they genetically based (perhaps after all explained by the diffusion of English out of the British Isles), is there a sociocognitive dimension (e.g. a preference for more regular features or an avoidance of socially or regionally marked ones), or does the answer lie in language universals? Second, sociolinguistic theory holds that no linguistic feature is frozen in time and space. Consequently, English universals must be subject to language variation and change as well; the question is when and under what conditions their use increases or decreases and what intralinguistic and language-external parameters condition their variability. Given what we know about ‘non-universal’ features, it must be possible to identify a set of parameters that govern their variation and to then compare these across varieties. Such an approach seems promising because it allows studying English universals from a broad comparative perspective, tracing parallels (and possibly some diagnostic differences as well). Some of these questions have not been dealt with prominently so far, and this chapter is a first step towards gaining insights into how English universals are subject to variation and change and what this can tell us for sociolinguistic theory in general. It addresses the following questions: — What can sociolinguistic theory benefit from an in-depth analysis of the manifestation of English universals in varieties around the world? What are the qualitative and quantitative differences, and what conclusions can we offer to account for the spread and continuing diversification of English as a world language? — To what extent is the variable usage of these features governed by language-internal and languageexternal factors? — To what extent are English universals contact-sensitive (so that there are different patterns in varieties derived from language and/or dialect contact)? If they are, how do variety-specific differences encourage us to rethink the validity of the concept ‘vernacular universals’ as a whole? The present chapter assesses and evaluates data from English around the world (spoken in North America, the British Isles, the South Atlantic, the Caribbean, India, and New Zealand). These varieties differ for a number of reasons: time depth (colonial involvement and settlement history), contact history (in contexts of transplanted British English varieties and subsequent interaction between those, leading to koinéisation vs. large-scale language contact, involving English and other European or indigenous languages, giving rise to pidgin and creoles varieties), and the general function of English (e.g. as an institutionalised second language [ESL] or a foreign language [EFL]). The cross-varietal analysis concentrates on two of the features singled out by Chambers (1995), namely: 1) consonant cluster reduction, and
< previous page
page_58
next page >
< previous page
page_59
next page >
Page 59 2) the levelling of the past be paradigm with pivot form was (“default singulars” in Chambers’ terminology). The two variables are discussed and evaluated in terms of their importance for sociolinguistic research and then compared in several varieties of English around the world. A final section assesses the relevance of the findings and discusses their implications for the study of universal features. 2. UNIVERSAL 1: CONSONANT CLUSTER REDUCTION English belongs to the minority of the world’s languages whose phonotactic systems admit groups of consonants (C) in a single syllable structure (cf. Blevins 1995). Consequently, clusters are typologically marked (in the sense of Lass 1984).This means, among other things, that they are less frequent than other structures, that they tend to be historically unstable and are vulnerable to change in contact conditions, particularly when competing with more frequent and regular types (i.e. CV structures, found in all of the worlds language, or CVC, found in most), or that they appear late during language acquisition (discussion in Lass 1984:132). Crucially for our purposes, consonant clusters (henceforth CCs) are subject to variation and change even in phonotactic systems that admit them, and the process referred to as Consonant Cluster Reduction (CCR) is one of the most extensively researched variables in variationist sociolinguistics (Schreier 2005). The consensus is that the application of a phonological rule to delete cluster-final consonantal segments (so that the total number of cluster segments is reduced: CCC#>CC#, CC#>C#) is a true universal of spoken English, a process found in all speakers of English, irrespective of bi- and/or multilingualism, regional and social origins, or language competence. There is an extensive literature: CCR has been investigated in English varieties all over the world, most prominently in U.S. American English (e.g. Tejano English, Bayley 1995; Appalachian English, Wolfram and Christian 1976; White Philadelphia English, Guy 1980; African American English, Labov 1972; Wolfram and Fasold 1974; Lumbee Native American English, Torbert 2001), Indian English (Khan 1991), in the Caribbean (mesolectal Jamaican Creole English, Patrick 1991; Bahamian English, Childs 2000), in Southern Hemisphere English (New Zealand English, Holmes and Bell 1994; Schreier 2003b), as well as in immigrant groups with non-English-language backgrounds immersed in an Englishspeaking community (Vietnamese English, Wolfram, Christian, and Hatfield 1986). All these studies attest to the universal character of this phonological rule and provide us with a wealth of insights into the nature (and possible differences) of its application. Wolfram and Fasold (1974:130 ff.), in an early and highly detailed analysis of CCR, suggest an extensive list of potential candidates that may undergo reduction and emphasise that its complexity derives from the interplay of various phonetic, morphosyntactic, and sociopsychological criteria, all of
< previous page
page_59
next page >
< previous page
page_60
next page >
Page 60 which contribute to the social and linguistic conditioning of this variable rule. (Table 3.1) Although some issues remain controversial, the findings postulated in the pioneering studies of Labov, Cohen, Robins, and Lewis (1968), Wolfram (1969), and Fasold (1972) have been confirmed with remarkable consistency. CCR is subject to two major types of constraints. First, a functional one, in that monomorphemic clusters (such as past , desk, or find ) are more prone to undergo reduction than bimorphemic clusters (such as passed, stopped , or knocked), where the cluster-final plosive represents an independent morpheme.1 Clusters are thus much more likely to remain intact when one of the segments carries morphological meaning or, to put it differently, when a CC includes more than one morpheme. The second factor to govern CCR is the immediately following phonetic segment. Due to the operation of another phonological rule, progressive assimilation, CCs followed by another consonant are more readily reduced than CCs followed by a vowel: “The less natural canonical sequence of a preconsonantal phonetic environment (e.g. west side or find time) favors CCR over the more natural sequence preceding a vowel (e.g. west end or find out)” (Wolfram, Childs, Table 3.1 Phonetic, Morphosyntactic and Sociopsychological Effects on English CCR Phonetic Preceding Environment nasal > lateral > sibilant > plosive ( ‘wind’ > ‘wild’ > [wεs] ‘west’ > [æk] ‘act’) Following environment plosive > sonorant > pause > vowel ( > > [bεs] > [bεs æt] ‘best kid’ ‘best name’ ‘best’ ‘best at’ Stress [- stress] > [+ stress] ( ‘contract n.’ > ‘contract v.’ Morphosyntactic Status of cluster (segments) monomorphemic > redundant bimorphemic > bimorphemic ( ‘guest’ > [slεp] ‘slept’ > ‘guessed’) Sociopsychological lower social class > higher social class (language-external) casual style > formal style language-contact derived > dialect-contact derived Adapted from Guy 1991; Wolfram and Thomas 2002:134.
< previous page
page_60
next page >
< previous page
page_61
next page >
Page 61
Figure 3.1. CCR in Hyde County NC English (adapted from Wolfram and Thomas 2002:142). and Torbert 2000:18). Therefore, both the morphemic status of cluster segments and phonetic environment affect the variable nature of CCR in English. Figure 3.1 and Table 3.2 illustrate the intricate patterning of the two constraints with data from White NC American English, adapted from Wolfram and Thomas (2002:142). Such is the regularity of CCR in English around the world that it has been labelled the “showcase variable for variationist sociolinguistics” (Patrick Table 3.2 CCR Rates in Hyde County NC English Pre-C Pre-P Pre-V Mono (n=210) Bi (n=117) Mono (n=70) Bi (n=48) Mono (n=143) Bi (n=169) reduced 123 48 23 3 14 7 realised 87 69 47 45 129 162 Total % CCR 58.6 41.0 32.9 6.3 9.8 4.1 Adapted from: Wolfram and Thomas 2002:142.
< previous page
page_61
next page >
< previous page
page_62
next page >
Page 62 1991:171) or “the paradigm case of systematic variability in social dialectology” (Wolfram et al. 2000:17). Consequently, then, CCR certainly qualifies as a ‘vernacular root’, and both the manifestation and the systematic conditioning of this variable application has been demonstrated amply in English around the world. CCR operates in all varieties of English; as such, it offers little qualitative information on dialect distinctiveness and has no diagnostic value whatsoever. If all varieties have CCR, then it simply cannot be listed as a feature that distinguishes them. The point adopted here, however, is that it is the manifestation, not the presence, of CCR that is important. The focus must consequently be on possible differences between varieties so that CCR, although omnipresent, operates in different ways. Assuming we can pinpoint quantitative and variationist differences, we must ask what the governing factors and internal and external correlates would be and what these can tell us about variation in English universals. In the following, I present data from English varieties around the world to demonstrate differences and, by doing so, show how even universals may carry diagnostic value. I focus on two variationist aspects of CCR: quantitative (How often are clusters reduced?) and qualitative (Does the established British and American pattern always hold? Do all varieties reduce the same clusters?). Then I separate them into three areas of possible differentiation: quantity (2.1), contact sensitivity (2.2), and generality of rule application (2.3). In anticipation of criticism, I am fully aware that this simplifies the complex nature of CCR because these are at times closely related and they overlap considerably. I discuss possible implications in situ . 2.1 Quantity First of all, we need to ask how often CCR applies. Is its application quantitatively stable across varieties or are there differences in how often it occurs? This question cannot be easily explained for methodological reasons: there is no consistent practice as to what clusters should be extracted for analysis. Whereas it was common practice in early works to count all CCs, regardless of language-internal considerations such as syllable weight and word structure, frequency, stress, or total cluster length, later research has shown that CCR is much more context sensitive than previously thought. Neu (1980:53) found that clusters in highfrequency items (such as and or the adverb just) were much more likely to undergo reduction (particularly because they are unstressed) than those in low-frequency ones like sand or mind. By the same token, Santa Ana’s (1992, 1996) research on Los Angeles Chicano/a English showed that cluster length was a determining factor and that CCs consisting of three segments were more likely reduced than clusters with two Cs. Consequently, a comparative analysis of global reduction values is certainly insightful, but we have to bear in mind that the data analysed influence the results considerably (most obviously when large amounts of
< previous page
page_62
next page >
< previous page
page_63
next page >
Page 63 Table 3.3 Some Global Reduction Values of English(es) Around the World Variety Total % CCR Source White York (UK) English 24Tagliamonte and Temple (2005:12) White Philadelphia English 33Guy (1991) Pakeha (White) New Zealand English 27.8Schreier (2005:200) White American English (Hyde County NC) 28.8Wolfram and Thomas (2002:136) African American English 40–65Poplack and Tagliamonte (2001) Hispanic varieties ca. 50Bayley (1995), Santa Ana (1996) English-based Caribbean creoles 50–97Poplack and Tagliamonte 2001 Mesolectal Jamaican English 75Patrick (1999:152) St Helenian English 87Schreier (2008) high-frequency words are included), and we are well advised to give our findings some leeway for methodological considerations. A first look at English varieties around the world (summarised in Table 3.3) reveals that quantitative differences are on a large scale. All varieties have CCR, and the minimal amount appears to be around 25 percent (e.g. in White British and New Zealand English), which is to say that at least one in four clusters is reduced. On the other side of the spectrum, some varieties have generalised CCR to the extent that it is almost a categorical process. A look at the literature reveals that English-based Caribbean Creoles, for instance, have almost 97 percent reduction rates. Word-final clusters are thus practically absent from these phonotactic systems. Consequently, frequency of rule application is an important criterion here. CCR is omnipresent, in all speakers and all varieties, but there is considerable variation as to how often clusters are reduced. These quantitative differences may thus represent a valuable tool to distinguish between varieties and do certainly carry diagnostic value. As we see, this has all kinds of consequences on variation and change in English CCR around the world. 2.2 Contact Sensitivity The next step is to look into the sociohistorical, sociolinguistic, and linguistic characteristics of these varieties. How can we account for the differences? For one, CCR appears to be a contact-dependent variable; its application
< previous page
page_63
next page >
< previous page
page_64
next page >
Page 64 is lowest in varieties of White British, American, and New Zealand English (roughly ranging between 25 and 33 percent, perhaps to some extent depending on the methodologies adopted). The higher the CCR rates, the more likely it is that the variety in question has undergone extensive contact with languages other than English or, alternatively, that it developed in a context of sustained multilingualism (one could mention Hispanic or African- American English here). Put differently, the heavier the impact of contact-induced restructuring or the more rudimentary the L2 competence in early phases of language shift, the higher the amount of CCR and the more canonical are CVC syllable structures (e.g. in English-based Creoles). CCR thus increases considerably during language-contact conditions in all its facets (bilingualism, pidginisation, creolisation, ESL, EFL, etc.) and is a reliable indicator of contact histories and contact-induced language change. It is equally clear that the phonotactic systems in contact play a major role. At the risk of simplifying complex processes, one should state that contact as such has no enhancing effect on CCR. A quantitative analysis of CCR in White NZE, for instance, a variety that originated in dialect contact and mixing in the second half of the nineteenth century (Gordon et al. 2004), yields almost exactly the same reduction (ca. 28 percent) and precisely the same internal constraints on variability as varieties of White British or American English (see Table 3.4 and Figure 3.2). The vernacular root CCR is thus not affected by dialect contact and new-dialect formation, which can only mean that contact between structurally and phonotactically similar systems does not lead to an increase of this vernacular root (in contrary to large-scale language contact between English and languages with other syllable structures, most prominently CV or CVC, which, as argued in Schreier [2005], entails a direct transfer of CV(C) syllable structures onto English targets). There is yet another contact-induced effect that carries diagnostic value. In varieties of English shaped in scenarios of extensive language contact, most prominently in creoles, the morphemic constraint may undergo complete reversal, so that - ed marked forms of weak verbs are effectively more Table 3.4 CCR in Nineteenth Century White New Zealand English Pre-C Pre–P Pre–V Mono (n=348) Bi (n=181) Mono (n=101) Bi (n=54) Mono (n=335) Bi (n=235) reduced 225 59 22 5 31 7 realised 123 122 79 49 304 228 % CCR 64.7 32.6 21.8 9.3 9.3 3.0 From Schreier 2005:151.
< previous page
page_64
next page >
< previous page
page_65
next page >
Page 65
Figure 3.2 CCR in New Zealand English (from Schreier 2005:152). likely to undergo CCR than monomorphemic clusters. This is witnessed in varieties such as St Helenian English and mesolectal Jamaican Creole English, where internal constraints operate differently so that bimorphemic clusters have in fact higher reduction rates than monomorphemic ones (see Figure 3.3). This is most plausibly explained by the interplay of phonotactic transfer from the substrate and reduction of inflectional morphology in the lexifier (cf. Patrick 1991). This unusual reversal is a consequence of the lack of morphological tense marking in these varieties and its compensation with a system of preverbal TMA markers (such as done, did, been , yuustu , etc.). In other words, it is explained by coinciding and mutually reinforcing effects of grammatical restructuring and phonological transfer. It is certainly interesting to note that there are parallels between two varieties that have no sociohistorical connections or similarities whatsoever, other than that they developed in contexts of extensive language contact, of course. One could make a case in point here that this is an exclusive process of creolisation, and thus highly diagnostic. This appears to be the case, but I would hesitate to answer in the affirmative simply because we need quantitative evidence from more varieties (particularly West African and Pacific varieties) to show that this is a general by-product of creolisation indeed. For the time being, we maintain that whereas one of the two well-established constraints holds here (following segment, although it is
< previous page
page_65
next page >
< previous page
page_66
next page >
Page 66
Figure 3.3 CCR in St Helenian English (from Schreier 2008). not very strong), the status of segment clusters, however, shows a reverse—and highly diagnostic—effect. 2.3 Generality of Rule Application The next point to consider in this context concerns the phonetic properties of the clusters that undergo (variable) reduction. The literature on CCR shows that one of the most important phonetic constraints on ‘reducibility’ is whether the individual cluster segments are homo-voiced or not. In AmE and BrE, clusters can only be reduced if the two adjoining Cs are either both voiced or both nonvoiced. Clusters with heterovoicing (/-lt/, /-mp/, / k/, etc.) can under no circumstances be reduced, not even in varieties that originated in extensive language contact. Some have gone as far as to claim that this is a universal constraint. Wolfram and Christian (1976), for instance, claim that Clusters such as mp (jump, ramp), lt (colt, belt), k (crank, rank) and lp (gulp, help) are not affected by this [CCR] process…. While linguists disagree as to the reason for the failure of some clusters to be affected by this process of simplification, they are in basic agreement as to those clusters that can and cannot undergo simplification. (Wolfram and Christian 1976:34)
< previous page
page_66
next page >
< previous page
page_67
next page >
Page 67 Whereas Wolfram and Christian’s claim is undoubtedly correct for well-established native-speaker varieties with no recent histories of language contact, such as NZE, it does not apply to some ESL varieties, such as Indian English (IndE). Khan (1991) noted that IndE has a particularly prominent tendency towards CCR; moreover, she found that speakers reduce a wider variety of clusters. Coda-final plosives in hetero-voiced clusters, such /-lt/, /-lp/, or /-lk/, were frequently reduced by most of the subjects studied. This is strong evidence that IndE is not subject to the voicing constraint, or at least not in the same way as varieties of AmE or BrE are. Khan’s findings suggest that ESL (and other) varieties may have different constraints on CCR than ‘inner-circle’ varieties (Kachru 1985). This is not an isolated phenomenon: similar observations have been made in Pacific creoles (Bislama), Singapore English (Lim 2004:33), and I can also support this with data from my own research on StHE (Schreier 2008). How can we account for the fact that the voicing constraint is categorical in BrE and AmE and absent (or at least weakened) in IndE or StHE? The overall amount of CCR increases in contact conditions, which invites the implication that contact-induced language change has an additional effect on the total amount of ‘reducible’ clusters. Speakers of IndE apply the reduction rule to a larger number of CCs and, hence, also to more CC segments with different properties. This is indicative of a process well studied in second-language learning: rule generalisation. A constraint present in the target (superstrate) variety is adopted, yet generalised and applied to a larger number of potential candidates. As a result, the voicing constraint is not operative, or it is at least less rigid, in ESL or EFL varieties and in English-derived pidgins and creoles (such as Melanesian Pidgin English). Native-speaker constraints are weakened due to substratal influence and phonotactic transfer of non-English structures, and this leads to yet another qualitative difference. Finally, it should also be mentioned in this context that non-native ESL or EFL varieties use an arsenal of competing techniques to cope with marked linguistic structures (such as CCs). For instance, they display tendencies not found in varieties with well-established speaker groups—for instance, the systematic restructuring to CV by inserting a (usually short checked) V in inter-C position (VCC# > VCVC#). A strong trend towards epenthesis of /i/ or /ə/ has been noted in Korean EFL (Lee 2000), and epenthesis yields insights into alternative strategies of breaking up word-final CCs; although less documented, this process may be phonologically conditioned. The question is why some varieties would favour reduction and others epenthesis, but this discussion has to be continued elsewhere. 2.4 Some First Conclusions Importantly, there exist individual and dialect-specific differences as to the frequency and conditioning of CCR in English varieties around the world. Dialects such as African-American English (Bailey and Thomas 1998; Labov
< previous page
page_67
next page >
< previous page
page_68
next page >
Page 68 1972; Labov et al. 1968; Wolfram 1969) or Hispanic varieties of English (Bayley 1995; Santa Ana 1996) have more CCR than varieties like New Zealand English (Holmes and Bell 1994) or Appalachian English (Wolfram and Christian 1976). CCR variation is thus quantitatively significant. Not only do dialects differ in their overall percentages, CCR is also operative in distinct phonetic contexts. Of particular relevance here is reduction in prevocalic environments (stan’ up , pos’ office), which arguably is particularly diagnostic of language contact and adaptation processes. Varieties differ extensively in their tendencies to reduce syllable-coda CCs in prevocalic or prepausal environments. For instance, prevocalic CCR is much more prominent in non-native varieties of English (e.g. Indian English, Khan 1991, or Pueblo Native American English; Wolfram 1980), dialects that may have adopted substratum features as a result of language shift (e.g. Chicano/a English; Bayley 1995; Santa Ana 1996), or varieties of English that have undergone creolisation (e.g. mesolectal Jamaican Creole English; Patrick 1991). Accordingly, this is a first indication that the extent and operation of CCR (not its existence) is highly diagnostic (e.g. indicative of language contact and phonological transfer of non-English features). 3. UNIVERSAL 2: ‘DEFAULT SINGULARS’, OR PAST BE LEVELLING WITH PIVOT FORM WAS The third universal feature of English listed by Chambers (1995) concerns the past-tense paradigm of be— namely, the usage of was in the context of were . Crucially, with five allomorphs, be stands out for its irregularity: it is the only verb that has maintained some former morphological complexity and personnumber concord in the past tense. As Wolfram, Hazen, and Schilling-Estes (1999:75) put it, “the irregular status of be is without parallel in the current configuration of subject-verb concord.” This irregularity is explained by the historical development of the verb, which can be traced to no less than three separate Indo-European verbs (Brunner 1963; Curme 1931; Luick 1964). Past be levelling makes an appearance in English around the world. Because these regularisation processes operate in many (if not all) vernacular dialects of English (cf. Chambers 1995), be is a prime candidate for analogical change, and paradigm regularisation involving was and were has a long-standing continuity and is historically well attested. Levelling to was in particular has historical continuity, and Quirk and Wrenn (1960) provide evidence that analogical language change operated in Old English already. This usage continues in Middle English, and Visser (1963–73) and Curme (1931) show regional tendencies for was instead of were from the fourteenth century on, particularly with plural persons. For example, Visser (1970, 3:72) cites the following example from Richard Coeur de Lion, written around 1300:
< previous page
page_68
next page >
< previous page
page_69
next page >
Page 69 (1) Thrytty knyghtes, withouten lye, forsothe was in that companye The historical literature suggests that the second person singular was subject to considerable regional variation. Forsström (1948), for instance, finds a sharp division between the south of England, where were was predominantly used for the second person singular, and northern varieties, where was was favoured. He also points out that there were considerable changes inasmuch as “preterit indicative was (or wes ) [was] frequently introduced into the plural in the North” (Forsström 1948:22). It is more important to trace this feature in contemporary varieties of English. An ample literature scrutinises the manifestation of past be levelling in varieties of English around the world, for instance, in the British Isles (Anderwald 2001; Britain 2002; Cheshire 1982; Tagliamonte 1998), the United States (Feagin 1979; Labov 1972; Wolfram and Christian 1976), Australia (Eisikovits 1991), Canada (Meechan and Foley 1994), the Caribbean (Tagliamonte and Smith 2000), and in the South Atlantic (Schreier 2002, 2003a). The breadth of data allows us to compare the universal in Englishes around the world and to pinpoint possible differences. The following discussion centres around three points: quantitative and variationist aspects of past be (3.1), competition with alternative paradigms (3.2), and potential for ongoing change (3.3). These are discussed in turn before they are evaluated and contextualised in section 3.4. 3.1 Quantity and Conditioning Two questions are of relevance here: namely, 1) how often levelling to was occurs quantitatively, and 2) in what linguistic environments it occurs predominantly (i.e. to what extent it is governed by language-internal factors such as grammatical person, singular vs. plural, etc.). Figure 3.4 presents data from a variety of Englishes around the world: Samaná English (Tagliamonte and Smith 2000), Sydney English (Eisikovits 1991), English Fens English (Britain 2002), Anniston/Alabama English (Feagin 1979), West Virginian Appalachian English (Christian et al. 1988), Buckie Scots (Smith and Tagliamonte 1998), and York English (Tagliamonte 1998). The figure presents global percentages of was in the context of were for all these varieties in all environments (second singular; first, second, and third person plural; plural NPs and existentials). Of course, it is always problematic to compare these results strictu sensu; for Samaná and Buckie, for instance, data are only available for the oldest members of the respective communities. Whereas we have a representative sample of social classes from Anniston/Alabama, there are only data for workingclass speakers from the West Virginian Appalachians. As for internal constraints, there are no data on existential plurals in Samaná English, and the subject pronouns are collapsed into a single category in the Appalachian English study, making a more fine-grained analysis impossible. We thus have to take care here as well, and the methodological
< previous page
page_69
next page >
< previous page
page_70
next page >
Page 70
Figure 3.4. A cross-dialectal comparison of past be levelling to was (adapted from Schreier 2002:85). and analytical inconsistencies warn against jumping to conclusions. Nevertheless, the picture that emerges allows some first speculations. Figure 3.4 shows that varieties differ drastically in how often regularisation to was occurs. It is found in all varieties of English, although in distinct environments and with varying frequencies. It seems to be least common in Yorkshire and Australian English (with total percentages of 17.0 and 12.9, respectively) and more common in the other varieties, most strikingly in TdCE (93.8). Thus, these findings in a sense mirror what was reported for the variety-specific manifestation of CCR above. As for the manifestation of languageinternal parameters, a number of studies have looked into language-internal constraints (Britain 2002; Hazen 1996; Montgomery 1989; Tagliamonte 1998; Wolfram et al. 1999), offering evidence that the type of subject is a major factor. Levelling with the pivot form was , for instance, may occur with second singular or plural persons, as well as with plural NPs and plural existential constructions, but some persons co-occur with levelled forms more frequently than others (Christian, Wolfram, and Dube 1988; Montgomery 1989; Smith and Tagliamonte 1998). Montgomery (1994), for instance, argues that singular versus plural marking of present-tense verbs and past be in Middle and early Modern Scots was subject to a type of subject constraint (first formulated by Murray 1873). The ‘subject-
< previous page
page_70
next page >
< previous page
page_71
next page >
Page 71 type rule’ holds that the co-occurring personal pronouns or NPs condition variation and the constraint hierarchy manifests itself clearly in the cross-dialectal comparison. Therefore, we must distinguish thirdperson subjects according to whether they are personal pronouns (such as ‘ they was down fishing’), lexical nouns (‘the kids was away for the weekend’), or the subject of an existential (‘there was about 600 cows left’). Not all of these environments favour leveling equally. Eisikovits (1991) and Feagin (1979) find that third-person plural pronouns are less prone to leveling than other persons. Feagin (1979:203) reports that, compared with other environments, they is the least frequent personal pronoun to co-occur with regularised was (particularly in her urban and rural working-class informants). However, existential plural constructions have “by far the strongest effect on the use of was in contexts of standard were ” (Tagliamonte and Smith 2000), a finding which is reproduced in all varieties that have been subject to linguistic scrutiny (Christian et al. 1988; Eisikovits 1991; Schilling-Estes and Wolfram 1994). This allows speculations that regularisation from were to was is most common with existential plural constructions and plural NP subjects and less frequent with personal pronouns (see e.g. Montgomery 1989, 1994). This emerges clearly when looking at Figure 3.4 again. All the varieties show a preference for levelled was forms with plural existentials (i.e. when data are available), and past be levelling occurs less regularly with plural NPs and subject pronouns. The Exist. Plural > Plural NP > Subj. Pron. hierarchy holds by and large, making internal constraints less diagnostic here. 3.2 Competition with Alternative Paradigms An important criterion not mentioned so far is that the past be paradigm may give rise to qualitative differences as well. Crucially, was is only one of three possible pivot forms that may be adopted for analogical regularisation, were and weren’t being the other two. Why was is so common (at the expense of the other two) is a hotly debated issue, resting on a range of cognitive, morphosyntactic, and social criteria. Consequently, when interested in the diagnostic value of such patterns, one needs to focus on alternative mechanisms simply because they compete with universal phenomena and are thus much more diagnostic. A cross-varietal comparison indicates that there are no one-dimensional tendencies: the dialect landscape in the British Isles is characterised by a high degree of diversity in the regional distribution of the two forms. The Linguistic Atlas of England (LAE; Orton, Sanderson, and Widdowson 1978) shows a highly intricate patterning of regional differentiation in regional varieties of British English. This is a clear indication of the flexibility and fluctuation of past be in different areas, and distinct past-tense paradigms continue to exist in contemporary British English (to some
< previous page
page_71
next page >
< previous page
page_72
next page >
Page 72 extent specific to the type of subject). There is a patchwork of isoglosses all over the country—what David Britain (2002:33) has called “a myriad of locally differing systems.” Alternations in was/were levelling transcend the community level. Variability at both inter- and intrapersonal levels is commonly reported in London English (Laura Wright, p.c. 2001), Fens English (Britain 1991), and in York English (Tagliamonte 1998). Tagliamonte (1998:155), for example, reports “alternation between the two morphological forms was and were even in the speech of the same individual in the same sentence”, as in (2) a. There was a lot of us that were sort of seventeen. b. She were a good worker, she was a helluva good worker.” Levelling of past be is consequently not a unilateral process, and alternative regularisation trends may occur in one and the same variety. Such variability makes it difficult to pinpoint overall trends in the directionality of past be regularisation in British English, but its variation is regionally distributed (although it has to be stressed that extension of was [as in ‘they was quite disappointed’] is more common than extension of were [as in ‘I were quite disappointed’]). A number of varieties have patterns that are remarkably intricate and complex—namely, levelling to was for positive and to weren’t for negative contexts (Britain 1991, 2002; Cheshire 1982; Schilling-Estes and Wolfram 1994). ‘Two-fold’ levelling techniques effectively result in the realignment of the two paradigms, leading to a “remorphologization of the was and were allomorphs of past be along positive/negative, rather than person-number, lines” (Schilling-Estes and Wolfram 1994:280). As a result, both was and weren’t allomorphs may occur in all subject contexts, with both numbers and with subjects of all persons. This is confirmed by Britain (2002:17), who reports that “[y]oung people in the [English] Fens increasingly use was in affirmative clauses (irrespective of person/number) and almost exclusively use weren’t in negative contexts.” Consequently, the fact that more than one allomorph may survive regularisation is regionally and varietally diagnostic, and the adoption and survival of alternative techniques outside the British Isles may yield vital insights into founder effects and patterns of dialect diffusion (e.g. in varieties that developed along the Eastern U.S. seaboard, such as Ocracoke, NC, or Smith Island, VA, which inherited this pattern from the founders’ inputs and maintain it to the present day). Notwithstanding, there is a considerable difference in varieties of English in the British Isles and in postcolonial varieties around the world. Whereas levelling to were is a robust feature in nonstandard British English and the adoption of weren’t for all persons is a dynamic change in progress in Fens English (Britain 2002), transplanted Englishes around the world show an overwhelming trend towards was as a pivot form, strongly supporting Chambers’ assessment.
< previous page
page_72
next page >
< previous page
page_73
next page >
Page 73 3.3 Ongoing Change The last point I would like to take up here concerns the diachronic stability of English universals. We saw earlier that past be variation is well attested historically and that the differences are to some extent varietyspecific. A particularly striking development is found in Tristan da Cunha English (TdCE), a variety of South Atlantic English that formed in the nineteenth century (Schreier 2003a). Figure 3.4 indicated that levelling to was was highly advanced in this variety, and that this pivot featured in more than 90 percent of cases where a standard variety would have were . In Schreier (2002), I studied the relevance of this finding for language change and contact dynamics in new-dialect formation processes, uncovering an unusual pattern of crosscommunity distribution. There is a change in progress in TdCE, and individual subgroups of the Tristan community participate differently in this process. Figure 3.5 shows that TdCE speakers born before WWII, when a British naval battalion was installed on the island and dozens of outsiders came to the island, have categorical levelling rates, exclusively using was for past be and thus having a perfectly regular verb paradigm (‘I/you sg./he, she, it/we/you pl./they was’). Later generations had lower levelling rates, increasingly having were with second singulars and the plural persons, and this can only be interpreted as an adoption of nonlocal features and thus as the sociolinguistic consequence of the community’s opening-up and emergence from endocentricity. In other words, the alternation between was and were was levelled out completely in
Figure 3.5. Past be levelling in Tristan da Cunha English (adapted from Schreier 2002:87).
< previous page
page_73
next page >
< previous page
page_74
next page >
Page 74 early twentieth-century TdCE (perhaps triggered by high rates in the inputs, but this is speculation), and this represents an unusual and—to the best of my knowledge—unprecedented process of categorical language change (discussion in Chambers 2004). By the same token, women were found to have consistently higher levels of were ; the men, in contrast, showed no variation whatsoever, and categorically use was in the entire past be paradigm. Even young Tristanian males, born in the mid-1980s, have categorical was , and this mirrors effects from social network theory (Milroy 1987)—namely, that the widening of social contacts (mainly by women) makes them particularly prone to pick up innovations. This is by no means an isolated phenomenon, and other studies attest to variation and change in past be (to name but Hazen 1996 on NC Ocracoke English or Gordon et al. 2004 and Hay and Schreier 2004 on NZE; the latter showed that levelled was featured quite frequently in nineteenth-century NZE and that it decreased with existential constructions between 1860 and 1890, only to increase again in the twentieth century). In this sense, English universals seem to operate no differently than other variables. 3.4 Some First Conclusions A cross-comparison shows that varieties differ both quantitatively and qualitatively in their manifestations of past be. Past be levelling has not advanced at the same rate in individual varieties, and there is also a difference in the directionality of the levelling process, explained by the fact that three different allomorphs compete for adoption in a more regular (or, more precisely, in a less irregular) past-tense paradigm. Consequently, it is the frequency of past be levelling and the choice of the pivot form that is particularly diagnostic here; in terms of internal conditioning, variation in past be levelling is not particularly insightful (at least not in what regards the three major categories of subject pronoun, plural NPs, and plural existentials). 4. CONCLUSION This chapter has discussed findings from more than a dozen of English varieties around the world, shaped by different communities in various contact settings. The analysis fully supports Chambers’ (1995) assessment: both CCR and past be levelling to was occur in all the varieties analysed and are thus true universals of English, representing “variables [that] appear to be primitives of vernacular dialects in the sense that they recur ubiquitously all over the world” (Chambers 1995:242). It is of great importance that they are found in varieties that have no sociohistorical connections whatsoever, such as Indian and New Zealand English, or St Helenian and West Virginian Appalachian English. This emphasises their importance for sociolinguistic theory as well as for the interaction of sociolinguistics with other disciplines
< previous page
page_74
next page >
< previous page
page_75
next page >
Page 75 (cognitive and historical linguistics, dialect typology, etc.). Now, on the one hand, one could argue that these features carry no diagnostic value whatsoever, simply on account of the fact that they are so common and pervasive in English everywhere. A statement such as “Final consonant clusters are generally reduced to one spoken consonant [in Singapore English], such as ‘juss’ for just or ‘toll’ for told ” (McArthur 1993:938) does no more than state the obvious and provides no information on dialect differentiation, completely failing to give insights into characteristic features of the variety under examination. This chapter has argued from another perspective. Although making an appearance everywhere, English universals can nevertheless carry diagnostic value. Many of their variety-specific characteristics cannot be uncovered by qualitative research, however, so global assessments are of little or no use. The prior analyses have shown that universals can be of immense value to sociolinguists, variationists, and contact linguists alike, provided that they are analysed in depth, with databases and corpora that allow researchers to tease out parameters of variation and internal constraints. If these are available, then CCR and past be (and all other English universals) move into the centre of interest of a wide range of research areas and turn out to be highly useful for cross-dialectal comparisons. A quantitative analysis shows that English universals are important variables for contact linguistics (yielding insights into contact histories, prior pidginisation, and/or creolisation); that they are sensitive to language learning and transfer effects (e.g. of non-English syllable structures), language variation and change, as well as social stratification, and that they yield insights into historical diffusion and founder principles. For instance, CCR is a reliable indicator of creolisation, both in terms of amount and internal conditioning effects, past be levelling yields information on founder effects, dialect affinities, contact dynamics in colonial scenarios, and so on. If analysed quantitatively, these variables can yield a wealth of synchronic and diachronic information. These are so prominent that one might well ask whether the attested quantitative differences might not be important enough to obtain quasi-qualitative significance. This is suggested by Wolfram et al. (2000), who claim that an increase of CCR and its application in certain contexts equals a qualitative difference, reduction in prevocalic environments being particularly diagnostic. By the same token, categorical levelling of past be in male TdCE speakers is indicative of processes found everywhere else, but is so advanced that the line between quantitative and qualitative differences is blurred. This is a manifestation of categorical language change because were no longer features in the past be paradigm at all (at least in the early twentieth century, based on apparent time principles). As for causation, finally, it is clear that contact histories and speaker characteristics (language contact, sustained multilingualism, etc.) have an effect on the variable usage of English universals. The further we move away from native-speaker norms and Kachru’s (1985) inner circle varieties (North America, Great Britain,
< previous page
page_75
next page >
< previous page
page_76
next page >
Page 76 Australian, and New Zealand) and the more influence through language contact we take into account, the more diversity there is with regard to these features. This, in turn, strengthens the ‘diagnosticity’ of universals and may be of importance for the continuing diversification of English as a global language. To conclude, the in-depth study of English universals offers crucial insights into a variety of research questions and is of relevance for various linguistic disciplines. The omnipresence of these features in unrelated varieties attests to the strength and vitality of ‘vernacular roots’. By the same token, more finegrained and extremely diagnostic differences only emerge in a detailed quantitative analysis—namely, when researchers adopt a ‘Sherlock Holmesian’ approach, concentrating on details rather than the big picture. NOTES 1. Past-tense formation of regular verbs in English involves - ed suffixation, which leads to bimorphemic CCs (e.g. /vd/ in lived, /st/ in passed, or /zd/ in praised), but of course not when an - ed suffix is preceded by a coronal stop, such as in grant or mend, where we find a surface realisation [•d] or [əd]. REFERENCES Anderwald, L. 2001. Was/were variation in non-standard British English today. English World-Wide 22: 1–21. Bailey, G., and E. Thomas. 1998. Some aspects of African American phonology. In African-American English: Structure, History and Use, edited by S. Mufwene, J. Rickford, G. Bailey, and J. Baugh, 85–109. London: Routledge. Bayley, R. 1995. Consonant cluster reduction in Tejano English. Language Variation and Change 6: 303–27. Blevins, J. 1995. The syllable in phonological theory. In The Handbook of Phonological Theory , edited by J. Goldsmith, 206–44. Oxford: Blackwell. Britain, D. 1991. Dialect and Space: A Geolinguistic Study of Speech Variables in the Fens . Unpublished thesis, University of Essex. ———. 2002. Diffusion, levelling, simplification and reallocation in past tense be in the English Fens. Journal of Sociolinguistics 6: 16–43. Brunner, K. 1963. An Outline of Middle English Grammar . Oxford: Blackwell. Chambers, J. 1995. Sociolinguistic Theory . Oxford: Blackwell. ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-linguistic Perspective , edited by B. Kortmann, 127–45. Berlin, New York: Mouton de Gruyter. Cheshire, J. 1982. Variation in an English Dialect. Cambridge: Cambridge University Press. Cheshire, J., ed. 1991. English Around the World: Sociolinguistic Perspectives . Cambridge: Cambridge University Press. Childs, B. 2000. A Hyde County clusterfest: The role of consonant cluster reduction in a historical isolated African-American community . Unpublished MA thesis, North Carolina State University, Raleigh, NC. Christian, D., W. Wolfram, and N. Dube. 1988. Variation and Change in Geographically Isolated Communities: Appalachian English and Ozark English (Publication
< previous page
page_76
next page >
< previous page
page_77
next page >
Page 77 of the American Dialect Society, PADS 74). Tuscaloosa: University of Alabama Press. Curme, G. 1931. A Grammar of the English Language. Vol. 3: Syntax . Boston: D.C. Heath. Doyle, A.C. 1890. The Sign of Four. Philadelphia: Lippincott’s Monthly Magazine. Eisikovits, E. 1991. Variation in subject–verb agreement in Inner Sydney English. In Cheshire, ed. 1991. 235– 55. Fasold, R. 1972. Tense Marking in Black English: A Linguistic and Social Analysis. Arlington VA: Center for Applied Linguistics. Feagin, C. 1979. Variation and Change in Alabama English. Washington, Georgetown University Press. Forsström, G. 1948. The Verb ‘to be’ in Middle English: A Survey of the Forms . Lund: C.W.K. Gleerup. Gordon, E., L. Campbell, J. Hay, M. Maclagan, A. Sudbury, and P. Trudgill. 2004. The Origins and Evolution of New Zealand English. Cambridge: Cambridge University Press. Guy, G. 1980. Variation in the group and in the individual: The case of final stop deletion. In Locating Language in Time and Space, edited by W. Labov, 1–36. New York: Academic Press. ———. 1991. Explanation in variable phonology. Language Variation and Change 3: 1–22. Hazen, K. 1996. Dialect affinity and subject–verb concord: The Appalachian Outer Banks connection. The SECOL Review 20: 25–53. Hay, J., and D. Schreier. 2004. Reversing the trajectory of language change: Subject-verb agreement with be in New Zealand English. Language Variation and Change 16: 209–35. Holmes, J. 1997. Maori and Pakeha English: Some New Zealand social dialect data. Language in Society 26: 65–101. ———, and A. Bell. 1994. Consonant cluster reduction in New Zealand English. Wellington Working Papers in Linguistics 6: 56–82. Kachru, B.B. 1985. Standards, codification, and sociolinguistic realism: The English language in the outer circle. In English in the World, edited by R. Quirk and H.G. Widdowson, 11–30. Cambridge: Cambridge University Press. Khan, F. 1991. Final consonant cluster simplification in a variety of Indian English. In Cheshire, ed. 1991. 288–98. Labov, W. 1972. Language in the Inner City . Philadelphia: The University of Philadelphia Press. ———, P. Cohen, C. Robins, and J. Lewis. 1968. A study of the non-standard English of Negro and Puerto Rican speakers in New York City . Final report, cooperative research project 3288. Washington DC: Educational Research Information Center. Lass, R. 1984. Phonology: An Introduction to Basic Concepts. Cambridge: Cambridge University Press. Lee, H. 2000. The pronunciation of English consonant clusters by Koreans. Revised version of a paper presented at the 2nd International Conference of Phonetic Sciences , held in San Francisco CA. Online: http://www.ksps.or.kr/data/papers/40/79.pdf. Lim, L., ed. 2004. Singapore English: A Grammatical Description (Series Varieties of English Around the World). Amsterdam and Philadelphia: Benjamins. Luick, K. 1964. Historische Grammatik der Englischen Sprache. Stuttgart: Bernhard Tauchnitz. McArthur, Tom. 1993. The Oxford Companion to the English Language. Oxford: Oxford University Press.
< previous page
page_77
next page >
< previous page
page_78
next page >
Page 78 Meechan, M., and M. Foley. 1994. On resolving disagreement: Linguistic theory and variation—There’s bridges. Language Variation and Change 6: 63–85. Milroy, L. 1987. Language and Social Networks . 2nd edition. Oxford: Basil Blackwell. Montgomery, M. 1989. Exploring the roots of Appalachian English. English World-Wide 10: 227–78. ———. 1994. The evolution of verbal concord in Scots. In Proceedings of the Third International Conference on the Languages of Scotland , edited by A. Fenton and D.A. MacDonald, 81–95. Edinburgh: Canongate Academic Press. Murray, J. 1873. The Dialect of the Southern Counties of Scotland: Its Pronunciation, Grammar and Historical Relations . London: Philological Society. Neu, H. 1980. Ranking of constraints on /t, d/ deletion in American English: A statistical analysis. In Locating Language in Time and Space, edited by W. Labov, 37–54. New York: Academic Press. Orton, H., S. Sanderson, and J. Widdowson, eds. 1978. The Linguistic Atlas of England. London: Croom Helm. Poplack, Shana, and Sali Tagliamonte. 2001. African American English in the Diaspora . Oxford: Blackwell. Quirk, R., and C.L. Wrenn. 1960. An Old English Grammar . London: Methuen. Patrick, P. 1991. Creoles at the intersection of variable processes: -t, d deletion and past marking in the Jamaican mesolect. Language Variation and Change 3: 171–89. ———. 1999. Urban Jamaican Creole: Variation in the Mesolect (Varieties of English around the World G17). Amsterdam and Philadelphia: John Benjamins. Santa Ana, O. 1992. Chicano English evidence for the exponential hypothesis: A variable rule pervades lexical phonology. Language Variation and Change 4: 275–88. ———. 1996. Sonority and syllable structure in Chicano English. Language Variation and Change 8: 63–91. Schilling-Estes, N., and W. Wolfram. 1994. Convergent explanation and alternative regularization patterns: Were/weren’t leveling in a vernacular English variety. Language Variation and Change 6: 273–302. Schneider, E. 2003. The dynamics of New Englishes: From identity construction to dialect birth. Language 79: 233–81. Schreier, D. 2002. Past be in Tristan da Cunha: The rise and fall of categoricality in language change. American Speech 77: 70–99. ———. 2003a. Isolation and Language Change. Houndmills, Basingstoke, and New York: Palgrave Macmillan. ———. 2003b. Convergence and language shift in New Zealand: Consonant cluster reduction in 19th century Maori English. Journal of Sociolinguistics 7: 378–91. ———. 2005. Consonant Change in English Worldwide: Synchrony Meets Diachrony . Houndmills, Basingstoke, and New York: Palgrave Macmillan. ———. 2008. St Helenian English: Origins, Evolution and Variation (Varieties of English Around the World G37). Amsterdam and Philadelphia: John Benjamins. Smith, J., and S. Tagliamonte. 1998. We Was all Thegither, I Think We Were all Thegither. Was Regularization in Buckie English. World Englishes 17: 105–26. Tagliamonte, S. 1998. Was/were variation across the generations: View from the city of York. Language Variation and Change 10: 153–91. ———, and J. Smith. 2000. Old was , new ecology: Viewing English through the sociolinguistic filter. In The English History of African American English, edited by S. Poplack, 141–71. Oxford: Basil Blackwell. ———, and R. Temple. 2005. New perspectives on an ol’ variable: (t, d) in British English. Language Variation and Change 17: 281–302.
< previous page
page_78
next page >
< previous page
page_79
next page >
Page 79 Torbert, B. 2001. Tracing Native American language history through consonant cluster reduction: The case of Lumbee English. American Speech 76: 361–87. Visser, F. 1963–73. An Historical Syntax of the English Language. 3 vols. Leiden: Brill. Wolfram, W. 1969. A Sociolinguistic Description of Detroit Negro Speech. Arlington VA: Center for Applied Linguistics. ———. 1980. Dynamic dimensions of language influence: The case of American Indian English. In Language: Social Psychological Perspectives , edited by H. Giles, W. Robinson, and P. Smith, 377–88. Oxford and New York: Pergamon Press. ———, B. Childs, and B. Torbert. 2000. Tracing language history through consonant cluster reduction: Comparative evidence from isolated dialects. Southern Journal of Linguistics 24: 17–40. ———and D. Christian. 1976. Appalachian Speech. Arlington VA: Center for Applied Linguistics. ———, D. Christian, and D. Hatfield. 1986. The English of adolescent and young adult Vietnamese refugees in the United States. World Englishes 5: 47–60. ———and R. Fasold. 1974. The Study of Social Dialects in American English. Englewood Cliffs, NJ: Prentice Hall. ———, K. Hazen, and N. Schilling-Estes. 1999. Language Change and Maintenance in Outer Banks English (Publications of the American Dialect Society). Tuscaloosa: University of Alabama Press. ———, and E. Thomas. 2002. The Historical Development of African American English. Oxford: Basil Blackwell.
< previous page
page_79
next page >
< previous page
page_80
next page >
Page 80 4 Number Agreement in Existential Constructions A Sociolinguistic Study of Eighteenth-Century English Terttu Nevalainen 1. INTRODUCTION In general terms, this chapter contributes to the discussion on the loss of inflectional agreement in the English language. This typological change is unprecedented among the Germanic languages: in the course of its documented history, English has changed from a synthetic language to a largely analytic one. Little remains of the grammatical system of agreement today apart from number marking—the distinction between singular and plural, which is shown, for instance, in determiner–head and subject–verb agreement. However, as Matthews (1987:156) remarks, “[subject–verb] agreement in English is a notoriously variable matter”. It is this inherent variability in subject–verb agreement and attempts to regulate it that constitute the dual focus of this chapter.1 The outcome of this variability is parallel (sub)systems of subject–verb agreement in social and regional varieties of English. Be is the only verb to provide more choice because it can also record person contrasts in the singular (Kortmann and Schneider 2006; Trudgill 1999:101–8). However, although often presented as invariable paradigms, in actual use these parallel systems are typically probabilistic in that speakers and writers have access to alternative forms, the choice of which can stem from not only syntactic but also semantic selection criteria. Chambers (2004) includes one of these subsystems, ‘default singulars’, in his inventory of ‘vernacular universals’, which consists of features that are found in vernacular varieties of English around the world but not in the standard language. As a case in point, Chambers singles out the generalisation of was , the past-tense form of the verb be, to both singular and plural subjects. The agreement-controller type that has been found particularly to favour was -generalisation is the plural notional subject in existential there constructions.2 Research on earlier periods shows that existential there constructions provided the most typical site for ‘default singulars’ from Late Middle English onwards (Martínez-Insua and Pérez-Guerra 2006; Nevalainen 2006a,
< previous page
page_80
next page >
< previous page
page_81
next page >
Page 81 2006b). In purely grammatical terms, this can be seen as a consequence of the conflict between two possible agreement controllers in these constructions: the dummy there and the plural notional subject. In this study, I concentrate on the form of the copula be with plural existential NP subjects in the Corpus of Early English Correspondence Extension, which covers a range of social and regional variation. I also consider the impact that prescriptive grammar might have had on the diminishing use of ‘default singulars’ in this period. The comments of normative grammarians recorded in Sundby, Bjørge, and Haugland (1991) suggest that heightened awareness of grammatical agreement was common among the literate social ranks in the late eighteenth century. If the rise and fall of the collocation you was is anything to go by (TiekenBoon van Ostade 2002), this heightened awareness is likely to have impacted on other agreement patterns as well, reinforcing and maintaining grammatical agreement in the standard and other mainstream varieties of English. In section 2, I discuss some earlier findings concerning the various systems of subject–verb agreement found in English, with particular attention to ‘default singulars’. I introduce my data in section 3, and give an overall view of the eighteenth-century normative tradition in section 4. Normative grammars serve as the backdrop for the discussion of agreement variation in my corpus in section 5, where I define the scope of my enquiry. In section 6, I report my quantitative findings by analysing both the diachronic development and tense variation of default singulars in existentials and correlating them with such social factors as the writer’s gender and social status. I return to the issue of agreement mismatches in section 7, placing my findings on eighteenth-century agreement variation in a broader grammatical perspective. I look at both default singulars and notional plurals, and reconsider the types of syntactic and semantic motivation that give rise to subject–verb agreement mismatches in a minimally inflected language such as English. 2. PATTERNS OF SUBJECT–VERB AGREEMENT 2.1 Present-Day Variation Kortmann and Schneider’s (2006) survey provides a broad overall view of nonstandard agreement patterns in 46 varieties of English around the world. It maps eight commonly attested agreement patterns, all but one of which represent simplifications of various kinds when compared with the modern standard and mainstream varieties of English. Two of these patterns consist of invariant present-tense forms due to zero marking for the third-person singular or, alternatively, the generalisation of the third-person - s in all persons. Two patterns involve verb deletion: of the copula be and of the auxiliary have . One pattern involves the dummy subject in existential clauses, which can be realised by there or it or be absent altogether (all three are historically attested; see section 2.2).
< previous page
page_81
next page >
< previous page
page_82
next page >
Page 82 Two patterns simplify number marking with the copula be. One of these is the topic of this chapter (i.e. the use of there’s , there is, and there was with plural subjects), and the other is the generalisation of either was or were to both singular and plural. A more complex pattern, the Northern Subject Rule, constitutes a feature of its own in the survey. The use of there’s , there is, and there was with plural subjects shows a wide distribution across the colloquial and nonstandard varieties included in Kortmann and Schneider (2006), occurring in about half of them. It is pervasive or frequent throughout the British Isles, North America, Australia, New Zealand, and South Africa, but occurs in other Southern Hemisphere varieties of English as well. As suggested by the survey materials discussed above, singular agreement is found with existentials even in mainstream varieties in conversational language. In their corpus-based grammar of present-day British and American English, Biber, Johansson, Leech, Conrad, and Finegan (1999:186) analysed the contracted forms there’s , here’s, and where’s followed by plural subjects. They found that, in conversation, plural subjects occur more frequently with there’s than with there followed by a plural copula. The American English corpora studied by Crawford (2005) show a similar trend: with plural subjects, there’s is almost as frequent as there are in conversation and academic lectures. The mean frequency of nonstandard agreement is 51% of the total instances of there’s , there is, and there are with plural subjects in conversations (N = 2,294), and 45% in lectures (N = 1,408).3 Both Crawford (2005) and Biber et al. (1999:191) account for these findings by considering the contracted form to be a fixed expression in which individual elements are not chosen independently. Huddleston and Pullum (2002:242) note that, in terms of agreement, the cliticised form there’s is often treated like a pronoun in speech. 2.2 Historical Variation There is some diachronic research on New Zealand English suggesting that agreement patterns can change in a speech community over a short period of time. Hay and Schreier (2004) found that singular agreement virtually disappeared from existentials in spoken New Zealand English in the late nineteenth century, but that it has since regained its position and become an established part of colloquial New Zealand English. Over a longer time span, much more variation emerges, showing that (a) in some cases present-day variation may have a long history, but (b) not all patterns attested today necessarily have documented histories, and (c) new patterns can emerge with time. Old and Middle English had three basic patterns of existential constructions: those with the dummy subject there , those with the dummy subject it , and those without a dummy subject. That they were genuine alternatives is evidenced, for example, by parallel translations of the same text. Those
< previous page
page_82
next page >
< previous page
page_83
next page >
Page 83 in (1) come from the various manuscripts of Cursor Mundi (line 2210, ed. Morris 1873; cited from Breivik 1991:32). (1a) Þat tim it was bot a langage,/ Hebru þe first þat adam spac; (Cotton) (1b) Þat time was bot an langage/ hebreu þe first at adam spac. (Fairfax) (1c) Þat tim was þer but o langage/ Ebrew þe furste þat adam spak. (Trinity) (1a) and (1c) have dummy subjects, either without subject–verb inversion (1a) or with it (1c); in (1b), the subject slot filler is empty. As Breivik (1990, 1991) shows, the most common existential construction in Old English consisted of the zero pattern with no dummy subject. Breivik’s corpus of 1,653 existential clauses indicates that the dummy it formed a minor option throughout the period from Old English to Early Modern English, constituting some 2 percent of the cases with dummy subjects in Old English, about 5 percent in Middle English, and only 1 percent in Early Modern English. However, the relative proportions of the majority patterns underwent a radical reweighting in the course of the Middle English period. Figure 4.1 presents the distribution of the zero and existential there patterns in the four periods in Breivik (1991:37). By Early Modern English, there existentials had become the dominant pattern found in the written data sources that Breivik investigated. He argues that the insertion of there , an empty topic, was associated with the verb-second constraint: in Old English, the dummy subject appeared preverbally in the vast majority of the cases. The dummy there was reanalysed
Figure 4.1. Diachronic changes in the two major patterns of existentials according to Breivik (1991:37). (OE = Old English, EME = Early Middle English, LME = Late Middle English, EModE = Early Modern English)
< previous page
page_83
next page >
< previous page
page_84
next page >
Page 84 as a subject NP when Middle English developed into a verb-medial language (Breivik 1991:35–7). Even after this reanalysis, a good deal of agreement variation prevailed with postverbal NP subjects, for example (Fischer 1992:366). There are a couple of recent corpus-based studies on the history of English existentials which discuss subject–verb agreement. Martínez-Insua and Pérez-Guerra (2006:195–97) report low overall frequencies of subject–verb nonagreement with existentials in two multigenre corpora, the Helsinki Corpus and ARCHER, which extends from the mid-seventeenth century to the 1990s. Their figures, based on the 4,007 existential there -sentences, display a declining trend from the overall rate of 5.3% nonagreement in Late Middle English and about 6% in Early Modern English down to 3.2% in Late Modern English and 2.7% in presentday English. Martínez-Insua and Pérez-Guerra (2006:196) consider the total number of existential sentences with verb and notional subject (non)agreement, including those with verbs other than be, but the overwhelming majority of the instances are with be. Despite the declining diachronic trend, default singulars typify the present-day results to the extent that the authors suggest that the ‘ there + singular verb pattern’ has undergone grammaticalisation and become a formula for introducing new information into the discourse. In Nevalainen (2006b), I studied the general use of was with plural subjects in the Corpus of Early English Correspondence before the prescriptive era. My results are based on 1,821 instances of the was/were variable, and they indicate that was typically occurred with plural subjects in the Northern English data in the fifteenth and sixteenth centuries, but that this construction was not restricted solely to the North. In the course of the seventeenth century, this pattern of agreement mismatch declined and levelled dialectally, but continued to be used to some extent by the literate social ranks throughout the country. Figure 4.2 shows the regional distribution of the use of was with plural subjects in the letter corpus between 1440 and 1681. In the fifteenth and sixteenth centuries, was was particularly favoured with plural existential and ordinary NP subjects, but the subject NP factor ceased to be statistically significant from the late sixteenth century onwards. The fact that was persisted with plural existential subjects can be taken as a sign of the diachronic robustness of the existential there construction as a conditioning factor. Interestingly, there are only a handful of instances in my data of was being triggered by pronouns, you in particular, until the latter half of seventeenth century. I come back to the subject types associated with agreement mismatches in section 7. The eighteenth-century data discussed by Tieken-Boon van Ostade (2002) and Laitinen (forthcoming) support my findings on the earlier centuries. In combination with singular you, the use of was only gained momentum in the course of the eighteenth century, picking up until about 1760, but receding in the written language—in novels and letters, for
< previous page
page_84
next page >
< previous page
page_85
next page >
Page 85
Figure 4.2. The frequency of was with plural subjects, 1440–1681. Redrawn from Nevalainen (2006b: 359). instance—towards the end of the century. Tieken-Boon van Ostade suggests that this arrested development “may very likely have been the result of the influence of normative grammarians such as Lowth” (2002:97– 8). This U-turn in the eighteenth-century development of you was runs against the normal pattern of diffusion of linguistic changes, such as those discussed in Nevalainen and Raumolin-Brunberg (2003). It can be taken as an indication that the change took place above the level of social awareness. One of the questions I investigate in the rest of this chapter therefore is whether a similar decline can be detected in agreement mismatches with existentials. 3. MATERIAL OF THE STUDY The data for this study come from the Corpus of Early English Correspondence Extension (CEECE; 1681– 1800). Its current version includes more than 4,400 letters by 300 writers in 72 collections and amounts to 2 million running words.4 As periodisation is needed for the tracing of frequency distributions through time, I further divided the 120-year period into three 40-year subperiods: 1680–1719, 1720–1759, and 1760–1800. Not every writer in the CEECE contributed to the data, which yielded 631 instances of be with plural existential subjects (there is/are , was/were + NP pl). This linguistic variable is discussed in more detail in section 5. It is worth noting that the cliticised there’s appeared throughout the period studied and was included in the raw data to be analysed. However,
< previous page
page_85
next page >
< previous page
page_86
next page >
Page 86 I follow the literature and limit my discussion to the simple present- and past-tense forms of be in existentials, and thus exclude the few instances of compound tenses that appeared in the data, although they could display number variation (as in examples (2a) and (2b)). (2a)It is a pity there have been so many accidents of this kind here-abouts. (A 1759 T THUGHES 38) (2b)whither obligers are apt to exact too large returns, or whither human Pride naturally hates to remember obligations, but I have seldom seen Freindships continue long where there has been great Benefits conferr’d, (A 1755? FN MMONTAGU III,99) 4. THE EIGHTEENTH-CENTURY NORMATIVE TRADITION 4.1 Prescriptive Grammars Before moving on to agreement patterns with existentials, a few words need to be said about prescriptive grammars. It should be noted that the first grammars of English, 32 published in the seventeenth century, were descriptive and did not proscribe usage. The earliest even listed alternative expressions when language change was in progress (Nevalainen 2006a: 16–20). A widening concern, and market, for grammatical correctness in the eighteenth century is reflected in the increase in the number of grammar books published in the latter half of the century. Figure 4.3, calculated on the basis of Michael (1970), shows the sharply rising numbers between 1760 and 1800. The ‘commodification’ of language in the eighteenth century included many bookseller’s projects (McIntosh 1998:8–9, 169–94; Tieken-Boon van Ostade 2000). The personal correspondence of the bookseller Robert Dodsley, for example, reveals that the idea for Robert Lowth’s A Short Introduction to English Grammar (1762) did not come from Lowth, but from Dodsley. The same applies to Samuel Johnson’s Dictionary (1755), which, according to Johnson’s biographer Boswell, was similarly initiated by Dodsley (Tieken-Boon van Ostade 2000:28). Broadly speaking, eighteenth-century grammars had three main aims: (1) to codify the facts of English grammar; (2) to decide on variant forms, showing one of them to be the correct one; and (3) to pin down what were, in the opinion of the grammarians, common errors of language use. As a case in point, Lowth’s Short Introduction condemns, for example, double negatives and constructions such as between you and I and different than/to, as well as you was , which it calls ‘an enormous solecism’. At least 22 editions of the grammar came out in the eighteenth century, and it had a multitude of imitators. Because grammar norms were to a large extent shared on both sides of the Atlantic at the time, Lowth’s grammar was influential throughout the
< previous page
page_86
next page >
< previous page
page_87
next page >
Page 87
Figure 4.3. English grammars published in the 18th century (absolute figures; based on Michael 1970:588– 94). English-speaking world. It was used by Harvard students into the 1840s, and it exerted strong influence on the grammatical thinking of Noah Webster and other eminent proponents of American English (Finegan 2001:365, 371; Nevalainen 2003). Eighteenth-century grammars took their illustrations of erroneous usage from seventeenth- or early eighteenth-century literary sources. A Dictionary of English Normative Grammar, 1700–1800 , compiled by Sundby, Bjørge, and Haugland (1991), contains a sample of nearly 200 normative works. Their most frequently cited sources of grammatical errors are, in this order, Swift, The New Testament, Hume, Addison, Pope, The Spectator, The Old Testament, Shakespeare, Dryden, and Milton (1991:35). The attitude labels given to proscribed forms and usages range from simply disapproving (“bad”, “censurable”, “inadmissible”, “unpardonable”) and some more or less elusive qualities of decorum (“affected”, “barbarous”, “harsh”, “inelegant”) to social and register variation (“cant”, “colloquial”, “vulgar”), regional provenance (“Scotticism”), and many other aspects of linguistic variation (“new”, “obsolete”, “rare”; see Sundby et al. 1991:38–53). 4.2 Prescriptive Views on Agreement Subject–verb (non)agreement was a topic regularly treated in eighteenth-century grammars. A number of patterns of ‘false concord’ in existential constructions were distinguished and proscribed. The following lists of what I refer to as ‘default singulars’ and ‘notional plurals’ derive from Sundby,
< previous page
page_87
next page >
< previous page
page_88
next page >
Page 88 Bjørge, and Haugland (1991:156–58). After each item, one or two cases of ‘wrong’ usage are illustrated. I have indicated the offending construction in bold type. I will use the first list, items (i) to (vi), in delimiting the linguistic variable for this study (see section 5). Default Singulars: (i) a/some/few: There is some/a few of them very merry. (ii) many: When their Vices forsake them, there’s many flatter themselves, that they have forsaken their Vices. (iii) plural N: There was taken up of Fragments, twelve baskets. In the British Army, in Time of Peace, there is seventy two Regiments of Foot. (iv) coordinated Ns (and): Derham says there is seas and rivers in the moon. (v) numeral: There’s two or three of us have seen strange things. There was twenty. (vi) coordinated pronouns (and): There was him and her and me. With default singulars, plural nouns, numerals, and coordinate structures as the notional subject attracted the most comment. The contracted form there’s appears in the illustrations, as do both present- and pasttense forms of be. With notional plurals, cases (vii) to (xiv), the picture looks more varied. Notional Plurals: (vii) every thing: There are every Thing done that can be expected from a prince who removes Mountains. (viii) collective N of pl. N: There are a variety of virtues to be exercised. (ix) partitive N: There are a Kind of Magnetism in Goodness. (x) singular N: there are no conversation so agreeable as that of the men of integrity. (xi) disjunctive coordinated Ns (neither … nor ): There were neither Thomas nor me, nor thee there. (xii) exclusive coordinated Ns (or ): There was much Genius in the World, before there were Learning, or Arts to refine it. (xiii) none: There are none have less than I. (xiv) nothing: There are nothing so delightful, says Plato, as the hearing or the speaking of truth. These illustrations reflect the central role that parts of speech played in these grammars. Interestingly, only 6 of the 14 cases are concerned with default singulars and the rest with notional plurals. This preponderance of plurals where singular concord is expected on syntactic grounds suggests that the issue is broader than would be predicted on the basis of paradigm levelling (cf. section 2.1).
< previous page
page_88
next page >
< previous page
page_89
next page >
Page 89 Some cases were a matter of controversy. They include, for example, the treatment of and-coordinated nouns. In his The Grammatical Art Improved, Richard Postlethwaite (1792) suggested that coordinated nouns may sometimes “agree with a Verb Singular; because it may be considered as applied to them severally: As … ‘Sand, and Salt and a Mass of Iron, is easier to bear than a Man without Understanding’”. Other grammarians, however, including Lindley Murray (1795), gave the same sentence as an illustration of false concord (Sundby et al. 1991:149). As they divided contemporary normative opinion, in actual use NP subjects consisting of coordinate singular noncount nouns could obviously follow either grammatical or notional concord. The singular verb in (3), for instance, corresponds to Postlethwaite’s distributive view. (3)In Dr Priestly’s Liquor of conviction there is much nitrous acid, some marine acid, and tenfold the quantity of water. What follows? (A 1788 TC EDARWIN 181) The tension between syntactic and semantic criteria in subject–verb agreement is a long-standing one in English. Fischer (1992:364–69), for example, discusses agreement variation in Middle English with respect to collective nouns, coordinate, and postverbal NP subjects. Most of these issues are raised in modern descriptive grammars such as Quirk, Greenbaum, Leech, and Svartvik (1985:757), who distinguish the principles of grammatical and notional concord as well as the principle of proximity. In the next section, I compare the distinctions they make with those of eighteenth-century ‘false concord’. 5. DELIMITING THE LINGUISTIC VARIABLE The linguistic variable constructed for this study of agreement patterns in existential clauses containing the dummy there consists of the choice between singular and plural forms of the copula be accompanied by a notional subject in the plural. This basic choice is relevant to both the present- and past-tense forms of be; see the illustrations in (4) and (5). (4a) There are two little things I should have a Curiosity to see. (A 1746? T JBROWN 104) (4b)… and there is two children dead, one before I received the things and one since. (A 1759 T JKENTING 54) (5a)and tell her I was last night at St Jeames, and ytther was but a few dancers. (A 1699? FN A2HATTON 241)
< previous page
page_89
next page >
< previous page
page_90
next page >
Page 90 (5b)The perjured priest read prayers att his own church, of Sunday morning, where there were such members of the mobb, that those near him kisst his hand. (A 1710 FO ACLAVERING 77) In order to be able to follow the diachronic development of these patterns systematically, however, I need to analyse some borderline cases of the linguistic variable more closely. As shown by the above lists of ‘false concord’, coordinate subjects presented particular problems for normative grammarians, who tended to adhere to purely syntactic criteria. However, the use of the singular with some coordinated NP subjects can be justified in semantic terms. Quirk et al. (1985:760–1) discuss these cases under the heading of coordinative apposition . When the coordinated units of the subject NP share the same referent, the use of the singular is justified if each of the units is in the singular. Similar views were expressed by eighteenthcentury grammarians, although they were contested by the more rigorous, including Murray (1795:92). In view of this semantic motivation and considering the contemporary controversies, I excluded cases like (6) and (7) from my quantitative analysis. (6)And since there is likewise a manifest Rotation and Circling of Words and Phrases , which go out and in, like the Mode and Fashion: (A 1689 T JEVELYN 209) (7)He farther observes, that when there was any Commotion or Rebellion in the parts of Italy, or Gaule, the General of Horse carried a Blew Cornet; (A 1685 T JEVELYN 156) Taking a step further in the notional direction, there is a degree of indeterminacy in the interpretation of abstract nouns, which depends on, as Quirk et al. (1985:761) put it, “whether qualities are seen as separate or as a complex unity.” They refer to cases such as fairness and impartiality and law and order, which can appear with a singular or a plural verb. Similar ‘latitude of application’ (Murray 1795:92) applies to Late Modern English. The example in (8) is an illustration of this pattern, also omitted from the frequency analysis.5 (8)When he express’d himself upon these Subjects, there was a Weight and a Dignity in his Manner, such as I never saw before. (A 1770 TC WCOWPER I,228) The grammatical process of ellipsis gives rise to coordinate structures which on the surface are much like the ones discussed previously. As shown by the adverbial complements in (9) and (10), however, cases like this are syntactically rather than semantically determined. In (9), there is is ellipted from the & more luck clause; the case in (10) may also be interpreted
< previous page
page_90
next page >
< previous page
page_91
next page >
Page 91 syntactically as an afterthought. They were not included in my quantitative analysis. (9) there is much merit in ye Play, & more luck in ye Choice of the Subject – I rejoice at the success, for Monsr de Belloy the Author of it is a most ingenious Modest & deserving man (A 1765 TC DGARRICK II,450) (10)If I did not mention Tristram to you, it was because I thought I had done so before. there is much good fun in it, & humour sometimes hit & sometimes mist. (A 1760? TC TGRAY 681) Both eighteenth-century and modern grammars agree that subjects with singular members coordinated with or and nor take a singular verb on syntactic grounds (Quirk et al. 1985:760, 762–3). Cases such as (11) and (12) therefore do not qualify as instances of ‘default singulars’. (11)If there is any preface or note or any thing relating to it please to copy the whole, and tell nobody about it. (A 1780 T SAJOHNSON III,293) (12)Doctor Collet at Newbury informed me the minister in whoes parish he inspect’d never demand’d anything which, I think, is a proof there is no lawful nor customary fee . (A 1759 T JJCLARK 27) With noncoordinate subject NPs, my choice of data to include in the default singular category was largely guided by form. I therefore agreed with Prince Augustus Hanover, one of the royal contributors to the CEECE, that hopes is plural in (13a). Adopting this analysis made the usage of his brother Prince Adolphus in (13b) a default singular. (13a)I have constantly delayed writing in hopes of being able to inform your Majesty of my sailing. There are hopes now of the wind coming fair, in which case not a moment will be lost. (A 1800 FN AHANOVER III,459) (13b)sAs the weather has a little changed these last three days and therefore there is some small hopes of the opening of the Elbe, I take the liberty of profiting of this first opportunity to return your Majesty my most humble thanks for your two last letters which I received at the end of January. (A 1800 FN ADHANOVER III,326)
< previous page
page_91
next page >
< previous page
page_92
next page >
Page 92 6. RESULTS OF THE STUDY 6.1 Overall Findings The linguistic variable constructed according to the above principles yielded 631 instances in the CEECE, consisting of 151 ‘default singulars’ with is and was and 480 plurals with are and were . Figure 4.4 shows the distribution of the singular forms in the three periods studied. At no point in time do default singulars make up more than 40 percent of the cases. Moreover, the figure indicates a steady fall in the frequency of singular agreement with plural existentials across time. In contrast to the pattern with you was , discussed in section 2.2, there is no dramatic drop between the last two periods. However, these figures are much higher even in the third subperiod than those Martínez-Insua and Pérez-Guerra (2006) obtained using other corpora.6 6.2 Present and Past Let us next move on to the tense distribution of the default singulars. According to the original definition of the term, ‘default singular’ was used only with reference to was , the past tense form of be. Figure 4.5 presents the breakdown of my data according tense. Two parallel developments emerge here—both statistically significant on the basis of the chi-square test— the decline of the default singular is from 33 to 12 percent of the cases (p < .001, N = 489), and the decline of the default singular was from 54 to 28 percent (p < .01, N = 142). With was , the decline takes place between the second and third periods. The figures
Figure 4.4 Overall distribution of singular concord with plural existentials (pooled). Percentages of is and was .
< previous page
page_92
next page >
< previous page
page_93
next page >
Page 93
Figure 4.5 Past- and present-tense forms. Percentages of is and was . confirm the emphasis that previous studies have placed on past-tense forms as prototypical ‘default singulars’ (cf. section 2), but they also show that this pattern applies equally to the present-tense forms, which are more frequent than past-tense forms in running text. Similarly, it is worth pointing out that almost all the cases of is with plural existentials in the data are full forms, rather than instances of contracted there’s , which typically appears with singular subjects. 6.3 Gender Variation Even more dramatic variation emerges when we analyse the external variable of gender. Figure 4.6 indicates the gender distribution in the data over the three subperiods. The overall figures for women’s usage of the past-tense
Figure 4.6 Gender distribution. Percentages of is and was .
< previous page
page_93
next page >
< previous page
page_94
next page >
Page 94 forms are low, but the diachronic trend is statistically significant (p < .001, N = 48), whereas the corresponding trend in male usage is not (N = 94). The falling frequencies of the present-tense forms are significant for both sexes at the p < .001 level (N = 117 for women and N = 372 for men). In both cases, women show systematically higher frequencies of the vernacular forms than men, which could be an indication that the change spread from above the level of social awareness. Women rarely had the same formal education as men at the time when grammar books and review magazines were set up as vehicles for social evaluation (cf. Nevalainen and Raumolin-Brunberg 2003:131, 205–6). In principle, it would therefore be possible to interpret these figures in terms of the growing influence of normative grammar towards the end of the eighteenth century: with was , the decline is more pronounced between the second and third subperiods than between the first and second. With is, the decline is more moderate, but nevertheless statistically significant. 6.4 Variable Grammars To find some qualitative support for these statistical findings, I reconstructed the basic types of probabilistic grammar that individual writers displayed in their agreement patterns. For this purpose, I selected all writers with ten or more instances of the variable and calculated their degree of ‘vernacular entrenchment’. Predictably, three categories emerged: (1) those who never used singular forms with plural existentials, (2) those who employed them to varying degrees, and (3) those who always used them. These three groups are listed in Table 4.1 for the three periods. Table 4.1 Individual Variation in the Use of Singular Forms. Percentages of is and was . Period A (1680–1719) Period B (1720–1759) Period C (1760–1800) Elizabeth Clift Isabella Wentworth 100% 100% William Clift Peter Wentworth 92% 37% Mary W. Montagu 28%Josiah Wedgewood 20% Thomas Twining John Evelyn 6% 12% Jane Austen 0% Jeremy Bentham 0% William Cowper 0% Thomas Gray Humfrey Wanley 0%Richard Hurd 0% 0%
< previous page
page_94
next page >
< previous page
page_95
next page >
Page 95 Table 4.1 suggests that there was a good deal of individual variation throughout the period. This variation was partly connected with education. Those men who never used default singulars had a university education or legal training: for example, Wanley and Gray were scholars, and Hurd was a bishop. The men with occasional default singulars, Evelyn and Twining, were also university educated, but those with higher frequencies, Peter Wentworth and William Clift, were not. Wentworth was a royal equerry, originally from Yorkshire, and Clift, a Cornish miller’s upwardly mobile son, worked in London in the household of John Hunter, the eminent surgeon and anatomist. There were two categorical default singular users, both of them women. Lady Isabella Wentworth was a Yorkshire gentlewoman who served as a bedchamber woman to the Queen of James II; she wrote her letters between 1705 and 1713. By contrast, Elizabeth Clift, writing between 1794 and 1799, was the eldest sister of William Clift and scraped a meagre living as a domestic servant in Cornwall. Their letters are illustrated in (14), with modern punctuation added to Isabella Wentworth’s letter in (14a) by the editor of her correspondence. (14a)My dearist and best of children, I never remember such colde wether at this time of the year, which I fear will make corn grow dear; it is rissen a little. I told you a great lye in a former letter, that thear was a great many of the young men of this town in Newgate, but thear was but twoe, most of the rest run away; (A 1710 FN IWENTWORTH 121) (14b)there was a misforting hapned it was a Good friday witch I think is very unproper day to work but Master had ordred his work Peopel all down to margret to work that day ther was three Carpenters an two Masons an Eight Labrours and whin I Came home from milkin in the morning they told me that master was Gone to margret (A 1794 FN ECLIFT 95) The passage of time is nicely captured by these two women. In the prenormative era, a gentlewoman at Court could be a consistent there is/was user, whereas less than a hundred years later, we have to go much lower down on the social scale to find one. 6.5 Default Singulars and Notional Plurals The structural criteria that have been found to promote subject–verb nonagreement in existentials from Late Middle English on also apply to my eighteenth-century data. In addition to those discussed in section 4.2, they include proximity of the verb and the dummy subject, a coordinated NP subject with the first member in the singular, and distance between the
< previous page
page_95
next page >
< previous page
page_96
next page >
Page 96 notional subject and the verb (Corbett 2000:188; Fischer 1992:366; Nevalainen 2006b: 364–5; Quirk et al. 1985:757). The use of singular was with a plural subject in (15) illustrates a context in which a coordinate NP subject with its first member in the singular is separated from the verb by an adverbial complement and from the second member by the parenthetical remark I think his name is. (15)Young Mr. Rodbeard being to goe out of the Country the Monday following sent to me desiring to drinke with me that day as the last I should see him, and comeing to him there was in the Roome of his Company One Mr. Amery I think his name is that Marryed Pitts his sister and two other Taunton men whome I know not nor never enquired after. (A 1699 FN NPINNEY 106) If structural factors typically promoted the use of default singulars, both the syntactic and semantic factors listed by normative grammars motivated notional plurals. My data do not show any noticeable levelling to were or are with singular subjects in existentials (cf. section 2.1), but there were certain singular contexts in which plural verbs occurred. These commonly involved a singular determinative noun associated with a plural noun, as in (16) and (17). The grammatical status of these determinative nouns varies. Some, such as abundance in (16), which occurs without an article, and number in (17), assume postdeterminer functions as quantifiers (Quirk et al. 1985:261–4). Partitive expressions are illustrated by set in (18), which provides a minimal pair showing both plural (18a) and singular concord (18b). (16) Tho‘ there are abundance of coaches and waggons that pass this road, yet but one carriage comes by at a reasonable hour … (A 1761 T THUGHES 131) (17) We were at a bal Masqué on Thursday night, where there were a very considerable number of French and German women ; the beauty of the former exceeded much that of the latter. (A 1792 FN GLGOWER 44) (18a) There are a New set of Puppets as big as life, chief Part of which have been brought from ye Country, at a very great Expence. (A 1735 FN DGARRICK I,15) (18b) There is a set of Italian Comedians who act twice a week at the Opera house, but they are very little approv’d off, for the Harlequin is very indifferent, so that they find but small encouragement. (A 1726 T JGAY 63) Finally, there are cases in which the purely structural links between a plural verb and a notionally plural subject are more tenuous; see (19) and (20).
< previous page
page_96
next page >
< previous page
page_97
next page >
Page 97 The example in (20) is interesting in that the writer, Lady Isabella Wentworth, was one of the categorical users of default singulars in my data sample. (19)When I left him the vessel swarmed with Jews: but these were either left at Gravesend, or at least are not Cabbin Passengers: for of these there are but one besides himself; a decent civil well-behaved man; with whom he agrees very well. (A 1779 FN JBENTHAM II,275) (20)I wish Betty could be gott in made of Honour, thear are lykly very soon to be room for twoe or thre, as its said. Your sister W. father has setled sumthing upon the eldest girle, … (A 1706 FN IWENTWORTH 58) 7. DISCUSSION 7.1 Vernacular Continuity Both default singulars and notional plurals have a role to play in English grammar, both in existentials and more generally. Most of the criteria that promote these constructions in existentials apply to agreement patterns in general, with the exception of the dummy there , which may be interpreted as the agreement controller with default singulars. With notional plurals, semantically motivated agreement mismatches may be placed on a cline from more to less vernacular. As pointed out in section 4.2, notional plurals such as collective nouns are not only vernacular features, but belong to the grammatical common core of English; moreover, they show regional differences in frequencies of use (cf. Hundt 2006). In contrast, usages similar to those proscribed by eighteenth-century normative grammarians appear in mainstream spoken English corpora; (21) comes from Crawford (2005). (21)now I mean there are a whole industry dealing on people who who spend their time worrying about the calculation of energy bands. (Crawford 2005:42) Cases like (21) are likely to be dismissed by grammarians as performance errors. Not surprisingly, selfcorrections and hesitations may be found in contexts where a plural verb is followed by a singular subject, as in (22) and (23), both cited from the British National Corpus (BNC). (22)… but in the next twenty years so there are a problem with schools, there are problems, I think (BNC J3W 10) (23)… relaxation, romance, having a good time <pause> there are a whole se--sequence of different reasons (BNC F88 320)
< previous page
page_97
next page >
< previous page
page_98
next page >
Page 98 However, (23) constitutes a determinative phrase, similar to those discussed in section 6.5. The following illustrations, also from the BNC, show more lexical variety. (24)Note that this is possible even if there are a minority of individuals … (BNC CM2 333) (25)… women in their twenties and early thirties, there are a significant proportion who never experience “retirement”. (BNC FST 1033) (26) There are a staggering amount of pictures here. (BNC K1D 1520) Modern descriptive grammars discuss a variety of cases in which syntactic agreement rules are overriden. Biber et al. (1999:190, 255–8) recognise the “occasional use of plural concord with species nouns” and note that “the expressions with species nouns [kind of , form of , type of ] behave in some respects like determiners.”7 Examples such as those in (22) to (26) suggest that this use is not limited to species nouns, but includes a wider set of what Huddleston and Pullum (2002:503–504) call number-transparent nouns, which require a complement. These determinative constructions provide potential sites for grammaticalisation. This is also the case with the contracted form there’s , which is now more frequent than there is in many varieties of English, including the mainstream colloquial British and American English data studied in Biber et al. (1999:186). 7.2 Subject-Type Hierarchies Individual studies of regional and social varieties of English propose slightly different implicational hierarchies for subject types that trigger the use of is and was with plural subjects. In comparative accounts, constructions with the dummy subject there come first and those with you second. Chambers (2004:139) places we third, Tagliamonte (1998:158) plural NPs; both agree that the pronoun they is the least likely subject type to induce an agreement mismatch (see also Tagliamonte 2002:742–5). Hierarchies such as these justify the treatment of there -existentials as a separate construction, as in Kortmann and Schneider (2006). As we saw in section 2.2, there are also historical reasons for doing this. However, subject-type hierarchies are not historically stable, but interact with other developments that take place in the language. In English, the loss of the second-person singular pronoun thou is a case in point. Given that singular agreement with singular you is semantically justified in Modern English, we could exclude it from the inventory of plural subject types.
< previous page
page_98
next page >
< previous page
page_99
next page >
Page 99 Considering you as a case apart, Tagliamonte’s general hierarchy looks very much like the one proposed for Late Middle and Early Modern English by the VARBRUL analysis in Nevalainen (2006b): existential constructions with plural notional subjects come first, followed by other plural NP subjects, leaving pronoun subjects as the least likely triggers of singular concord. In my sixteenth- and seventeenth-century data, the few pronoun subjects were mostly relative pronouns, whereas personal pronoun subjects were even rarer. This generalised hierarchy is reminiscent of Corbett’s (2000:190) Agreement Hierarchy, in that both suggest that personal pronouns stand out as the category most likely to show agreement with notional, semantic justification. In subject–verb agreement, pronoun controllers have this property despite the paradigmatic pressures exerted by regularised agreement patterns such as generalisation to was . A controller-type hierarchy that predicts agreement mismatches with plural subjects can now be set up: there’s + pl. NP < there + pl. NP < pl. NP < pl. personal pronoun As we move from left to right, the likelihood of agreement forms with greater semantic justification increases. Corbett (2000) shows that similar semantic and syntactic criteria are found with noun-phrase controllers cross-linguistically. Semantic criteria regulate collective nouns such as committee, which may take either singular or plural agreement depending on the intended meaning and the language variety in question.8 By contrast, syntactic criteria are often more important with coordinated noun phrases, for instance, where principles such as proximity come into play not only in English, but other languages as well. 8. CONCLUSION The factors, both syntactic and semantic, that control the use of default singulars in existentials today also operated in the normative eighteenth century. This is in evidence in the examples prescriptive grammars give of what they call ‘false concord’. A significant downward trend in the use of default singulars was found in the correspondence data as the century wore on, paralleling other researchers’ findings on the collocation you was in written data. The difference between these two cases lies in the fact that, whereas the use of default singulars in existential there constructions has a long history in English, you was represents a late innovation in the language. It is therefore tempting to ascribe at least part of this downward trend to the impact of normative grammars. I would like to stress, however, that the supralocal decline in the use of default singulars in general did not begin with eighteenth-century prescriptive grammars. Similarly to many other supralocal processes, it started much
< previous page
page_99
next page >
< previous page
page_100
next page >
Page 100 earlier and also affected the North, where default singulars were current in Middle and Early Modern English in the data that have come down to us. Normative grammarians were nevertheless influential in stigmatising subject–verb nonagreement, thus making certain concord patterns a conscious choice for the educated. From a typological perspective, this awareness reinforced syntactic criteria in the determining of subject– verb agreement and worked against the levelling of the number contrast in those mainstream varieties that had preserved it. At the same time, it is noteworthy that the converse of default singulars, notional plurals, have in some cases resisted the influence of normative pressures. Collective nouns and nouns with determinative functions, for instance, can display semantically motivated agreement patterns even in standard English. These are an interesting extension and sign of the vitality of notional concord in English across time. Default singulars, structurally motivated though they are particularly in existential constructions, account for only one chapter, albeit a long one, in the grammar of number agreement in English. NOTES 1. The research for this chapter was supported by the Academy of Finland National Centre of Excellence funding for the Research Unit for Variation, Contacts and Change in English (VARIENG), which I gratefully acknowledge. 2. My terminology here follows that of Corbett (2000:178), who distinguishes between agreement controllers and targets. In the case of subject–verb agreement, it is the subject that determines the form of the verb, which hence is the agreement target. See section 7.2. The term existential also covers so-called presentative constructions, but these constitute a small minority of the cases in my corpus. 3. I constructed the linguistic variable and calculated its frequencies on the basis of the raw data given in Table 2 in Crawford (2005:43). 4. The team compiling the CEECE includes: Terttu Nevalainen (leader), Samuli Kaislaniemi, Mikko Laitinen, Minna Nevala, Arja Nurmi, Minna Palander-Collin, Helena Raumolin-Brunberg, Anni Sairio, and Tuuli Tahko. For more information, see Raumolin-Brunberg and Nevalainen (2007). 5. In all, controversial cases of the kind illustrated in (6)–(8) amounted to fewer than a dozen instances in the whole corpus. 6. One should bear in mind, however, that the principles of calculating the frequencies are different in these studies. Mine are based on constructing the linguistic variable of ‘alternative ways of saying the same thing’, whereas Martínez-Insua and Pérez-Guerra (2006) calculate the relative frequency of nonagreement with respect to all existentials, singular and plural. 7. Discussing the grammaticalisation of sort of , kind of , and type of , Denison (2006:296–7) makes a useful distinction between binomial constructions (three kinds of cheese) and postdeteminers (those sort of people). 8. Interesting findings on ‘mixed concord’, mismatch between syntactic and semantic selection criteria, are reported for collective nouns which display singular agreement with the verb, but plural agreement with personal pronouns (e.g. Biber et al. 1999:192; Hundt 2006:210–211).
< previous page
page_100
next page >
< previous page
page_101
next page >
Page 101 REFERENCES Biber, D., S. Johansson, G. Leech, S. Conrad, and E. Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Breivik, L.E. 1990. Existential There: A Synchronic and Diachronic Study (Studia Anglistica Norvegica 2). 2nd edition. Oslo: Novus Press. ———. 1991. On the typological status of Old English. In Historical English Syntax , edited by D. Kastovsky, 31–550 Berlin and New York: Mouton de Gruyter. (The ) British National Corpus , Version 2 (BNC World) 2001, Distributed by Oxford University Computing Services on behalf of the BNC Consortium, URL: http://www.natcorp.ox.ac.uk/ Chambers, J.K. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-linguistic Perspective , edited by B. Kortmann, 128–445 Berlin and New York: Mouton de Gruyter. Corbett, G.G. 2000. Number . Cambridge: Cambridge University Press. (The ) Corpus of Early English Correspondence . 1998. T. Nevalainen, H. Raumolin-Brunberg, J. Keränen, M. Nevala, A. Nurmi, and M. Palander-Collin (comps). Helsinki: University of Helsinki. Crawford, W. 2005. Verb agreement and disagreement: A corpus investigation of concord variation in existential there + be constructions. Journal of English Linguistics 33(1): 35–661 Denison, D. 2006. Category change and gradience in the determiner system. In The Handbook of the History of English, edited by A. van Kemenade and B. Los, 279–304. Malden, MA and Oxford, UK: Blackwell Publishing. Finegan, E. 2001. Usage. In The Cambridge History of the English Language, Vol. 6, English in North America, edited by J. Algeo, 358–4421 Cambridge: Cambridge University Press. Fischer, O. 1992. Syntax. In The Cambridge History of the English Language, Vol. 2, 1066–1147 , edited by N. Blake, 207–4408 Cambridge: Cambridge University Press. Hay, J., and D. Schreier. 2004. Reversing the trajectory of language change: Subject–vverb agreement with be in New Zealand English. Language Variation and Change 16(3): 209–335 Huddleston, R., and G.K. Pullum. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Hundt, M. 2006. The committee has/have decided …: On concord patterns with collective nouns in innerand outer-circle varieties of English. Journal of English Linguistics 34(3): 206–332 Kortmann, B., and E. Schneider. 2006. Varieties of English. Berlin: Mouton de Gruyter. Online material available at: http://www.varieties.mouton-content.com/ (accessed 28 December 2006). Laitinen, M. Forthcoming. You was/were variation and English normative grammars in the eighteenth century. McIntosh, C. 1998. The Evolution of English Prose, 1700–11800: Style Politeness, and Print Culture . Cambridge, U.K. and New York: Cambridge University Press. Martínez-Insua, A.E., and J. Pérez-Guerra. 2006. On there-sentences in the recent history of English. In ‘ These Things I Write vnto Thee …’ Essays in Honour of Bjørg Bækken, edited by L.E. Breivik, S. Halverson, and K.E. Haugland, 189–2211 Oslo: Novus Press. Matthews, P.H. 1987. Syntax . Cambridge: Cambridge University Press. Michael, I. 1970. English Grammatical Categories and the Tradition to 1800. Cambridge: Cambridge University Press. Morris, R., ed. 1873. Cursor Mundi (EETS, O.S. 57). London: Trübner.
< previous page
page_101
next page >
< previous page
page_102
next page >
Page 102 Murray. L. 1795. English Grammar, Adapted to the Different Classes of Learners. York: Wilson, Spence & Mawman. Nevalainen, T. 2003. English. In Germanic Standardizations: Past to Present (Impact: Studies in Language and Society 18), edited by A. Deumert and W. Vandenbussche, 127–556 Amsterdam and Philadelphia: John Benjamins. ———. 2006a. An Introduction to Early Modern English. Edinburgh: Edinburgh University Press. ———. 2006b. Vernacular universals? The case of plural was in Early Modern English. In Types of Variation: Diachronic, Dialectal and Typological Interfaces (SLCS 76), edited by T. Nevalainen, J. Klemola, and M. Laitinen, 351–669 Amsterdam and Philadelphia: John Benjamins. Nevalainen, T., and H. Raumolin-Brunberg. 2003. Historical Sociolinguistics: Language Change in Tudor and Stuart England (Longman Linguistics Library). London: Pearson Education. Postlethwaite, R. 1792. The Grammatical Art Improved: In Which the Errors of Grammarians and Lexicographers are Exposed. London: J. Parsons. Quirk, R., S. Greenbaum, G. Leech, and J. Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman. Raumolin-Brunberg, H., and T. Nevalainen. 2007. Historical sociolinguistics: The Corpus of Early English Correspondence . In Creating and Digitizing Language Corpora. Volume 2: Diachronic Databases , edited by J. Beal, K. Corrigan, and H. Moisl, 148–71. Houndsmills: Palgrave-Macmillan. Sundby, B., A.K. Bjørge, and K.E. Haugland. 1991. A Dictionary of English Normative Grammar 1700–1180 . Amsterdam and Philadelphia: John Benjamins. Tagliamonte, S. 1998. Was/were variation across the generations: View from the city of York. Language Variation and Change 10: 153–91. ———. 2002. Comparative sociolinguistics. In The Handbook of Language Variation and Change, edited by J.K. Chambers, P. Trudgill, and N. Schilling-Estes, 729–663 Malden, MA and Oxford, UK: Blackwell Publishing. Tieken-Boon van Ostade, I. 2000. Robert Dodsley and the genesis of Lowth’s Short Introduction to English Grammar. Historiographica Linguistica 27(1): 21–336 ———. 2002. You was in eighteenth-century normative grammar. In Of Dyuersitie & Chaunge of Langage: Essays Presented to Manfred Görlach on the Occasion of his 65th Birthday, edited by K. Lenz and R. Möhlig, 88–1102 Heidelberg: C. Winter. Trudgill, P. 1999. The Dialects of England. 2nd edition. Oxford: Blackwell.
< previous page
page_102
next page >
< previous page
page_103
next page >
Page 103 5 There Was Universals; Then There Weren’t A Comparative Sociolinguistic Perspective on ‘Default Singulars’ Sali A. Tagliamonte 1. INTRODUCTION The variable use of default singulars, particularly for present and past tense be as in example (1), is a frequent, widely attested, and extensively studied feature of English. (1) a. There are secret rooms … there’ s lots of places. (Canada/N/f) b. They was alright but they were nae great big herring. (Northern Ireland/PVG/007) According to Chambers’ theory of ‘Vernacular Roots’, certain variables appear to be primitives of vernacular dialects in the sense that they recur ubiquitously all over the world. Default singulars are perhaps the prototypical exemplar of these roots because they are mentioned in virtually every discussion (Chambers 2000, 2001, 2003, 2004). Thus, they present the quintessential site for exploring the nature of this theory. The fundamental question is what explains the widespread, ubiquitous presence of this variable feature. Is it a particularly robust example of the inherent variation typical of all languages? Is it the product of regularization? Is it a typological breaking point? Is it “regression to primitive linguistic instincts” (Chambers 2004:139)? Where did it come from? Why does it exist? At least part of the answer resides in the more compelling question: How did varieties of English from locations all over the world get them? One hypothesis is that default singulars are the result of general ‘regularization’ processes in language (e.g. Fries 1940). This explanation is based on the idea that the verb to be is gradually becoming more like the other (regular) verbs in English in having the same form throughout the verbal paradigm, rather than the more complex distinction pitting and third person singular in opposition to second person singular and first, second, and third person plural.
< previous page
page_103
next page >
< previous page
page_104
next page >
Page 104 However, if blanket regularization processes were the only explanation, we might expect that the variation reported to be rampant in the sixteenth to eighteenth centuries (The Oxford English Dictionary 1989) would have gone to completion by the twentieth century and that the variation would have disappeared. Nothing could be further from the truth. In the last couple of decades, there has been a plethora of studies on colloquial default singulars, from the most remote islands on the planet (e.g. Britain and Sudbury 2002; Schreier 2002) to the most populated, from older varieties of English (Nevalainen 2006) to newer ones (e.g. Cheshire 1982; Christian, Wolfram, and Dube 1988; Feagin 1979; Schilling-Estes and Wolfram 1994). Another way to approach this question is to put it in context with typology (e.g. Kortmann 2002). There are two perspectives—typological change and markedness. Default singulars fit this scenario. First, as English gradually drifts from synthetic to an analytic language, there is an overarching tendency to strip away inflections. Part of this process can involve regularization and extension. Second, the third singular - s is typologically marked, which is yet another reason that it is vulnerable to change. Another approach is to invoke a population genetic explanation whereby a variety’s characteristics can be traced back to the founders who happened to be people speaking varieties which contained the same feature (cf. Weinreich, Labov, and Herzog 1968). As Mufwene (1996) has argued, founders exert an indelible mark on the future of the varieties they establish, although ethnographic and structural forces later converge in shaping and reshaping the new variety according to the local ecology. Chambers argues against a standard diffusionist scenario given the all-encompassing geographic spread of the default singular. Further, he suggests that it does not explain how the same feature ended up in pidgins and creoles, child language, and interlanguage (Chambers 2000, 2001). The theory of Vernacular Roots provides an answer by suggesting that the occurrence of default singulars is not the result of diffusion or even simple regularization, but instead represents the more general tendency in all nonstandard varieties of English to gravitate toward more primitive (i.e. not learned) linguistic patterns. In other words, default singulars are not simply a universal of the language per se, but of the vernacular language by which Chambers means the variety found in working-class and rural vernaculars. In addition, the vernacular is perpetually in flux and always in conflict with prescriptive norms. Chambers (2001:130) notes that “[s]ocially, the vernacular universals appear to fall into well-defined patterns in the acrolect-basilect hierarchy.” Therefore, the factors involved must necessarily be linguistic as well as social. The next step is to contextualize all this with the historical record. 2. HISTORY Although contemporary Standard English requires strict subject–verb agreement according to person, the historical record confirms that this has, in fact,
< previous page
page_104
next page >
< previous page
page_105
next page >
Page 105 always been variable (Curme 1977; Forsström 1948; Jespersen 1909/1949; Pyles 1964; Visser 1970). Indeed, as you can see from example (2), default singulars have existed in every century of the language. (2)a.Thyrtty knyghtes … Forsoth was in that companye. [c. 1300 Rich. Coeur de Lion] b.Whan thay came togeders, thay was . . occupyyd with their own maters. [c. 1460 - Fortescue Abs. & Lim. Mon. 108] c. This womans wordes was well harde. [c. 1523–5 Ld Berners, Froiss III 78] d.There was many Dukes, Erles and Baron. [c. 1533 Ld. Berners Huon 2 22.] e.I suppose you was in a dream. [c. 1684 Bunyan Pilg. II 76] f. What was you reading when I came in? [c. 1749 Fielding Tom Jones VI v.] g.You was to come to him at six o’clock. They was here. [c. 1837 Dickens Pickwick Papers xxxiii] Reports on usage from the sixteenth and seventeenth centuries confirm that this type of variability was not just anomalous, but is described as ‘frequent’ and ‘regular’ (The Oxford English Dictionary 1989; Traugott 1972; Visser 1970). 3. DATA Indeed, over the last 15 years, I have engaged in fieldwork and data-collection projects spanning North America, the Caribbean, and the United Kingdom, and I have found hundreds of default singulars. In the analyses that follow, I revisit each of the data sets and tackle some of the questions that Chambers raised in his 2001 article. These are listed below. a. How does the process work cross-linguistically? Or cross-dialectally? b. What constraints are imposed on the process hierarchically? c. What constraints govern the highly constrained variety known as the standard? d. What principle guides it? Such questions are all part of the more general goal shared by many linguists to figure out what is universal and what is not. Chambers words it in this way: “to discern what is primitive as opposed to what is learned —what belongs to the bioprogram and what is an experiential excrescence on it” (Chambers 2001:9). Another formulation is “to differentiate processes that are general/universal from those that are particular” (Horvath and Horvath 2003:144).
< previous page
page_105
next page >
< previous page
page_106
next page >
Page 106
Figure 5.1 Geolinguistic scale—Canada. To focus my discussion and to concentrate on the area of grammar where this phenomenon is most robust, most frequent, and unencumbered by independent phenomena (such as contraction and the possible formulaic nature of presentational there’s ), I focus on the past-tense forms of the verb to be only. All of the data come from corpora which comprise sociolinguistic interviews conducted at the community level among speakers born and raised in the locales in question. In North America, the data come from Nova Scotia, the eastern seaboard of Canada (Poplack and Tagliamonte 1991, 2001), and Toronto, the largest urban centre in Canada (Tagliamonte 2003–2006). In the Caribbean, the data come from Samaná in the Dominican Republic (Tagliamonte and Smith 1998, 2000). In the United Kingdom, the data come from York, a small city in the northeast (Tagliamonte 1998), as well as innumerable data sets from fishing villages on the shores of Scotland and Northern Ireland, to the farming hamlets of the Southwest (Tagliamonte 2000– 2001, 2001–2003). The communities represented by these corpora may be differentiated on a number of geographic scales and linguistic dimensions as you can see in Figures 5.1, 5.2, and 5.3. The smallest geolinguistic continuum involves each
Figure 5.2 Geolinguistic Scale—Samaná.
< previous page
page_106
next page >
< previous page
page_107
next page >
page_107
next page >
Page 107
Figure 5.3 Geolinguistic scale—UK.
< previous page
< previous page
page_108
next page >
Page 108
Figure 5.4 Local differences—Canada. community, the county or province in which the community is situated, the region or country, and finally the continent. Further, as readily observable in Figures 5.4 and 5.5, even within the smallest geographic locality, there are local differences based on ethnicity, urbanity, history, and relative locality (i.e. north vs. south). In Canada, for example, the data come from two different provinces—Ontario and Nova Scotia. Toronto is the only urban context. However, both Toronto and Guysborough Village are ethnically European, whereas North Preston and
Figure 5.5 Local differences—UK.
< previous page
page_108
next page >
< previous page
page_109
next page >
Page 109 Guysborough Enclave are African. At the same time, all the communities in Nova Scotia are alike in that they are all rural. In the UK, for example, the data come from three different countries—Scotland, England, and Northern Ireland. In Scotland, there are two communities—one in Banffshire and the other in Ayrshire which distinguish northern Scotland from Lowland Scotland. In northern English, there are three communities—one is urban (York), one is a small town (Wheatley Hill), and one is rural (Maryport). On yet another level, the individuals who are using these default singulars are all basically the same on two important dimensions. First, most of the people are the oldest generation in each of the communities; and second, the data were collected using the same methods in the same 10-year time frame (1991–2003). In only two cases am I able to explore a cross-generational perspective which reveals the critical importance of speaker age and the scale of ‘era’ (or ‘generation in time’). Finally, the dialects represented in these communities may also be differentiated on an important linguistic dimension—the presence or absence of regularization to were . There are reports of were in contexts of standard was in widely separated areas of England dating back to Ellis (1869–1889). Due to the fact that the studies of variation employed here also tabulated variation between was and were in standard was contexts, the extent of were regularization in each community may be juxtaposed with the results for the default singular cases. I return to this issue on page 116. In sum, these materials provide an unprecedented opportunity for consistent cross-variety research under varying external conditions and hence for an in-depth consideration of the universal status of default singulars—linguistically and sociolinguistically. As Chambers notes: As sociolinguistics becomes less restricted to local events, it becomes comparative and, as the comparative aspect gains weight, cross-linguistic generalizations not only become possible but inevitable. (Chambers 2004:128) My goal is to focus in on default singulars in particular as a test case, with the overall goal of trying to differentiate processes that are general or universal from those that are particular. I also offer an innovative methodological approach to these materials. This technique was spear-headed by Horvath and Horvath (2003:144), who were interested in making the link between linguistics and geography with a similar goal— that is, to investigate the potential universality—or not—of conditioning constraints. The method focuses on order, contrast, and scale. In this method, the notion of constraint hierarchy is particularly relevant. Evidence is drawn from the following:
< previous page
page_109
next page >
< previous page
page_110
next page >
Page 110 a) the consistency of the order of the constraints across analyses b) the strength of the favouring and disfavouring constraints c) the degree of contrast between the factors that promote and inhibit the variants. (Horvath and Horvath 2003:144) In addition, a critical dimension is the effect of ‘scale’. Scale can be the community, the region, the nation, or the supranational level. Attending to ‘scale’ will help me to distinguish those patterns of variation in the condition of a linguistic variable that are universal and those that are local (Horvath and Horvath 2003:163). Evidence for a universal constraint will be: consistent order across scales, consistent strength across scales, and scale independence. In contrast, evidence for a local constraint will be: varying order across scales, varying strength across scales, and scale dependence. 4. DISTRIBUTIONS, OVERALL Figure 5.6 shows an overall distribution of default singulars across communities. This broad view reveals tremendous variability. There is clear support for the basic pattern—what Chambers (2000) refers to as ‘Vernacular Pattern I’—namely, that “ was occurs variably for standard were .” In every locality, there are plenty of default singulars.
Figure 5.6 Overall distribution of was by community.
< previous page
page_110
next page >
< previous page
page_111
next page >
Page 111 However, notice that some locales exhibit a lot more default singulars than others. Samaná, for example, has near categorical use of was . In an earlier paper, I argued that the explanation for this was that these speakers learned their English in a context far removed both geographically and psychologically from the influence of an imposing standard, essentially a result entirely consistent with ‘vernacular roots’ (Tagliamonte and Smith 1998). The fact that York—the only urban community in this cohort—has the least amount of default singular usage provides corroborating evidence. However, what explains the wide-ranging frequencies elsewhere? Overall distributions such as these may mask underlying patterns. The next step is to provide a consistent distributional analysis of the constraints underlying the production of default singulars. ‘Constraint hierarchy’ is the conditioned variability that underlies overall frequencies. More technically, the constraint ranking or constraint hierarchy as it is often called is the order (from more to less) of factors that exert an effect on the occurrence of one variant over another (Poplack and Tagliamonte 2001:93; Tagliamonte 2006:237). 5. GRAMMATICAL PERSON Grammatical person is the foremost grammatical constraint reported to constrain the occurrence of default singulars. Chambers (2004:141) refers to it as “the remarkable regular hierarchy of subject-types”, which are thought to provide “a kind of evolutionary scale” for default singulars …, “along which concord replaces basilectal nonconcord” (Chambers 2004:141). First, there is the “inherent numberlessness of the existential pronoun there ” which has been found to promote default singulars the most. Britain and Sudbury (2002:19– 20) refer to this pattern as ‘the existential constraint’. According to this constraint, “ was is most frequent after there ” (ibid.; Chambers 2004:132). Where are existential contexts positioned in the grammatical person hierarchy? Figure 5.7 shows a distributional analysis of was by the entire grammatical person paradigm across communities. The far right-hand side of the figure shows the use of was in plural existential constructions. Note that the frequency of the default singular is quite high in most communities. Figure 5.8 separates out the existential contexts and views the frequency of default singular usage with geographic scale now positioned on the x-axis. The use of there was is such a pervasive pattern that it even occurs in standard varieties. For example, the English spoken in Toronto is Canadian English, a variety well known to be a relatively monolithic ‘homogenized’ dialect (Chambers 1995), yet there is a healthy proportion of default singulars. Culleybacky in Northern Ireland is a tiny fishing village, yet it has even less use of default singulars than urban Toronto.
< previous page
page_111
next page >
< previous page
page_112
next page >
Page 112
Figure 5.7 Distribution of variable was by grammatical person across communities. In sum, although default singulars in existential subjects are widespread and generally robust, in some places they are very frequent—such as Wheatley Hill and Guysborough Village. In other contexts, they are very infrequent—such as in Culleybacky and in Toronto. So far, these discrepancies are not easily explained. The two communities where default singulars are frequent are in two different countries, in two different types of communities.
Figure 5.8 Distribution of was in existential contexts across communities.
< previous page
page_112
next page >
< previous page
page_113
next page >
Page 113 The same is true of the places where default singulars are not frequent. Moreover, in one of the most rural, isolated places of all—Samaná—default singulars in plural existential contexts do not exist because existential there constructions do not exist. The next most frequently attested grammatical category which influences the use of default singular use is second-person singular you. This context is often singled out as having a high degree of nonstandard was . The Oxford English Dictionary (1989), for example, reports that it was used “almost universally” in the sixteenth to eighteenth centuries. Chambers notes that the favoured use of default singulars with secondperson singular you may be due to the syncreticism between singular (you) and plural (you). However, in more recent research, second-person you has demonstrated what Chambers (2004:136) refers to as “unruly behaviour”. Figure 5.9 shows the distribution of was by we , you, and they . Some localities have frequent use of was in this context—North Preston, Guysborough Enclave, Samaná, Buckie, and Tiverton. However, in many other locales, this use is rare. The subjects— you, we , and they — are often reported to be ordered regularly so that you has the most of nonstandard was , then we , and the third-person plural pronoun they has the least. However, according to the theory of Vernacular Roots, such an overarching hierarchy is not predicted because, as Chambers (2004:141) argues, these categories “have no semantic or grammatical feature that distinguishes among them”. This means that these grammatical subjects should be expected to be ordered differently by geographic scale. Comparing you with we in Figure 5.9
Figure 5.9 Distribution of was by you, we , and they .
< previous page
page_113
next page >
< previous page
page_114
next page >
Page 114 reveals that you has higher rates of default singulars in North Preston, possibly in Guysborough Enclave, and in Buckie, Maryport, and Tiverton, but the reverse is visible in Samaná and Cumnock. Little further can be gleaned from the other locales because you is an infrequent category. Finally, they has fewer default singulars than we in the African Nova Scotian villages and in Wincanton, but more in Samaná and Tiverton. Elsewhere, it is rare. Once again, the data match the theory—there is no regular relationship such that you ranks above we . Nor is there a regular relationship such that we ranks above they . Nor is there any straightforward clarity or predictability at “the poles of the dialect continuum, the acrolect and basilect” (Chambers 2004:133). Although Samaná is perhaps not the ‘Galapagos’ that Tristan da Cunha is, it may well represent the extreme point on the dialect continuum here. Arguably, so might the small fishing village of Portavogie. However, these two varieties could not be more different. I return to this observation later. 6. THE NORTHERN SUBJECT RULE Perhaps the most famous constraint involving default singulars is the so-called “Northern Subject Rule”, a feature described by Ihalainen (1994) and Klemola (1996) based on earlier dialectological work described by Murray (1873). This refers to the effect of grammatical person as categorized into third-person noun phrases as opposed to personal pronouns as well as the proximity of subject to verb (e.g. Murray 1873). I only consider the contrast between NP and pronoun here. An interpretation of the Northern Subject Rule in variable terms would predict more use of default singulars with nouns than with pronouns, as in example (3) from Ireland. This prediction is consistent with Filppula’s (1999) work on Irish English. (3) And these boys is earning a mint. Okay they’re off in the summer. (PVG/1) In constraint terms, default singulars can be expected to be more frequent after a noun phrase than after pronouns. Figure 5.10 shows the distribution of was according to nouns and pronouns across communities. Most of the varieties exhibit this constraint or the categorical or near categorical use of singular forms of the verb with plural nouns, as you can see in Figure 5.10. Only one variety is an outlier: Samaná. 7. THE EXISTENTIAL CONSTRAINT Now let us come back to the existential constraint, which is claimed to be one of the strongest effects on the use of default singulars. Due to the
< previous page
page_114
next page >
< previous page
page_115
next page >
Page 115
Figure 5.10 Distribution of was in NP plural and they . overwhelming use of default singulars here and the lower usages in most locales in the other contexts, Figure 5.11 plots the use of was in existential constructions versus everything else. This is perhaps the best example of a scale-independent constraint yet. Every single locale shows this contrast with one exception only: Culleybacky.
Figure 5.11 Existential vs other.
< previous page
page_115
next page >
< previous page
page_116
next page >
Page 116 Before turning to the next phase of the analysis, the unique status of Culleybacky must be addressed. As mentioned earlier, some British dialects not only exhibit default singular with was , but also the opposite: default plurals with were , as in example (4). (4) a. For it were domestic trouble that made me blow the head of that engine. (CLB/2) b. She were a great worker, mi mother. (CLB/2) c. There were no doors locked. There weren’t even a lock on mi Granny’s door. (CLB/2) It may be that the differences in the use of default singulars across varieties are the result of mixing different regional dialects. A separate test of the use of were in contexts of standard was in the British and Northern Ireland communities in shown in Figure 5.12. Figure 5.12 reveals that all the dialects have a certain amount of were regularization; however, of all the dialects, Cullybackey has the most. The high frequency of were (as opposed to was ) in plural existential constructions is due to the fact that speakers are regularizing towards a different form for past-tense be, as in the examples in (4). To sum up, the idea that there is a specific universal (vernacular) hierarchy according to grammatical person does not hold out. However, there are fairly consistent scale-independent contrasts within the grammatical person hierarchy. The clearest one is between existential constructions and everything else. There also seems to be a fairly regular contrast between NPs and pronouns. The next step is to perform an analysis of contrast following the Horvath and Horvath (2003) method.
Figure 5.12 Distribution of were regularization across communities.
< previous page
page_116
next page >
< previous page
page_117
next page >
Page 117
Figure 5.13 Contrast analysis (Horvarth and Horvarth 2003). Each contrast analysis plots the degree of contrast between the promoting category (like noun phrase) and the inhibiting category (like pronoun), as in Figure 5.13. The degree of contrast is measured by the difference in their probability weights when the data are subjected to a multiple regression analysis (Horvath and Horvath 2003:152). The way to interpret the results is basically this: if the value on the graph is close to the x-axis, the contrast (and hence the constraint) is weak. The higher they are, the stronger the contrast (and hence the constraint) is. Comparison across the geographic scale reveals any variation in the consistency of the strength of the contrast. Figure 5.14 shows a contrast analysis between you or we and they . Overall, the strength is moderate at best. Some points are as high as .4, but most are well below, indicating a moderate to weak effect. Figure 5.15 shows a contrast analysis of noun phrases and pronouns, the Northern Subject Rule. Here something completely different is observed. The degree of contrast is great in some locales and weak in others. Many varieties are positioned at .6 or above, indicating a very strong effect. There are two low at .4—Maryport and Wheatley Hill—indicating a strong effect. Finally, there are two where the effect is minimal —the two African Nova Scotian villages. Once again, Samaná stands apart as the only community where the quintessential vernacular universal pattern holds up—no difference between nouns and pronouns. Some varieties simply do not have any nonstandard was with third-person plural pronouns—like Buckie. In this case, the effect is categorical. This indicates considerable scale dependence.
< previous page
page_117
next page >
< previous page
page_118
next page >
Page 118
Figure 5.14 Contrast analysis for you, we , and they .
Figure 5.15 Contrast analysis for noun phrases versus pronouns.
< previous page
page_118
next page >
< previous page
page_119
next page >
Page 119
Figure 5.16 Contrast analysis for existential constructions. Figure 5.16 shows a contrast analysis between existentials and everything else. This shows yet another pattern. There are no extremes. Instead, there is a fairly regular scale-independent effect. In most locales, this contrast is strong or very strong, whereas in only two locales, the effect is weak. In sum, the analysis of contrast adds an important dimension to the building picture of default singulars across communities. Although the constraint ranking shows the order of the constraints, the contrast analysis enables a view of the strength of the effect and an assessment of its scale independence. 8. NEGATION The next most important constraint reported for default singulars is the effect of negation, as in example (5). The extent of variability is notable. (5) a. They wasn’t close family, you know, they was far family. (SAM/D) b. They was out here, wasn’t they? (GYE/k) c. They weren’t seats. There was seats at t’front, but we used to go up t’back. (MPT/P) d. They was nae supposed to get onything, but they were very good to them here. (PVG/8)
< previous page
page_119
next page >
< previous page
page_120
next page >
Page 120 Vernacular Pattern I, which is the basic vernacular system identified by Chambers, is when the was variant occurs regardless of type of sentence, as in examples (5a–5b). This is thought to be the simpler and more basic pattern. A second pattern, labeled Vernacular Pattern II, Chambers predicts is derivative. In this case, weren’t occurs in negatives, but the default singular was occurs in affirmatives (Chambers 2004:131), such as in example (5c). This pattern turns up in North American dialects (North Carolina), in southwest England (Reading) (Cheshire, Edwards, and Whittle 1989), in the Fens in southeast England (Britain and Sudbury 2002) and elsewhere in Britain (Anderwald 2002). However, a third pattern has also been reported (Tagliamonte and Smith 2000). This is a tendency towards wasn’t for negatives and were for affirmatives, as in example (5d). Figure 5.17 shows a distributional analysis of negative vs.versus affirmative contexts by community. Vernacular Pattern I is identified by an equal proportion of default singulars for negative and affirmative contexts. Notice that this exists in three locales. Once again, Samaná is one of them, but this is also true of Guysborough Village, one of the small towns in Nova Scotia, as well as Wincanton in Somerset. What all of these places share is the advanced age of the speakers, isolation, rurality, and lack of mobility. However, they also differ dramatically on other criteria: in the latter villages, the residents either are British or have British ancestry, whereas in Samaná, the residents are African. Yet note the comparable results from Toronto. Although these are existential contexts only, it is interesting to see that Vernacular Pattern I prevails.
Figure 5.17 Distribution of was by negation.
< previous page
page_120
next page >
< previous page
page_121
next page >
Page 121 Vernacular Pattern II is identified by a contrast between a higher proportion of default singulars for affirmatives than negatives. Notice that this ranking of the negative constraint turns up in four communities —Tiverton, Devon in the southwest, York, possibly Wheatley Hill in the northeast, and Culleybacky in County Antrim in Northern Ireland. There is also a case for Vernacular Pattern III. In yet another four communities, the constraint is ordered in the other direction—more default singulars in negative contexts over affirmative contexts. This constraint hierarchy is found in North Preston, Guysborough Enclave, Portavogie, Cumnock, and Maryport. What do all these localities have in common? They are rural, nonurban, and peripheral. The fact that at least the UK locales are all in the north or in Northern Ireland suggests ‘a northern Englishes’ effect. However, recall that the same contrast is also present in both of the African Nova Scotian locales in Canada. Figure 5.18 shows a contrast analysis of the negation effect. Despite a possible Northern English explanation, the real contenders for the negative effect are the African Nova Scotian communities and, interestingly, York in England. Recall, however, that the effect goes in the opposite direction here. In every other case, the effect is weak, regardless of direction, hovering below or well below .3. On the one hand, the negative effect is scale-dependent—that is, local, by constraint ranking. On the other hand, it exhibits characteristics of a universal by the measure of contrast.
Figure 5.18 Contrast analysis of negative versus affirmative.
< previous page
page_121
next page >
< previous page
page_122
next page >
Page 122 9. SOCIAL EMBEDDING Finally, I consider the social embedding of these constraints and what impact this may have on the patterning of forms. Britain and Sudbury (1999:220–1, Figures 1 and 2) analyzed both past and present existential constructions in New Zealand and the Falkland Islands. They found higher levels of default singular forms among younger speakers and higher levels among males. In Schreier’s (2002) research on Tristan da Cunha, once the existentials were separated from the nonexistentials, two trends were apparent. In non-existential contexts, elderly, traditional speakers had the most default singulars, and younger speakers had fewer. Females led in the adoption of the standard form. Existential contexts, however, had an entirely different pattern by age. There was a leveled system among the old (57/57) and young (14/14), with the middle-aged speakers slightly lower at 86.2% (21/25). This suggests contrasting social embedding for these two grammatical contexts. Given that Tristan da Cunha is the ideal test site for Vernacular Roots, the patterns there can be considered representative. If so, there are at least two sociolinguistic indicators pointing to an explanation of Vernacular Roots: first, among younger speakers, females should lead in the adoption of the standard form. Second, older speakers should have more nonexistential default singulars than younger speakers. Figure 5.19 examines the contrast between male and female speakers across communities. In some communities, males have more default singulars than females, and in others, the opposite effect is visible: males have more than females.
Figure 5.19 Distribution of was by sex.
< previous page
page_122
next page >
< previous page
page_123
next page >
Page 123 Note, for example, the similarity between Buckie and Guysborough Village. Thus, although there is some consistency in this effect, it may pattern either way. A contrast analysis confirms this. Figure 5.20 shows a contrast analysis of the sex effect. The strength of this effect in every locale is moderate at best and for most part weak.1 Recall, however, that these data all come from the oldest generation in each locale. The sex effect reported by Schreier was heavily underlain by other external factors, such as age. Moreover, there appears to be a worldwide diminution of nonexistential default singulars generally. A further step, therefore, is to examine the existential versus nonexistential difference between older and younger speakers in the same community, as in Figure 5.21. The two communities where these data are available are York, England, and Toronto, Canada. First, consider the York materials where the contrast between existential and nonexistential contexts was one of the strongest contrasts (Tagliamonte 1998). Existential default singulars are frequent, whereas nonexistential ones are exceedingly rare. In contrast, Tristan da Cunha has near-categorical default singulars for nonexistential constructions. Moreover, this use is increasing among the younger speakers. Toronto English has precisely the opposite pattern. There, nonexistentials are virtually nonexistent and existential constructions are increasing. Thus, in one of the most remote islands on the planet, the middle-aged speakers veer
Figure 5.20 Contrast analysis of sex effect.
< previous page
page_123
next page >
< previous page
page_124
next page >
Page 124
Figure 5.21 Comparison of existential and nonexistential contexts in York, England, and Tristan da Cunha (Schreier 2000). towards the prestige norms, but in England, middle-aged to younger speakers are tending towards the vernacular. Two urban localities, one in Canada (Toronto) and the other in England (York), provide the possibility for a direct cross-dialect comparison, shown in Figure 5.22. Because Toronto English does not have default singulars anywhere else but in existential contexts, only existential contexts are included. The trend is the same in both locales: the younger the residents are, the more they use default singulars. Moreover, the trend observable in Toronto and York is just as reported for New Zealand and the Falkland Islands (Britain and Sudbury 2002): Default singulars in existentials are increasing in contemporary varieties. To push this comparison even deeper, Figure 5.23 compares male and female speakers in the youngest generation. In York, the young women are spearheading a shift towards default singulars in existentials. In Toronto, however, young people are not distinguished in this way. In sum, when speaker age (era) is brought into the picture, the social embedding of default singulars points to a general increase in the use of default singulars in existential constructions, at least among the current generations of speakers in urban mainstream communities. This appears to contrast with rural, isolated, peripheral communities, where default singulars in
< previous page
page_124
next page >
< previous page
page_125
next page >
Page 125
Figure 5.22 Comparison of default singulars in existential constructions, Toronto, Canada, and York, England. both existential and nonexistential constructions may be a long-time feature of the traditional vernacular. However, it appears that the nonexistential contexts are receding and giving way to prestige norms and standardization processes. Such results are provocative and suggest that default singular usage in existential constructions is a variable linguistic phenomenon on the move. It will thus be informative to track this variation into the coming
Figure 5.23 Comparison of existential constructions in Toronto, Canada, and York, England, by sex.
< previous page
page_125
next page >
< previous page
page_126
next page >
Page 126 generations and across different varieties of English to see how its use plays out in contrastive geographic and sociocultural situations. The methodology I have outlined here provides a consistent way for elucidating these explanations further through ongoing comparative study. The fact that new contact varieties studied by Sand (2005), such as Jamaican English, Indian English, Kenyan English, and Singapore English do not seem to have the same array of default singulars by grammatical person as more traditional varieties of English makes this endeavour all the more exciting. This may provide a partial explanation for the lack of existential constructions in Samaná and perhaps for Samaná’s anomalous behaviour elsewise. Perhaps it patterns along with varieties which have coexisted with one or more indigenous languages (in Samaná’s case, this would be Spanish). This is Samaná’s one unique contrast with the other communities studied here which are all varieties entirely surrounded by local varieties of English. Further, the social embedding of the existential change—if it is a change—by geographic scale warrants further investigation because it seems clear that default singulars in this context have varying prestige value across the varieties of English. Thus, it offers us yet a new window on how vernacular universals play out in practice. 10. DISCUSSION This chapter has considered the results of 39 multivariate analyses across 13 different dialects, an analysis of constraint hierarchy, of strength of the individual constraints (i.e. contrast analysis), and of geographic scale. In so doing, I have discovered that the frequency of default singulars as well as the well-known constraints that underlie them often did not maintain the same relative position with regard to the probability weights, nor were they scale-independent. Like Horvath and Horvath (2003), I have also established that the measure of constraint ranking alone can hide some important patterns unless contrast is considered as well. Indeed, the strength of constraints as well as the strength of the relationship between pairs of oppositions within a given hierarchy uncovers ‘variability within universality’. An analysis that takes into account the interplay among constraint hierarchy, contrast, and scale, as presented here, reveals that, within the constraint hierarchy, pairs of relationships are often strongly ordered, whereas it is the third or fourth category that varies (see also Horvath and Horvath 2003). A case in point is the contrast between ‘unruly’ behaviour of second-person singular you and opposition pairs like noun phrase versus pronoun and existential versus nonexistential. The latter are more scale-independent, and by the criteria set at the beginning, this gives them status as universals. In contrast, for constraints where there is a binary opposition, like negative versus affirmative, the constraint ranking may go one way or the
< previous page
page_126
next page >
< previous page
page_127
next page >
Page 127 other. Thus, for example, some varieties have grammaticalised the category as wasn’t versus were , others as weren’t versus was . At the same time, the contrast analysis shows scale independence. In other words, the constraint exists nearly everywhere, but the ranking order fluctuates. I conclude that the universal patterns are there, but they manifest differently under varying local conditions. Indeed, there seem to be ‘soft spots’ in the grammar that avail themselves to change and interpretation. As I argued in earlier research: minor perturbations in the broader context of linguistic variability […] appear to form part of the available resources that may or may not be taken up by a group of speakers on a community-by-community basis to assert their local vernacular identity (Tagliamonte 1998:187) Horvath and Horvath (2003) come to a similar conclusion: “Features that are less controlling open up the possibility of social intervention”. To conclude and return to the original questions, it seems apparent that universal principles underlie variation in language. This is particularly notable in the use of default singulars, which are pervasive among varieties of English in wide-ranging geographic and sociocultural settings. However, even if a feature is universal, it will be subject to constraints. Such constraints reflect universal tendencies as well, but they will manifest differently—by pattern, contrast, and strength—according to the varying constellation of influences in the local setting. NOTES 1. Unfortunately, this information cannot easily be recuperated from the African Nova Scotian data files. REFERENCES Anderwald, L. 2002. Negation in Non-Standard British English. London and New York: Routledge. Britain, D., and A. Sudbury. 1999. There’s tapestries, there’s photos and there’s penguins: Variation in the verb BE in existential clauses in conversational New Zealand and Falkland Island English. Paper presented at Methods X. Memorial University of Newfoundland, St. John’s, Newfoundland, Canada. ———. 2002. There’s sheep and there’s penguins: Convergence, ‘drift’ and ‘slant’ in New Zealand and Falkland Island English. In Language Change: The Interplay of Internal, External and Extra-linguistic Factors, edited by M.C. Jones and E. Esch, 209–40. Berlin: Mouton de Gruyter. Chambers, J.K. 1995. Sociolinguistic Theory: Linguistic Variation and Its Social Significance . Oxford: Blackwell Publishers.
< previous page
page_127
next page >
< previous page
page_128
next page >
Page 128 ———. 2000. Universal sources of the vernacular. In Special Issue of Sociolinguistica: International Yearbook of European Sociolinguistics , edited by U. Ammon, P.H. Nelde, and K.J. Mattheier, 11–15. Tübingen: Max Niemeyer Verlag. ———. 2001. Vernacular universals. In Proceedings of ICLaVE 1, The First International Conference on Language Variation in Europe , edited by J.M. Fontana, L. McNally, T.M. Turell, and V. Enric, 52–60. Barcelona: Universitate Pompeu Fabra. ———. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Significance . Malden and Oxford: Blackwell Publishers. ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar From a Cross-Linguistic Perspective , edited by B. Kortmann, 127–45. Berlin and New York: Mouton de Gruyter. Cheshire, J. 1982. Variation in an English Dialect: A Sociolinguistic Study. Cambridge: Cambridge University Press. ———, V. Edwards, and P. Whittle. 1989. Urban British dialect grammar: The question of dialect levelling. English World-Wide 10(2): 185–225. Christian, D., W. Wolfram, and N. Dube. 1988. Variation and Change in Geographically Isolated Communities: Appalachian English and Ozark English. Tuscaloosa, AL: American Dialect Society. Curme, G.O. 1977. A Grammar of the English Language. Essex, CT: Verbatim. Ellis, A.J. 1869–1889. On Early English Pronunciation: With Special Reference to Shakespeare and Chaucer. New York: Greenwood. Feagin, C. 1979. Variation and Change in Alabama English: A Sociolinguistic Study of the White Community. Washington, DC: Georgetown University Press. Filppula, M. 1999. The Grammar of Irish-English: Language in Hibernian Style. London: Routledge. Forsström, G. 1948. The Verb ‘to Be’ in Middle English: A Survey of the Forms . Lund: C.W.K. Gleerup. Fries, C.C. 1940. American English Grammar . New York: Appleton, Century, Crofts. Horvath, B.M., and R.J. Horvath. 2003. A closer look at the constraint hierarchy: Order, contrast, and geographical scale. Language Variation and Change 15(2): 143–70. Ihalainen, O. 1994. The dialects of England since 1776. In The Cambridge History of the English Language. Vol. 5: English in Britain and Overseas: Origin and Development , edited by R. Burchfield, 197–270. Cambridge: Cambridge University Press. Jespersen, O. 1909/1949. A Modern English Grammar on Historical Principles. Part VI: Morphology. London: George Allen & Unwin. Klemola, K.J. 1996. Non-standard periphrastic do: A study of variation and change. PhD dissertation, Essex. Kortmann, B. 2002. New prospects for the study of English dialect syntax: Impetus from syntactic theory and language typology. In Syntactic Microvariation , edited by S. Barbiers, L. Cornips, and S. van der Kleij. Amsterdam. http://www.meertens.knaw.nl/books/synmic/ (accessed 5.8.08). Mufwene, S.S. 1996. The founder principle in creole genesis. Diachronica 13(1): 83–134. Murray, J.A. H. 1873. The Dialect of the Southern Counties of Scotland: Its Pronunciation, Grammar and Historical Relations . London: Philological Society. Nevalainen, T. 2006. Vernacular universals? The case of plural was in Early Modern English. In Types of Variation: Diachronic, Dialectal and Typological Interfaces, edited by T. Nevalainen, J. Klemola, and M. Laitinen, 351–69. Amsterdam: John Benjamins. The Oxford English Dictionary . 1989. 2nd edition. Oxford: Oxford University Press.
< previous page
page_128
next page >
< previous page
page_129
next page >
Page 129 Poplack, S., and S.A. Tagliamonte. 1991. African American English in the diaspora: Evidence from old-line Nova Scotians. Language Variation and Change 3(3): 301–39. ———. 2001. African American English in the Diaspora: Tense and Aspect . Malden: Blackwell Publishers. Pyles, T. 1964. The Origins and Development of the English Language. New York: Harcourt, Brace and World. Sand, A. 2005. Angloversals? Shared morpho-syntactic features in contact varieties of English. Albertludwigs-Universität Freiburg: Unpublished Habilitationsschrift Schilling-Estes, N., and W. Wolfram. 1994. Convergent explanation and alternative regularization patterns: Were/weren’t leveling in a vernacular English variety. Language Variation and Change 6(3): 273–302. Schreier, D. 2002. Past be in Tristan da Cunha: The rise and fall of categoricality in language change. American Speech 77(1): 70–99. Tagliamonte, S.A. 1998. Was/were variation across the generations: View from the city of York. Language Variation and Change 10(2): 153–91. ———. 2000–2001. Vernacular Roots: A database of British dialects . Research Grant B/RG/AN 6093/APN11081. Arts and Humanities Research Board of the United Kingdom (AHRB). ———. 2001–2003. Back to the roots: The legacy of British dialects . Research Grant. Economic and Social Research Council of the United Kingdom (ESRC). #R000239097. ———. 2003–2006. Linguistic changes in Canada entering the 21st century. Research Grant. Social Sciences and Humanities Research Council of Canada (SSHRC). #410–2003–0005. http://individual.utoronto.ca/tagliamonte/. ———. 2006. Analysing Sociolinguistic Variation. Cambridge: Cambridge University Press. ———and J. Smith. 1998. Analogical levelling in Samaná English: The case of was and were . Journal of English Linguistics 27(1): 8–26. ———and J. Smith. 2000. Old was ; new ecology: Viewing English through the sociolinguistic filter. In The English History of African American English, edited by S. Poplack, 141–71. Oxford & Malden: Blackwell Publishers. Traugott, E.C. 1972. A History of English Syntax: A Transformational Approach to the History of English Sentence Structures. New York: Holt, Rinehart & Winston. Visser, F.T. 1970. An Historical Syntax of the English Language. Leiden: E. J. Brill. Weinreich, U., W. Labov, and M. Herzog. 1968. Empirical foundations for a theory of language change. In Directions for Historical Linguistics, edited by W. P. Lehmann and Y. Malkiel, 95–188. Austin: University of Texas Press.
< previous page
page_129
next page >
< previous page
page_130
next page >
page_130
next page >
Page 130 This page intentionally left blank.
< previous page
< previous page
page_131
next page >
Page 131 III Universals and Contact in Varieties of English
< previous page
page_131
next page >
< previous page
page_132
next page >
page_132
next page >
Page 132 This page intentionally left blank.
< previous page
< previous page
page_133
next page >
Page 133 6 Irish Daughters of Northern British Relatives Internal and External Constraints on the System of Relativisation in South Armagh English (SArE)1 Karen P. Corrigan 1. INTRODUCTION 1.1. Focus and Background to the Research The focus here is on variability in the sociolinguistic conditioning of relative marking strategies used by speakers of a contact vernacular spoken in South Armagh, Northern Ireland. Variation within this community is compared with findings from recent research on Englishes spoken in northern regions of the British Isles by Beal and Corrigan (2005), Herrmann (2005) and Tagliamonte, Smith, and Lawrence (2005). This is an important objective because South Armagh was originally settled by speakers of relic forms of these dialects. As such, although one would not necessarily predict congruity with respect to social constraints across the varieties, one might expect certain generalities regarding the linguistic constraints curtailing the variation. This is because, in general terms, the use of a relativisation strategy of any kind is an attempt to minimise potential ambiguity between root and embedded clauses (Temperley 2003). Natural languages have variously solved this problem, although I argue that there is a universality in the types of strategy available and in the ways in which specific strategies are internally constrained. In addition, due attention is paid to the dynamics of relativisation in Ulster Irish as described by Adger and Ramchand (2006), Duffield (1996), Goodluck, Guilfoyle, and Harrington (2006), McCloskey (1985, 1990, 2000, 2001), Neilson (1808/1990), and Ó Siadhail (1984, 1989). This is because it acted as the substrate during the initial bilingual phase arising from the linguistic contact induced in County Armagh by the Jacobean Plantation. Irish may, therefore, have imposed restructuring effects on the Northern British superstrates from which the modern South Armagh English (SArE) vernacular descends. Such an outcome is not without precedent because Mesthrie (1987, 1991, 1992:75), for instance, has previously claimed that relative strategies in South African Indian English (SAIE) are inherited from
< previous page
page_133
next page >
< previous page
page_134
next page >
Page 134 Indic and Dravidian substrates. Given that certain other Irish-English (IrE) constructions have been demonstrated to result from this type of language transfer (Filppula 1999 inter alia ), this chapter outlines what the Ulster Irish possibilities are and then assesses the degree to which these differ from SArE and the Northern British superstrates. Harris (1993), Finlay (1994), Geisler (2002), and Policansky (1982) give preliminary accounts of relativisation in Northern Irish-English (NIrE) from a largely synchronic, sociolinguistic perspective. Herrmann (2005) and Tagliamonte et al. (2005) provide fuller treatments of data from the Northern Ireland Transcribed Corpus of Speech (Kirk 1992) and the Roots Corpus (Tagliamonte 2001–2003), respectively, and they also add a diachronic dimension. Crucially, their research recognises the importance of attending to internal constraints and, as such, shares the orientation of attempts to use the generative framework to account for IrE ‘zero relativisation’ (see Corrigan 1997a, 1997b; Doherty 1993; Henry 1995). The research reported here differs from all of these because the relative system of SArE in its entirety is to be reviewed, and consideration is given to: (1) the origin and diachrony of particular strategies within a universalist framework, and (2) synchronic language-internal and language-external variation and diachronic change. 1.2 The Community and Data Sets The community under study is situated in the southernmost reaches of County Armagh, Northern Ireland, bordering the Republic of Ireland. Its geographical setting is both rural and insular; like many of these communities globally, it is prone to cultural and linguistic conservatism of various kinds. South Armagh Irish (SArI), for instance, was preserved in this region into the early twentieth century, and certain speakers would have been natively bilingual until just prior to the Second World War. There is evidence to suggest that, from the later twentieth century onwards, this community has become more open to standardising influences both socioculturally and linguistically. As such, this investigation also seeks to ascertain whether the system of relativisation used by speakers of SArE reflects these changes to the peripherality of the region.2 To test this hypothesis, the research findings presented here are based on two distinctive corpora collected almost 60 years apart. The first, referred to subsequently as ‘The Murphy Corpus’, is a 52,000-word written corpus collected by a local folklorist between 1942 and 1974, which was digitised by me in 1995. The second data set that I refer to is the Questionnaire Survey, comprising a set of native speaker judgments collected in 2001–2003 as questionnaire responses from 44 local informants. These included 22 male and female school children whose average age was 15 years and the same number of male/female adults with an average age of 44.3 Respondents whose personal network ties were of the dense/multiplex type were targeted in both collections. This is crucial because it is a good
< previous page
page_134
next page >
< previous page
page_135
next page >
Page 135 indicator that speakers will be loyal to local vernacular norms because access to power and advantage in such rural communities is not normally determined by class status. In this way, the region is rather similar to the Irish and Scottish areas targeted by Herrmann (2005) and Tagliamonte’s 2001–2003 Back to the Roots project, as well as those Celtic English regions described by Johnston (1997:445), in which ‘strict class-tying’ is also absent. 2. A HISTORY OF RELATIVISING STRATEGIES IN THE TARGET/SOURCE LANGUAGES 2.1. The Target Language With a view to reconstructing what the Early Modern English (EModE) target might have been, this section reviews the development of ancillary relative strategies in English restrictive (RR) and nonrestrictive (NRR) relative postmodification in addition to that of the three principal relative marking strategies, viz. WHrelative pronoun (WH- ); THAT- complementizer (TH- ), and omission (Ø ). By convention, RR/NRR are differentiated on the assumption that heads in the former can only be successfully identified by recourse to the given modification, as in (1), whereas the head of an NRR (2) is either unique or has been identified independently in the preceding discourse (Quirk, Greenbaum, Leech, and Svartvik 1985:858).4 (1) the swans which are white are in that part of the lake (2) the swans, which are white, are in that part of the lake (Fabb 1990:57) In spoken language, these types are distinguished by intonation and in writing by punctuation, and it seems clear, as we shall see, that their historical development, contemporary reflexes, and syntactic functions are somewhat different. Nevertheless, it is important to note that there are no such intonational cues in either the early English texts on which my diachronic account is based nor in the SArE Murphy Corpus. Moreover, the potential for ambiguity between RR/NRR modification is increased by the fact that punctuation in both of these sources is not reliable either.5 Following Quirk et al. (1985: §13.10 and §13.14), I assume that there are no ancillary strategies available in the contemporary Standard. The latter are defined here as any option available in the grammar of a language for relative postmodification that does not involve operator movement of the sort argued to account for examples (1) and (2) in traditional generative accounts of relativisation like Haegeman (1991). A number of such strategies have been identified, and I argue that they were extant in earlier stages of English and that their reflexes are still available in contemporary
< previous page
page_135
next page >
< previous page
page_136
next page >
Page 136 vernaculars. A case in point would be the so-called ‘Resumptive Pronoun’ (RP) strategy noted in Chomsky (1982), which is possible in colloquial French (3b) and American English (4b). Examples (3a) and (4a) are derived in the same fashion as example (1), whereas in example (3b), the relative clause is headed by the complementizer que and the IP contains an RP lui which is co-indexed with the relativised DP l’homme. Although example (4b) is not identical, it also contains an RP which, on this occasion, is identified with the WH- element in the SPEC of CP. The assumption here too is that operator movement is not involved because the pronoun object argument occupies its base position indicating that WH- must be base-generated in CP, rather than having moved there from the object slot. (3a) voici [DP l’homme [CP à quii [C’ ø here is the man to whom [IP Marie a parlé ti ]]]] Marie talk.Past Part. ‘here is the man to whom Marie has talked’ (3b) voici [DP l’hommei [CP ø [C’ que here is the man that [IP Marie RP lui i a parlé ]]]] Marie to-him.RP talk.Past Part. ‘here is the man that Marie has talked to’ (Zribi-Hertz 1984:27) (4a) here is [DP the man [CP whomi [C’ ø [ IP John saw ti ]]]] here is the man who John see.PAST ‘here is the man whom John saw’ (4b) here is [DP the man [CP whoi [C’ ø here is the man who [IP John saw RP himi ]]]] John see.PAST him.RP ‘here is the man who John saw him’ (Chomsky 1982:11) Turning now to the types of ancillary relativisation attested in the history of English, it is unsurprising that Old English (OE) would have made greater use of these because it did not inherit any relative markers from Indo-European. Although subordinating relative constructions are by no means rare in OE, most scholars agree that the system of parataxis in this period was more fully developed (Fischer, van Kemenade, Koopman, and van der Wurff, 2000:58ff.). The options are summarised and exemplified in (i)–(iii) below.
< previous page
page_136
next page >
< previous page
page_137
next page >
Page 137 (i) Pronouns/Demonstrative Adverbs with Ambivalent Status As has been noticed in research on child language, pidgins, and creoles, there are certain types of linguistic category (particularly the deictics) that have the potential to perform the function of a relative marker in the absence of such a category in the grammar (Romaine 1984a: 275; 1984b, 1988, 1992). We might conceive of this usage as ‘incipient hypotaxis’ (Stockwell and Minkova 1991:370), and it is affirmed in OE by structures in which the demonstrative adverb seems to have an ambivalent function in the language. (ii) Correlative Particles The use of ‘copy-correlatives’ with identical spelling at the head of both the root and subordinate clauses is another option available for the purpose of minimising ambiguity, which is reminiscent of the RP strategy exemplified in example (3b/4b). The strategy is attested in OE, and it would seem to have persisted into the EModE period (Stockwell and Minkova 1991). (iii) ‘Subordinating and ’ with Relative Function And has had a subordinating function since OE times, and there is some evidence in the Helsinki Corpus , for example, that this construction could also operate as an ancillary relativising strategy up to the EModE period (Filppula 1991), particularly when the DP that is to be relativised is especially complex (Herrmann 2005:21; Mesthrie 1992:93; Romaine 1992:162). There are a further three strategies for relative marking of restrictive and nonrestrictive relatives, namely, ‘ Ø- ’, ‘ WH- ’, and ‘ TH- ’. Jespersen (1933/1983:360–1) used the term ‘contact’ relative to define the first of these (i.e. sentences which clearly participate in a matrix-embedding relationship in which a DP is relativised without recourse to overt marking). Fischer (1992:17) states that ‘the employment of subjectless relatives in Middle English (ME)’ has been attributed to post-conquest borrowing from French, although she notes that there is no ‘consensus of opinion on [this]’. Indeed, some scholars believe that, although it was rare and restricted to the subject position, the omission strategy originated in colloquial usage in OE and continued into ME, increasing in frequency during the EModE period. In support of these views, the existence of Januslike structures in OE, in which the relativised DP subject functions just as it does in contemporary English vernaculars that have the Ø option, are held to have been crucial (Fischer et al. 2000:93–4). By 970, the indeclinable relative particle, which, amongst its other functions, has a clear relative and subordinating force, had become reasonably frequent in relativised structures (Seppänen 2004). Moreover, possibly as a reflex of the correlative type, the ‘doubly-filled COMP’ or RP-type relative
< previous page
page_137
next page >
< previous page
page_138
next page >
Page 138 (‘he…. etc. who’) was common in emphatic contexts or where the marking of agreement features was required. In ME, these two seem to have coalesced into a weak form and a corresponding strong form, which eventually disambiguated their functions into article (the ) versus demonstrative/ relative pronoun (that ), respectively. A similar differentiation in status seems to have occurred with the ancestors of the modern WH- forms, which are derived from the OE interrogative class (Fischer et al. 2000:91–4). Prior to the work of Romaine (1982) and Meurman-Solin (2000) on Medieval/Renaissance Scots, especially pertinent to the evolution of SArE, accounts of relativisation (RR and NRR) in the history of English could be described as monolithic. The latter, for instance, demonstrates very clearly that the picture is much more complex. Similarly, Romaine (1982) has argued that, despite the high rates of variability that she also finds, changes to the relativisation strategies of northern speakers since the ME period are in some sense universally motivated and hence predictable because they correlate, for instance, with the Case Accessibility Hierarchy (CAH).6 In discussions of the evolution of these relativisation strategies, she argues that the changes were incremental and that relative omission was the norm in the most complex positions until WHmarking acquires its relativising functions, becoming more frequent from the fourteenth century onwards (see also Fischer et al. 2000:93). Gradually, subject-relative omission became disfavoured to the extent that, in the Modern Standard, as we have seen, WH- relative and TH- markers have become categorical in this function, although omission continues with objects and indefinites in certain registers (Dekeyser 1984; Fischer et al. 2000:94). Genitive of which and oblique object to which are thought to have appeared first and to have become established in the period 1400–1500. The pronouns whom and whose surface in the fifteenth century while the nominative who only ousts the subject-contact unmarked type in the sixteenth century. Moreover, the adoption of the WH- strategy appears to have been sensitive to both stylistic stratification, in the sense of Labov (1966),7 and RR versus NRR function (Meurman-Solin 2000). 2.2 The Source Language As noted in section 1.1, Mesthrie (1992:75) has claimed that SAIE incorporates strategies inherited from its substrates. Because certain other SArE constructions have been demonstrated to be the result of this type of language transfer, I outline what the Irish possibilities are and assess in section 2.3 the degree to which these differ from the English ones described previously. What are traditionally termed ‘direct’ and ‘indirect’ relatives in Irish are instances of what have been defined previously as marked and ancillary relative strategies, respectively. The former is accessible to positions high on the CAH, i.e. relativised subject (example [5]) and direct object DP slots. It is signalled by the proclitic a which lenites8 the following verb (aL is
< previous page
page_138
next page >
< previous page
page_139
next page >
Page 139 the conventional designation) and the construction is derived by operator movement. Thus, the operator aL is assumed to be co-indexed with an anaphoric gap (ti ) at the extraction site (see Adger and Ramchand 2006; Duffield 1996: §2.2.2 and Ch.3; Goodluck et al. 2006; McCloskey 1985, 1990, 2000, 2001; Neilson 1808/1990; Ó Siadhail 1984, 1989; Scott 2003). (5) An bheani aL dhíol ti na páistí the woman who sold the children ‘The woman who sold the children’ In general, the indirect strategy, which uses an RP alongside the morphophonological process of eclipsing the following verb (a N ), is the preferred option when the marking of DP positions low on the CAH, such as genitive case (example [6]), is required. (6) An bhean i aN bhfuil RPai páistí díolta the woman whose are her children sold ‘The woman whose children are sold’ The Irish subordinating relative strategy illustrated in examples (7) and (8) is described as the equivalent of a relative clause in Ulster Irish by Ó Siadhail (1984:127) and (1989: §11.1.8(i)2(b)). Moreover, the grammaticality of examples (7) and (8) indicates that, unlike the direct and indirect types, there is no restriction on the accessibility of this strategy, so that relativisation can be achieved at both the upper and lower ends of the CAH. (7) [CP Bhí [DP bean i ann] AGUS (SC) [IP [DP íi [I’ ag coladh]]]] BE-PAST woman there and her SLEEP-VN ‘There was a woman who was sleeping’ (8) [DP Píosa de chlári cearnógach [CP bhí mar mharc aice]] Piece of board square BE-PAST as (a) target at-her AGUS (SC) [PP [DP fáinne beag [P’ ina i lár and ring small in-its middle ‘Her target was a square piece of board in the middle of which was a small ring’ [Adapted from Ó Siadhail (1989:11.1.8(i)2:197)]9 2.3 A Review of Relativisation Processes in the Target and Source Languages To summarise, then, if we compare the relativisation strategies available in Irish with those that are extant in contemporary Standard English, it will be observed that they share only those that involve operator movement. Thus, were we to discover that the grammar of SArE incorporates an RP or
< previous page
page_139
next page >
< previous page
page_140
next page >
Page 140 ‘subordinating and’ relativising strategy, we might postulate that this lends further support to the substratum position. However, I have argued elsewhere (Corrigan 1997a, 2000a, 2000b) that comparisons such as these may turn out to be false because the trigger for extraterritorial varieties is rarely the Standard language. Indeed, even in cases where the historiographic evidence might conceivably point to a Standard English model (or where the textual evidence is such that we have no direct access to any variety other than the Standard), there is no necessary affinity between the contemporary Standard and earlier varieties. In contrast, if the comparison is made between Irish and the evidence reviewed here, it becomes increasingly difficult to decide whether possible SArE forms are the result of transfer, convergence, or the universal principles governing such processes in all human languages. Until the relativisation strategies that are available in SArE have been introduced more fully, these issues are merely conjectural. Hence, I return to them in section 4 after providing an account of the variant forms of relative clauses attested in vernacular Englishes globally. 3. RELATIVE FORMATION STRATEGIES IN SArE AND IN VERNACULAR ENGLISHES GLOBALLY The considerable research available on relativisation in nonstandard British and extraterritorial Englishes suggests that the traditional scenario for the history of Standard English cannot be generalised, particularly when referring to spoken data.10 In many of these varieties (as examples [9–21] show), the typical ratio of relative pronouns lags behind. This necessitates greater reliance by their speakers on Ø and TH- strategies in order to relativise heads at both the upper and lower positions of the CAH. Moreover, there is also support for the view that contemporary vernaculars have recourse to complex relative marking strategies, including versions of the ancillary types mentioned earlier (14/16 and 17–21) which were extant in EModE.11 (9) Leck is a young boy [ ø was coming home from school] ‘.........who ..........’ [Beal (1993:208): Tyneside/Northumbrian] (10) He’s got one boy [ ø is twelve, and the other is nine] ‘.........who ..........’ [Mesthrie (1991:466): SAIE] (11a) [DP the magistrate [CPOpi [C’ ø [IP ti was to try him was an old soldier too]]]] ‘.........who ..........’ [U1409/L2325–2328/1946F/MS1215: SArE]12
< previous page
page_140
next page >
< previous page
page_141
next page >
Page 141 (11b) That was given to [DP the cow [CP Opi [C’ that [IP ti was bad]]]] [U66/L98–99/1974M/MS1861: SArE] (11c) [DP Dr McDonald [CP whoi [C’ ø [IP ti was the parish priest in Kilkenny]]]] [U6/L8–9/1974F/MS1861: SArE] (12) The man [that went] [Edwards (1993:228): Southern British English] (13) [DP Matha Locklin [CP Opi [C’ that [IP ti lived under the roof with Padgy Bug]]]] [U2976/L4255–4257/1945F/MS976: SArE] (14) She had a child [and the child died in the ship] ‘.........who ..........’ [Mesthrie (1992:78): SAIE] (15) If I feel energetic I do roll-overs, an exercise [ø taught me by Ben Travers [ and he lived to 94]] ‘......... who ..........’ [Policansky (1982:45): Belfast English] (16) got in tow with [DP this girli ] an’ [SC [DP her father i [XP had a bit of ground ]]] ‘.........whose ..........’ [U2102/L3156–3157/1945M/MS976: SArE] (17) The person [that his foot is touched] ‘......... whose ..........’ [Romaine (1988:237): Scots] (18) This is the man [ ø his horse was stolen] ‘.........whose ..........’ [Ihalainen (1985:66): Somerset English] I thought they would have put a steel door on [that they couldn’t have opened it] ‘.........which ........’ (19) [Policansky (1982:45): Belfast English] (20) Starter is a thing [that it gets hot quickly] ‘.........which ..........’ [Mesthrie (1991:467): SAIE]
< previous page
page_141
next page >
< previous page
page_142
next page >
Page 142 (21) [DP This mani [CP ø [XP RP hei lived over there where this woman was reared]]] ‘............who ..........’ [U448/L709–713//1970M/MS1784: SArE] As can be seen from examples (11a/b/c), SArE preserves the tripartite relativisation strategies of Standard English (with Ø- marking being confined to DPs which function as subject and direct object in the matrix clause). TH-/WH- marking is used for these slots and for DP positions low on the CAH. Given my earlier observation that the WH- strategy entered English in the lowest positions and was first used in NRRs, the occurrence of this type of marking in the SArE corpus may imply a phylogenetic parallel. Thus, indirect objects with the Ø strategy are not attested, and WH- marking predominates in those clauses that may reasonably be construed as NRRs, as Herrmann (2005:39–40) also finds. Moreover, although it is not strictly my concern at this juncture, I might add that the usage of the three types in SArE-embedded relatives is rather like that of other nonstandard vernaculars (i.e. marking by WH- is comparatively rare) (Beal and Corrigan 2005; Herrmann 2005; Tagliamonte et al. 2005) and, as I demonstrate later, is subject to both internal and external constraints. As examples (16) and (21) demonstrate, both the RP and ‘subordinating and’ strategies are also possible in the SArE corpus, and the constructions are used (much as they are in the other vernaculars exemplified earlier) to relativise heads whose syntactic functions in their root clauses range from subject to genitive. Moreover, although the relative-marking strategies are preferred for the relativisation of simple DPs in the Murphy Corpus, complex DPs favour strategies not entailing extraction. 4. INTERNAL AND EXTERNAL CONSTRAINTS ON THE SYSTEM OF RELATIVISATION IN SArE Previous research has indicated quite clearly that the relativisation system of vernacular Englishes is constrained internally and externally. I intend to demonstrate this for SArE in the next sections by providing quantitative analyses of the distribution of relative clauses in the Murphy Corpus data set and more recent Questionnaire Survey. 4.1 Transfer, Convergence, or Linguistic Universals? Let’s start by taking a closer look at a topic introduced in section 2.3—namely, the extent to which the relative systems of Irish and/or the target contributed to the development of the SArE strategies. Given the Early/vernacular English and Scots evidence that I have presented, it would seem inappropriate to suggest that SArE inherited its system of relativisation simply by negative
< previous page
page_142
next page >
< previous page
page_143
next page >
Page 143 transfer from the substrate—especially because the marked strategies, for example, take WH-/TH-/Ø forms rather than aL or aN . On the face of it, the existence of operator movement and ancillary relative types in Irish, English, and the ensuing dialect may, instead, lend support to the convergence hypothesis advocated, for instance, in Corrigan (2000b). However, as far as relativisation is concerned, the interaction between Irish and English relatives does not resemble that which has been ascribed to other convergent situations. Aitchison (1992:312), for instance, argues that successful relatives in Tok Pisin are ‘those which mimic the surface structure of both Oceanic languages and English’. Because the surface structure marking of Irish and English relatives is somewhat different, this comment on the outcome in another language contact variety may lend further weight to a universal explanation because the only other source for the parallels observed here would seem to be the underlying mechanics of operator movement and RP relative strategies. The resultant SArE system may, therefore, owe far more to the influence of linguistic universals in conditions of language contact/ shift than to the integration of superstratal relativisation which is motivated by the presence of equivalent relatives in Ulster Irish. There are two means by which the validity of this claim could be tested: (1) assess the extent to which SArE relativisation resembles that which is found in members of neither the Germanic nor Celtic language families (should the resultant system be similar, then it becomes more difficult to support the idea that SArE inherited its relatives from contact per se), and (2) investigate whether the claims made by the CAH, ostensibly a universal model, are supported by the evidence from SArE. If they are, then perhaps SArE relative postmodification takes the form that it does, not because of the unique genesis of this vernacular, but because it is a natural language and, as such, is governed by the same general principles that constrain all human languages. Because I have already explored evidence of this kind from Zribi-Hertz (1984) in connection with the contrasts between colloquial and literary French in section 2.1 (examples [3a and 3b], respectively), I do not wish to labour the argument. However, such data are easily obtainable from, for instance, Tarallo’s (1983, 1996) sociolinguistic interviews of Brazilian Portuguese (BP) informants. His findings add further weight to the existence of marked and ancillary strategies in other languages apparently controlled by the same general principles of syntactic distance/position and so on. 4.2 SArE and the CAH Regarding point (2) in section 4.1, it was remarked in (iii) of section 2.1 that the CAH is intended to capture the availability of DP argument slots for relative clause formation. Informally, the CAH predicts that the subject position is the most accessible to relativisation, whereas the object of comparison slot is the least. Languages can, therefore, vary as to whether they have the full range of six possible slots (Standard English) or just subject
< previous page
page_143
next page >
< previous page
page_144
next page >
Page 144 like Malagasy (Cook 1991:28). Keenan and Comrie (1977:67) also refer to constraints on the operation of the CAH, such that any relative clause formation strategy must apply to a continuous segment of it. Furthermore, strategies that occur at any one point of the CAH may in principle cease to apply further down. Keenan and Comrie (1979b: 653) propose that the CAH represents a characterisation of relative clauseformation strategies that reflects their semantic and cognitive dimensions. Thus, subjects are the most frequently relativised because this is arguably the most semantically prominent position in discourse. The lower positions on the CAH are less accessible because relativisation here increases the degree of perceptual/ processing difficulty and, as such, relative-marking strategies are favoured as an aid to comprehension (Keenan 1987:61). Moreover, as Romaine (1984c: 468, fn. 15) reminds us, the CAH ‘also correlates with syntactic complexity’ which I have interpreted within the generative framework adopted in this chapter to mean whether the DP argument position is a syntactic island. Previous analysis of the Murphy Corpus along these lines (Corrigan 1997a) has suggested that an Accessibility Hierarchy does exist in SArE with respect to the syntactic positions that are relativiseable in line with the universal scheme outlined in Keenan and Comrie (1977) inter alia . This suggestion is examined in detail after a brief discussion of the interaction in relative clause formation between the factors of ‘embeddedness’ and ‘focus’. Certain aspects of the process of L1/L2 acquisition appear to follow a regimented developmental sequence (Sorace 2005). Romaine (1988:232; 1992:149) reports that relative clauses appear relatively late in this process on account of their complexity, claiming that the dynamics of relative embeddedness/focus are responsible. The former varies according to the function in the matrix sentence of the antecedent DP (‘DPa’) which the relative clause postmodifies. Focus, in contrast, relates to the function in the embedded clause of the relativised argument DP (‘DPr’). A combination of all the possible values for these two criteria produces the four major types illustrated next. Examples (22) and (24), in which the DPa is object, are usually termed object-embedded relatives, and examples (23) and (25), where it is subject, are known as subjectembedded relatives. In terms of the focus criterion, then, examples (22) and (23), in which DPr is subject, are subject-focus relatives, and examples (24) and (25) can be referred to as object-focus relatives. (22) OS she knew people [ø was with them] DPa DPr [U1180/L1996–1997/1951M/MS1220] (23) SS all the spuds [potatoes] [that grew in the fields round here] DPa DPr [U2773/L4000–4004/1945M/MS976]
< previous page
page_144
next page >
< previous page
page_145
next page >
Page 145 (24) OO do you know that bridge [ø you’ll cross at Peter Hanratty’s?] DPa DPr [U2246/L3310–3311/1945M/MS976] (25) SO the man [that I had working with me was reared up there] DPa DPr [U508/L838–840/1970M/MS1784] As the previous discussion noted, many studies report that developing linguistic systems impose a continuum on the order in which these are acquired so that constructions with object antecedents (examples [22] and [24]) appear to be more accessible than those with subjects (examples [23] and [25]). Because Mesthrie’s (1992:91) analysis of SAIE also revealed that ‘object embedding is overwhelmingly more frequent than subject embedding’, I would expect SArE relativisation to conform to this continuum if it were to be used in support of a universal explanation. Evidence that it does exactly that is offered in Table 6.1 which combines the figures for the distribution of object- and subject-embedded relativisation in SArE with those for L1/L2 acquisition data from Romaine (1988:234) and Schumann (1980), respectively; figures for Tok Pisin (TP) (Romaine 1992:152), and, finally, Mesthrie’s findings on SAIE. The proposal that has emerged as a result of these observations is that the embeddedness effect is explicable in terms of the syntactic distance prevailing between the head DP and the relative marker. In OS and OO types, the head DP is nearer to the marker than is the case with subject-embedded relatives largely because the former are inherently paratactic, whereas the latter entail centre-embedding. Turning now to the effect of focus, which is the factor that the CAH is driven by (although the term is never actually used in the research), recall that the basic premise of the CAH is that subject(-focus) DPs are more accessible than object(-focus) DPs, meaning that we would expect the Murphy Corpus to contain more instances of the OS/SS than the OO/SO types. This focus effect is also quite strong in this data set with the frequency of occurrence (73% vs. 27%) being slightly less marked than that for embeddedness, although still indicating a radical difference. With regard to the CAH, there are a number of studies which verify the differential that is also supported here between subject positions and those Table 6.1 The Effects of Embeddedness on Relative Clause Production SArE L1 L2 TP SAIE OS+OO 80 28 88 62 80.1 % SS+SO 20 11 10 38 19.9 %
< previous page
page_145
next page >
< previous page
page_146
next page >
Page 146 Table 6.2 Relativization and the CAH Adult SArE Child Scottish English Adult SAIE Adult Brazilian Portuguese SU 67.8 51.1 55.4 58.4 % DO 21.1 39.7 32.2 22.6 % IO/OBL 9.3 8.7 5.2 18.1 % GEN 1.9 0.5 0.4 1.0 % After Corrigan (1997a:345), Romaine (1988:235), Mesthrie (1992:93), and Tarallo (1996:206), respectively. lower down, some examples of which follow in Table 6.2. In each case, relativisability decreases as the CAH is descended so that genitive is the least frequently relativised position.13 The aspect of relativisation to which I have paid greatest attention thus far is the relationship between the type of strategy used and the position on the hierarchy (which is arguably a reflex of the general principles of locality and binding). There is also the related issue of the fact (as Keenan and Comrie 1979a remark) that, cross-linguistically, RP and other [+Case]-coding strategies are normally reserved for relativising DPs on the lower positions of the CAH. Romaine (1992:171) supports this position on the basis of findings from Dreyfuss (1977) with respect to Tok Pisin, as well as a range of other L1 and L2 Englishes. Furthermore, from a historical perspective, Romaine (1982) reports that [+Case] WH- relatives in English appeared first in the lowest slots reaching subject position last. Similarly, Miller (1988:114–17) notes that RP strategies are maintained as an adult Scottish norm, and they are mentioned in Miller (1993:111) as typical ‘if the relative clause contains a long constituent or another clause’. With respect to the latter, there are two findings in the previous SArE Folklore data which are relevant: (1) genitives appear not to be relativiseable without recourse to ancillary strategies because there are no examples at all in the Murphy Corpus, and (2) a number of the DPs lower down on the CAH that were relativised by ancillary strategies in this data set were often promoted to subject focus so that they were hardly distinguishable as indirect objects or obliques.14 The first of these inferences may be explained on account of the fact that the possessive WH- forms of Standard English (whose, of which) are categorically absent in the SArE Murphy Corpus, so that the only [+Case]-coding strategy available is either by RP or ‘subordinating and’ (see also Herrmann 2005:57–61; Sigley 1997:230). This would also seem to be the function of RP/‘subordinating and’ in EModE and of Irish RP/‘subordinating agus’ described in section 2. This usage appears to indicate that both of
< previous page
page_146
next page >
< previous page
page_147
next page >
Page 147 Table 6.3 The Distribution of [+/–Case]-Coding Strategies in the SArE Murphy Corpus Case Accessibility Hierarchy + Case % – Case % UPPER 80 89 LOWER 20 11 these alternative [+Case] strategies operate in all these varieties as Romaine (1984b: 268) suggests they might (i.e. they ‘take up the slack in the system’). Given the constraint that ancillary strategies cluster at the lower end of the CAH, my finding vis-à-vis promotion to subject, which consequently entails a greater number in the Murphy Corpus of subject-focus relatives constructed using these than one might expect, is far from unique. For instance, it has been shown by McCloskey (1990) inter alia that Irish permits the use of an RP strategy to relativise on the subject slot (provided that this is not the Highest Subject Position), and Mesthrie (1992:94) remarks that all the RP types in his SAIE corpus except one were subject focus. Indeed, quite early on, Keenan and Comrie (1979b: 652) mentioned the possibility of promotion to subject by passivisation in cases where languages have direct objects that cannot normally be relativised.15 Finally, I consider the issue of whether there is the expected correlation in SArE between [+/-Case]-coding strategy and the position of the argument DP. Because there is no general consensus about the complementiser/pronominal status of English that , the discussion is restricted to comparisons between the undisputed [-Case] Ø strategy and the [+Case] WH- /RP/and strategies. I also exclude the category adverbial used in a relativising function on the basis that they are not strictly pronominal either.16 Table 6.3 illustrates my findings for the remaining categories and strategies. Although both [+Case] and [-Case]-coding strategies are preferred in the upper positions of the CAH (the lack of differentiation being due largely to the number of promoted subjects discussed earlier), they diverge as the hierarchy predicts for the lower positions. Thus, Ø- marking is favoured for relativising only 11% of the available positions at the lower end of the CAH, whereas the percentage of [+Case]-coding strategies occurring in these positions is almost twice that figure. 4.3 Sentence Structure as an Internal Constraint in SArE Other notable internal constraints have been isolated in previous research as governing the usage of Ø , TH- , and WH- marking, including the sentence type in which the relativised subject occurs. Tagliamonte et al. (2005), for instance, have noted that the zero variant is favoured by their Northern Irish
< previous page
page_147
next page >
< previous page
page_148
next page >
Page 148 and North Western English communities in existential constructions of a similar kind to the SArE example given in example (26). (26) there’s an old big one ø caught and pulled down [U123/L199–202/1973M/MS1810] This seemed worth exploring for both of the SArE databases, especially given their suggestion that it may be a reflex of older patterns. The Murphy Corpus contains over three times the number of existentials isolated in the Tagliamonte et al. Northern Irish communities (51 vs. 14, respectively). As such, the concern expressed in Tagliamonte et al. (2005) that the trend observed for zero preference was not statistically significant in their sample is less of an issue here. That said, 90 percent of the existentials observed in the SArE Corpus were of the type illustrated in example (26), whereas the rest resembled example (27), all of which were produced with that as the relative marker. (27) there was a Peter Mallon that was in it and they were borrowed from our Parish [U721/L1187–1190/1968M/MS1747] With a view to investigating the extent to which a similar constraint can be observed in the judgments of younger and older speakers in a time period more consistent with that of Tagliamonte et al. (2005), a test sentence containing an existential with zero marking such as example (26) was offered to all informants taking part in the Questionnaire Survey. Informants were asked to rate this on a scale from 1 to 4: • 1 = I use this type of sentence myself . • 2 = I frequently hear this type of sentence in South Armagh. • 3 = This type of sentence is not common in South Armagh—but it doesn’t seem too odd . • 4 = This type of sentence would never be used in South Armagh—it seems very odd . Responses from all 44 informants are presented in Table 6.4. Although there is a marginal difference between adult males and females, the scores for younger speakers are identical and differ only minimally from the adult males. Despite these nuances, all speakers—even the relatively conservative older women—judge these structures to be high on the acceptability scale provided (i.e. between ratings 1 and 2). This offers independent support for the conclusions based on the earlier data set from the same community as regards the favouring effect of existentials. It also bolsters the suggestion by Tagliamonte et al. (2005) that peripheral dialects of NIrE may well have
< previous page
page_148
next page >
< previous page
page_149
next page >
Page 149 Table 6.4 Acceptability Judgements for the Zero Marking of Relatives in Existential Constructions Respondent’s Demographic Characteristics Average Group Scores Total Number of Responses Older Females 2.05 11 Younger Females 1.9 11 Older Males 1.79 11 Younger Males 1.9 11 a propensity towards Ø usage in such constructions. What is more, as the subsequent analyses confirm, this internal constraint delimits the effect of external factors with respect to Ø usage, which appear robustly when tested in other kinds of syntactic construction, and it is to these that I now turn. 4.4 Language-External Variation and Change in SArE In addition to uncovering at least some of the internal mechanisms operating on relativisation in SArE, I was also keen to find evidence to support previous research indicating that certain relative markers had become associated with particular social groups and, if these patterns also existed in this dialect, whether they were stable across time. 4.4.1 External Constraints in the Murphy Corpus Many scholars have already established that, aside from any internal linguistic constraint, the choice of relative pronoun may also have external correlates. Hence, Herrmann (2005:55) notes that the ‘frequency of zero subject relative clauses can be tied to traditionality and/or colloquiality of speaker’. Similarly, Tagliamonte et al. (2005) remark that across their enclave communities, ‘WH forms were highly restricted to certain individuals, not surprisingly those with more education’. Hence, although one might reasonably assume that Ø relative marking and WH- marking for my SArE informants could represent nonprestigious and prestigious variants, respectively, the choice may well be governed by the internal constraints noted in the various subsections of section 4, as well as external factors. These facts make the decision as to which marker to choose as the standard variant much less straightforward than the usual case. Moreover, there is also the status of TH- marking to consider. For all intents and purposes, it appears to be a Standard variant, but because TH- can lack the prestige of WH- types, it is generally confined to spoken rather than written language (Miller 1993:110). These factors prompted Romaine (1981:87), for instance, to
< previous page
page_149
next page >
< previous page
page_150
next page >
Page 150 construct a trinary variable index for the variants in order to achieve an ‘accurate measure’ of the stylistic constraints on the use of the variable in different text types. Indeed, in a more recent investigation of stylistic effects on relativisation in New Zealand English, Sigley (1997:230) goes so far as to claim that: the presence of categorical differences both in the strategies possible, and in the influences on their occurrence severely undermines the usual practice of modelling this variation as a single application of variable rules. Because my statistical methodology prohibits the use of weighted scores associated with Romaine’s strategy and because the data set is also problematic with respect to traditional variationist quantitative methods, it is not possible to investigate the social distribution of relative markers in SArE by either of these methods anyway (Corrigan 1997a). Consequently, I consider each of the marked variants found in the Murphy Corpus separately in case they are not constrained by real-time or gender in SArE in the way that previous research on other nonstandard and historical dialects of English might lead one to expect. Figure 6.1a17 is based on a distributional analysis of the Ø variant in the Murphy Corpus, and it demonstrates that it does seem to be affected by real time because there is a distinct pattern of decreased usage amongst speakers from 1942 to 1974. Figure 6.1b, which includes separate figures for males and females, illustrates the fact that this same pattern is, indeed, replicated across both genders, although there are grounds for proposing that the progress of the
Figure 6.1a Occurrence of zero variant for all informants in the Murphy Corpus (1942–1974) (N = frequency of occurrence per 1,000 words per year).
< previous page
page_150
next page >
< previous page
page_151
next page >
Page 151
Figure 6.1b Occurrence of zero variant for male/female informants in the Murphy Corpus (1942–1974) (N = Frequency of occurrence per 1,000 words per year). change amongst women may be proceeding at a faster rate than that for men. Figure 6.2a shows a clear opposition between the usage of Ø , illustrated in Figure 6.1a, and TH- marking in real time for all informants. In this case, rather than the sharp decrease of Figure 6.1a, there is relative stability between the periods, and this invariance is matched in Figure 6.2b for both genders. It should also be noted that the peak for the first age cohort of women in 1946 is unlikely to be representative as the number of words (312) is well below the mean for the early period of Murphy’s data
Figure 6.2a Occurrence of TH- variant for all informants in the Murphy Corpus (1942–1974) (N = frequency of occurrence per 1,000 words per year).
< previous page
page_151
next page >
< previous page
page_152
next page >
Page 152
Figure 6.2b Occurrence of TH- variant for male/female informants in the Murphy Corpus (1942–1974) (N = frequency of occurrence per 1,000 words per year). collection. Ignoring this distortion suggests that the female group may even have begun a trend towards increased usage, perhaps as a consequence of the loss of the Ø option posited earlier. Turning now to the WH- variant, it can be observed from Figure 6.3a that, like usage of the TH- strategy and unlike that of Ø- marking, there is no discernible pattern of change in real time for the corpus overall. Figure 6.3b, in contrast, shows a striking contrast between the genders. The male data arguably depict a trend towards a decrease in the WH- strategy, although there is some inconsistency in that male usage for the first
Figure 6.3a Occurrence of WH- variant for all informants in the Murphy Corpus (1942–1974) (N = frequency of occurrence per 1,000 words per year).
< previous page
page_152
next page >
< previous page
page_153
next page >
Page 153
Figure 6.3b Occurrence of WH- variant for male/female informants in the Murphy Corpus (1942–1974) (N = frequency of occurrence per 1,000 words per year). age cohort in 1945/1951 and that for the second in 1972, for example, are not dissimilar which precludes any definitive statement in this regard. I can be considerably more confident about the scores for females presented in Figure 6.3b. It shows a dramatic increase in the number of occurrences of WH- marking by women in exactly the period, where it was noted in Figure 6.1b that their usage of the Ø marker appears to have ceased and there is the lowest proportion of female words in the corpus overall. This finding (and that for the females in Figure 2b) is important because the correlation between the loss of the Ø marker, the marginal increase of TH- amongst females, and the significant shift to WH- marking appears to offer confirmation of an emerging trend with respect to much of the quantitative data collated from the Murphy Corpus with respect to other morphosyntactic variants (i.e. there is evidence to postulate an ongoing shift by female speakers of this dialect away from local norms) (Corrigan 1997a). This, therefore, seemed to be an important line of enquiry to follow in the more recent Questionnaire Survey and, therefore, is explored in some detail in section 4.4.2. 4.4.2 External Constraints in the Questionnaire Survey Nonexistential, subject, and object relatives either identical to or reminiscent of examples (22) and (23) were offered to all 44 informants. They contained various relative marking options and were presented in different orders and more than once to minimise what Labov (1996:100) termed ‘social intervention’.18 This was felt to be particularly crucial in the case
< previous page
page_153
next page >
< previous page
page_154
next page >
Page 154 of relativisation given previous findings with respect to the close association between WH- marking and educational attainment. In addition, given the fact that corpus-based studies of relativisation (cf. Beal and Corrigan 2005; Herrmann 2005; Poussa 2002; Tagliamonte et al. 2005) have demonstrated the importance of distinguishing between different antecedent and sentence types, as well as between restrictive and nonrestrictive relatives,19 these issues were considered during the compilation of the questionnaire and when the responses were analysed. As such, the following discussion is confined to highlighting patterns of variation among WH- , TH- , and Ø in restrictive relatives only, and the analysis is sensitive to grammatical category, as well as to the nature of the antecedent and sentence type. Generally speaking, the overall results for speaker judgments relating to Ø subject relatives when they do not occur in existential constructions appear to indicate that this is the least favoured position. As Table 6.5 demonstrates, there was also some gender and age variation (particularly between the older females and younger males). In general, however, the choice for this variant ranges primarily between options three and four (i.e. the structure is uncommon/odd). Judgments relating to similar test sentences in all respects, bar the fact that the DP being relativised was in object position, are also given in Table 6.5. They indicate a greater tolerance for Ø when relativising this particular grammatical category, although it is only younger males who score within the positive evaluation range. Table 6.6, especially when compared with the results for Ø- marking in Table 6.5, indicate that the THvariant as a marker of both relativised subject and object DPs is generally preferable to the Ø option in nonexistentials. In fact, as Table 6.7 demonstrates, although females clearly prefer WH- marking with subject relatives over the Ø variant, their preferred marker for objects is, in fact, TH- . This contrasts particularly with the male pattern, the younger group accepting both markers for object relatives equally well and their older peers showing a distinct preference for WH- marking of DPs in this position. Table 6.5 Acceptability Judgements for Zero Marking of Subject and Object Relatives in Non-Existentials Respondent’s Demographic Average Group Scores Average Group Scores Total Number of Characteristics Subject Relative Object Relative Responses Older Females 3.55 2.61 11 Younger Females 3.27 2.27 11 Older Males 3.23 2.30 11 Younger Males 2.59 1.72 11
< previous page
page_154
next page >
< previous page
page_155
next page >
Page 155 Table 6.6 Acceptability Judgements for TH- Marking of Subject and Object Relatives in Non-Existentials Respondent’s Demographic Average Group Scores Average Group Scores Total Number of Characteristics Subject Relative Object Relative Responses Older Females 3.16 1.77 11 Younger Females 2.72 1.45 11 Older Males 2.23 2.00 11 Younger Males 2.18 1.72 11 These results marry well with the quantitative analysis of the Murphy Corpus presented earlier, as well as those of other northern British vernaculars on a number of counts. In the first place, Tagliamonte et al. (2005:101) revealed Ø subject usage to be in and around 20 percent, leading them to suggest that, despite the dominance of this type of marking in earlier Englishes, it retains only a minimal foothold in the peripheral dialect areas they surveyed which resemble South Armagh in a number of other respects. In addition, the trend identified in the Murphy Corpus with respect to the predominance of WH- marking in female speech and the relative stability of TH- in the community as a whole is echoed in the Questionnaire Survey and may well support the views of Herrmann (2005:28) that ‘the frequency of WH- pronouns serves as a yardstick for the degree of standardisation or traditionality of speakers’. Indeed, Tagliamonte et al. (2005:106) go so far as to suggest that ‘WH forms are an overlay from outside the vernacular grammar’ which is why the THvariant is ‘still holding its own as the universal relative marker’ in Northern British dialects and their descendants in Northern Ireland. Table 6.7 Acceptability Judgements for WH- Marking of Subject and Object Relatives in Non-Existentials Respondent’s Demographic Average Group Scores Average Group Scores Total Number of Characteristics Subject Relative Object Relative Responses Older Females 2.05 2.05 11 Younger Females 2.02 2.00 11 Older Males 2.30 1.61 11 Younger Males 2.22 1.72 11
< previous page
page_155
next page >
< previous page
page_156
next page >
Page 156 5. CONCLUSION Taking all the evidence presented in the various subsections of section 4 together, conclusions of various kinds can be drawn from the analyses. In the first place, the case for convergence between the EModE and Irish triggers appears to be much less tenable than it first appeared to be. SArE relativisation processes bear a remarkable resemblance to those that occur in French/BP, for instance, which cannot be accommodated simply by recourse to the substratum hypothesis and seem instead to support a universal explanation rather similar to that invoked for the small-clause analysis of ‘subordinating agus/and’ articulated in Corrigan (2000a).20 Furthermore, many aspects of relative postmodification in SArE are predicted by the CAH. Although this model is limited in certain respects (see Herrmann 2005; Maxwell 1979; Romaine 1992; Tallerman 1990), and it is clear that discourse considerations can override mere considerations of Case (see Fox and Thompson 1990), there does appear to be a congruence between the type of relative marker used and issues of grammatical complexity/cognitive processing (Fodor 1998). Whether this is interpreted as the CAH (Herrmann 2005:49–62), falling from principles of locality/binding (Haegeman 1991:420–9), or more simply as ‘length’ (Tagliamonte et al. 2005:97), the evidence suggests that there are major internal constraints on relativisation in SArE. Crucially, the same trends identified in SArE are also echoed in the historical development of English/Scots, as well as in contemporary (contact) vernaculars and in the process of acquisition. Moreover, this research supports the hypotheses recently suggested in Cornips and Corrigan (2005), in particular, that syntactic changes can be imposed by E-language factors such as real/apparent time and gender, which feature prominently in this investigation, but that these are always constrained in certain respects. They can, for example, be construction-specific (speakers can favour Ø strategies in existential structures, but disfavour them in other contexts). Moreover, the entire system of relativisation is subject to other natural (cognitive) processes, such as position within the clause. Similar principles have been demonstrated to exist cross-linguistically by Comrie and Kuteva (2005) and cross-dialectally by Herrmann (2005) and Tagliamonte et al. (2005). This investigation of Northern British Englishes and their Northern Irish relatives lends further support to this research, as well as to our understanding of the dynamics of relativisation in a single peripheral community during a critical period in its recent social history. NOTES 1. Warm thanks are due to the participants of the World Englishes Symposium for their comments and to the editors for useful feedback on an earlier draft. Any remaining errors or infelicities are my own. I am also grateful to the Dept. of Irish Folklore, University College, Dublin, for granting access to Murphy’s
< previous page
page_156
next page >
< previous page
page_157
next page >
Page 157 Folklore manuscripts. I would, likewise, wish to express my gratitude to staff and pupils at St. Brigid’s/St. Joseph’s High Schools; The Cardinal Tomás Ó Fiaich Heritage Centre; The Rural Health Partnership; Cullyhanna Women’s Group; Women on Rural Development; and all who assisted in the administration/completion of the Questionnaire Survey in 2001–2003. The latter was made possible by my Research Fellowship from the Leverhulme Trust (RF&G/1/RFG/2000/0410). 2. Corrigan (1997a, to appear) details the rationale for these social changes. 3. Full details of the methodology of this survey can be found in Corrigan (to appear). 4. For an analysis of this distinction within different theoretical frameworks, see Arnold (2007), Borsley (1992), Fabb (1990), and Peterson (2004). 5. Corrigan (1997a) contains a fuller discussion of the difficulties with the Murphy Manuscripts in this regard, and see Fischer et al. (2000:58), Fox and Thompson (1990:298), Herrmann (2005:38–41), Hope (1994:32– 33), Mesthrie (1991:465), and Romaine (1982:82). 6. For fuller accounts of this arguably universal constraint on relative systems than will be possible here, see also Corrigan (1997a) and Herrmann (2005), as well as the seminal works of Keenan and Comrie (1977, 1979a, 1979b) and Keenan (1987). 7. Romaine (1980, 1981, 1982) views this type of stratification as an index of complexity, which it may well be, but it is important to remember that the complexity involved is syntactic in nature. 8. See Duffield (1996: Ch.2 and §3.2); McCloskey (2001), and Ó Siadhail (1989: §6.2) for analyses of the mutation processes in Irish associated with relative formation strategies, and see Goodluck et al. (2006) and Scott (2003) for a comparison of L1/L2 speakers’ ability to successfully utilise these distinctions. 9. The abbreviated clausal analyses in examples (7) and (8) signal that I continue to assume, as argued in Corrigan (2000b), that these are small-clause adjuncts. 10. For relativisation in other vernaculars, as well as in the history of Germanic languages more widely, see Poussa (2002). Comrie and Kuteva (2005) provide an account of relative structures across other language families, including Germanic. 11. This list is not inclusive because I ignore relative marking by at/as/what , which are common in nonstandard, usually southern British, and certain extraterritorial varieties (Beal 1993, 2006; Beal and Corrigan 2005; Comrie 1999; Herrmann 2005; Mesthrie 1991; Seppänen 1999). Although Filppula (1999) notes that they are not generally reported for dialects of IrE, Herrmann (2005:27) finds that ‘while what is unknown’, as does occur in the NITCS, although the frequency of occurrence is extremely low at 0.5 percent. 12. These codes specify the locations of these constructions in the original Murphy Corpus. 13. It is interesting in this regard that Goodluck et al. (2006:659) claim that ‘Irish-speaking children find relativization of subject and object positions equally easy … [because] movement of either the subject or object in Irish is movement across the verb’. By contrast, in the SVO languages, which Romaine (1984b, 1988) focuses on, for example, it is only object movement that would entail movement across the verb. 14. Herrmann (2005) as well as Seppänen and Kjellmer (1995) provide fuller accounts of the relativisation strategies associated with noun phrases low on the CAH than is possible here, and their findings largely agree with these for the SArE data.
< previous page
page_157
next page >
< previous page
page_158
next page >
Page 158 15. Tallerman (1990:301–4) also argues for the possibility of ‘demotion’ from subject position in Welsh so that a [+Case] RP strategy can be used that will disambiguate a [-Case] word-order strategy. 16. For details of these anomalies and the pronominal/nonpronominal status of that , see Miller (1988, 1993), Romaine (1982:58 and 214ff.), and Seppänen (1997). 17. All figures in these graphs are expressed as occurrences per thousand words following Labov (1982:87) on account of the problematic dimensions of the database (see Labov 1994:26; Corrigan 1997a: Ch.5, §3 for a rationale). 18. As Schütze (1996) and Cornips and Poletto (2004) also argue, the exact relationship between grammaticality tasks and intuitions is not straightforward, and it is thought that presenting the same task more than once improves the reliability of the exercise. 19. WH- is favoured in the latter, for instance, even in the otherwise highly nonstandard dialects of Tyneside and Sheffield (Beal and Corrigan 2005). 20. The review by Comrie and Kuteva (2005) of the relative structures found across all human languages is also instructive in this regard. REFERENCES Adger, D., and G. Ramchand. 2006. Dialect variation in Gaelic relative clauses. In Rannsachadh na Gàidhlig 3, edited by W. McLeod, J.E. Fraser, and A. Gunderloch, 1–15. Edinburgh: Dunedin Academic Press. Aitchison, J. 1992. Relative clauses in Tok Pisin: Is there a natural pathway? In Internal and External Factors in Syntactic Change , edited by M. Gerritsen and D. Stein, 295–316. Berlin: Mouton. Arnold, D. 2007. Non-restrictive relatives are not orphans. Journal of Linguistics 43(2): 271–309. Beal, J.C. 1993. The grammar of Tyneside and Northumbrian English. In Milroy and Milroy, eds. 1993. 187– 213. ———. 2006. Language and Region. Abingdon: Routledge. ———, and Corrigan, K.P. 2005. A tale of two dialects: Relativisation in Newcastle and Sheffield. In Dialects Across Borders, edited by M. Filppula, J. Klemola, M. Palander, and E. Penttilä, 211–29. Amsterdam: John Benjamins. Borsley, R.D. 1992. More on the differences between English restrictive and non-restrictive clauses. Journal of Linguistics 28: 139–48. Chomsky, N. 1982. Some Concepts and Consequences of the Theory of Government and Binding . Cambridge, MA: MIT Press. Comrie, B. 1999. Relative clauses: structure and typology on the periphery of Standard English. In The Clause in English: In Honour of Rodney Huddleston (Studies in Language Companion Series 45), edited by P. Collins and D. Lee, 81–91. Amsterdam: John Benjamins. ———, and T. Kuteva. 2005. Relativization strategies. In The World Atlas of Language Structures, edited by M. Haspelmath, M.S. Dryer, D. Gil, and B. Comrie, 494–501. Oxford: OUP. Cook, V. 1991. Second Language Learning and Language Teaching. London: Edward Arnold. Cornips, L., and K.P. Corrigan. 2005. Toward an integrated approach to syntactic variation: A retrospective and prospective synopsis. In Syntax and Variation, edited by L. Cornips and K.P. Corrigan, 1–30. Amsterdam: John Benjamins. Cornips, L., and C. Poletto. 2004. On standardising syntactic elicitation techniques, PART I. Lingua 115. Corrigan, K.P. 1997a. The syntax of South Armagh English in its socio-historical perspective . Unpublished thesis, National University of Ireland, Dublin.
< previous page
page_158
next page >
< previous page
page_159
next page >
Page 159 ———. 1997b. The acquisition and properties of a contact vernacular grammar. In Dán do Oide , edited by A. Ahlqvist and V. Capkovà, 75–94. Dublin: Linguistics Institute of Ireland. ———. 2000a. ‘What bees to be maun be’: Aspects of deontic and epistemic modality in a northern dialect of Irish-English. English World-Wide 21(1): 25–62. ———. 2000b. What are small clauses doing in South Armagh English, Irish and Planter English? In Celtic Englishes II, edited by H.L.C. Tristram, 75–96. Heidelberg: Carl Winter. ———. To appear. Parametric Variation within a Socially Realistic Linguistics: Syntactic Variation and Change in South Armagh English. Oxford: Blackwell. Dekeyser, X. 1984. Relativisers in Early Modern English: A dynamic quantitative study. In Historical Syntax , edited by J. Fisiak, 61–87. Berlin: Mouton de Gruyter. Doherty, C. 1993. The syntax of subject contact relatives. Unpublished ms., University of Southern California. Dreyfuss, G. 1977. Relative clause structure in four creole languages . Unpublished thesis, University of Michigan. Duffield, N. 1996. Particles and Projections in Irish Syntax . Dordrecht: Kluwer. Edwards, V. 1993. The grammar of Southern British English. In Milroy and Milroy, eds. 1993. 214–38. Fabb, N. 1990. The difference between English restrictive and non-restrictive relative clauses. Journal of Linguistics 26: 57–78. Filppula, M. 1991. Subordinating and in Hiberno-English syntax: Irish or English origin? In Language Contact in the British Isles , edited by P.S. Ureland and G. Broderick, 617–31. Niemeyer: Tübingen. ———. 1999. The Grammar of Irish-English. London: Routledge. Finlay, C. 1994. Syntactic variation in Belfast English. Belfast Working Papers in Language and Linguistics 12: 69–97. Fischer, O. 1992. Syntactic change and borrowing: The case of the accusative and infinitive construction in English. In Internal and External Factors in Syntactic Change , edited by M. Gerritsen and D. Stein, 17–88. Berlin: Mouton. ———, A. van Kemenade, W. Koopman, and W. van der Wurff. 2000. The Syntax of Early English. Cambridge: Cambridge University Press. Fodor, J. 1998. Learning to parse? Journal of Psycholinguistic Research 27(2): 285–319. Fox, B., and S. Thompson. 1990. A discourse explanation of the grammar of relative clauses in English conversation. Language 66: 297–316. Geisler, C. 2002. Relativisation in Ulster English. In Poussa, ed. 2002. 135–46. Goodluck, H., E. Guilfoyle, and S. Harrington. 2006. Merge and binding in child relative clauses: The case of Irish. Journal of Linguistics 42(3): 629–61. Haegeman, L. 1991. Introduction to Government and Binding Theory . Oxford: Blackwell. Harris, J. 1993. The grammar of Irish English. In Milroy and Milroy, eds. 1993. 139–86. Henry, A. 1995. Belfast English and Standard English. Oxford: Oxford University Press. Herrmann, T. 2005. Relative clauses in English dialects of the British Isles. In A Comparative Grammar of British English Dialects: Agreement, Gender, Relative Clauses, edited by B. Kortmann, T. Herrmann, L. Pietsch, L., and S. Wagner, 21–124. Berlin and New York: Mouton. Hope, J. 1994. The Authorship of Shakespeare’s Plays. Cambridge: Cambridge University Press. Ihalainen, O. 1985. Synchronic variation and linguistic change: Evidence from British English dialects. In Papers from the 4th International Conference on English
< previous page
page_159
next page >
< previous page
page_160
next page >
Page 160 Historical Linguistics, Amsterdam, 10–13 April 1985 (Current Issues in Linguistic Theory, Vol. 41), edited by R. Eaton, O. Fischer, W. Koopman, and F. van der Leek, 61–72. Amsterdam: John Benjamins. Jespersen, O. 1933. 2nd edition 1983. Essentials of English Grammar . New York: Holt & Company. Johnston, P. 1997. Regional variation. In The Edinburgh History of the Scots Language, edited by C. Jones, 443–513. Edinburgh: Edinburgh University Press. Keenan, E.L., ed. 1987. Universal Grammar . London: Croom Helm. ———, and B. Comrie. 1977. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8: 63–99. ———. 1979a. Data on the noun phrase accessibility hierarchy. Language 55: 332–52. ———. 1979b. Noun phrase accessibility revisited. Language 55: 649–55. Kirk, J.M. 1992. The Northern Ireland transcribed corpus of speech. In New Directions in English Language Corpora: Methodology, Results, Software Development , edited by G. Leitner, 65–73. Berlin: Mouton. Labov, W. 1966. The Social Stratification of English in New York City . Washington DC: Center for Applied Linguistics. ———. 1982. Building on empirical foundations. In Perspectives on Historical Linguistics, edited by W.P. Lehmann and Y. Malkiel, 17–92. Amsterdam and Philadelphia: John Benjamins. ———. 1994. Principles of Linguistic Change. Oxford: Blackwell. ———. 1996. When intuitions fail. Papers from the 32nd Regional Meeting of the Chicago Linguistics Society 32: 76–106. Maxwell, D. 1979. NP accessibility and universal grammar. Language 55:352–71. McCloskey, J. 1985. The Modern Irish double relative and syntactic binding. Ériu 36: 45–84. ———. 1990. Resumptive pronouns, A-bar binding and levels of representation in Irish. In The Syntax of the Modern Celtic Languages, edited by R.A. Hendrick, 199–248. San Diego: Academic Press. ———. 2000. Resumption, successive cyclicity and the locality of operators. Unpublished ms., University of California at Santa Cruz. ———. 2001. The morphosyntax of WH- extraction in Irish. Journal of Linguistics 37(1): 67–100. Mesthrie, R. 1987. From OV to VO in language shift. English World-Wide 8: 263–76. ———. 1991. Syntactic variation in South African Indian English: The relative clause. In English Around the World, edited by J. Cheshire, 462–73. Cambridge: Cambridge University Press. ———. 1992. English in Language Shift. Cambridge: Cambridge University Press. Meurman-Solin, A. 2000. Geographical, socio-spatial and systemic distance in the spread of the relative who in Scots. In Generative Theory and Corpus Studies: A Dialogue from 10ICEHL (Topics in English Linguistics 31), edited by R. Bermúdez-Otero, D. Denison, R.M. Hogg, and C.B. McCully, 417–38. Berlin: Mouton de Gruyter. Miller, J. 1988. That a relative pronoun? Sociolinguistics and syntactic analysis. In Edinburgh Studies in the English Language, edited by J.M. Anderson and N. MacLeod, 113–19. Edinburgh: John Donald. ———. 1993. The grammar of Scottish English. In Milroy and Milroy, eds. 1993. 99–139. Milroy, J., and L. Milroy, eds. 1993. Real English: The Grammar of English Dialects in the British Isles . London: Longman. Neilson, W. 1808. 2nd edition 1990. An Introduction to the Irish Language in Three Parts . Dublin: Wogan. (1990 Facsimile by Ultach Trust, Belfast.) Ó Siadhail, M. 1984. Agus (Is)/And: A shared syntactic feature. Celtica 16: 125–37.
< previous page
page_160
next page >
< previous page
page_161
next page >
Page 161 ———. 1989. Modern Irish . Cambridge: Cambridge University Press. Peterson, P. 2004. Non-restrictive relatives and other non-syntagmatic relations in a Lexical-Functional Grammar. In Proceedings of the LFG04 Conference, University of Canterbury , edited by M. Butt and T. Holloway King. Stanford: CSLI Publications. Policansky, L. 1982. Grammatical variation in Belfast English. Belfast Working Papers in Language and Linguistics 6: 37–66. Poussa, P., ed. 2002. Relativisation on the North Sea Littoral . Munich: Lincom Europa. Quirk, R., S. Greenbaum, G. Leech, and J. Svartvik, eds. 1985. A Grammar of Contemporary English. London: Longman. Romaine, S. 1980. The relative clause marker in Scots English: Diffusion, complexity and style as dimensions of syntactic change. Language in Society 9: 221–49. ———. 1981. Syntactic complexity, relativisation and stylistic levels in Middle Scots. Folia Linguistica Historica 2: 56–77. ———. 1982. Socio-historical Linguistics. Cambridge: Cambridge University Press. ———. 1984a. The Language of Children and Adolescents. Oxford: Blackwell. ———. 1984b. Relative clauses in child language, pidgins and creoles. Australian Journal of Linguistics 4: 257–81. ———. 1984c. Towards a typology of relative clause formation strategies in Germanic. In Historical Syntax , edited by J. Fisiak, 437–70. Berlin: Mouton. ———. 1988. Pidgin and Creole Languages. London: Longman. ———. 1992. The evolution of complexity in a creole language. Studies in Language 16: 139–82. Schumann, J.H. 1980. The acquisition of English relative clauses by second language learners. In Research in Second Language Acquisition, edited by R.C. Scarcella and S.D. Krashen, 118–31. New York: Newbury House. Schütze, C.T. 1996. The Empirical Base of Linguistics. Chicago: University of Chicago Press. Scott, S. 2003. Second language acquisition of relative clauses in Irish. In Proceedings of the Sixth Generative Approaches to Second Language Acquisition Conference (GASLA2002) , edited by J.M. Liceras, H. Zobl, and H. Goodluck, 260–8. Cascadilla Proceedings Project, Sommerville, MA. Seppänen, A. 1997. Relative that and prepositional complementation. English Language and Linguistics 1: 111–33. ———. 1999. Dialectal variation in English relativisation. Lingua 109: 15–34. ———. 2004. The Old English relative thorn+e . English Language and Linguistics 8(1): 71–102. ———, and G. Kjellmer. 1995. The dog that’s leg was broken: On the genitive of the relative pronoun. English Studies 38: 389–400. Sigley, R. 1997. The influence of formality and channel on relative pronoun choice in New Zealand English. English Language and Linguistics 1(2): 207–32. Sorace, A. 2005. Selective optionality in language development. In Syntax and Variation, edited by L. Cornips and K.P. Corrigan, 46–111. Amsterdam: John Benjamins. Stockwell, R.P., and D. Minkova. 1991. Subordination and word order change in the history of English. In Historical English Syntax , edited by D. Kastovsky, 367–409. New York: Mouton. Tagliamonte, S.A. 2001–2003. Back to the Roots: The Legacy of British Dialects . Research Grant. Economic and Social Research Council of the United Kingdom (ESRC). #R000239097. ———, J. Smith, and H. Lawrence. 2005. No taming the vernacular! Insights from the relatives in Northern Britain. Language Variation and Change 17(1): 75–112. Tallerman, M. 1990. Relativisation strategies. Journal of Linguistics 26: 291–314.
< previous page
page_161
next page >
< previous page
page_162
next page >
Page 162 Tarallo, F. 1983. Relativisation strategies in Brazilian Portuguese . Unpublished thesis, University of Pennsylvania. ———. 1996. Turning different at the turn of the century: 19th century Brazilian Portuguese. In Towards a Social Science of Language Papers in Honor of William Labov . Volume 1: Variation and Change in Language and Society (Current Issues in Linguistic Theory, Vol. 127), edited by G.R. Guy, C. Feagin, D. Schiffrin, and J. Baugh, 199–220. Amsterdam: John Benjamins. Temperley, D. 2003. Ambiguity avoidance in English relative clauses. Language 79(3): 464–86. Zribi-Hertz, A. 1984. Orphan prepositions in French and the concept of null pronoun. Recherces Linguistiques 12: 46–91.
< previous page
page_162
next page >
< previous page
page_163
next page >
Page 163 7 The Case of Bungi Evidence for Vernacular Universals Elaine Gold 1. INTRODUCTION The Bungi dialect arose in the mid-eighteenth century along trade routes south of Hudson Bay, in what is now the Canadian province of Manitoba. This English vernacular resulted from contact and intermarriage between Scottish English-speaking Hudson Bay Company traders and the indigenous Cree. During its development, Bungi had little, if any, contact with standard Canadian English, as the speakers were both socially and geographically isolated from mainstream Canadian culture. Bungi continued to be spoken into the twentieth century, but by the 1990s, only a few elderly speakers remained (Blain 1992). Because of its isolation and unique roots, Bungi is an excellent test case for the theory of vernacular universals. If certain traits are indeed universals, then one would expect to find strong evidence of their presence in Bungi, just as they are found in other English vernaculars around the world. Despite the lack of contemporary speakers, there are sufficient Bungi data for research of this kind. Interview transcripts from the 1980s provide some examples of native Bungi speech (Blain 1992). There are also stories and anecdotes transcribed from Bungi (Blain 1992; Scott and Mulligan 1951; Stobie 1968, 1971; Walters 1993). Walters’ stories are particularly helpful. Walters was born in 1898 in England, moved to the Red River Settlement in 1904, and became a fluent speaker of Bungi. His Tales of the Old Red River are stories of the lives of Red River farmers in the late nineteenth century. Although not a native speaker of Bungi, his stories give insight into dialect characteristics from the early twentieth century. An excerpt from Walters’ story This Is What I’m Thinkin can be found in Appendix A. Section 2 of this chapter presents evidence that four traits that have been identified as vernacular universals by Chambers (2003, 2004) are found frequently in Bungi. The four traits discussed are: alveolar realization of final unstressed -ing, final consonant cluster simplification, subject–verb nonconcord, and final obstruent devoicing. The presence of these features in Bungi, as well as in such distant and distinct vernaculars as Newfoundland English (Clarke 1997) and African Nova Scotian English (Poplack and
< previous page
page_163
next page >
< previous page
page_164
next page >
Page 164 Tagliamonte 2001), provides strong support for the theory of vernacular universals. Section 3 analyzes three other Bungi features from the vernacular universal approach: the lack of distinction between he and she, the extended use of progressive constructions, and the use of the auxiliary be in perfect constructions in place of standard English have . These traits have not previously been identified as vernacular universals; however, I present evidence that these traits are found in a range of English vernaculars where their presence cannot be accounted for by diffusion. I argue that this approach both expands the catalogue of vernacular universals and provides a new and valuable means to account for the presence of these features in Bungi. Section 4 outlines previous claims that the features discussed in sections 2 and 3 are present in Bungi as a result of language interference, either through contact or shift. The discussion then weighs the relative merits of the language contact and vernacular universal accounts. I conclude that the two approaches are not antithetical: the theory of vernacular universals is valuable in explaining the persistence of such traits in vernaculars worldwide, but should be considered together with the effects of language contact and shift when analyzing the specific dialect features of Bungi. 2. VERNACULAR UNIVERSALS IN BUNGI 2.1 -in’: Alveolar Realization of Final Unstressed -ing There is strong evidence that the alveolar realization of final unstressed -ing is virtually categorical in Bungi: “the full -ing ending does not occur in Bungee” (Blain 1992:236). Walters’ story This Is What I’m Thinkin includes 95 tokens of - in’ and none of the velar nasal (Blain 1992). In Mulligan’s The Shtory of Little Red Ridin Hood , 23 of 26 tokens, or 88 percent, are spelled -in (Scott and Mulligan 1951). Examples from the interview transcripts include sweatin , listenin , a-comin, and a-gettin from Mrs. Adams and a-singin from Mrs. Barnes (Blain 1992). Thus, despite the occasional written transcription as -ing, it appears that the alveolar pronunciation was the norm in Bungi. 2.2 Final Consonant Cluster Simplification There are many examples of final consonant cluster simplification in Walters’ story This Is What I’m Thinkin (Blain 1992:238–45). For example, and is always transcribed as an’, and old is usually transcribed as ol’. Chambers (2003) claims that in standard speech, final consonant cluster simplification is most common before a consonant or semi-vowel and more common after obstruents than after sonorants. The Bungi data show consonant cluster simplification occurring in more environments than in the standard dialects:
< previous page
page_164
next page >
< previous page
page_165
next page >
Page 165 the following phrases from Walters (Blain 1992) show final consonant cluster simplification before vowels (1) and after sonorants (2). (1) an’ it across the fiel’ an’ over the fence (2) aroun’ the, secon’ daughter secon’ plate bes’ roun’-toed moccasins Not all clusters are simplified: there are occurrences of the words hand and end with no apparent simplification in the Walters story, and there is no evidence of final cluster simplification in the Mulligan story. Further, simplification does not appear to be common in the informants’ speech, although Mrs. Adams did display simplification in her pronunciation of Donald and McDonald as Donal’ (or Dondal) and McDonal’ (Blain 1992:65). Because all of the twentieth-century speakers of Bungi were bidialectal with standard Canadian English, there was likely influence of standard English on their speech, as well as some influence of standard English orthography on their writing. Thus, mid-twentieth-century Bungi might not have kept nonstandard features as strongly as earlier speakers of the dialect would have. Certainly the evidence from Walters suggests widespread consonant cluster simplification in early twentieth-century Bungi. 2.3 Subject–Verb Nonconcord Lack of concord between the subject and verb is widely found in Bungi. In Walters’ story, more than 50 percent of verbs in first person have the third singular inflection - s. There appears to be some correlation between nonconcord and specific verbs. For example, the verbs say and take have high rates of nonconcord in first person: 19 out of 20 instances of I say appear as I says (sez ), and 4 out of 5 instances of I take appear as I takes. Walters writes I gets , I talks, I gives , I turns , and I hits , but I hope, I see, I mind, I jog, and I make (Blain 1992:238–45). Nonconcord with plural subjects, resulting in default singulars, is also very common in Bungi. Chambers (2004) proposes a hierarchy for subject constraints on default singulars, from the most inhibiting subject to the least: they > plural NP > we > you > there . Bungi examples can be found from all levels of the hierarchy. The examples in (3), from Walters’ stories, show default singulars with they , which is the most inhibiting subject according to the hierarchy (Blain 1992:153, 176, 173). A default singular with a plural NP from the informant Mrs. Adams is shown in example (4) (Blain 1992:164), and examples of default singulars from Walters with the subjects we and there are given in examples (5) and (6) (Blain 1992:238, 153). These examples of default singulars with subjects from the most inhibiting to the
< previous page
page_165
next page >
< previous page
page_166
next page >
Page 166 least inhibiting on the hierarchy illustrate a significant occurrence of nonconcord in Bungi. (3) And when they save that up, they buys shovels and stuff When they gets away there Maybe they was there yesterday (4) Times is changed my girl (5) we comes, we keeps (6) There’s two more places 2.4 Final Devoicing Stobie (1971:20) notes that final devoicing is characteristic of Bungi speech: “a dock may mean either a dog or a place to tie up a boat”. Blain describes “considerable devoicing of final obstruents” in Mrs. Adams’ speech, with examples of the following words undergoing final devoicing: shindig, hearted, could, you’d , and cottage (Blain 1992:57–9). In the Mulligan story, the wicked wolf is always wickit (Scott and Mulligan 1951), and in the Walters story, one finds Jacop for Jacob and I hits for I heads (Blain 1992:238). Blain (1992:90– 1) further notes that Walters occasionally devoiced final stops and fricatives. It is evident that final devoicing is very common, although not categorical, in Bungi. 3. VERNACULAR UNIVERSALS APPROACH TO OTHER BUNGI FEATURES The data in section 2 strengthens the theory of vernacular universals by providing evidence that these traits are found in Bungi, an English vernacular never before studied from this perspective. At the same time, the theory of vernacular universals provides insight into the Bungi dialect in providing an account for the presence of these features in Bungi. In the following section, I consider three Bungi traits from the approach of vernacular universals. These three traits have previously been noted as characteristic of Bungi, but none has been analyzed from a viewpoint of universal linguistic processes. This approach both expands the theory of vernacular universals and provides a novel account for the presence of these traits in Bungi. 3.1 Nondistinction he/she The substitution of he for she (and occasionally vice versa) is a very striking feature of Bungi. Constructions like my daughter he is coming and my wife
< previous page
page_166
next page >
< previous page
page_167
next page >
Page 167 he didn’t hit me once (Stobie 1971) are common. Walters regularly substitutes he for she, as do the informants interviewed by Blain. The following examples are from Mrs. Adams (7) and Mrs. Hart (8) (Blain 1992:155–6). (7) My father-in-law … she I love listenin er (male priest) John … she she goes by himself he’d (grandmother ) say the same (8) his (f ) hands his (f ) name He’s a widow woman Mesthrie (1992:174) lists “a tendency not to distinguish between male and female for third person pronouns” as one of the “overwhelming similarities among New Englishes of widely divergent L1 backgrounds”; he includes Amerindian Englishes such as Bungi within the category of New Englishes. Because this nondistinction is a widespread feature of vernaculars, this trait can be considered to be a vernacular universal; its presence in Bungi can be attributed to a universal pattern that does not distinguish thirdperson pronouns according to gender. 3.2 Extended Use of the Progressive Constructions be + V+ing In Bungi, the progressive construction is used in many more contexts than in standard English. The following examples from Mrs. Adams’ speech show that the progressive construction can be used for statives (9), habituals (10), and single past events (11) (Blain 1992:182, 186). (9) I’m not wanting a shabby looking purse, my dear (10) We’d be always seeing each other at school We’d be covering up with wool coats (11) The old man’s been passin away (passed away) Oh, somebody been givin (gave) my name An expanded use of progressives is found in many English vernaculars. Blain (1992:183) notes similarities to Ottawa Valley English and Black English, and Mesthrie (1992:174) notes that the use of the progressive for statives is very common in New Englishes, including South African Indian English. Platt, Weber, and Ho (1984) note this extended use of progressives in the English of India, Singapore, Papua New Guinea, West Africa, East Africa, and the Philippines.
< previous page
page_167
next page >
< previous page
page_168
next page >
Page 168 The example in (12) from Black South African English shows the progressive used for past reference (Wade 1997). This can be compared to the Bungi examples from Mulligan’s story in examples (11) and (13), where the progressive is used for single past events (Scott and Mulligan 1951). (12) I was laughing that time when she was saying it (meaning: ‘I laughed when she said it.’) (13) Little Red Ridin Hood’s bin thinking (thought) it’s a good idea sees bin fallin asleep (fell asleep ) the owld wife’s bin dzuympin out (jumped out) Gachelin (1997:34) gives the examples in (14) from Scottish English of the progressive used for the stative present. These are very similar to the Bungi construction in example (9). (14) I am wanting to be present I’m needing a cup of tea The extended use of progressives in English vernaculars around the world provides strong evidence that this feature should be categorized as a vernacular universal. 3.3 Perfect Constructions with be There are very few examples of perfect tenses—that is, constructions including the auxiliary have plus a past participle, in the Bungi data overall, and none at all in Walters’ stories. Rather, the auxiliary be is found with past participles. This is seen both in the perfect progressives described in section 3.3.1 and in the nonprogressive perfect constructions described in section 3.3.2. 3.3.1 Perfect Progressive: be + been + V+ing A typical example of a Bungi perfect progressive with the auxiliary be is shown in example (15) from the Mulligan story (Scott and Mulligan 1951). (15) I’m been taking some lunce to my owld gradmother With third-person singular subjects, it is impossible to determine which perfect auxiliary is being used because the contracted forms of both is and has are ’ s. This is illustrated in example (16) from Mulligan’s story, where the -s could represent either is or has. I found only one unambiguous have auxiliary with the perfect progressive in the Bungi data, in the same story,
< previous page
page_168
next page >
< previous page
page_169
next page >
Page 169 given in example (17). As in the examples (11) and (13), the progressive construction is used here to express simple past-tense meaning. (16) the dahrs bin flying open he’s bin so greedy (17) Red Riding Hood’s mother has been putting (put) a bannock 3.3.2 be + Past Participle Section 3.3.1 shows the auxiliary be used with perfect progressive constructions, where the auxiliary is followed by the past participle been and a present participle. The following examples show the inflected auxiliary be in nonprogressive perfect constructions. Example (18) from Walters has the past participle been (Blain 1992:177); in example (19), the past participle is the Scottish English slocked ‘put out’ or ‘snuff’ (Scott and Mulligan 1951). (18) they’re not been up to get any (19) Awe Willie, I’m just slocked it the light The past participle found most frequently with be in Bungi is got, occurring in all persons: I’m got, you’re got, we’re got, and they’re got. As noted earlier, it is not possible to tell whether he’s got represents he is got or he has got. However, given that all other persons, including third person plural, are conjugated with be, it is reasonable to conclude that be is also the auxiliary for third person singular. These constructions are illustrated with positive sentences in example (20) (Scott and Mulligan 1951:43) and negative sentences in example (21) (Blain 1992:177). The negative sentences differ further from standard English in that there is no do-insertion and there is the occasional double negative. (20) I’m got to get back hom I’m got Jane Mary’s bodice on I’m got the horse tied upset the Hotel (21) What if they’re not got no dolly? I’m not got that big money. You’re not got your fine boyish figure. It appears that paradigm leveling has occurred with Bungi auxiliaries, leading to an almost universal use of be as the auxiliary for periphrastic verbal constructions. This could be the result of overgeneralization from the very common progressive constructions for which be is the auxiliary. A misinterpretation of the contracted form -s of has as representing is might
< previous page
page_169
next page >
< previous page
page_170
next page >
Page 170 also have contributed to the spread of be to perfect constructions. The use of be with perfects has not been identified as a common feature in English vernaculars and so cannot be considered as a possible vernacular universal at this time. However, the approach of attributing dialect features to universal linguistic processes is valuable here because the presence of this feature in Bungi can be attributed to the universal process of paradigm simplification. 4. VERNACULAR UNIVERSALS VERSUS LANGUAGE INTERFERENCE The Bungi features described in sections 2 and 3 have been presented as resulting from universal tendencies found in vernaculars around the world. However, one can also argue that the presence of these traits in Bungi is a result of language interference: shift from, or contact with, other languages or dialects. The Bungi dialect arose from the contact between Scottish English-speaking traders and Native American languages, primarily Cree. Thus, the development of the dialect involved both a shift from Scottish English and contact with Cree. Further, many of the traders had Scottish English as their second language: Gaelic was their first language and continued to be spoken in Manitoba into the twentieth century. Thus, Gaelic could have influenced Bungi both through shift, because the speakers may have come to Bungi directly from Gaelic, and through contact. Further, Scottish English had been influenced by Gaelic long before it arrived in the New World. -in’: Alveolar Realization of Final Unstressed -ing Blain attributes the presence of this trait in Bungi to its earlier presence in Scottish English dialects. She gives the example of maken for making found in a letter written by an Orkney man in Canada in 1822 (Blain 1992:72). Final Consonant Cluster Simplification This feature, although recognized as widely occurring in English vernaculars, has not been mentioned elsewhere as a particular trait of Bungi; there has been no discussion attributing its presence in Bungi to Scottish English, Gaelic, or Cree influence. Subject–Verb Nonconcord/Default Singulars Nevalainen (2006) claims that the use of is and was with plural subjects was a typical Northern English Dialect feature of the fifteenth and sixteenth centuries. This feature could therefore have been present in the Scottish English spoken by the Hudson Bay traders and could have continued into Bungi through shift. Final Devoicing The presence of final devoicing in Bungi has been attributed to the influence of both Scottish English and Cree (Blain 1992; Stobie 1971). Stobie claims that Cree does not distinguish between voiced and voiceless stops, and that this lack of distinction affected Bungi voicing. She supports her claim of Scottish English influence on devoicing with an example from ‘the Gaelic-English speaker’ in Shakespeare, Sir Hugh Evans, who says Got
< previous page
page_170
next page >
< previous page
page_171
next page >
Page 171 pless you (Stobie 1971:20). Blain (1992:68) claims that devoicing was typical of Highlands, Lowlands, and Orkney English and gives the examples sprink ‘spring’ and thinkes ‘things’ from the same Orkneyman’s letter (Blain 1992:72). Blain (1992:61, 125), like Stobie, suggests that the devoicing in Cree and Scots English reinforced each other in Bungi. Nondistinction He/She Both Stobie (1971) and Blain (1992:156) attribute the lack of distinction between he and she to Cree interference. Stobie claims that it results from interference from Cree grammar which contrasts animacy rather than gender, and Blain notes similar Cree influence in Michif. Progressive Constructions be + V+ing Blain (1992:183) claims that the extended use of progressive constructions in Bungi arose from the influence of Gaelic periphrastic forms. Similarly, Gachelin (1997) attributes the worldwide phenomenon of extended use of the progressive construction to the influence of substrata languages, including Gaelic: Aspect is more important than tense in many substrata that may influence varieties of English, ranging from the Celtic fringe of Western Europe to African and Eastern languages like Hausa Swahili, Indo-Aryan, Chinese and Malay. (Gachelin 1997:33) Lamont (2005) notes that the progressive was frequently used for habitual meanings in Early Modern English and provides example (22). He argues that there was aspectual cross-over between the simple and progressive forms until the early twentieth century, giving example (23) from Keats, written in 1819. Therefore, the extended use of the progressive construction to express habitual actions and simple past tense may have come into Bungi through language shift; such extended use may have already been present in the eighteenth-century Scottish English from which Bungi developed. (22) She is always seeing Apparitions and hearing Death-Watches (23) What I should have lent you … was belonging to poor Tom. Perfect Constructions with be Blain (1992:195) claims that the Bungi construction of be got “clearly reflects Scots English” and gives example (24) from the letters written by Orkneymen in Canada (1992:190). (24) I am got two geldron ‘children’ I am got not noues to inform you of Rydén and Brorström (1987) claim that the be perfect was very common in eighteenth- and nineteenthcentury English, forming 80 percent of the
< previous page
page_171
next page >
< previous page
page_172
next page >
Page 172 perfects in 1700 and 60 percent in 1800. They give example (25) taken from early eighteenth-century letters (Rydén and Brorström 1987:103). (25) I am now got as far as my Mare will carry me. They are got married. It is possible then that the Bungi be perfects represent not an innovation, but a retention of an older form, carried into the dialect through shift. However, it should be noted that all of the English examples in Rydén and Brorström (1987) are intransitive, where the perfects with be represents a state, result, or process. Filppula (1999:116) also notes that the be perfect in Hebridean English occurs only with intransitive verbs. In Bungi, by contrast, these constructions are often transitive, as in the sentences showing possession and completed action in examples (20) and (21). Therefore, shift alone cannot account for the Bungi be perfect constructions. 5. DISCUSSION Table 7.1 summarizes the seven Bungi features discussed in this chapter. The first four traits have previously been described as vernacular universals (Chambers 2003:265–6). The next three are those I suggest benefit from a vernacular universal approach because their presence in Bungi can be attributed to universal processes. Two of these, he/she nondistinction and the extended use of progressives, are widely reported in New Englishes, and Table 7.1 Summary of Bungi Features Vernacular Universals* Proposed Vernacular Universals Language Interference Cree Scots Gaelic -in’ x x CC simplify x non-concord x x final devoicing x x x he <-> she x x extended progressives x x be perfect x *see Chambers (2003:265–6)
< previous page
page_172
next page >
< previous page
page_173
next page >
Page 173 I propose that they should be added to the roster of vernacular universals. The final column of Table 7.1 shows the language interference that has been associated with each of these traits, reflecting either language contact or language shift. The question then remains: are these characteristics best accounted for through a theory of vernacular universals or through processes of language interference? Are there universal processes of phonology, semantics, or syntax that can account for these traits’ presence in Bungi, or is it best to look at the history of language contact and language development to explain their presence? For many of these traits, these two influences are tightly intertwined, and it is difficult to separate the strands. For example, one could account for final devoicing in Bungi on the basis of language shift because final devoicing was a feature of the Scottish English that formed the basis for the Bungi dialect. However, this does not settle the question because Scottish English is a vernacular: how can we account for the final devoicing in that dialect? Here, the theory of vernacular universals provides a good account for devoicing in Scottish English: it reflects a universal tendency to devoice final consonants. This means that final devoicing in Bungi results from both a universal tendency to final devoicing and language shift. Devoicing in Cree might also be a result of the universal tendency to devoice; through language interference, the Cree devoicing could also have reinforced the tendency to devoice in Bungi. This interaction of universal tendencies and language interference also could have given rise to the extended use of progressives in Bungi. Extended progressives are found widely in New Englishes around the world, suggesting a universal pattern of aspectual expression. Yet there is also evidence that the Gaelic substratum has influenced the wider use of progressives in Irish and Scottish English. Perhaps the extended use of progressives in Bungi results both from Gaelic influence, either directly through contact with Gaelic speakers in North America or indirectly through earlier Gaelic influence on Scots English in Britain, and from a universal process of aspectual expression. Filppula, Klemola, and Paulasto (Chapter 10, this volume) attribute stative and habitual uses of progressives in regional British and Irish dialects of English to both vernacular universals and Celtic influence. Thomason (Chapter 15, this volume) also argues that these two approaches are not exclusive and proposes a theory of multiple causation. 6. CONCLUSION The theory of vernacular universals proposes that certain traits appear frequently in a wide range of English vernaculars on account of universal linguistic tendencies. The evidence given here from Bungi provides further support for this theory, in that four of the traits proposed to be vernacular universals have a strong presence in this isolated dialect.
< previous page
page_173
next page >
< previous page
page_174
next page >
Page 174 The theory of vernacular universals can be extended to account for the presence of other linguistic features in Bungi. I have argued that such an approach is valuable in analyzing be perfects, the extended use of progressive constructions, and the lack of gender distinction in third-person pronouns. The latter two traits are found in many other English vernaculars, including other Amerindian languages (Flanigan 1987:350), and I have proposed that they be added to the roster of vernacular universals. The theories of language interference and vernacular universals are not mutually exclusive. The theory of vernacular universals can explain the appearance of the same trait in different English dialects around the world; to account for characteristic traits of a single dialect, theories of language contact and shift are invaluable. A specific linguistic trait can have more than one source: the universal processes reflected in vernacular universals and the forces of language interference can interact in ways that reinforce one another. In Bungi, the Scottish English, Cree, and Gaelic influences have interacted with universal tendencies to create a unique set of dialect features. REFERENCES Blain, E. 1992. The Bungee Dialect of the Red River Settlement . Unpublished M.A. thesis. University of Manitoba. Chambers, J.K. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Significance . Oxford: Blackwell. ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective , edited by B. Kortmann, 127–45. Berlin: Mouton de Gruyter. Clarke, S. 1997. Language in Newfoundland and Labrador: Past, present and future. Journal of the CAAL 19(1–2): 11–34. Filppula, M. 1999. The Grammar of Irish English: Language in Hibernian Style. London: Routledge. Flanigan, B.O. 1987. Language variation among Native Americans: Observations on Lakota English. Journal of English Linguistics 20(2): 181–99. Gachelin, J.-M. 1997. The progressive and habitual aspects in non-standard Englishes. In Englishes Around The World. Volume 1: General Studies, British Isles, North America, edited by E.W. Schneider, 33–46. Amsterdam: John Benjamins. Lamont, G. 2005. The progress of English verb tenses and the English progressive. Online. Available http: <www.chass.utoronto.ca/~cpercy/courses/6362-lamont.htm> (accessed 27 November 2006). Mesthrie, R. 1992. English in Language Shift. Cambridge: Cambridge University Press. Nevalainen, T. 2006. Default singulars with existentials in the normative eighteenth century. Paper presented at World Englishes: Vernacular Universals vs. Contact-Induced Change, Mekrijärvi, September 2006. Platt, J., H. Weber, and M.L. Ho. 1984. The New Englishes . London: Routledge and Kegan Paul. Poplack, S., and S. Tagliamonte. 2001. African American English in the Diaspora . Oxford: Blackwell. Rydén, M., and S. Brorström. 1987. The be/have Variation with Intransitives in English. Stockholm: Almqvist and Wiksell.
< previous page
page_174
next page >
< previous page
page_175
next page >
Page 175 Scott, S. O., and D.A. Mulligan. 1951. The Red River Dialect. The Beaver 282: 42–5. Stobie, M. 1968. Backgrounds of the dialect called Bungi. Historical and Scientific Society of Manitoba Transactions III-24: 65–75. ———. 1971. The dialect called Bungi. Canadian Antiques Collector 6: 20. Wade, R. 1997. Arguments for Black South African English as a distinct ‘new’ English. Fourth International Conference on World Englishes . Online. Available http: <www.und.ac.za/und/ling/archive/wade-03.html> (accessed 27 November 2006). Walters, F. 1993. Pieces of the Past: A Collection of Tales of the Old Red River. Winnipeg: Bindery Publishing House. APPENDIX This Is What I’m Thinkin [Excerpts from a story told by F. J. Walters, transcribed into phonetic symbols and annotated by Blain (1992:228–45), here transcribed into an English version that retains Bungi features] It’s kinda darty out this marnin, so after feedin hup at the stable, I was settin along the stove havin a warm and dunin mi siken plate of bustin* and thick cream. It’s a little on the barnt side, I thought. I hope see has better luck with ’er nets batch an’ not to barn it so muts. This is what I’m thinkin—when all ass once there comes a knock on the dur, so I sez to the Missus, “see who’s that.” An’ this is my neighbour, Jamesie, now, so I sez to him, “Come, sit in, there’s lots of tea.” But right away ’e sez to me, “I’m got no time, b’y. I’m got a sick horse an’ I can’t get him op, b’y.” That’s ’at ’e said when he said that to me. So I sez to him, “Hol’ on now. I’ll git on me hat an’ coat an’ come along with yuh an’ take a looksee.” Now all this is in the late springtime, you’ll see. The wind’s a-whistlin an’ fairly blowin us offen our feet and nigh on takin our brith away out of us. Now we’re up to our knees in gutter in places and it’s tough goin— but we keeps to the lee of the bush as best we kun but, till we comes to the gully leadin op to Jamesie’s place and aboot op’sit aunt, ol’ Aunt Jennock’s. An then, as we’re passin, see calls out to me to call along on my way back and take a fres bannock for me an’ the Missus—made with hard grease. B’y, that’s ‘at’s gud along with nipi* jam. Now ’e comes polin across now an’ when ’e gits close, I see this is ol’ Buru. “Where’d ya git that ol’ tub?” I sez to ’im. “I expected it t’ turn apichekwani* an’ ya go chimuck* into the water.” He only sez, “Daa,” but, an ’e hol’s up the back of ’is hand to me and nods ’is head. Well, ’at’s a habit ’e’s got to say ’e’s too smart with a boat for a thing like that to happen. Well, we talks for a little ’boot the catfis ’e’s bin catchin lately,— then ’e puts me
< previous page
page_175
next page >
< previous page
page_176
next page >
Page 176 straight on the news aroun’. Seems like old Jicup’s (Jacob’s) siken’ dorter’s been thinkin on gittin married, ’e sez. ’is fers’ doter’s man not can git ’head so good since ’is coo died. “Too bad to lose a good animal like that,” I sez. “Well, ’iz siken’ dorter’s makin a better match, but,” ’e sez. “oredi ’er man has a cow, an’ a sack of flour, an’ a house. An’ even, he’s got a winda in ’is upstairs. * apichekwani —upside down (Plains Cree) * bustin—cracked wheat roasted and cooked as porridge * chimuck —The splashing sound made when something is dropped in the water. (Cree) * nipi —jam made from high bush cranberries (Plains Cree)
< previous page
page_176
next page >
< previous page
page_177
next page >
Page 177 8 The Regularisation of the Hiatus Resolution System in British English A Contact-Induced ‘Vernacular Universal’?1 David Britain and Sue Fox 1. INTRODUCTION Jack Chambers’ research on ‘vernacular universals’ (e.g. 2000, 2003, 2004) has highlighted the fact that sociolinguistic dialectology hasn’t really fully contributed to the debate on typology and language universals, but that it should, and that it has a wealth of evidence at its disposal to do so. Rather than restricting our vision to our own speech communities, he says, we should be more prepared to “expand the domain of enquiry across national borders … and especially across language borders” (2000:11) and to seek out and explore variables which “have universal, cross-linguistic counterparts, identifiable with structurally equivalent linguistic constraints in language after language” (2000:13). As a first step in this endeavour, Chambers highlights a number of features which appear to be used in many varieties across the Anglophone world (these include alveolar nasals being used in unstressed -ing; consonant cluster simplification and multiple negation) and which cannot possibly be linked through diffusion from one common source. He suggests that these forms are most likely to be found in child language, interlanguage, pidgins and creoles, and in working-class vernaculars, that they tend to be suppressed in standard varieties, and that they “appear to be natural outgrowths … of the language faculty, that is the species-specific bioprogram that allows (indeed, requires) normal human beings to become homo loquens ” (2004:128). In his most detailed account of vernacular universals to date, Chambers (2004) suggests that contact and ‘worldliness’ militate against the preservation of such universals—“put simply, the more urban and mobile the social setting, the more standard the speech” (2004:137). To support this claim, he contrasts, for one vernacular universal—the use of WAS for standard WERE as past-tense BE forms—(1) older speakers on Tristan da Cunha, a long isolated remote island of the South Atlantic; (2) the Fens in England and Anniston in the United States, both at the heart of large rural areas; and
< previous page
page_177
next page >
< previous page
page_178
next page >
Page 178 (3) York (England) and Sydney (Australia), “large cities in highly urbanized regions” (2004:138). Tristan da Cunha shows most use of nonstandard WAS, and the larger cities the least, with the Fens and Anniston in between, showing a correlation between ‘urban complexity’ (2004:138) and avoidance of the vernacular universal. In this chapter, we argue that contact (which, after all, was central to the genesis of contemporary Tristan da Cunha [Schreier 2003] and Fenland [Britain 1997, 2002] Englishes), rather than being viewed as antithetical to the use of linguistic forms that are used widely across unrelated dialects of English, should in fact (sometimes at least) be seen as a generator of them. Pidgins and creoles, suggested by Chambers as one of the sites of vernacular universals, are, of course, contact varieties par excellence, and so perhaps we should expect to find common forms emerging where we find (heavy) contact in language generally, especially where the standard language has had little influence in the development of the variety. Of course, it is not unusual in the literature to find connections between language contact, child and second-language acquisition, and the appearance of forms that are considered widespread, unmarked, and even universal (see e.g. DeGraff 1999). We examine here the example of how varieties of English traditionally resolve vowel–vowel hiatus. We highlight changes afoot in the hiatus breaking system, most notably in varieties that have undergone heavy sociocultural and language contact. These changes appear to be moving the language towards a levelled system that is shared not just by extremely diverse (and demographically unconnected) varieties of English, but, it seems, by language(s) more generally. 2. HIATUS RESOLUTION IN TRADITIONAL VERNACULAR ENGLISH ACCENTS Vernacular varieties of English have a complex system of resolving vowel– vowel hiatus.2 Most have a range of strategies at their disposal—strategies which are determined by a range of linguistic contextual factors (such as the quality of the first vowel and the grammatical status of the lexical item containing that first vowel), which, as we will see, are often the result of relics and residues of (often unconnected) historical processes that have been underway in the varieties in question for centuries. There has been very little empirical investigation of these strategies, however, and most research which has been carried out has considered just one of the hiatus-resolving mechanisms, rather than viewing them as a global system (although see Stene 1954 for an early overview). We begin, therefore, by outlining the different components of the hiatus-resolution system in vernacular accents of English before examining evidence that this system is undergoing change and speculating about what may be triggering that change.
< previous page
page_178
next page >
< previous page
page_179
next page >
Page 179 In contexts of V [+high] # V, such as in examples (1) and (2), the literature mostly uncontroversially accepts that a glide—either [w] if the high vowel is back and/or rounded or [j] if it is unrounded and front—eases the transition from vowel to vowel. Cruttenden (2001:289) calls this “intrusive [j w]” (see also Thomas 2001:54–5): (1) Go inside (2) Jelly and ice-cream Harris (1994:104) argues that “the hiatus glide can be straightforwardly explained as the spreading of an element from the first nucleus into the vacant onset”. Heselwood (2006:80) considers these segments to be “low-level articulatory transitional phenomena”, unlike other hiatus-resolving phenomena such as linking and intrusive /r/ (see below). As we will see, this distinction has consequences not only for the acquisition of the hiatus-resolution system, but also for the route along which the system appears to be currently changing. In contexts of V [-high] # V, such as in examples (3)–(9), most nonrhotic vernacular varieties of English insert /r/. This phenomenon is often called linking /r/ (if the first vowel is the consequence of the historical loss of rhoticity) (as in examples [3]–[5]) or intrusive /r/ (if there are no etymological traces of in the word) (as in examples [6]–[8]). Intrusive /r/ can occur after / / in nonrhotic accents of English, but also in other contexts in some nonstandard varieties, for example, /au/ realised as a non-high-long monophthong (e.g. in London [Wells 1982] and the East Anglian Fens [Britain 1991, 2003]) can trigger intrusive /r/ before a following vowel (see example [9]): (3) cider apple (4) far away (5) more apples (6) vodka and tonic (7) awe inspiring (8) keep the window open (non-standard Southern British English) (9) now and then (East Anglian Fens)
< previous page
page_179
next page >
< previous page
page_180
next page >
Page 180 The phonological status of intrusive and linking /r/ has been subject to a very vigorous debate in the theoretical phonological literature (e.g. Halle and Idsardi 1997; McCarthy 1993, 1999; McMahon 2000; Orgun 2001; Sebregts 2001; Uffmann 2007a, 2007b; Vennemann 1972), but there has been little empirical investigation into the frequency of its occurrence, its exact phonetic quality, or the linguistic constraints that encourage or discourage its use (see Bauer 1984; Foulkes 1997; Hay and Sudbury 2005; Heselwood 2006 for salient exceptions to this). One context in which intrusive /r/ is not found, except occasionally in early child language, is after the word ‘the’. The definite article is one of three contexts (along with the indefinite article and a group of often unstressed small function words—see below) where allomorphy can be triggered to help resolve hiatus— different forms are found before vowels than before consonants. In the case of the definite article, preconsonantal /ðə/ alternates with prevocalic /ði/3 as in examples (10) and (11), with the high front vowel of the prevocalic allomorph triggering an intrusive [j] before the following vowel: (10) the pear [ðə pεː] (11) the apple [ðijæpł] What is today the definite article was, at an earlier stage of English, a demonstrative pronoun. Over time, it lost its demonstrative function and became a marker of definiteness. The present-day /ðə/ pronunciation was derived from /ði/ via a process of vowel reduction and has become grammaticalised in most modern vernacular varieties of English before consonant-initial but not vowel-initial words (Raymond, Fisher, and Healy 2002). The prevocalic form represents relative conservatism, therefore, possibly in order to avoid the hiatus that would result were it to continue on the same track as preconsonatal /ðə/.4 A more unusual and (today) isolate form of hiatus resolution through allomorphy can be found in the indefinite article system, where a [ə] is used preconsonantally but an epenthetic [n] leads to an [ən] in prevocalic contexts, as in examples (12) and (13): (12) a pear [ə pεː] (13) an apple [ən æpł] This is the only hiatus-breaking element in English which is represented in the written form. The forms a and an are both derived from the Old English numeral an, equivalent to modern ‘one’, through a process of grammaticalisation (Hopper and Traugott 1993). The form an as an indefinite article lost its final /n/ in preconsonantal position, resulting in the distribution of
< previous page
page_180
next page >
< previous page
page_181
next page >
Page 181 a and an in Standard Modern English. This alternation is, as with the definite article allomorphy, somewhat of a historical fossil and parallels other cases found in Middle English. These include such alternations as my/mine, thy/thine, and no/none, where the shorter form was originally used before consonants and the longer retaining form was used before vowels (and phrase finally) so that my book but mine eyes exemplify earlier usage. The distribution of these pairs then ultimately came to be determined not by the nature of the following segment, but simply by whether it was final (or absolute) or not (Stene 1954:56). The final context in which allomorphy is triggered in order to avoid hiatus is a group of small, usually unstressed function words which include to, my, I, by , of , and you. These words often, when unstressed and in preconsonantal position, end in a non-high vowel, as in examples (14)–(18): (14) go to London (15) blown down by the wind (16) he can help you lift it (17) get my bag (18) I went home (19) cup of tea However, in many varieties, when in prevocalic position, they have allomorphs which are high-vowel final in order to trigger an intrusive [j] or [w], as in examples (14a)–(18a) (of provides an interesting case where [v] blocks hiatus in prevocalic positions—see examples [19] and [19a]5)6: (14a) go to Ipswich (15a) blown down by a tornado (16a) he can help you eat it (17a) get my apron (18a) I ate at home (19a) cup of Earl Grey A few varieties, notably ones which have been resolutely nonrhotic for a considerable period, can break hiatus in some of these small function words in a different way, using intrusive /r/—examples (14b)–(16b).
< previous page
page_181
next page >
< previous page
page_182
next page >
Page 182 (14b) go to Ipswich (15b) hit by a bus (16b) he can help you eat it So vernacular Englishes have a complex hiatus-breaking system involving in different but sometimes overlapping linguistic contexts, often facilitated by allomorphy in the articles and in a number of small but frequently occurring function words. 3. EVIDENCE OF CHANGE IN THE HIATUS-RESOLUTION SYSTEM Perhaps not too surprisingly given its complexity, however, the system described earlier is undergoing reorganisation in some varieties of English. Evidence for this reorganisation is quite patchy, and, like hiatus resolution, is very rarely presented for the system as a whole. We therefore draw together here the evidence from across the hiatus-breaking system before presenting a holistic analysis comparing rural adolescent Fenland English and the traditional Cockney London English of older speakers—varieties which retain the traditional system almost totally intact—with adolescents from the East End of London where radical reorganisation is in progress. Variable lack of allomorphy in the article system has caught the attention of a number of scholars, although few have conducted detailed empirical investigations nor explained how hiatus is resolved in such contexts. Wright (1905:71) claimed that “few dialects follow the rule of the literary language according to which ‘an’ is used before a vowel”, and across traditional regional varieties of English in England there is some evidence to support this view. Ojanen’s (1982) research on traditional dialects of Southern Cambridgeshire in the East of England provides example (20): (20) ‘when we went in there … in that cottage, there’d be a old chap (Ojanen 1982:186) and she adds that “the indefinite article ‘a’ is commonly used instead of the Standard English ‘an’ ” (1982:186fn). Staying in the East, both Peitsara (1996) and Claxton (1954) find a lack of allomorphy for the definite article in Suffolk and Britain (2003) finds a + vowel used by a speaker of Gypsy ethnicity in the Fens (21) (21) ‘there’s a agency in the local pub’ (Britain 2003:203) Prevocalic indefinite article a is also found in Bolton in the North-West (Shorrocks 1999:45), in Peasmarsh in Sussex in the South-East (Lodge
< previous page
page_182
next page >
< previous page
page_183
next page >
Page 183 1984), and across the South-West of England (Wagner 2004:155). Shorrocks (1999:45) is one of few who mention how hiatus may be avoided in such contexts, when he claims “when /ə/ is used before a vowel/diphthong, a glottal stop may well be interposed between the two”. Beyond the UK, Taylor (2003) finds a + vowel in written court transcripts of defendants in 1850s Sydney, Australia, Wolfram and Fasold (1974), Mufwene (2001), and Craig, Thompson, Washington, and Potter (2003) mention it as a characteristic of AAVE, and Labov (1972) finds it in New York. Watermeyer (1996:118) notes that for speakers of Afrikaans English in South Africa, “there is usually no distinction between a and an (a energetic class; a absolute miracle )”. Lass (1995:104) suggests that, although prevocalic a is restricted only to working class, especially Afrikaaner speakers, prevocalic [ðə] is also found among middle-class speakers in South Africa. Prevocalic [ðə] is reported for Singaporean English by Fraser Gupta (personal communication). More detailed and systematic studies of this absence of allomorphy are few, with several originating in the psycholinguistic literature and experimental in nature, rather than based on the more informal datacollection methods typical of sociolinguistics (see Fox 2007 for more details). Healy and Sherrod (1994) tested, in an experimental task, whether article use is governed by a rule-based distinction sensitive only to the onset of the following word. In one experiment, participants were presented visually with nouns and adjectives that began with either a consonant or vowel. They were then asked to select the form of the articles, both definite and indefinite, that they preferred to use with the visually presented words. For the indefinite article, the participants chose a/an in accordance with the standard rule but for the definite article, 25 percent of the responses deviated from the normative rule. However, this figure included both instances of reduced vowels in prevocalic position and unreduced vowels in preconsonantal position, and no figures were provided that would enable the reader to disambiguate the two contexts. Raymond, Fisher, and Healy (2002), in another experiment, examined the factors influencing the production of articles, this time in tasks which did not involve exposing the participants to written stimuli. Each participant was asked to listen to words preceded by the possessive pronoun ‘his’ which they then had to replace with an article, either ‘the’ or ‘a’/‘an’, followed by the word they had heard with the possessive pronoun. For example, the participant might hear ‘his business’ and respond by saying ‘the business’. Results showed that indefinite article choice corresponded closely with the standard rule. For definite articles, however, particularly before vowels, production did not consistently follow the standard rule. For example, more than 40 percent of the responses before vowel initial words did not fit the standard rule of using an unreduced vowel in this context. Gaskell, Cox, Foley, Grieve, and O’Brien (2003), investigating the use of prevocalic /ði/ by children found that this form was only used in 34 percent of all relevant contexts by 8-year-olds, but that this rose to 61 percent among 10-year-olds, suggesting that the allomorphy is picked up relatively late
< previous page
page_183
next page >
< previous page
page_184
next page >
Page 184 during language acquisition. We will return to matters of acquisition later. Todaka (1992) examined the use of prevocalic /ðə/ in the TIMIT corpus,7 finding it in around a quarter of all tokens. In cases where a reduced vowel was used in prevocalic position, Todaka noted significant variation among speakers based on age and dialect. Speakers from the South and New York City were shown to have the highest percentages of prevocalic /ðə/. Age was also significant, with no speaker over 50 years old pronouncing a prevocalic definite article with /ðə/. The highest use of the definite article with /ðə/ in prevocalic position was among speakers aged 20–30, leaving Todaka to conclude that “there is a definite trend among the young that the general distinction, i.e. [ðə] before consonants and [ðiː] before vowels, is becoming less obvious” (Todaka 1992:43). Keating, Byrd, Flemming and Todaka’s (1994:136–7) observations of Californian undergraduates suggest that “the norm seems to be [ðə] before a consonant and [ðəː] before a vowel” and that there “appears to be an ongoing change in a pronunciation norm”. We are aware of two variationist studies of allomorphy loss in the article system. Anderson et al. (2004) investigated change in progress towards prevocalic /ðə/ in New Zealand English. They found around 20 percent of prevocalic definite article tokens realised as [ðə] and found that the change was being led by young nonprofessional women and that it was most advanced preceding repeated head nouns, which contain stressed, front vowels. Ash and Myhill’s (1986) study investigated lack of allomorphy in the context of social network ties in Philadelphia. They found that African Americans who had little contact with Whites in Philadelphia realised prevocalic the as [ðə] 75 percent of the time compared with 30 percent among Whites with considerable contact with African Americans. Varieties which avoid or have low levels of hiatus-breaking intrusive and linking /r/, although they are nonrhotic, are fairly widely reported, especially outside England. Foulkes’ (1997) variationist study of linking and intrusive /r/ in Newcastle is an extremely important one not only because it finds these forms to be much less categorical than had previously been assumed, but also for the somewhat unexpected social patterning of results. Linking /r/, across apparent time, was found to be on the decrease in Newcastle and used more by middle-class speakers (Foulkes 1997:263). Intrusive /r/ is rare in natural conversation, and Foulkes (1997:264) found that even when the appropriate contexts occurred, it tended to be avoided. When it was used in conversation, it was largely by working-class speakers. In reading passage style, however, the use of intrusive /r/ increased and was used much more by middle-class than working-class speakers (Foulkes 1997:266), although these findings were based on just one sentence per speaker. Foulkes (1997:262fn) mentions that when linking and intrusive /r/ are not used “in some cases … speakers insert a glottal stop”. In South African English (SAfEng), a glottal stop is also reported in place of /r/ (Bowerman 2004; Lass 1995). L1 varieties of SAfEng are said to “have
< previous page
page_184
next page >
< previous page
page_185
next page >
Page 185 a greater tendency to avoid /r/ sandhi by using a glottal stop than non-rhotic dialects” (Lass, cited in Watermeyer 1996:7, see also Lass 1995:103). Trudgill and Hannah (2002) give the example four o’clock , confirming the use of glottal stop and agreeing that “very many varieties of SAfEng lack both intrusive r and linking r ”. Hay and Sudbury (2005) report that linking /r/ was frequent (85 percent), but that intrusive /r/ was relatively infrequent (21 percent) in nineteenth-century New Zealand English, and in modern NZE, Bauer and Warren (2004) suggest that intrusive /r/ is (still) a minority option. Tay (1982:138) claims that “linking r is hardly ever found” in nonrhotic Singaporean English (see also Trudgill and Hannah 2002) and adds “other forms of liaison were also noticeably absent … the absence of liaison is perhaps what gives the layman the impression that SgE is very ‘staccato’”. Low levels of linking /r/ are also reported by Labov (1972:13, 39) for New York African Americans, and Cutler (1999:431) has shown that lack of /r/ was adopted by a White middle-class teenager as a ‘crossing’ (Rampton 1995) strategy, as in example (22): (22) Mike: yo, she still looks her age (Cutler 1999:431) In the U.S. South (where the definite article is also reported to be commonly pronounced with a reduced vowel in prevocalic position—see above), Thomas (2004:317) reports that “linking r … has historically been absent for a large number of Southerners, though some speakers showed it, often variably. Intrusive linking r in other hiatus positions … is virtually unknown in the South” (see also Kurath and McDavid 1961:171–2, Maps 157–159; Trudgill and Hannah 2002). The literature on hiatus breaking after non-high vowels, therefore, suggests quite strongly that, in many varieties, particularly those that have developed in circumstances of high levels of sociocultural and language contact, glottal stops can either replace or covary with /r/. Beyond Stene’s (1954) comprehensive work on sandhi phenomena, few other scholars have examined the allomorphy of unstressed function words from the perspective of hiatus resolution. A number of varieties of English, however, again those where language contact has been central to their development, have been examined from the perspective of changes from stress timing to syllable timing and a general avoidance of ‘weak forms’. Most research in this regard has been carried out on New Zealand English (NZE), where it is particularly associated with speakers of Maori English (Bauer 1995; Benton 1966; Warren 1998, 1999; Warren and Britain 1999). Research by Helen Ainsworth and Janet Holmes has examined the use of ‘full’ vowels where ‘reduced’ vowels might be expected in Maori and Pakeha NZE (Ainsworth 1993; Holmes 1997; Holmes and Ainsworth 1996,
< previous page
page_185
next page >
< previous page
page_186
next page >
Page 186 1997) and found full vowels to be more frequently occurring in Maori than in Pakeha (=European) NZE, but also in more slow speech styles than in informal fast speech.8 Crystal (1995:176) argues that ‘World English’ shows evidence of syllable timing and proposes that it is due to English’s “contact with a range of languages of diverse structural types”. Evidence of this is found, he claims, in the Englishes of South Africa, India, and the people of Italian and Hispanic ethnicity in the United States, as well as in English-lexicon pidgins and creoles. Tay’s (1982:138) work on Singaporean English supports this view. She maintains that “Weak forms are also strikingly absent in SgE”, adding, interestingly, that the only weak forms consistently realised as weak were the articles /ðə/ and /ə/. Similarly, Deterding (2006:183) reports an almost total absence of ‘weak’ forms and a syllable-based rhythm in Chinese English. The ‘full’ forms reported for NZE and beyond, of course, can occur in preconsonantal positions as well as prevocalic, but none of the studies reported above appears to have investigated whether such full forms are more likely to occur in prevocalic contexts than elsewhere, nor reported how hiatus is resolved if such full forms trigger a -V#V- context. Generally, then, the hiatus-resolution system of English appears to be extremely complex and variable. In Britain, especially, for example, /j w r n v/ can all act as hiatus breakers in different contexts,9 and the complexity of this system triggers, in most places, following segment-sensitive allomorphy in the article system and among a set of frequently occurring small-function words. Outside Britain, especially, we find evidence of a relatively more levelled system, notable within which is the presence, albeit variable in some places, of glottal stops as an all-purpose hiatus breaker. We aim to consider now the extent to which SouthEastern England is undergoing such changes to its patterns of hiatus resolution. 4. HIATUS RESOLUTION IN THE FENS In order to examine the extent to which vernacular varieties of English adhere to the complex traditional hiatus-breaking system described by Stene (1954) and others, and presented previously, we first analysed a corpus of data from the Fens in South-Eastern England. The Fens are located 150 km north of London and are an area of low-lying former marshland, reclaimed from the mid-seventeenth century onwards, separating East Anglia from the Midlands. In terms of demography, they are, in comparison with the rest of South-East England, very sparsely settled, and with an almost entirely White population (98.6 percent at the 2001 Census).10 We examined recordings of informal conversation from 10 adolescents (aged 15–19) of British White ethnicity from the Fens, extracting the definite and indefinite articles,
< previous page
page_186
next page >
< previous page
page_187
next page >
Page 187
Figure 8.1 Use of prevocalic definite article in the Fens (N = 118). contexts where linking and intrusive /r j w/ may occur, and small-function words. Our results for the Fens show an almost totally intact traditional hiatus-resolution system. The indefinite article was consistently a before consonants and an before vowels. There was a small amount of variation for the definite article, presented in Figure 8.1. Linking /r/ was found between non-high vowels and a following vowel in 95.6 percent of all relevant tokens (with [ ] accounting for the remainder), and intrusive [j] and [w] in all cases where a high vowel was followed by a vowel. Allomorphy sensitive to following segment is also almost categorical in the Fens data, as Table 8.1 exemplifies. These data show only small traces of incipient glottal stops being used to break hiatus. We now move 150 km farther south, to the East End of London, to observe how this urban, ethnically heterogeneous community is resolving hiatus. Table 8.1 Allomorphy in Small Function Words in the Fens Function Word Preconsonantally Prevocalically OF [ə] 100% 100% TO [tə] 100% [tu:] 100% YOU [jə] 95.2% [ju:] 100% [ju:] 4.8% (esp. before [w])
< previous page
page_187
next page >
< previous page
page_188
next page >
Page 188 5. HIATUS RESOLUTION IN LONDON Our London corpus of data comes from the ‘traditional’ East End, an area which corresponds to the presentday Tower Hamlets, the inner London borough directly to the east of the City of London. Traditionally, this area has been associated with the Cockney dialect of what was the predominantly White working-class population, but in the last 50 years or so, the area has undergone rapid social and economic change and is now one of great social and ethnic diversity. More than 55 percent of the population comes from a large range of ethnic minority groups, with the largest single group being the Bangladeshis, who represent 33.43 percent of the total population. The Bangladeshi community, however, is not spread evenly across the borough, but tends to be situated in the neighbourhoods to the west, and in some of these neighbourhoods the Bangladeshi population reaches 75– 90 percent of the total. Another defining feature of the borough’s population is the high percentage of young people. The ‘under 16’ age group makes up 22.9 percent of the population (compared with 20.2 percent nationally); of this total, 38 percent are Bangladeshis. The ‘20–29’ category comprises 24.1 percent of the total population (compared with 12.6 percent nationally).11 More than 57 percent of all school-age pupils in Tower Hamlets come from Bangladeshi backgrounds, and more than 70 percent are from ethnic minority groups. The percentage of pupils who have English as an additional language is 65 percent, compared with 8 percent of pupils nationally.12 The data for this part of the analysis come from the recorded informal conversations of 39 adolescents, either born in Tower Hamlets or settled there before the age of 3. They comprise 9 White British girls, 11 White British boys, 19 Bangladeshi boys, and 2 boys of mixed race.13 For the purpose of the analysis, the mixed-race boys were included with the White British boys on the basis that they both lived with one White British parent and they were part of a largely White British friendship network. In addition, to provide a real-time comparison with today’s East End of London, we were able to examine transcripts from a study conducted by Sivertsen (1960) in Bethnal Green, also part of Tower Hamlets. These results show that the indefinite article was consistently a before consonants and an before vowels and similarly that the definite article was consistently [ðə] before consonants and [ð•] or [ði] before vowels at that time. The results from the present-day study of Tower Hamlets show a very different picture. For the indefinite article, a is consistently used before consonants, but there is variation in the use of a and an in prevocalic position. There were 94 tokens of the indefinite article in prevocalic position; The Bangladeshi boys produced 27 tokens, the White British/Mixed race boys produced 47 tokens, and the White British girls produced 20 tokens. The results are shown in Figure 8.2.
< previous page
page_188
next page >
< previous page
page_189
next page >
Page 189
Figure 8.2 Use of prevocalic indefinite article in Tower Hamlets, London. The Bangladeshi boys are clearly the most frequent users of a in prevocalic position (76 percent), but the White British/Mixed race boys also display variation, using a before vowels 25 percent of the time. The girls show only a small amount of variation, the use of a before a vowel occurring in only 1 token out of a possible 20 (5 percent). In those cases where a occurs before a vowel, hiatus is resolved by inserting a glottal stop. The same pattern emerges with the definite article. Although [ðə] occurs consistently before consonants, there is variation in the use of [ðə] and [ði] in prevocalic position. There were 312 tokens of definite article in prevocalic position; the Bangladeshi boys produced 100 tokens, the White British/ Mixed race boys produced 132 tokens, and the girls produced 80 tokens. The results are shown in Figure 8.3. Lack of allomorphy again occurs most frequently among the Bangladeshi boys who use [ðə] prevocalically 81 percent of the time. Following the same pattern as the indefinite article, the White British/mixed race boys also show variation with the use of [ðə] occurring in 35 percent of tokens. In all cases where [ðə] is used before a vowel, hiatus is resolved by inserting a glottal stop. As before, however, the girls use the traditional allomorphy sensitive to following segment. To a lesser degree, the same pattern emerges with the use of linking /r/. From a subset of 21 speakers, there were 341 contexts where ‘r’ occurred word-finally after a non-high vowel and before a following vowel: 103 tokens produced by the Bangladeshi boys, 101 tokens produced by the white British/mixed race boys, and 137 tokens produced by the girls. The results are shown in Table 8.2.
< previous page
page_189
next page >
< previous page
page_190
next page >
Page 190
Figure 8.3 Use of prevocalic definite article in Tower Hamlets, London. Among the Bangladeshi boys, the use of a glottal stop as the epenthetic consonant between the two vowels is a high figure of 45 percent. There is also a small amount of variation among both the White British/mixed race boys and the White girls. In function words, there is also a degree of variation, particularly among the Bangladeshi boys, presented in Table 8.3. Furthermore, what is clear from Table 8.3 is that the Bangladeshi boys show a general avoidance of the ‘weak forms’ in these function words, which seems to support the idea that the variety used by the Bangladeshi boys shows similarities with other varieties where language contact has been central to their development. The degree to which this pattern of use has been adopted by the White/mixed race adolescents in the study is much smaller, but nevertheless the pattern that emerges is the same as for the previous features analysed, with the boys showing more frequent use of the full forms than the girls. In other cases, where a high vowel was followed by a vowel, there was also variation in the epenthetic consonant selected to resolve hiatus, primarily among the Bangladeshi boys, as presented in Table 8.4. Table 8.2 Linking /r/ in Tower Hamlets White British girls [ ] 97% [ ] 3% White British/Mixed race boys [ ] 94% [ ] 6% Bangladeshi boys [ ] 55% [ ] 45%
< previous page
page_190
next page >
< previous page
next page >
page_191
Page 191 Table 8.3 Allomorphy in Small Function Words in Tower Hamlets, London Preconsonantally White British girls White British/mixed race boys Bangladeshi boys
[ə] 95% [ə] 94% [ə] 39%
White British girls White British/mixed race boys Bangladeshi boys
[tə] 100% [tə] 98% [tə] 27%
White British girls White British/mixed race boys Bangladeshi boys Table 8.4 Hiatus Resolution in White British girls White British/mixed race boys Bangladeshi boys White British girls White British/mixed race boys Bangladeshi boys
< previous page
OF 5% 6% 61% TO
[tuː] 2% [tuː] 73%
YOU 89% [juː] 11% 88% [juː] 12% [jə] 15% [juː] 85% V(+high)#V Contexts in Tower Hamlets, London V (+ high + front) # V: [j] 100% [j] 96% [j] 62% V (+ high + back) # V: [w] 100% [w] 100% [w] 85%
page_191
Prevocalically 100% 100% 100% [tuːw] 100% [tuːw] 100% [tuːw] 82% 18% [juːw] 100% [juːw] 100% [juːw] 100%
4% 38%
15%
next page >
< previous page
page_192
next page >
Page 192 We were able, for the indefinite and definite articles, to examine the relationship between social network ties and the use of glottal stops to break hiatus among the adolescents studied in Tower Hamlets. The data collected for this project were the result of several months’ ethnographic participant observation and recording in a youth club (see Fox 2007), and consequently ongoing close friendships within the club could be observed. As a result of this ethnography, Fox developed a ‘sociogram’ modelling the peer groups within the larger youth club setting which enabled her to identify any relationship between language use and friendship group membership. The advantage of this model is that it is able to graphically demonstrate how variants of different variables are used differentially across the club’s membership, and how close friendships and connections between friendship groups may act as conduits for the diffusion of particular linguistic variants (see Fox 2007 for further details). Figure 8.4 shows the sociogram for the use of [əː] as a prevocalic indefinite article and Figure 8.5 for the use of [ðəː] as a prevocalic definite article. The figures make it clear that it is the older Bangladeshi boys that lead in the use of [ə] and [ðə] accompanied by glottal stops to break hiatus
Figure 8.4 Prevocalic [ə?] as an indefinite article among different friendship groups in the Tower Hamlets Youth Club (based on Fox 2007). Key: each speaker is represented by a circle; circled groups of speakers represent small friendship groups; the lines between the friendship show intergroup connections. A cross through a circle shows there were no tokens for that speaker.
< previous page
page_192
next page >
< previous page
page_193
next page >
Page 193
Figure 8.5 Prevocalic [ðə?] as a definite article among different friendship groups in the Tower Hamlets Youth Club (based on Fox 2007). Key: each speaker is represented by a circle; circled groups of speakers represent small friendship groups; the lines between the friendship show intergroup connections. A cross through a circle shows there were no tokens for that speaker. before following vowels, with younger Bangladeshi boys following closely behind. The younger White and mixed-race boys and older White boys also show relatively significant levels of [ə] and [ðə], although as the sociogram importantly shows, this is because their friendship groups maintain close links with others whose members use high levels of these forms too. The White girls rarely, if at all, use [ə] and [ðə] in this way. Notice, however, that the sociograms point to the fact that they have no direct close connections with the Bangladeshis who use them most—the distribution patterns for this variable feature, therefore, are likely due to interaction levels between individuals and their friendship groups, rather than being essential characteristics of particular gender or ethnic groups. Importantly, very similarly patterned sociograms to these were found for several other linguistic variables that Fox studied, notably front and almost monophthongal realisations of the PRICE diphthong and close, also almost monophthongal realisations of the FACE diphthong (see Fox 2003, 2007 for further details). Close contact within the youth club setting, then, can be seen to be the route by which the use of these nonallomorphic forms is spreading from one group to another.
< previous page
page_193
next page >
< previous page
page_194
next page >
Page 194 Table 8.5 Use of Glottal Stops in Hiatus Contexts and “Full” Vowels in Hackney, London Example IPA Transcription Gender Ethnicity “like a animal” Male Kuwaiti/Moroccan “the older lot” Male Kuwait/Moroccan “do their own thing” Male Kuwait/Moroccan “fifteen to sixteen” Male Black African “of your energy” Male Black African “used to go” Male Black Caribbean “their own hood” Male Black Caribbean “gonna end up” Male White Anglo “my older sister” Female Moroccan “the African” Female Black Caribbean The data presented clearly suggest robust variation in the hiatus-resolution system in the English of this part of London. The Bangladeshi adolescents seem to favour the use of glottal stop more than other speakers, but it is also being adopted by the adolescent White boys too. There is evidence that glottal stop use is widespread in inner London more generally. Recent research on the English of Hackney in London14 has found a parallel use of glottal stops for hiatus breaking, as is exemplified in Table 8.5. Although the glottals are used predominantly by members of London’s ethnic minority groups, as the table shows, they are also found among White Anglo speakers. Although the rural Fens, with an overwhelmingly White population, have maintained the traditional, complex hiatus-resolution system, multicultural London has witnessed the beginnings of its simplification and regularisation into a system that uses glottal stops to block hiatus and avoids ‘weak’ forms of function words. Further evidence that sociocultural contact seems implicated in these developments can be seen in the English of another community which has a substantial White but non-Anglo minority. Bedford, lying 65 km to the southwest of the Fens and 90 km north of London, has a large Italian minority which began to settle in the town in the early 1950s to work in the local brick works (see Guzzo 2005, 2007). Early analysis of recordings of third-generation Bedford Italian adolescents shows a significant use, predominantly by young men, of the very same hiatus-avoiding forms as we found in Tower Hamlets and Hackney (see Guzzo, Britain, and Fox, forthcoming). Examples can be seen in Table 8.6.
< previous page
page_194
next page >
< previous page
page_195
next page >
Page 195 Table 8.6 Use of Glottal Stops in Hiatus Contexts and “Full” Vowels in Bedford Italian English Example IPA Transcription “hard to socialise” “her address” “cheaper in Italy” “River Island” “my auntie” “fifth of August” “going by aeroplane” “me and Kirsty” “a apple” “the other one” “gone to Leicester” Guzzo, Britain and Fox forthcoming. 6. SUMMARY AND DISCUSSION The traditional complex hiatus-resolution system of British English is exemplified by our data from the Fens. Allomorphy is found in the definite article system, where the high front vowel of the prevocalic allomorph [ði] facilitates a [j] glide to the following vowel, in the indefinite article, where a relic [n] is inserted to block hiatus, and in a number of small function words, which either have a prevocalic allomorph ending in a high vowel which can then trigger a glide to the following vowel (as in ‘to’: [tə] before consonants, [tu] + [w] before vowels) or have an allomorph which ends in a consonant to block hiatus (as in ‘of’, which is [ə] before consonants and [əv] before vowels). Linking and intrusive /r/ are usually triggered where a non-high vowel is found in prevocalic position and [j] or [w] where the prevocalic vowel is high. In London (and in Bedford), this system appears to be breaking down and regularised, however. Avoidance of ‘weak’ forms for function words means that allomorphy is less often resorted to generally, although the ‘full’ forms are beginning to adopt a glottal stop as a hiatus breaker instead of a glide. Typically, preconsonantal forms of the articles are being used in prevocalic contexts, however, with a glottal stop again blocking hiatus. Glottal stops are also replacing [j] and [w] after high vowels and replacing /r/ after non-high vowels.
< previous page
page_195
next page >
< previous page
page_196
next page >
Page 196 One might reasonably ask why it is a glottal stop that is being adopted to replace the many different forms of hiatus resolution traditionally found in British English. Lombardi (2002), reviewing examples of epenthesis from a wide range of languages, has argued that glottal stops, as pharyngeals, have the least marked place of articulation for a consonant and hence are to be expected as epenthetic consonants (e.g. to break hiatus) “all things being equal” (2002:246–7). Coronals, she argues, are “the next best choice” of epenthetic consonant (Lombardi 2002:223) and so will appear when “for some reason constraint conflict results in glottal stop being impossible” (2002:223). She continues: “glottal stop has the least marked place, but conflicting requirements may force the choice of slightly more marked, but still relatively unmarked Coronal … careful analysis shows that interacting facts about position, inventory, etc. in a given language can explain why the least marked … is not chosen” (2002:246–7). English English, as we have seen, provides some evidence of coronals being used to resolve hiatus—/r/15 after non-high vowels, [n] after the indefinite article,16 but both contexts are now keenly adopting the glottal stop to replace these historical artefact coronals, in our London data, in Bedford, as well as further afield in South Africa and Singapore, for example. Language and sociocultural contact are what seem to unite the locations undergoing a shift to glottal epenthesis to break hiatus.17 Lombardi (2002:224) also addresses the possibility that glides may be epenthesised instead of glottals. She says that a “common approach is to epenthesise a glide that agrees in features with an adjacent vowel, often a high vowel. Assuming some kind of spreading or multiple correspondence, insertion of an agreeing glide may incur fewer markedness violations, so a language may choose this where possible—where a corresponding glide exists” and adds “with non-high vowels, for most languages a corresponding glide does not exist, and glottal stop will be inserted instead”. Although this is not straightforwardly the case in our data, if we see this claim as a variable constraint rather than as a categorical statement, it shows notable parallels with the use of glottal stop in Tower Hamlets. Yes, some [j] and [w] glides are being replaced by glottal stops in London, but less so than are the strategies traditionally used to break hiatus following nonhigh vowels—linking and intrusive /r/. Uffmann (2007a), although agreeing that glottal stops make good intrusive consonants, argues against approaches, like that of Lombardi, that suggest markedness can operate in a context-free way with respect to epenthesis. He proposed (2007a, 2007b) that we need to distinguish between intrusion at word edges, on the one hand, and word-internally, on the other hand, with glottal stops being optimal epenthetic consonants at word edges because they provide a maximally steep sonority rise to the neighbouring sound. He considers (personal communication) that the emergence of the glottal stop as an all-purpose hiatus breaker in the Tower Hamlets data is perhaps symptomatic of a more deep-rooted change—namely, the basic unit of prosodification changing from the phonological phrase to the word.
< previous page
page_196
next page >
< previous page
page_197
next page >
Page 197 The use of glottal stops to break hiatus, then, could validly be seen as an example of the ‘emergence of the unmarked’ (McCarthy and Prince 1994).18 One place in which the emergence of the unmarked is most clearly visible, according to Gnanadesikan (1996, 2004), amongst others, is in child language, where initially markedness takes precedence, but gradually declines in importance as the adult norm is acquired. As we will see, there are remarkable parallels between our London data and the published evidence on hiatus breaking among children. The London data suggested, as we saw earlier, that the shift towards glottal stops was contextually conditioned by linguistic factors. Glottal stops were used instead of article allomorphy in 75–80 percent of all possible cases; instead of linking /r/ in 45 percent of all cases and instead of the glides [j] and [w] in 15–38 percent of all cases. Why might this patterning occur? Why is glottal stop more likely for the articles than as a replacement for linking and intrusive /r/, and why is it more likely there than as a replacement for the glides [w] and [j]? Interesting parallels with child language acquisition may help us understand this better. Newton and Wells (1999, 2002) investigated children’s use of linking /r/, [j], and [w], as well as article allomorphy, and found that children acquired the (traditional) adult system of hiatus resolution first in those contexts where glottal stops are replacing the traditional breakers in the London data least . Newton and Wells (2002) is a longitudinal study of a boy, CW, between the ages of 2;4 and 3;4. They show that “liaison with [j] occurs in CW’s speech right from the beginning of the study” and, very importantly for our argument here, “where /j/ liaison is not reported to have occurred open juncture is produced, with a glottal stop inserted at the word boundary, for example, ‘he in’: ” (2002:286, our emphasis). Liaison with /r/— intrusive or linking—however, is variably acquired later by CW, from around his third birthday, and again glottal stop is inserted if /r/ is not present (2002:288). They argue that [j] and [w] may be produced for “ ‘low level’ phonetic reasons” (like Heselwood 2006:80, see above) because one of the simplest ways of getting from a high vowel to a following vowel is via a glide, leading them to reject accounts (such as those of Broadbent 1991; Gick 1999) which suggest /r/ can be seen as being in the same group of glides as [j] and [w]. If this were so, they argue, “one would expect to observe /r/ appearing at around the same time as the other types of liaison. In fact, it does not; it emerges later” and that this unified view of glides “is not born out by the developmental data” (2002:292). Our London data confirm Newton and Wells’ (2002) and Heselwood’s (2006) scepticism about the possibility that /r/ behaves as a glide just like [j] and [w] because the adolescents in our study are more resistant to using a glottal stop to replace the glides [w] and [j] than they are to replace /r/. Newton and Wells (1999) consider [j w r] as hiatus breakers along with the adoption of article allomorphy among 94 children aged 3–7 years in the Western English city of Hereford. In this study of older children, they find that:
< previous page
page_197
next page >
< previous page
next page >
page_198
Page 198 Table 8.7 Parallelism Between Children’s Acquisition of Adult Hiatus Breakers (Based on Newton and Wells [1999, 2002]) and Adolescents’ Adoption of New Hiatus Breakers in Tower Hamlets Children acquiring English Adolescents in East End of London
Definite and indefinite articles
V[-high]#V
V[+high]#V
• [r] as a hiatus breaker is consistently produced less than [w] and [j] as breakers from the age of 3 right through to the age of 7 (1999:73); • Allomorphy in the definite and especially the indefinite articles is acquired really quite late. Allomorphy in the definite article does not reach an average of 75 percent of all examples until the children are aged 5 and for indefinite articles not until they are 6 years old; • Before the adult forms are acquired, glottal stops are used to break hiatus. We can see, therefore, that Newton and Wells’ results neatly parallel our own from Tower Hamlets—the later the ‘adult’ forms are acquired among children, the more likely glottal stops will be used (instead of traditional adult forms) to resolve hiatus in this contact-influenced variety of English spoken in London (see Table 8.7). 7. CONCLUSION In this chapter, we have been able to demonstrate that the well-established yet highly complex hiatusbreaking system of British English is beginning to undergo radical reorganisation. In a number of communities witnessing language and sociocultural contact—Tower Hamlets and Hackney in London and Bedford in the East Midlands—glottal stops are beginning to replace the traditional paradigm of hiatus breakers. A widespread vernacular system once dominant in England appears to be undergoing changes that have already affected other contact-influenced Englishes, such as those of Singapore, South Africa, and New Zealand. That it is the glottal stop taking on this role is perhaps not surprising—phonologists have argued
< previous page
page_198
next page >
< previous page
page_199
next page >
Page 199 that it is the least marked epenthetic consonant and have shown that it is the default form in children’s English until adult forms are acquired. Just as the shift from the unmarked, on the one hand, to a faithfulness to the adult norm, on the other hand, is linguistically conditioned, so, in exactly the same way, is the adoption of the glottal stop as a replacement for the complex hiatus-breaking system among London’s adolescents, suggesting an orderly heterogeneity in the move towards a feature of the type that Chambers labels a ‘vernacular universal’. It is found in child and adult varieties, both in countries where English has long been the dominant and first language, as well as those where it is a relatively recent arrival, or a second or additional language. Contact is the factor that unites those varieties which are undergoing these changes first and most enthusiastically. Rather than being a disruptor of ‘vernacular universals’, therefore, contact may well be crucial to the very emergence of widely distributed nonstandard forms in the first place. NOTES 1. We’d like to thank the audience at the Workshop on World Englishes: Vernacular Universals vs. ContactInduced Change, Mekrijärvi Research Station, Joensuu University, Finland, for their comments on an early oral presentation of the work presented here. We’d also like to thank the following people who read an earlier draft of the chapter and provided us with extremely helpful comments: Paul Foulkes, Barry Heselwood, Caroline Newton, Erik Thomas, Dick Hudson, Christian Uffmann, Raj Mesthrie, and Bill Wells. 2. Most varieties of English do insert some consonantal element between two vowels in these contexts. We are grateful to Peter Trudgill for drawing our attention to an apparent lack of such an element in parts of the North and Mid-West of the United States. It would be interesting if such varieties were investigated acoustically to observe the phonetic transition between the vowels there. 3. In addition, the prevocalic pronunciation /ðiː/ has been noted in preconsonantal position when it is stressed or used for emphasis as in ‘this is the (/ðiː/) hairstyle of the season’ or when it is used to signal problems in speech production (Fox Tree and Clark 1997). 4. We are aware, of course, that there exist other forms of (nonstandard) allomorphy in the definite article system in some varieties of English. Many northern English varieties exhibit definite article reduction (DAR), where is reduced to an alveolar or glottal stop (Beal 2004; Jones 1999, 2002; Rupp and PageVerhoeff 2005), and Viereck (1995) reports that traditional varieties of South-West England, part of East Anglia, and the north also delete the vocalic element, but retain an interdental consonant (e.g. ‘in th’oven). What is interesting about both of these forms is that they conform to a subtle rule-governed system of hiatus avoidance. As Jones (2002) points out, “the reduced articles of DAR are always vowel-less” and so by dropping the vowel in the definite article, the pronunciation eradicates the need to resolve hiatus when DAR is prevocalic. 5. Of, in this sense, appears to be behaving similarly to the indefinite article. In both cases, the vowel is reduced, and in both cases, the following consonant, appearing only in prevocalic position, represents a historically prior, less
< previous page
page_199
next page >
< previous page
page_200
next page >
Page 200 lenited form than that used in the more frequently occurring preconsonantal position. 6. Jurafsky, Bell, Fosler-Lussier, Girand, and Raymond (1998) provide evidence of the robustness of this allomorphy in the American English Switchboard corpus (see also Bell et al. 2003). 7. The TIMIT corpus consists of digitalised recordings of 6,300 sentences, 10 sentences each read by 630 speakers from eight different dialect regions of the United States (see Garofolo et al. 1993). 8. This is also noted for the formal English of BBC newsreaders by Allerton (2000:580). 9. In parts of the United States, /l/ can also be used as a hiatus breaker after 9. non-high vowels (Gick 2002). This is not, as far as we are aware, found in England. However, the prevalence of /l/ vocalisation in Britain has, in some varieties but not all, triggered the use of a ‘linking /l/’ in prevocalic positions (e.g. ‘bottle bank’ and ‘bottle opener’) (but only in contexts of an etymological /l/) (see Johnson and Britain 2007). Also, of course, intrusive /l/ is a stereotypical feature of the traditional variety of the West of England city of Bristol, but here it is not restricted to hiatus contexts. 10. This figure should not, however, be construed as suggesting the area is ethnically homogeneous. The Census of 2001 for the most part combined the populations claiming ‘British White’, White Irish’, and ‘Other White’ ethnicities. The Fens are home to a sizeable Traveller community (it is claimed to be the largest ethnic minority group in Cambridgeshire, for example (http://www.cambridgeshire.gov.uk/community/travellers/ [accessed 03/10/2008]), as well as not insignificant populations of other White ethnic groups, especially from Poland and the other EU accession states of Eastern Europe. 11. Source : Census 2001. 12. Borough Profile: www.towerhamlets.gov.uk (July 2004). 13. Fieldwork took place at a mixed-gender youth club; for cultural reasons, Bangladeshi girls are discouraged from social mixing in this context. Therefore, no data were sampled from this section of the population (see Fox 2007 for more details). 14. Kerswill, P., J. Cheshire, S. Fox, and E. Torgersen. Linguistic Innovators: the English of Adolescents in London . ESRC Award 000-23–0680. 15. Admittedly, however, linking/intrusive /r/ could actually be realised as a noncoronal (e.g. as [υ]). Many thanks to Paul Foulkes for reminding us of this fact. 16. But both of these forms are the result of historical relics from unrelated language changes: the loss of rhoticity in most of England for the emergence of intrusive and linking /r/, and the inability of the indefinite article to follow thy/ thine , my/mine, and no/none to mark the distinction between prenominal and final/absolute respectively. 17. Allerton (2000) provides some examples of glottal stops being used to break hiatus in the formal English of BBC news reporters (where language contact is not likely to account for their occurrence). In his conclusion, he speculates whether there is a relationship between these glottal stops and those used, in more informal styles, to replace [t] in many varieties of British English. He goes on to wonder whether there will be “a massive increase in the frequency of the glottal stop” (2000:581), but suggests that it is unlikely because “the glottal stop as part of liaison phenomena is clearly limited to a relatively careful formal pronunciation” (2000:581). Our London and Bedford data clearly contradict this claim of formality for the glottal stop in hiatus-breaking contexts. Stene (1954:43) provides examples from an analysis of the
< previous page
page_200
next page >
< previous page
page_201
next page >
Page 201 radio speeches of King George V which show the use of glottal stops as hiatus breakers before stressed syllables (but not unstressed ones) in slow and careful speech. An empirical quantitative analysis of such speech would be highly interesting given that some forms of RP are supposed to resist, for example, intrusive and linking /r/ as hiatus breakers. 18. See also Johnson and Britain (2007) for another example of the ‘emergence of the unmarked’, exemplified by data from nonstandard varieties of English from Southern England. REFERENCES Ainsworth, H. 1993. Rhythm in New Zealand English. Unpublished BA dissertation. Wellington: Victoria University of Wellington. Allerton, D. 2000. Articulatory inertia vs ‘Systemzwang’: Changes in liaison phenomena in recent British English. English Studies 6: 574–81. Anderson, N., J. Arnold, M. Barron, C. Evans, J. Hay, D. McConnel, J. McKenzie, J. Nielsen, and A. Walker. 2004. To thee or not to thee. Paper presented at the New Zealand Language and Society Conference. Massey University, Palmerston North, New Zealand. Ash, S., and J. Myhill. 1986. Linguistic correlates of inter-ethnic contact. In Diversity and Diachrony , edited by David Sankoff, 33–44. Philadelphia: John Benjamins. Bauer, L. 1984. Linking /r/ in RP: some facts. Journal of the International Phonetic Association 14: 74–9. ———. 1995. Spelling pronunciation and related matters in New Zealand English. In Studies in General and English Phonetics Essays in Honour of Professor J D O’Connor, edited by J. Windsor Lewis, 320–5. London: Routledge. ———, and P. Warren. 2004. New Zealand English: Phonology. In Schneider et al., eds. 2004. 580–602. Beal, J. 2004. English dialects in the North of England: Morphology and syntax. In Kortmann et al., eds. 2004. 114–41. Bell, A., D. Jurafsky, E. Fosler-Lussier, C. Girand, M. Gregory, and D. Gildea. 2003. Effects of disfluencies, predictability, and utterance position on word form variation in English conversation. Journal of the Acoustical Society of America 113(2): 1001–24. Benton, R. 1966. Research into the English Language Difficulties of Maori School Children 1963–1964 . Wellington: Maori Education Foundation. Bowerman, S. 2004. White South African English: Phonology. In Schneider et al., eds. 2004. 931–42. Britain, D. 1991. Dialect and Space: A Geolinguistic Study of Speech Variables in the Fens . Unpublished PhD dissertation. Colchester: University of Essex. ———. 1997. Dialect contact and phonological reallocation: ‘Canadian Raising’ in the English Fens. Language in Society 26: 15–46. ———. 2002. Diffusion, levelling, simplification and reallocation in past tense be in the English Fens. Journal of Sociolinguistics 6(1): 16–43. ———. 2003. Exploring the importance of the outlier in sociolinguistic dialectology. In Social Dialectology, edited by D. Britain and J. Cheshire, 191–208. Amsterdam: Benjamins. Broadbent, J. 1991. Linking and intrusive r in English. University College London Working Papers in Linguistics 3: 281–302. Chambers, J. K. 2000. Universal sources of the vernacular. Sociolinguistica 14: 1–15. ———. 2003. Sociolinguistic Theory . Oxford: Blackwell.
< previous page
page_201
next page >
< previous page
page_202
next page >
Page 202 ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-linguistic Perspective , edited by B. Kortmann, 127–45. Berlin: Mouton de Gruyter. Claxton, A. 1954. The Suffolk Dialect of the 20th Century. Ipswich: Adlard. Craig, H., C. Thompson, J. Washington, and S. Potter. 2003. Phonological features of child African American Vernacular English. Journal of Speech, Language and Hearing Research 46: 623–35. Cruttenden, A. 2001. Gimson’s Pronunciation of English. London: Arnold. Crystal, D. 1995. Documenting rhythmical change. In Studies in General and English Phonetics: Essays in Honour of Professor J D O’Connor, edited by J. Windsor Lewis, 174–9. London: Routledge. Cutler, C. 1999. Yorkville Crossing: White Teens, hip hop and African American English. Journal of Sociolinguistics 3(4): 428–42. DeGraff, M. 1999. Language Creation and Language Change: Creolization, Diachrony and Development . Cambridge: MIT Press. Deterding, D. 2006. The pronunciation of English by speakers from China. English World-Wide 27: 175–98. Foulkes, P. 1997. Rule inversion in a British English dialect: A sociolinguistic investigation of [r] sandhi in Newcastle. In A Selection of Papers from NWAVE 25, Special issue of University of Pennsylvania Working Papers in Linguistics, edited by C. Boberg, M. Meyerhoff, and S. Strassel, 259–70. Philadelphia: Penn Linguistics Club. Fox, S. 2003. Linguistic and Sociocultural Contact in London’s East End: The Bangladeshi Presence . Paper presented at the conference on Consequences of Mobility: Linguistic and Sociocultural Contact Zones, Roskilde, 23–24 May 2003. ———. 2007. The Demise of Cockneys? Language Change Among Adolescents in the ‘Traditional’ East End of London . Unpublished PhD dissertation. Colchester: University of Essex. Fox Tree, J., and H. Clark. 1997. Pronouncing “the” as “thee” to signal problems in speaking. Cognition 62: 151–67. Garofolo, J., L. Lamel, W. Fisher, J. Fiscus, D. Pallett, N. Dahlgren, and V. Zue. 1993. TIMIT AcousticPhonetic Continuous Speech Corpus . Philadelphia: Linguistic Data Consortium. Gaskell, M., H. Cox, K. Foley, H. Grieve, and R. O’Brien. 2003. Constraints on definite article alternation in speech production: To ‘thee’ or not to ‘thee’? Memory and Cognition 31(5): 715–27. Gick, B. 1999. A gesture-based account of intrusive consonants in English. Phonology 16: 29–54. ———. 2002. The American intrusive L. American Speech 77: 167–83. Gnanadesikan, A. 1996. Child phonology in optimality theory: Ranking markedness and faithfulness constraints. In Proceedings of the 20th Annual Boston University Conference on Language Development , edited by A. Stringfellow, D. Cahana-Amitay, E. Hughes, and A. Zukowski, 237–48. Somerville: Cascadilla Press. ———. 2004. Markedness and faithfulness constraints in child phonology. In Constraints in Phonological Acquisition, edited by R. Kager, J. Pater, and W. Zonneveld, 73–108. Cambridge: Cambridge University Press, Guzzo, S. 2005. Language Shift: The Case of the Bedford Italian Community. Unpublished MA dissertation. Colchester: University of Essex. ———. 2007. Multilingualism and language variation in the British Isles: The case of the Bedford Italian community. In Discourse Analysis and Contemporary Social Change , edited by N. Fairclough, G. Cortese, and P. Ardizzone, 233–58. Frankfurt: Peter Lang.
< previous page
page_202
next page >
< previous page
page_203
next page >
Page 203 ———, D. Britain, and S. Fox. Forthcoming. From L2 to ethnic dialect: Hiatus resolution strategies across the generations in Bedford Italian English. Essex Research Reports in Linguistics. Halle, M., and W. Idsardi. 1997. R, Hypercorrection, and the Elsewhere Condition. In Derivations and Constraints in Phonology, edited by I. Roca, 331–48. Oxford: Clarendon. Harris, J. 1994. English Sound Structure. Oxford: Blackwell. Hay, J., and A. Sudbury. 2005. How rhoticity became /r/ sandhi. Language 81: 799–823. Healy, A., and N. Sherrod. 1994. The/Thee Pronunciation Distinction: A Local Model of Linguistic Categories. Paper presented at the 35th annual meeting of the Psychonomic Society, St. Louis, Missouri. Heselwood, B. 2006. Final schwa and R-sandhi in RP English. Leeds Working Papers in Linguistics 11: 78–95. Holmes, J. 1997. Maori and Pakeha English: Some New Zealand social dialect data. Language in Society 26: 65–102. Holmes, J., and H. Ainsworth. 1996. Syllable-timing and Maori English. Te Reo 39: 75–84. ———. 1997. Unpacking the research process: investigating syllable-timing in New Zealand English. Language Awareness 6: 32–47. Hopper, P., and E. Traugott. 1993. Grammaticalisation. Cambridge: Cambridge University Press. Johnson, W., and D. Britain. 2007. L – Vocalisation as a natural phenomenon: Explorations in sociophonology. Language Sciences 29: 294–315. Jones, M. 1999. The phonology of definite article reduction. In Dialectal Variation in English: Special Issue of Leeds Studies in English 30, edited by C. Upton and K. Wales, 103–21. Leeds: University of Leeds. ———. 2002. The origin of definite article reduction in northern English dialects: Evidence from dialect allomorphy. English Language and Linguistics 6: 325–45. Jurafsky, D., A. Bell, E. Fosler-Lussier, C. Girand, and W. Raymond. 1998. Reduction of English function words in switchboard. ICLSP–98 , 7: 3111–14. Keating, P., D. Byrd, E. Flemming, and Y. Todaka. 1994. Phonetic analyses of word and segment variation using the TIMIT corpus of American English. Speech Communication 14: 131–42. Kortmann, B., K. Burridge, R. Mesthrie, E. Schneider, and C. Upton, eds. 2004. A Handbook of Varieties of English: Morphology and Syntax . Berlin: Mouton de Gruyter. Kurath, H., and R. McDavid. 1961. The Pronunciation of English in the Atlantic States . Ann Arbor: University of Michigan Press. Labov, W. 1972. Language in the Inner City: Studies in Black English Vernacular, Philadelphia . University of Pennsylvania Press. Lass, R. 1995. South African English. In Language and Social History: Studies in South African Sociolinguistics , edited by R. Mesthrie, 89–106. Cape Town: David Phillip. Lodge, K. 1984. Studies in the Phonology of Colloquial English. London: Croom Helm. Lombardi, L. 2002. Coronal epenthesis and markedness. Phonology 19: 219–51. McCarthy, J. 1993. A case of surface constraint violation. Canadian Journal of Linguistics 38: 169–95. ———. 1999. Review of Iggy Roca, ed., 1997, Derivations and Constraints in Phonology. Phonology 16: 265– 71. ———, and A. Prince. 1994. The emergence of the unmarked: Optimality in prosodic morphology. In Proceedings of the North East Linguistics Society 24, edited by M. Gonzàlez, 333–79. Amherst: GLSA.
< previous page
page_203
next page >
< previous page
page_204
next page >
Page 204 McMahon, A. 2000. Change, Chance and Optimality. Oxford: Oxford University Press. Mufwene, S. 2001. African American English. In The Cambridge History of the English Language. Vol. 6: History of American English, edited by J. Algeo, 291–324. Cambridge: Cambridge University Press. Newton, C., and B. Wells. 1999. The development of between-word processes in the connected speech of children aged between 3 and 7 years. In Pathologies of Speech and Language Advances in Clinical Phonetics and Linguistics, edited by B. Maassen and P. Groenen, 67–75. London: Whurr Publishers. ———. 2002. Between-word junctures in early multi-word speech. Journal of Child Language 29: 275–99. Ojanen, A. 1982. A syntax of the Cambridgeshire dialect . Unpublished licentiate dissertation. Helsinki: University of Helsinki. Orgun, C. 2001. English r-insertion in Optimality Theory. Natural Language and Linguistic Theory 19: 737– 49. Peitsara, K. 1996. Studies on the structure of the Suffolk dialect. In Speech Past and Present: Studies in English Dialectology in Memory of Ossi Ihalainen, edited by J. Klemola, M. Kytö, and M. Rissanen, 284–307. Frankfurt am Main: Peter Lang. Quirk, R., S. Greenbaum, G. Leech, and J. Svartvik. 1985. A Comprehensive Grammar of the English Language. Longman: London. Rampton, B. 1995. Crossing: Language and Ethnicity Among Adolescents. London: Longman. Raymond, W., J. Fisher, and A. Healy. 2002. Linguistic knowledge and language performance in English article variant preference. Language and Cognitive Processes 17(6): 613–62. Rupp, L., and J. Page-Verhoeff. 2005. Pragmatic and historical aspects of Definite Article Reduction in northern English dialects. English World-Wide 26: 325–46. Schneider, E., K. Burridge, B. Kortmann, R. Mesthrie, and C. Upton. 2004. A Handbook of Varieties of English: Phonology. Berlin: Mouton de Gruyter. Schreier, D. 2003. Isolation and Language Change: Contemporary and Sociohistorical Evidence from Tristan da Cunha English. London: Palgrave. Sebregts, K. 2001. English [r] Liaison: Rule-Based Theories, Government Phonology and Optimality Theory . Unpublished MA dissertation. Leiden: University of Leiden. Shorrocks, G. 1999. A Grammar of the Dialect of the Bolton Area. Part II: Morphology and Syntax . Franfurt: Peter Lang. Sivertsen, E. 1960. Cockney Phonology. Oslo: Oslo University Press. Stene, A. 1954. Hiatus in English. Copenhagen: Rosenkilde and Bagger. Tay, M. 1982. The phonology of Educated Singapore English. English World-Wide 3: 135–45. Taylor, B. 2003. Englishes in Sydney around 1850. Australian Journal of Linguistics 23: 161–83. Thomas, E. 2001. An Acoustic Analysis of Vowel Variation in New World English. Durham: Duke University Press/American Dialect Society. ———. 2004. Rural Southern white accents. In Schneider et al., eds. 2004. 300–24. Todaka, Y. 1992. Phonetic variation of the determiner ‘the’. UCLA Working Papers in Phonetics 81: 39–47. Trudgill, P., and J. Hannah. 2002. International English: A Guide to the Varieties of Standard English. 4th edition. London: Arnold. Uffmann, C. 2007a. Intrusive [r] and Optimal Epenthetic Consonants. Language Sciences 29: 451–76. ———. 2007b. A new idea-r about intrusive [r]. Paper presented to the Research on Languages and Linguistics at Sussex Seminar Series, University of Sussex, 26 November 2007.
< previous page
page_204
next page >
< previous page
page_205
next page >
Page 205 Vennemann, T. 1972. Rule inversion. Lingua 29: 209–42. Viereck, W. 1995. Medieval Dialectology. Berlin: Mouton de Gruyter. Wagner, S. 2004. English dialects in the southwest: Morphology and syntax. In Kortmann et al., eds. 2004. 154–74. Warren, P. 1998. Timing patterns in New Zealand English rhythm. Te Reo 41: 80–93. ———. 1999. Timing properties of New Zealand English rhythm. In Proceedings of the 14th International Congress of Phonetic Sciences , 1843–48. University of California at Berkeley. ———, and D. Britain. 1999. Intonation and prosody in New Zealand English. In New Zealand English, edited by A. Bell and K. Kuiper, 146–72. Amsterdam: John Benjamins. Watermeyer, S. 1996. Afrikaans English. In Focus on South Africa, edited by V. de Klerk, 99–124. Amsterdam: John Benjamins. Wells, J. 1982. Accents of English. Cambridge: Cambridge University Press. Wolfram, W., and R. Fasold. 1974. The Study of Social Dialects in American English, Englewood Cliffs. New Jersey: Prentice Hall. Wright, J. 1905. The English Dialect Grammar . Oxford: Oxford University Press.
< previous page
page_205
next page >
< previous page
page_206
next page >
Page 206 9 The Interplay of ‘Universals’ and Contact-Induced Change in the Emergence of New Englishes Donald Winford 1. INTRODUCTION The field of contact linguistics has become more inclusive of late, with increasing attention being paid to the processes and principles of change that are shared across different types of contact languages. One significant development in this regard is the growing rapport between creole linguistics and studies of second-language acquisition (SLA), particularly cases of group SLA or language shift. The New Englishes fall into the latter category, and their relationship to creoles has long been a matter of discussion (see Harris [1986], Odlin [1997], etc. on Irish English; Ho and Platt [1993] on Singapore English). So far, however, we have had no comprehensive comparison of processes that led to the formation of creoles and New Englishes. One of the major themes in the study of creoles and the (other) New Englishes is the interplay among universals, substrate influence, and internal developments in the emergence of creoles. However, two separate traditions of scholarship seem to have emerged for the study of the two categories of contact languages. One paradigm, research on English as a world language, has been almost exclusively devoted to the ‘indigenized’ varieties of English spoken in the so-called ‘English as a second language (ESL)’ communities in Africa, South and Southest Asia, etc., which comprise one division of the ‘Outer Circle’ of English. The other paradigm, creole linguistics, has been devoted to the study of so-called ‘English as a second dialect (ESD)’ communities, such as those in the Anglophone Caribbean, where ‘creole’ varieties of English coexist with local varieties of Standard English. The division is arguably an artificial one given, first, that creoles emerged as second languages and, second, that some creoles have grammars that are so heavily influenced by West African languages that they can hardly be called ‘dialects’ of English. Given the similarities in the historical circumstances in which all of these contact Englishes were formed, it seems desirable to account for their origins and development within a unified theoretical framework. It has long been noted that there are significant typological/structural similarities between indigenized varieties such as Irish English and Singapore Colloquial English,
< previous page
page_206
next page >
< previous page
page_207
next page >
Page 207 on the one hand, and English-lexicon creoles, on the other hand. This would suggest that they were shaped by similar processes and principles of change. This chapter argues that all of these contact vernaculars are the result of ‘natural’ or ‘untutored’ SLA, and that the theoretical framework within which SLA has been studied is most relevant to a unified explanation of their origins. Such a framework allows us to test the differential impact on these vernaculars of two processes that are central to SLA, viz., universal language learning strategies and L1 transfer. To demonstrate this, I examine the emergence of tense-mood-aspect (TMA) systems in Irish English, Singapore Colloquial English, and Barbadian creole. I argue that there are basic similarities in the paths of development characteristic of these contact varieties, and that these paths are shaped by the interaction of L1 knowledge, input or more accurately intake from varieties of English, and universals of language creation. Before we proceed, it is necessary to clarify the ways in which the concept of ‘universals’ has been used and applied to cases of SLA because the term has been used in so many different senses. In its most general sense, we can use the term simply to refer to a set of principles that inform the acquisition process. Within the generative tradition, these principles are seen as innate properties of the human language faculty and referred to as Universal Grammar (UG). This notion is meant to explain the fact that children attain a level of linguistic ability that goes far beyond the input they have access to during first-language acquisition. Hence, it is argued, the gap between the input and the knowledge acquired must be bridged by innate properties of the human language faculty, which constitute UG. There is no need to discuss here all of the details of this theory or the kinds of evidence that have been adduced both for and against it in the generative literature. Suffice it to say that the principles of UG are simply those that underlie the design features of the human language faculty and, by extension, the design of human languages. Given this broad conception, it is difficult to see how one would deny a role for UG in either first-language acquisition or SLA because that role is simply to ensure that the end product of acquisition conforms to universals principles of language design. Thus, it is trivially true that UG constrains language acquisition. However, as is well known, researchers differ about the extent to which UG plays a role, both in firstlanguage acquisition and SLA. Henceforth, I focus my attention on the latter, which is more relevant to our discussion. The most contentious issue in SLA research has to do with the comparative contribution of UG and L1 influence to the SLA process. Some SLA researchers argue that “UG in its entirety constrains L2 acquisition” and assign little or no role to L1 influence (Epstein, Flynn, and Martohardjono 1996). In this view, SLA is subject to the same principles and constraints as first-language acquisition. The most controversial version of the full access to UG claim is Bickerton’s (1981, 1984) claim that UG, or an innate language bioprogram, directly dictates the grammatical features of creole languages. Because this
< previous page
page_207
next page >
< previous page
page_208
next page >
Page 208 claim has been thoroughly debunked in the creole linguistics literature, I do not consider it further here. Many other researchers argue that the course of L2 development is constrained by both UG and the L1 grammar to differing degrees (Schwartz and Sprouse 1996). According to this view, SLA is constrained by three factors: the L1 grammar, the L2 input, and UG. This seems to be a reasonable point of view, and it is the one I adopt here. In principle, the role of universals and L1 influence are not at all incompatible. In fact, the sharp divide created between the two by some scholars is an artificial one because universal principles guide all aspects of the acquisition process, including L1 influence and internal developments in the interlanguage system. The question then arises, how can we identify the relevant principles and the precise roles they play in the emergence of contact languages? This question is confounded by the fact that many different types of universals have been discussed in the literature. For instance, a broad distinction has been made between universals of language design (UG) and typological universals, which are empirically based generalizations about the similarities and differences across languages. Although the latter are of great value in determining taxonomies of linguistic phenomena and their distribution in the world’s languages, one can scarcely argue that they are explanatory principles. Chambers’ ‘vernacular universals’ are a subtype of such typological universals, as found in the various forms of nonstandard English that are spoken throughout the English-speaking world. They include phenomena such as morpheme-final consonant cluster simplification, leveling of irregular verb forms (John seen the movie), regularization of subject–verb concord ( They/you/we was ), multiple negation, and copula absence/deletion (Chambers 2004:129). Chambers argues: In so far as these processes arise naturally in pidgins, child language, vernaculars, and elsewhere, they are primitive features, not learned. As such they belong to the language faculty, the innate set of rules and representations that are the natural inheritance of every human being. (Chambers 2004:129) This sounds very much like a strong version of the full access to UG hypothesis, recalling Bickerton’s claim that features shared across creoles are in fact directly selected by the language bioprogram. Like Bickerton’s hypothesis, Chambers’ claim seems overstated and has not in fact been demonstrated to be true. It seems more reasonable to claim that such ‘vernacular universals’ arise through similar processes of language acquisition and change, guided by true universal principles. In other words, we need to keep the notions of typological and ‘innate’ universals separate and acknowledge that only the latter play a role in shaping the outcomes of language acquisition and change. As Kiparsky (2008:52) notes, we need to reconcile
< previous page
page_208
next page >
< previous page
page_209
next page >
Page 209 research that seeks to uncover typological generalizations, which are the result of recurrent processes of language change, with research directed at uncovering the universal principles underlying such processes. This kind of integration of approaches is what contact linguistics now lacks, although much progress has been made in the area of typological generalizations about certain classes of contact languages, such as creoles. I here adopt Siemund’s (2004:402) view that “universals of language architecture, language acquisition, processing and development—in short, language universals”—are what shape the grammars of contact varieties such as creoles and indigenized Englishes. But my primary focus is on universals of SLA as they apply to the three contact languages that are the subject of this chapter. I proceed first by trying to arrive at the kinds of typological generalizations we can make about the structure of the tense/ aspect systems of these languages, and I then consider how the similarities and differences we find might be explained in terms of universal processes interacting with other influences, such as the nature of the targeted input, and of the substrate languages involved. But before we proceed, it is important to note that there are many variables involved in contact situations that make it very difficult to compare contact languages. Such variables include the types of linguistic input present in the contact situation, the degree of access to the language being targeted, the degree of typological distance between the input languages, and the degree of homogeneity in the L1s (or substrate languages) of the learners. Thus, we would expect a great deal of similarity between the outcomes of contact situations that differ only in terms of one or two of these variables. One example might be the high degree of similarity in the circumstances in which more radical Caribbean creoles emerged, involving similar English dialectal and West African inputs, and conditions that allowed for relatively little access to the superstrate dialects of the colonizers. Such circumstances led to various shared characteristics of their TMA systems, and of certain other areas of their grammar, which testify that similar principles of contact induced change were at work. In comparing contact languages of very different circumstances of origin, our task of identifying common explanations for any similarities we find becomes far more difficult. It is not necessarily the case that such similarities emerged for the same reasons or because of similar processes of change. If we could find quite different contact situations that resulted in very similar innovations in the contact languages, such innovations would be prime candidates for explanation in terms of universal principles. There are some interesting similarities as well as differences among the situations in which the three languages discussed here emerged. Irish English and Singapore English were ‘endogenous’ varieties created in the home communities of their learners, whereas Bajan was an ‘exogenous’ language acquired by African slaves transported to plantations in Barbados. However, Irish English and Bajan were both created during the seventeenth century
< previous page
page_209
next page >
< previous page
page_210
next page >
Page 210 and shared somewhat similar superstrate inputs, including SW English dialects. Singapore English, in contrast, was a creation of the earlier twentieth century, and its learners had access to both standard and nonstandard varieties of English. Finally, although Irish English and Singapore English had relatively homogeneous substrates (Irish and Southern Chinese, respectively), Bajan was created by West Africans speaking languages of different families, although New Kwa languages were apparently the dominant ones. The differences go some way toward explaining why each language developed the way it did, yielding quite different tense-aspect systems. However, similarities in the ecology of the contact situations explain why certain similarities emerged. In sections 2, 3, and 4, I present a brief overview of the contact situation that led to the emergence of each contact variety, paying special attention to the linguistic inputs and the social circumstances in which the contact took place. I then describe the tense-aspect system and compare it with those of the superstrate and substrate languages, with a view toward determining the relative strength of influence from each. In each case, I point to both similarities and differences in the three tense-aspect systems. In section 5, I discuss the various factors that might account for the similarities and differences, and I assess the extent to which they might be explained in terms of universal principles of contact-induced change. 2. IRISH ENGLISH Irish English seems to have had its roots in the period 1500–1700, when the English-speaking population grew steadily, owing to migration of settlers from other parts of Britain. British colonial settlements were established in Ulster (Northern Ireland) and eastern Ireland. Large numbers of settlers were introduced to Ulster from Scotland and northern England, while eastern Ireland attracted settlers from various regions of England, especially the south and southwest midlands. There appears to be general consensus that the chief superstrate input to Irish English in its formative period (the sixteenth and seventeenth centuries) came from nonstandard varieties of Early Modern English, particularly those introduced by the English and Scottish settlers of that period. All varieties of Irish English share various characteristics that include continuities from these early Modern English dialects, as well as features due to Irish substratum influence and other processes of contact-induced change. 2.1 English and Irish Influences on Irish English Tense Aspect There seems to be general agreement that the overall organization of the tense-aspect system of Irish English is a blend of continuities from both of its source languages. Within the tense system, the categories of present, past,
< previous page
page_210
next page >
< previous page
page_211
next page >
Page 211 and future are modeled directly on those of English. In contrast, the rich aspectual system reflects influences from both Early Modern English and Irish. 2.1.1 Tense in Irish English In the present tense, Irish English uses concord -s with both singular and plural third-person subjects (Harris 1993:154). (1) Her grandchildren comes down. In the past tense, we find leveling of strong verb systems to one- or two-form patterns. Two form patterns include: do, done, done; go, went , went ; and come, come, come. Other verbs following this pattern include bite, hide , sing, drink, break , see, and so on. Verbs with just one form include come, run, give , beat, and loss ‘lose’ (Harris 1993:153). These are all well-known features of nonstandard English dialects. Finally, future is expressed by will, as in the English sources. 2.1.2 Aspect in Irish English Within the aspectual system, there is a conservative Progressive that may be modeled on Irish, although it also bears resemblance to the pattern found in earlier English and contemporary nonstandard varieties such as those in southwest England. (2) She is at the milking of the cow (Schlauch 1973:172, quoted in Siemund 2004:404) The more usual means of expressing Progressive is by V- ing, with both statives and nonstatives allowing ing. (3) Everyone was wanting to see. (Siemund 2004:404) The Habitual is expressed either by be + Pred in copula-type constructions or do + Verb, as in the following examples from Henry (1957:168–72, quoted in Kallen 1989:3). (4) There be’s a lot of people at the fairs of O’Boyle (5) It does rain a lot in winter. Habitual be appears to derive from so-called ‘finite’ be which is found both in southwest English dialects and earlier Scots. However, in the latter dialects, finite be is a simple copula and does not express the sense of habituality.
< previous page
page_211
next page >
< previous page
page_212
next page >
Page 212 The opposition between simple copula forms of be (am , is, are ) and habitual be in Irish English copula-type constructions seems to have its source in Irish, which makes a similar distinction. Compare the following: (6) a. She’s here now. b. She be’s here often. ‘She’s often here.’ The same distinction is found in Irish, as in: (7) a. Tá sí anseo anois. Be+nonpast she here now. b. Bíonn sí anseo go minic be+nonpast+hab. she here often An alternate to habitual be with progressive verbs is a complex tense consisting of do + be + V-ing, expressing the sense of a habitual progressive. (8) They do be fightin’ among other ‘They’re usually fighting among themselves’ Bliss (1984) observed that be(’s ) was more common than do + be in Northern Ireland, perhaps reflecting greater Scottish influence, although do be was ‘the more general.’ Present habitual do appears to derive from simple periphrastic do in seventeenth-century English dialects, especially those of the Southwest. The reanalysis of do as a habitual auxiliary seems to have been motivated both by the fact that it often conveyed habitual meaning in those dialects and the fact that Irish had a distinct habitual category that provided a model for its reinterpretation (Harris 1986:180). A similar reinterpretation of do occurred in Bajan (Barbadian English) for much the same reasons (see below). The most interesting area of the aspectual system, and one that has generated much debate, is the system of Perfect marking. There are two categories in this subsystem: the Resultative and the “hot news” Perfect. The Resultative displays two patterns: one using have with transitive verbs and the other using be with intransitives such as leave , die , go, come, and so on. These are illustrated in examples (9) and (10), respectively. (9) She’s nearly her course finished. (Harris 1984:307) ‘She’s nearly finished her course.’ (10) All of ‘em people are come down here now. (Filppula 1999:116) As has frequently been pointed out, both of these constructions have models in Early Modern English:
< previous page
page_212
next page >
< previous page
page_213
next page >
Page 213 (11) E.Mod.E. Thou hast thy father much offended (Siemund 2004:410). But there is also a somewhat similar construction in Irish, which may have also acted as a model. (12) Tá an leobhar leite aige. (Siemund 2004:410) Is the book read at him ‘He has read the book.’ The ‘hot news’ Perfect is illustrated in the following: (13) She’s after selling the boat. ‘She’s just sold the boat.’ Irish has a very similar construction, exemplified in the following: (14) Tá sí tréis an bád a dhíol. Be+nonpast she after the boat selling. (Harris 1984:319). As can be seen, the Irish English construction conforms to English rather than Irish word order, but the Irish influence is clear. Much attention has also been paid to the fact that Irish English expresses the sense of a ‘continuative’ or ‘extended’ Perfect with present-tense forms and the sense of an experiential perfect with the simple past, whereas Standard English uses have V-ed in both cases. Continuative: (15) We’re living here seventeen years (Harris 1993:161) ‘We’ve been living here seventeen years’ Experiential: (16) I didn’t hear him playing with years and years (Filppula 1997:52) ‘I haven’t heard him playing for years and years’ It has been noted that there are similar constructions in Irish as in the following: (17) Tá sé marbh le fada riamh (Harris 1984:318) Be + non past he dead with long-time ever ‘He’s been dead a long time’ The use of ‘temporal’ with , as in with years and years , directly reflects the Irish pattern. At the same time, it has been argued that such constructions also have models in Early Modern English.
< previous page
page_213
next page >
< previous page
page_214
next page >
Page 214 Table 9.1 Correspondences Between Irish English and Early Modern English Dialects in Tense/Aspect Marking Time Reference Irish Eng. 17th c. SW Eng. Dialects Simple present ø ~ -s do/does + V or V + pres. infl. Simple past V-ed did + V OR V + past infl. Future shall/will + V shall/will + V. Aspectual Reference Irish Eng. 17th c. SW Eng. Dialects Pres. Habitual Inflected do + V n/a ‘Finite’ be +VP Past habitual used to + V used to + V Progressive Cop. be+ V-ing (do) be (a) V- ing Resultative Perfect have NP V-ed have NP V-ed Cop. be+ V-ed be V-ed “Hot news’ Perfect after V-ing n/a Table 9.1 makes a comparison between the forms used for the tense-aspect categories of Irish English and those of Early Modern English, particularly the southwest English dialects. The general conclusion we can draw from this is that Irish English owes its tense-aspect system to the combined influence of both Early Modern English and Irish. It may be the case that, as Bliss (1984:143) claimed, “Southern Hiberno English has precisely the same range of tenses as Irish has, but the forms are built up of English material” (quoted by Filppula 1997:62). But this is because the tense-aspect categories of Irish and Early Modern English overlap to a considerable extent, except in the case of categories like the Present Habitual and the ‘hot news’ Perfect, which were modeled on Irish alone. 3. BARBADIAN ENGLISH (BAJAN) In the four decades after the English colonization of Barbados in 1624, the vast majority of British settlers came from southwest England, Somerset, Devon, Cornwall, and Dorset in particular (Le Page 1960:12; Niles 1980:25–30). On the whole, the British arrivals consisted of a minority of planters, merchants and proprietors, and a large majority of servants and poorer White laborers and small farmers. Niles (1980:24) tells us that the southwest region of England supplied 78.6 percent of the total number of
< previous page
page_214
next page >
< previous page
page_215
next page >
Page 215 servants shipped from Bristol to Barbados between 1655 and 1660. Because new migration of British settlers to Barbados had practically ceased by 1700 (Williams 1987:48), it is clear that we must look to the dialects introduced in the seventeenth century as sources of the features of Bajan. The linguistic evidence lends support to the previous scenario. Many of the distinctive features of Bajan, both in phonology and morphosyntax, are clearly derived from English regional dialects of the seventeenth century—particularly those of the southwest. Researchers such as Le Page and Tabouret-Keller (1985:41) argue that the Barbadian English vernacular “is much closer to British English in its grammar, and contains far fewer Africanisms … than does, for example, Jamaican or Belizean or Guyanese creole. Moreover, it contains some features of pronunciation and grammar which are closely associated with the West of England and Ireland”. 3.1 The Emergence of the Bajan TMA System As in the case of Irish English, the forms that express tense and aspect in Bajan have their sources in the English dialects introduced to Barbados in the seventeenth century, but there are often significant differences in the meanings they convey. 3.2 Tense in Bajan There are three tense categories in Bajan: Relative Past, Future, and Prospective Future. Bajan has no present or absolute past-tense category, but expresses the meanings of simple present and past-time reference in ways quite similar to other Caribbean creoles (i.e. by use of unmarked verbs, which have perfective aspect). Unmarked statives express simple present, whereas unmarked nonstatives convey simple past (in both cases, when the reference point is S). In the southwest English dialects, these meanings are expressed by present and past forms of do, respectively. The Relative Past category is expressed by preverbal did, which is used with both statives and nonstatives to convey the sense of a past in relation to S or some other reference point in the past (Winford 1993a). (18) I di see din fo bring yo broda. (Burrowes 1983:43) I PAST say NEG for bring your brother ‘I said not to bring your brother’ By contrast, most researchers seem to agree that periphrastic did conveyed simple past meaning in Early Modern English, sometimes alternating with the past-inflected verb in this function. This use of periphrastic did would have made it amenable to reanalysis as a Relative Past marker in the Bajan tense-aspect system.1 Another difference between Bajan did and its southwest English cognate can be seen in their use in conditional and
< previous page
page_215
next page >
< previous page
page_216
next page >
Page 216 temporal clauses. In Early Modern English, did seems to function in much the same way as the simple past in (unreal) conditional clauses (i.e. to convey hypothetical meaning). Elworthy (1877:50) points out in a footnote that If I did dig in Somerset is equivalent to If I should dig. In contemporary southwest English dialects, did seems to be used in a similar sense in conditional clauses, as in the following example from E. Somerset (Ihalainen 1976:617): (19) If you did buy up a load of peat in them days, it used to cost you ten shillings. By contrast, did in conditional clauses in Bajan conveys the sense of a counterfactual past or present, as in the following examples: (20) a. If I did run off the road there, ah di goin in a precipice. ‘If I had run off the road there, I would have gone over a precipice.’ b. If I did have money, I woulda go. ‘If I had money, I would go’ OR ‘If I had had money, I would have gone’ Both in meaning and pragmatics, then, Bajan did is similar to basilectal creole ben and quite unlike English periphrastic did. It is possible that the meaning of Bajan did is due to substrate influence, although the precise source of this influence is difficult to pinpoint. Some of the relevant substrates, including Akan and some Gbe varieties, have an opposition between a simple past expressed by the unmarked verb and an Anterior or Perfect category expressed by an overt morpheme. Another possibility is that Bajan did was reinterpreted on the model of Relative Past marker ben, which had emerged in the more ‘basilectal’ rural creole that developed on some plantations in the eighteenth century. This is speculative, however, and still leaves unanswered the question of the original source of the semantics of ben and did. Future is expressed by gon rather than by will as in the English dialects. Prospective Future, in contrast, is expressed by goin’ (to), the only difference from the English dialects being the absence of auxiliary be support. Note, however, that the expression of the two future categories reflects that found in other creoles. 3.3 Aspect in Bajan The Bajan aspectual categories of Progressive and Past Habitual seem to be modeled directly on those of English dialects. But there are sharp differences with regard to the expression of Present Habitual and Perfect aspect. Present Habitual is conveyed by preverbal does .
< previous page
page_216
next page >
< previous page
page_217
next page >
Page 217 (21)
He does catch fish pretty. ‘He catches fish nicely.’ Does clearly has its source in the third-person present form of periphrastic do as used in the seventeenthcentury southwest English dialects. Present forms of do in these dialects conveyed the same meaning as the simple present tense of contemporary southeast (Ellegård 1953:209). The habitual meaning of does seems to derive in part from the fact that, like the present tense, present periphrastic do often conveyed the sense of habituality, as illustrated in the following example from Elworthy (1886: xx, xlvi): (22) I du zay zom prayers now and again. (Devon) This habitual interpretation is, of course, typical of the Standard English present tense. But there are also significant differences between the use of does and that of periphrastic do. For example, like the Standard English present tense, present periphrastic do could be used with stative verbs to express simple presenttime reference, as in the following: (23) I do know (Cornwall: Jago 1882:57) ‘I know’ In addition, periphrastic do conveyed present time in open conditional and temporal clauses: (24) I shall pick it up whun I da goo whom. (Somerset: Elworthy 1877:52) Bajan never uses does in these contexts, preferring, instead, the unmarked verb. (25) a. I ø know you vex, but Barrie bring she dey. (Burrowes 1983) b. We gon pick it up when we ø go home/ Finally, does can be used with all verbs, including be. By contrast, periphrastic do is not, and has never been used with be in the southwest English dialects (Klemola 1996:68). (26) Wha does be wrong wid you though? ‘What’s always wrong with you, though?’ The evidence available to us indicates that most of the likely substrate languages introduced to Barbados in the seventeenth to eighteenth centuries had a distinct Habitual category. This applies to Kikongo (Mufwene 1988:38), Akan (Kós-Dienes 1984), and Gbe dialects like Gengbe (Kós-Dienes
< previous page
page_217
next page >
< previous page
page_218
next page >
Page 218 1984:40) and Fongbe (Lefebvre 1996:271). The reinterpretation of does would have been facilitated by the fact that present periphrastic do often expressed habituality, providing a basis for the interlingual identification that often triggers reinterpretation of TL forms in terms of L1 functions. Completive Perfect Completive Perfect is expressed by preverbal done, which had its source in English dialectal done, used both as a main verb meaning ‘finish’ and as a past participle in a perfect construction consisting of (be/have ) + done + V-en. The latter was apparently common throughout all English dialects up to the fifteenth century and was more confined to northern dialects after that (Traugott 1972:146, 193). Ellegard (1953:143) suggested that the construction was characteristic of uneducated usage because it was common in the fifteenth-century Paston letters and in the sixteenth-century Machyn’s diary. (27)a.I have …. done dewely examyned the instrument. (The Paston letters, ed. by J. Gardiner 1904: letter 12, p. 26, line 5). (Ellegard 1953:143). b.As I afore have done discuss … (William Lauder 1556. Scots. Office and Dewtie of Kyngis, 340). Unfortunately, we have no evidence from earlier grammars, dictionaries, and dialect surveys to confirm this use of be/have + done + V-en in the southwest dialects of the seventeenth century. Nor are there any attestations of it in the SED apparently. Despite this, it seems indisputable that this construction was the source of preverbal Perfect done in varieties like Southern (White) American English (including Appalachian English) and African American Vernacular English. In these cases, however, the source was most probably Ulster Scots, which may also have provided a model for Completive/Perfect done in Bajan and other creoles. Another possible source is the past-tense form done, which was (and still is) the past form of do in southern English dialects (Niles 1980:123). Like the other forms of do, preverbal done in Bajan differs in both semantics and syntactic properties from its English dialectal cognate. With nonstative predicates, done functions as a Resultative Perfect, whereas with statives (including progressives, adjectives, and locatives), it describes a state that has been in existence for some time, up to and including S. In both cases, it conveys the sense of ‘already.’ (28) a. The man done paint the car. ‘The man has already painted the car.’ b. Dat time I done know wo ii see aredi. (Borrowes 1983:43) ‘By that time I already knew what I had seen.’ This syntactic distribution is clearly different from that of English dialectal done, which, as far as we know, could not be used with statives in
< previous page
page_218
next page >
< previous page
page_219
next page >
Page 219 Table 9.2 Correspondences Between Bajan and SW English Dialects in Tense/ Aspect Time Reference BC 17th c. SW Eng. Dialects Simple present ø (statives) do/does + V or V + pres. infl. Simple past ø (non-statives) did + V OR V + past infl. Relative past did + V ——— Future go(n) + V shall/will + V Prospective goin to + V be going to + V Aspectual Reference BC 17th c. SW Eng. Dialects Pres. Habitual does + V n/a Past habitual useto + V did/used to + V Progressive V- in (do) be + (a) V- ing Perfect done + V be/have (done) + V- ed ? base form or with adjectives, locatives, and progressives. Moreover, English dialectal done conveys the sense of a perfect of result and does not always force the ‘already’ interpretation (Winford 1998). The reanalysis of done as a marker of Completive Perfect seems clearly due to West African substrate influence. All of the relevant substrates (Gbe, Kikongo, and Akan) have a Completive/Perfect category which is similar in semantics to done (Winford and Migge 2007). In general, these markers of Perfect derive from a verb meaning ‘finish’. It is therefore not surprising to find that done is used in all Caribbean English-lexicon creoles both as a verb meaning ‘finish’ and as a marker of Completive Perfect. It is also possible that the semantics of done was also modeled after preverbal done in the more basilectal variety of rural Bajan which had been shaped under West African substrate influence. Table 9.2 compares the inventory of tense-aspect categories in Bajan with those of southwest English dialects as described in the dialectal literature (Barnes 1886; Elworthy 1877, 1886; Klemola 1996). It is clear that the Bajan tense-aspect system is not a straightforward replica of its seventeenth-century source. 4. COLLOQUIAL SINGAPORE ENGLISH Colloquial Singapore English (henceforth SingE) emerged in the period 1930–1960. The English input came from English taught in schools by
< previous page
page_219
next page >
< previous page
page_220
next page >
Page 220 teachers from Britain, Ireland, India, and Sri Lanka, but also from second-language varieties spoken by speakers educated in Tamil-, Malay-, and Chinese-medium schools, as well as from pidginized forms of English spoken by the older generations. The dominant substrate influence came from southern Chinese, particularly Hokkien, and from Malay varieties, including Bazaar Malay, used earlier as a lingua franca, and Baba Malay, a hybrid of Malay and Hokkien (Ho and Platt 1993:8–9). SingE gradually became a lingua franca for the ethnically diverse population and is now increasingly being used as a first or primary language by younger generations of Singaporeans. Platt, Weber and Ho (1983:9) explain its origin as due to the fact that “children were using English in natural communication situations while still in quite early stages of acquisition,” and hence there was strong influence on their English from Chinese and, to a lesser extent, Malay. The extent of substrate influence in colloquial SingE has led some scholars (e.g. Ritchie 1986) to suggest that this contact vernacular is typologically closer to Chinese than to English. The strong influence of Chinese is evident in the tense and aspect system as a whole, particularly the latter. 4.1 Tense in SingE The categories of present and past tense in SingE show evidence of considerable reduction as well as variability in the use of inflection. Present-time reference, in particular, seems to be marked more by temporal adverbials such as today , now, and so on. Hence, there is variable use of third sing -s. Similarly, past-time reference tends to be marked by adverbials such as yesterday and the like. The patterns of tense marking seem to be ascribable partly to a universal tendency to simplify inflectional morphology in the acquisition process and partly to influence from the substrates, which lack such inflection. Ho and Platt (1993:88) found that past marking varied according to the morphological subtype of the verb, according to the following hierarchy: VC type (strong) > -id type > Vd > CC see want die love This hierarchy generally reflects the pattern of past marking found in other contact varieties of English, such as African American English and Trinidadian English (Winford 1993b). Ho and Platt also found that past marking was highest with punctual verbs (56.2 percent), then statives (36.9 percent) and finally nonpunctuals (14.7 percent). They claim that this reflects similar findings by Bickerton’s (1981:170) study of Hawai’i Creole English and Guyanese creole, as well as his claim that children tend to use a similar hierarchy of past marking in the early stages of L1 acquisition. Hence, they conclude that “the punctual-nonpunctual distinction is a
< previous page
page_220
next page >
< previous page
page_221
next page >
Page 221 language universal.” Whether this means that the pattern of tense marking they observe also represents a language universal is, however, debatable. At best, it seems to represent a general tendency in the acquisition of tense marking, and it might therefore be regarded as a ‘vernacular universal’ in the sense of Chambers (2004). But, in general, absence of tense inflection was almost certainly encouraged by the fact that the substrate languages are isolating in nature and do not mark tense via inflection (Ansaldo 2004:136). Future-time reference is often expressed by would + V, a pattern also found in contact Englishes such as Trinidadian English. Even so, future-time reference is also marked by adverbials like tomorrow . English will, in contrast, tends to be used for present habitual situations (see below). 4.2 Aspect in SingE The aspectual system of SingE has its sources in both English and the substrates, but the influence of the latter, especially Chinese, is much stronger. Progressive and Habitual, expressed by (be) V-in and by useto + V, respectively, clearly derive from English. The former, like progressive V-in in other contact Englishes, requires no auxiliary support and can be regarded as another vernacular tendency, if not universal. The fact that Chinese has a progressive category may have reinforced the emergence of this category. Choice of useto for present habitual, contrast, seems odd in view of the fact that the cognate used to expresses past habitual in English. Ho (2003:42) suggests that this use of useto may have been influenced by the fact that Chinese Pefective le cannot occur with verbs in habitual contexts. Because Chinese lacks a category of past habitual, selection of useto for present habitual would not have seemed incongruous. This may also reflect the fact that learners have a tendency to associate habitual action with present situations (Bardovi-Harlig and Reynolds 1995:119). Deterding, Ling, and Brown (2003:35) note that the use of will to describe habitual action is common in SingE and suggests that it may be due to influence from Chinese hui, and possibly from Malay akan . Ho (2003:43) also notes that speakers often use will + V rather than would + V for past habitual, although they also use would use to. In addition to possible substrate influence, this use of will may reflect a general tendency cross-linguistically for future-tense markers to be extended to habitual functions, a tendency that is well documented in the typological literature on tense aspect (Bybee, Perkins, and Pagliuca 1994). The categories that show strongest influence from the substrates are the Completive Perfect and Experiential Perfect. The former is expressed by already , which has a distribution and range of interpretations quite distinct from those of its English cognate. Bao (2005) demonstrates how the use and meaning of already parallels that of the Chinese Completive Perfect marker le . Thus, both already and le convey the sense of incompletive with a nonstative verb, as in the following examples (I label le as COMPL):
< previous page
page_221
next page >
< previous page
page_222
next page >
Page 222 (29) I wash my hand already (Bao 2005:239) ‘I (have) washed my hand’ (30) wŏmen chī le liúlián (Bao 2005:242) We eat COMPL durian “We ate durian’ Similarly, both already and le convey an inchoative sense with statives (including habituals). (31) The wall white already. (Bao 2005:239) ‘The wall (has) turned white/ *The wall was white’ (32) qiáng bái le (Bao 2005:242) wall white COMPL ‘The wall is whitened.’ Finally, both VP-final already and VP-final le convey the sense of ‘inceptive’ (the start of an event): (33) It rain already (Bao 2005:241) ‘It has started to rain.’ (34) xià yŭ le (Bao 2005:242) down rain COMPL “It started/is about to rain.’ Bao also demonstrates that the English adverbial ever is used in much the same way as Chinese marker guo to express the sense of an Experiential Perfect (i.e. the sense that some event occurred at least once in the past). (35) I ever try this type of fruit before (Bao 2005:244) ‘I have tried this type of fruit before.’ (36) wŏ yĭqián shì guo zhè zhŏng shuĭguŏ (Bao 2005:245) I before try EXP this type fruit ‘I have tried this type of fruit before.’ Bao concludes, with good reason: Given the convergence in aspectual meanings between Singapore English and Chinese, we conclude that already and ever are English words which have been grammaticalized to express the perfective aspects derived from Chinese. (Bao 2005:245)
< previous page
page_222
next page >
< previous page
page_223
next page >
Page 223 Table 9.3 Tense/Aspect Categories in Singapore English and Their Models Time Reference Sing Eng Model(s) Simple present ø (variable 3sg –s) English (with Chinese influence) Simple past ø (variable -ed) English (with Chinese influence) Future would + V English Or V + fut. adv. English/Chinese Aspectual Reference Sing Eng Model(s) Pres. Habitual useto + V English (with Chinese influence) Progressive (be) V- in English Completive Perfect VP already Chinese le Experiential Perf. ever + VP Chinese guo Table 9.3 presents the putative models on which the tense and aspect categories in SingE are based. 5. ON THE ROLE OF UNIVERSALS AND SUBSTRATUM INFLUENCE IN CONTACT ENGLISHES It seems clear from the overview provided earlier that the tense-aspect systems of Irish English, Bajan, and SingE are organized in very different ways, and that there are very few tense-aspect categories that match closely across all three languages. Hence, we cannot claim that these categories provide any evidence for typological universals shared by contact Englishes, far less for the hypothesis that such categories are in some sense predetermined by UG, in the sense that Bickerton’s language bioprogram hypothesis proposed. Recall that, according to this hypothesis, radical creole tense-aspect systems are shaped directly by UG and consist of three basic oppositions: punctual versus nonpunctual aspect, anterior versus nonanterior tense, and realis versus nonrealis mood. These labels are sufficiently vague that they can apply to just about any creole tense-aspect system or indeed, to the tense-aspect systems of other languages. Recent research (Singler 1990; Winford 2000; etc.) has demonstrated that there are significant differences across creoles in the inventory of their tense-aspect categories and in their overall organization. Such similarities as are found across, say, Atlantic creoles can be explained more feasibly in terms of shared substratal influence
< previous page
page_223
next page >
< previous page
page_224
next page >
Page 224 from West African languages or, indeed, in terms of the diffusion of particular creoles. It seems hardly likely, then, that the tense-aspect systems of contact Englishes formed in very different circumstances would display the kinds of similarities found across the creoles. Yet some scholars have claimed precisely this. For instance, Ansaldo (2004:136–7) claims that “SE [Singapore English] shows the classic TMA system that has been observed in Creole languages.” But this claim is based on a somewhat idiosyncratic analysis of the tense-aspect system as consisting of four aspectual classes: Anterior (marked by last time), Perfective (marked by already ), Nonpunctual (marked by ‘Durative’ still or habitual always ), and Irrealis (marked by would). This is an unfortunate instance of confusing notional meanings that appear in all languages, with TMA categories, which are the grammaticalized expression of a subset of such meanings. It is clear from the analysis of SingE presented earlier that its tense-aspect system bears little resemblance to those of Atlantic creoles. Because ‘innatist views are clearly suspect, we need to seek explanations in terms of universals “which have been claimed to pertain to the language acquisition process in general, and to language learning/acquisition in a language contact situation in particular” (Filppula 1990:48). Research on SLA involving learners of different L1s attempting to learn different L2s has shown that the development of L2 tense-aspect systems follows a very similar pattern of development in all cases. Bardovi-Harlig (2000:25 ff) summarizes the stages as follows: Stage 1: The pragmatic stage. This is characterized by use of bare verbs, reliance on chronical order, and the strategy of ‘scaffolding’ or reliance on the other interlocutor’s utterances. Stage 2: The lexical stage. In this stage, the use of bare verbs continues, and there is strong reliance on temporal and locative adverbs to convey time reference. Other strategies include the use of connectives (e.g. ‘and, then’), the use of dates or days of the week, and the use of temporal verbs like ‘start’ and ‘finish.’ Stage 3: The morphological stage. Again, use of bare verbs continues, but then verbal morphology begins to appear, usually in a fixed order, depending on the target involved. For instance, the (Perfective) Past tense emerges first in all cases, followed by the Imperfective Past in L2 varieties of Romance languages and the Perfect in L2 varieties of Germanic languages. Studies by Klein (1993, 1995) and his associates have demonstrated this general pattern of acquisition (see Bardovi-Harlig [2000:119] for a summary). We can assume that the first two of these stages were replicated in the construction of individual interlanguages that formed the basis for the tense-aspect systems of contact Englishes. These are the stages associated with processes of ‘simplification’, including loss/nonacquisition of inflectional
< previous page
page_224
next page >
< previous page
page_225
next page >
Page 225 morphology and regularization of simple strategies for conveying temporal and other meanings. These are what might be referred to as ‘vernacular universals’. For instance, the use of bare verbs and nouns is a universal feature of early interlanguage systems. Similar universal features include variable use of copula/auxiliary be, omission of subject pronouns, the use of a single all purpose negator, and so on. The universality of several of these features in contact Englishes, as well as in creoles and early interlanguage, suggests once more that all of these situations are characterized by very similar types of restructuring in the early stages of acquisition. By contrast, the developments that occur in Stage 3 of the acquisition of tense-aspect clearly differ across interlanguages and contact languages. The development of verbal morphology outlined by Bardovi-Harlig applies only to those learners who have close and continuing access to input from native models of the target language. At this stage of SLA, we do find some ‘universal’ features such as variable marking of past tense depending on predicate type (telic/nontelic), the stative/nonstative distinction, and the effects of phonetic environment. This pattern of past marking is quite similar across contact Englishes. In these cases, substratum influence or L1 retention appears to play a secondary role to ‘universal’ acquisitional processes such as simplification. This said, contact languages that arise from communal language shift may follow very different paths of development depending on the ecology of the contact situation, including demographics, patterns of interaction between speakers of the languages in contact, the nature of the superstrate and substrate inputs, the motivations of learners, the social functions of the contact language, and so on. It would therefore be simplistic to expect some ‘universal’ pattern of development in every stage of acquisition. The differences among Irish English, Bajan, and SingE are testimony to the fact that different sociolinguistic ecologies may result in both similarities and differences in the outcomes of contact. For instance, given the fact that southwest English dialects were a significant part of the superstrate input to both Irish English and Bajan, it is not surprising to find a certain degree of similarity in the tense-aspect systems of these languages, particularly with regard to the functions of forms of do. By the same token, correspondences between the relevant substrate languages can also result in similarities across contact languages. The best-known examples of this include Caribbean creoles (as discussed earlier) and the Melanesian Englishes of the Pacific (as discussed by Keesing 1988; Siegel 2000; and others). Another case in point is the similarities we find between SingE and Hawai’i Creole English because of their shared Sinitic substrate. For instance, the adverb already has been grammaticalized in both languages as a marker of Completive aspect with similar functions. In short, just as in the case of Atlantic creoles, such similarities as we find between HCE and SingE tense-aspect systems can be more readily explained in terms of shared substrate influence than in terms of some putative bioprogram.
< previous page
page_225
next page >
< previous page
page_226
next page >
Page 226 Similarities among different substrates can also lead to similar outcomes across contact situations. Thus, Ho and Platt (1993:1) note that ‘basilectal’ SingE shares many features with Caribbean creoles, including directional serial verb constructions, variable marking of past tense, variable absence of copula and auxiliary be, existential got, and so on. The fact is that these features are not uniquely ‘creole’, but rather result from the fact that the respective substrates have similar structures. At the same time, we can find subtle differences in the nature of the substrate influence despite their apparent similarities. For example, SingE displays a hierarchy of copula absence in which forms of be appear most often with locatives, then nominals, progressive V-ing, and Adjectives, in that order: Locatives > Nominals > V-ing > Adjectives As Ho and Platt (1993:55–68) argue, this pattern corresponds closely to that in Chinese. In contrast, Bajan and other intermediate creoles display the following hierarchy: Nominals > Adjectives/Locatives > V-ing This appears to reflect, in part, the pattern in West African languages. Such data are clear evidence that substratal inputs can lead to subtle differences in what might appear at first glance to be an innovation that is shared across contact languages. None of this is meant to imply, however, that the nature of the substrate influence alone explains the similarities and differences we find across contact Englishes or that universal principles played no role. There is an unfortunate tendency to treat universals and substrate influence as if they were opposing factors in the genesis of contact languages. It seems more feasible to view them as complementary, with universal principles acting as constraints on the role of substrate influence and other factors. Recent research on creole formation has attempted to provide a more explicit formulation of some of these principles. First, the research of Siegel (2003) builds on the insights of research in SLA to explore the various constraints on the role of transfer (L1 influence) in creole formation. Siegel argues that such constraints fall into two broad categories: ‘availability constraints’ and ‘reinforcement principles’. The former are based ultimately on Andersen’s (1983, 1990) “Transfer to Somewhere Principle” for SLA and have to do with whether there is a salient morpheme or string of morphemes in the superstrate input, which “can be used or reanalyzed according to the rule of the substrate” (2000:83). Reinforcement principles have to do with the degree of homogeneity among the substrates, which has the effect of ensuring the preservation of certain features rather than others. A complementary line of research is that of Lefebvre (1998) and her associates, which attempts to explain substratum influence in terms of the process
< previous page
page_226
next page >
< previous page
page_227
next page >
Page 227 of ‘relexification’, by which superstrate forms assume the functions and syntactic properties of semantically related items in the substrates. This approach provides a more formal framework in which to investigate such changes. A third line of research is that of Myers-Scotton (2002) and her associates, which appeals to psycholinguistic models of language production to account for how speakers combine aspects of the lemmas of superstrate items with those of corresponding substrate items. All three approaches are concerned with the same problem: how to account for the ways in which substratum influence is regulated by universal principles. These principles provide an explanation for how several tense-aspect categories emerged in contact Englishes. Cases in point include the after Perfect of Irish English, the already Completive of SingE, and the done Perfect of Bajan. Research on contact Englishes, therefore, has much to gain from these approaches to creole formation. In summary, we have to recognize that differences in both the linguistic inputs as well as the social settings of the contact play a major role in determining the nature and extent of the restructuring that takes place in the emergence of contact languages. The fact that the outcomes are not identical in all cases is clear evidence that such differences are crucial. At the same time, we should recognize that all such outcomes are shaped by similar principles that regulate contact-induced language change. Understanding how the different sociolinguistic ecologies, linguistic inputs, and universal principles interact with each other constitutes one of the central goals of a theory of contact-induced change. Students of contact Englishes can contribute much to this endeavor by documenting the details of the sociohistorical settings which gave rise to the new Englishes and creating taxonomies of the typological similarities and differences that they manifest. We can then achieve the goal of explaining our typological generalizations in terms of the general principles that underlie processes of contact-induced change. NOTES 1. It has also been claimed that did has habitual function in contemporary southwest English dialects (Ihalainen 1976:615). Klemola (1996:123) analyzed data from three southwestern dialect corpora (twentieth century) and concluded that, although “there is a clear tendency” for did to receive a habitual interpretation (67–85 percent of all tokens), did was also used to convey simple past (3–15 percent) and in conditional and temporal clauses, “where the form seems to function as a marker of irrealis modality rather than habitual aspect” (10–15 percent). It may well be that the tendency to use did for habitual meaning in some contemporary southwest dialects is a relatively recent development. REFERENCES Andersen, R.W. 1983. Transfer to Somewhere. In Language Transfer in Language Learning , edited by S. Gass and L. Selinker, 177–201. Rowley, MA: Newbury House.
< previous page
page_227
next page >
< previous page
page_228
next page >
Page 228 ———. 1990. Models, processes, principles and strategies: Second language acquisition inside and outside the classroom. In Second Language Acquisition/Foreign Language Learning , edited by B. VanPatten and J.F. Lee, 45–66. Philadelphia: Multilingual Matters. Ansaldo, U. 2004. The evolution of Singapore English: Finding the matrix. In Singapore English: A Grammatical Description , edited by L. Lim, 129–64. Amsterdam: John Benjamins. Bao, Z. 2005. The aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics 41: 237–67. Bardovi-Harlig, K. 2000. Tense and Aspect in Second Language Acquisition: Form, Meaning and Use. Oxford: Blackwell. ———, and Reynolds, D. 1995. The role of lexical aspect in the acquisition of tense and aspect. TESOL Quarterly 29(1): 107–31. Barnes, W. 1886. A Glossary of the Dorset Dialect with a Grammar of Its Word Shapening and Wording. London: Trübner and Co. (Reprinted 1970, Guernsey: Steven Cox, The Toucan Press.) Bickerton, D. 1981. Roots of Language. Ann Arbor: Karoma. ———. 1984. The language bioprogram hypothesis. The Behavioral and Brain Sciences 7: 173–88. Bliss, A.J. 1984. English in the south of Ireland. In Language in the British Isles , edited by P. Trudgill, 135– 51. Cambridge: Cambridge University Press. Burrowes, A. 1983. Barbadian creole: A note on its social history and structure. In Studies in Caribbean Language, edited by L.D. Carrington, 38–45. Society for Caribbean Linguistics, University of the West Indies, Trinidad. Bybee, J., R. Perkins, and W. Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago: The University of Chicago Press. Chambers, J.K. 2004. Dynamic typology and vernacular universals. In Kortmann, ed. 2004. 127–45. Deterding, D. 2003. Tenses and will/would in a corpus of Singapore English. In Deterding et al., eds. 2003. 31–38. ———, L.E. Ling, and A. Brown, eds. 2003. English in Singapore: Research on Grammar . Singapore: McGraw-Hill Education (Asia). Ellegård, A. 1953. The Auxiliary DO: The Establishment and Regulation of Its Use in English (Gothenburg Studies in English 2). Stockholm: Almqvist & Wiksell. Elworthy, F.T. 1877. An Outline of the Grammar of the Dialect of West Somerset. Illustrated by Examples of the Common Phrases and Modes of Speech Now in Use Among the People. London: Trübner and Co. ———. 1886. The West Somerset Word Book: A Glossary of Dialectal and Archaic Words and Phrases Used in the West of Somerset and East Devon . London: Trübner and Co. Epstein, S.D., S. Flynn, and G. Martohardjono. 1996. Second language acquisition: Theoretical and experimental issues in contemporary research. Behavioral and Brain Sciences 19(4): 677–714. Filppula, M. 1990. Substratum, superstratum, and universals in the genesis of Hiberno-English. Irish University Review 20: 41–54. ———. 1997. The influence of Irish on perfect marking in Hiberno-English: The case of the “Extended-now” Perfect. In Focus on Ireland (Varieties of English Around the World, Vol. 21), edited by Jeffrey Kallen, 51– 69. Amsterdam: John Benjamins. ———. 1999. The Grammar of Irish English: Language in Hibernian Style. London: Routledge. Harris, J. 1984. Syntactic variation and dialect divergence. Journal of Linguistics 20: 303–27.
< previous page
page_228
next page >
< previous page
page_229
next page >
Page 229 ———. 1986. Expanding the superstrate; habitual aspect markers in Atlantic Englishes. English World-Wide 7(2): 171–99. ———. 1993. The grammar of Irish English. In Real English: The Grammar of English Dialects in the British Isles , edited by J. Milroy and L. Milroy, 139–86. London: Longman. Henry, P.L. 1957. An Anglo-Irish Dialect of North Roscommon. Zurich: Aschmann & Scheller. Ho, M.-L. 2003. Past tense marking in Singapore English. In Deterding et al., eds. 2003. 39–47. ———, and J.T. Platt. 1993. Dynamics of a Contact Continuum: Singapore English. Oxford: Clarendon Press. Ihalainen, O. 1976. Periphrastic do in affirmative sentences in the dialect of east Somerset. Neuphilologische Mitteilungen 77: 608–22. Jago, F. 1882. The Ancient Language and the Dialect of Cornwall. Truro: Netherton and Worth. Kallen, J. 1989. Tense and aspect categories in Irish English. English World-Wide 10(1): 1–39. Keesing, R. 1988. Melanesian Pidgin and the Oceanic Substrate. Stanford: Stanford University Press. Kiparsky, P. 2008. Universals constrain change; change results in typological generalizations. In Language Universals and Language Change, edited by Jeff Good, 23–53. Oxford, UK: Oxford University Press. Klein, W. 1993. The acquisition of temporality. In Adult Second Language Acquisition: Cross-linguistic Perspectives , Vol. 2. The Results , edited by C. Perdue, 73–118. Cambridge UK: Cambridge University Press. ———. 1995. The acquisition of English. In The Acquisition of Temporality in a Second Language, edited by R. Dietrich, W. Klein, and C. Noyau, 31–70. Amsterdam: John Benjamins. Klemola, J. 1996. Non-standard Periphrastic DO: A Study in Variation and Change . PhD Dissertation, University of Essex. Kortmann, B., ed. 2004. Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective . Berlin: Mouton de Gruyter. Kós-Dienes, D. 1984. Sketch of the TMA system of Akan. In Selected Papers from the Tense-Mood-Aspect Project , edited by Ö. Dahl and D. Kós-Dienes, 82–7. Stockholm, Sweden: Institute of Linguistics, University of Stockholm. Lefebvre, C. 1996. The tense, mood and aspect system of Haitian Creole and the problem of transmission of grammar in creole genesis. Journal of Pidgin and Creole Languages 11: 231–313. ———. 1998. Creole Genesis and the Acquisition of Grammar: The Case of Haitian Creole. Cambridge: Cambridge University Press. Le Page, R.B. 1960. An historical introduction to Jamaican Creole. In Jamaican Creole: Creole Language Studies 1, edited by R.B. Le Page and D. DeCamp, 1–124. London: Macmillan. ———, and A. Tabouret-Keller. 1985. Acts of Identity . Cambridge: Cambridge University Press. Mufwene, S. 1988. Formal evidence of pidginization/creolization in Kituba. Journal of African Languages and Linguistics 10: 33–51. Myers-Scotton, C. 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes . Oxford: Oxford University Press. Niles, N.A. 1980. Provincial English Dialects and Barbadian English. PhD dissertation, University of Michigan. Odlin, T. 1997. Hiberno-English: Pidgin, creole or neither? CLCS Occasional Paper No. 49. Trinity College, Dublin: Centre for Language and Communication Studies.
< previous page
page_229
next page >
< previous page
page_230
next page >
Page 230 Platt, J.T., H. Weber, and M.L. Ho. 1983. Singapore and Malaysia (Varieties of English around the world, Vol. 4). Amsterdam: John Benjamins. Ritchie, W.C. 1986. Second language acquisition and the study of non-native varieties of English: Some issues in common. World Englishes 5: 15–30. Schlauch, M. 1973. The English Language in Modern Times (Since 1400). Oxford: Oxford University Press. Schwartz, B.D., and R.A. Sprouse. 1996. L2 cognitive states and the Full Transfer/ Full Access model. Second Language Research 12: 40–72. Siegel, J. 2000. Substrate influence in Hawai’i Creole English. Language in Society 29: 197–236. ———. 2003. Substrate influence in creoles and the role of transfer in second language acquisition. Studies in Second Language Acquisition 25: 185–209. Siemund, P. 2004. Substrate, superstrate and universals: Perfect constructions in Irish English. In Kortmann, ed. 2004. 401–34. Singler, J. 1990. On the use of sociohistorical criteria in the comparison of creoles. Linguistics 28: 645–69. Traugott, E. 1972. A History of English Syntax: A Transformational Approach to the History of English Sentence Structure. New York: Holt, Rinehart & Winston. Williams, J.P. 1987. Anglo-Caribbean English: A Study of Its Sociolinguistic History and the Development of Its Aspectual Markers. PhD dissertation, University of Texas at Austin. Winford, D. 1993a. Predication in Caribbean English Creoles. Amsterdam: Benjamins. ———. 1993b. Back to the past: The BEV/Creole connection revisited. Language Variation and Change. 4: 311–57. ———. 1998. On the origins of African-American Vernacular English—creolist perspective. Part II: Linguistic features. Diachronica 15: 99–154. ———. 2000. Tense and aspect in Sranan and the creole prototype. In Language Change and Language Contact in Pidgins and Creoles, edited by J. McWhorther, 383–442. Amsterdam: John Benjamins. ———, and B. Migge. 2007. Substrate influence on the emergence of the TMA systems of the Surinamese creoles. Journal of Pidgin and Creole Languages 22(1): 73–99.
< previous page
page_230
next page >
< previous page
page_231
next page >
Page 231 10 Digging for Roots Universals and Contact in Regional Varieties of English1 Markku Filppula, Juhani Klemola, and Heli Paulasto 1. INTRODUCTION This chapter explores the relationship between vernacular and other kinds of universals and language or dialect contacts through a corpus-based comparative method. We focus on three nonstandard syntactic features that can be assumed to represent different levels of universality and language/ dialect contact backgrounds and their distribution in two sets of regional or national varieties of English, which in turn can be positioned differently on the vernacular–standard continuum, and furthermore, have different sociohistorical backgrounds either as so-called ‘L1 varieties’ or as ‘contact’ or ‘L2 varieties’. It is hoped that this kind of comparison will shed light on the question of whether there are ‘vernacular universals’ in the sense of Chambers (2004) and, if so, what kinds of varieties of English constitute their proper locus. Crosslinguistic comparisons will also be made wherever possible to ascertain the ‘degree of universality’ of the investigated features. The three features are: (1) absence of plural marking with nouns of measurement, (2) nonstandard use of the definite article in certain kinds of contexts, and (3) wider use of the progressive form with stative verbs. We describe these features in detail in sections 3 to 5 and give an account of their distribution in the investigated varieties and globally. All three features represent nonstandard syntax, and they are found— with varying frequencies—in all of the varieties investigated here. They have a common denominator which functioned as the starting point for selecting them: they all occur in (at least one of) the Celtic-influenced varieties of English spoken in the British Isles and have structural parallels in one or the other of the Celtic languages. Thus, there is a basis for considering each of them contact-induced at least within their own regional settings. What adds a possible global and universal dimension to the question of their origins is the fact that these features are also found in several other varieties of English beyond the context of the British Isles. Thus, the question of origins cannot be solved without taking into account several possible sources of influence,
< previous page
page_231
next page >
< previous page
page_232
next page >
Page 232 including vernacular or other kinds of universals, dialect or language contacts, and universal features of second-language acquisition (SLA). 2. A DESCRIPTION OF THE DATABASES AND THE GENERAL BACKGROUNDS OF THE INVESTIGATED VARIETIES As mentioned, our database consists of data drawn from two sets of varieties of English. The first set consists of three varieties of nonstandard traditional vernacular English spoken in the British Isles, while the second set comprises five varieties representing (more or less) standardised or educated varieties of English spoken in Britain and in Asia and Africa. The nonstandard vernacular varieties are: Traditional English English Dialects (EngE), Welsh English (WE), and (southern) Irish English (IrE). The following is a description of the databases representing these varieties: • EngE: The Survey of English Dialects Tape-recordings (Klemola forthcoming). This corpus consists of informal, transcribed interviews of 298 informants from 286 localities. The majority of the recordings were made in the 1950s. The informants were of the NORM type and were born between 1863 and 1909; thus, their speech can be characterised as very conservative. The length of this corpus is 515,000 words. • WE: A corpus of speech collected by Heli Paulasto from Llandybie, Carmarthenshire, southwest Wales, between 1995 and 2000. This corpus, a part of a larger one, also consists of informal interviews with 12 informants born between 1915 and 1930. Nearly all the informants were first-language Welsh speakers and had a minimum education. This corpus comprises 61,400 words (see Paulasto 2006 for further details). • IrE: A corpus collected and compiled from various sources by Markku Filppula in the late 1970s and the early 1980s. This corpus contains informal interviews with 20 elderly speakers from four different regions in the south of Ireland: Counties Clare, Kerry, and Wicklow, and the city of Dublin. All informants had English as their first language, but those from the (south)-west of Ireland, in particular, had some knowledge of Irish, too. The length of this corpus is 158,000 words (for further details, see Filppula 1999). The educated varieties, which form the second set of varieties, can all be characterised as representing different national varieties which, however, are standardised to varying degrees and have very different sociohistorical backgrounds. They are: British English (BrE), Indian English (IndE), Singapore English (SingE), Philippines English (PhilE), and East African English (EAfrE). The Indian, Singaporean, Philippine, and EastAfrican varieties of
< previous page
page_232
next page >
< previous page
page_233
next page >
Page 233 English have certain characteristics in common which distinguish them from British English: they are nonnative, yet to a certain extent institutionalised varieties, having developed through the educational system, and are generally spoken as an L2 in their respective countries. All of the corpora representing the national varieties form part of the so-called International Corpus of English project. The details of the parts used for this study are as follows: • BrE: ICE—Great Britain . The spoken, unscripted components of the corpus, totalling 528,500 words. • IndE: ICE—India . Spoken, unscripted, total 553,400 words. • SingE: ICE—Singapore . Spoken, unscripted, total 496,200 words. • PhilE: ICE—Philippines . Spoken, unscripted, total 555,700 words. • EAfrE: ICE—East Africa . Spoken, unscripted, total 261,600 words. As in all of the ICE corpora, the ‘spoken’ sections contain transcripts of speech obtained from various types of spoken situations: dialogue and monologue, face-to-face conversations and radio interviews, lectures, and sportscasts. Stylistically, they range from the informal and colloquial to more formal situations. Of the five national varieties represented in the ICE corpora, educated BrE comes closest to our general conception of ‘Standard English’. As for the others, the picture is not so straightforward. In India, there are in fact two varieties of English used by the “educated members of the Indian speech community” (Bhatt 2004:1016–17): Standard IndE, which is essentially similar to Standard BrE, but differs in its phonetics, and vernacular IndE, which shows structural influence from local languages and is used in ‘low’ functions. Kachru (1996:907, cited in Bhatt 2004) states that the teachers of English were first and foremost locals from an early stage, but that they were formerly often overseen by teachers of Irish, Welsh, or Scottish backgrounds. The main substratum language in IndE is Hindi, which is the primary language in most of the country. Dravidian languages are spoken in southern India. SingE is, according to Zhiming (2005:237), “a contact language with a constant linguistic ecology”. The main substratum language is (Mandarin) Chinese, while smaller roles are played by Malay and Tamil, together with some other Indian languages. Platt, Weber, and Ho (1983:8–10) note that SingE was also influenced by Indian teachers of English in the nineteenth century: there are certain similarities in lexical and syntactic usage. Today, English is increasingly important as the language of commerce, science, technology, and media, but almost all Singaporeans speak another language too (Platt, Weber, and Ho 1983:12; see also Ansaldo 2004; Lim and Foley 2004). In the Philippines, educated Filipino speakers of English speak ‘Standard Filipino English’, which includes stylistic shift, code mixing, and code
< previous page
page_233
next page >
< previous page
page_234
next page >
Page 234 switching (Gonzales 1997:205–7). The less standardised, local varieties of English are very much influenced by a substratal Philippine vernacular, whether Tagalog, Ilocano, or some other ethnic language. Thompson (2003) writes that, historically, PhilE is more closely connected to AmE than IndE and SingE because the Philippines were a U.S. colony from the early twentieth century up until 1946. In the last couple of decades, the status of English has shifted from an L2 to ‘English as a foreign language’ with the expanded use of Tagalog/Filipino (Thompson 2003; see also Sibayan 1998). As for EAfrE, Schmied (2004:929) states that an independent EAfrE grammar is not easily distinguished. Dialect features occur infrequently, which may partly result from stigmatisation. English is a language of the high domains only, particularly in Tanzania and Malawi. Thus, African English exists in a continuum with StE (Schmied 1997). The main substratum language is Swahili, which is an intranational lingua franca of the eastern parts of Central and Southern Africa. In the following sections, we move on to a detailed discussion of the three syntactic features, their distribution in the mentioned varieties, and their origins. 3. ABSENCE OF PLURAL MARKING WITH NOUNS OF MEASUREMENT (APM) In this section, we first discuss APM in our set of British Isles vernacular varieties, with a view to establishing its extent of use in these varieties and the possible effect of their different sociohistorical backgrounds. The next step is to compare the findings with the data from the ICE corpora to see to what extent the occurrence of the APM is conditioned by the more standard type of language represented in these corpora. As said before, the ICE corpora investigated here also enable comparisons between L1 and L2 varieties of English. Finally in this section, we place the results of our inquiry in a global cross-linguistic perspective to ascertain the degree of universality of the APM. 3.1 APM in Vernacular British Isles Varieties On the basis of previous research, APM presents itself as a promising candidate for a vernacular universal at least in the context of the British Isles. In his English Dialect Grammar , Joseph Wright (1905, § 382) states that “[n]ouns expressing time, space, weight, measure and number when immediately preceded by a cardinal number generally remain unchanged in the plural in the dialects of Scotland and England”. In a similar vein, but in even stronger terms, Edwards and Weltens (1985) describe this feature of British dialects as “almost a universal rule”, which dictates that, “after numerals, nouns of measurement and quantity retain their singular form, e.g. twenty
< previous page
page_234
next page >
< previous page
page_235
next page >
Page 235 year, two dozen, twenty pound, three acre, six pint ” (Edwards and Weltens 1985:114). For our study, we have selected 10 nouns of measurement for closer examination: mile, foot/feet, inch , acre, stone , pound (both money and weight), year , kilo, kilometre, and metre. Although the last three can be expected to be rare in the vernacular British Isles Englishes for obvious reasons, they were included so as to take into account the metric system prevalent in some other varieties. By the same token, some of the imperial measures can be expected to be rare in these latter varieties. The unavoidable problem for quantitative comparisons was not, however, considered harmful in the light of the general scarcity of the APM in the ICE corpora (see below). The syntactic contexts considered here were ones in which these nouns were preceded by a numeral attribute, such as a cardinal number, or some other quantifying expression, such as a couple of or a few. APM in the mentioned contexts is illustrated by the following example sentences from the investigated varieties: (1)I once remember six inch of snow on t(he) fifth of November, at … (SED Y4: Egton, Yorkshire) (2)…. to start thy mowing. “It was three acre, and he went and mew [mowed] this … (SED Y4: Egton, Yorkshire) (3)Oh yes, well, after that I did a lot of cycling. [Mm.] Twenty or thirty mile a day. (WE, Llandybie: ML) (4)They offered me two hundred pound redundancy money. (IrE, Dublin: M.L.) Table 10.1 provides the frequencies of the APM in the British Isles corpora. In addition to the absolute and relative frequencies, the table gives in brackets the normalised frequencies per 10,000 words. Table 10.1 Plural Marking in EngE (SED), WE and IrE Corpus Ø -s Total %Ø SED 617 581 1198 51.5 (515,000) (11.98) (11.28) (23.26) WE 23 129 152 15.1 (61,400) (3.75) (21.01) (24.76) IrE 171 331 502 34.1 (158,000) (10.82) (20.95) (31.77)
< previous page
page_235
next page >
< previous page
page_236
next page >
Page 236 As the figures show, the highest percentage of the APM is found in SED, where over half of all instances lack the plural suffix. IrE presents the second highest percentage, at approximately one third of all instances, while WE trails somewhat behind the two others. As can be expected, the rate of occurrence of the APM varies between the corpora depending on the noun, but there are some clear tendencies as well. The nouns that show the highest rates of zero marking in these corpora are stone and pound , with foot being the next (except the WE corpus, where both of the two instances are marked for plural). By contrast, inch typically receives the plural suffix; year , mile, and acre are likewise infrequent in the singular form in corpora other than SED. Unsurprisingly, metric nouns are not part of traditional vernacular speech, and no instances were found in these corpora. The figures for the five most common nouns attracting APM in SED are: pound (89 percent), stone (86 percent), foot (71 percent), mile (66 percent), and year (44 percent). There are no instances of stone/s in the WE corpus, and even the two instances of foot/feet appear in the plural form. As in SED, APM is most common there with pound (67 percent), whereas it remains rather infrequent with the others: mile (8 percent) and year (5 percent). In the IrE corpus, APM is common with all items but acre: stone (100 percent), pound (64 percent), foot (35 percent), mile (32 percent), and year (25 percent). The relatively high rates of usage of the APM in the SED data could of course be explained by historicallinguistic reasons: Old English had a smallish set of measure nouns that were either uninflected in the nominative and accusative plural (e.g. gëar ‘year’; pünd ‘pound’) or had the genitive plural ending -a (twentig fota ), where, after the loss of unstressed vowels, the form became identical with the singular. Thus, APM could be considered first and foremost a conservatism, reflecting the old usage. The regional distribution of the APM within the SED data offers some support for this line of argumentation: the percentage of the APM was highest (65 percent) in the generally more conservative Northern counties, whereas the proportional use in the other SED regions ranged between 46 and 51 percent. This cannot, however, be used to account—at least not without reservations—for the almost equally prominent occurrence of the APM in IrE, not to mention some vernacular varieties spoken beyond the British Isles, which is discussed in the next section. For IrE, there is always the possibility of substratum influence from Irish Gaelic to be reckoned with. In this case, however, it does not offer any straightforward basis for ‘interlingual identifications’ and transfer because in most contexts the Irish nouns of measurement show plural marking. However, what could have provided a model for the Irish learners of English in the most intense times of language shift is the presence in some Irish dialects of a special older plural form which is only used with measure nouns and which is shorter in form and thus closer to the singular form than the normal plural form (for examples, see Ó Siadhail 1989:166–7). Irish also uses the singular of nouns with the numeral ten and multiples
< previous page
page_236
next page >
< previous page
page_237
next page >
Page 237 of ten , and this is in fact mentioned by Joyce (1910/1988:84) as a feature which may have affected the IrE usage. The likelihood of substrate influence should be even greater in WE because modern Welsh has no plural marking after any cardinal numerals (Thorne 1993:149–50).2 This expectation is only partially borne out by our WE data, which show only limited use of APM. That the APM is a feature of vernacular WE dialects is confirmed by the Survey of Anglo-Welsh Dialects (SAWD) data discussed in Parry (1999:107). They include examples like two year ago , two gallon, three acre big, two or three ton, four ounce, two inch , twenty-odd mile away, and eight foot long recorded from various localities. However, because this survey does not contain any quantitative data, it is impossible to assess the generality of the APM across different WE dialects. Our own searches through a fully transcribed section of the SAWD interviews, consisting of some 26,500 words, yielded five instances of inch , two instances of foot, and another two instances of pound , thus lending some more credence to the existence of this feature in vernacular WE dialects. To sum up the results from the British Isles vernacular varieties, APM is most common in EngE dialects and, on average, reaches the level of being the preferred choice in them. It is slightly less frequent in IrE and, perhaps surprisingly, least frequent in our WE data, despite the close substratal parallels. To round off the overall picture for the British Isles varieties, it is interesting to compare our results with those of Macaulay (1991:110), who has investigated number marking in measure nouns in the dialect of Ayr, Scotland. According to his findings, plural nouns of measurement lacked inflection in 54 percent of all 90 instances. This puts the rate of usage of the APM in Ayr dialect at an even higher level than the SED data, yet Macaulay considers the results from Ayr too low to justify APM being a “universal rule” (Macaulay 1991:110). 3.2 APM in ICE Varieties In this section, we look at APM as it appears in five national educated varieties, including BrE, and we rely our discussion on the spoken components of the ICE corpora representing each of these. We begin with a selection of examples from ICE-GB, ICE-India, ICE-Philippines, ICE-Singapore, and ICE-East Africa: (5) Five foot eight inches tall stocky lad about twelve and a half stone (ICE-GB: S2A-017) (6) A: How can it be man laughter He is not even thirty-eight thirty-five kilo <$A> B: Only five kilo of flesh is there <,> (ICE-Ind: S1A-017)
< previous page
page_237
next page >
< previous page
page_238
next page >
Page 238 (7)Uh we know that uh uh in Subic uh there are uh large areas for uh building spaces for storage spaces we know that there is a runway <.> the there uh nine <.> hu nine thousand uh uh foot long (ICE-Phil: S1B-059) (8)If you’re eighteen year old you don’t mind looking thirty because even if you look thirty cause even if you look thirty you’re not that old (ICE-Sing: S2A-028) (9)He was so worried that he was already six feet and he was uh fifteen year I mean sort of odd thing he was six feet up and he was sixteen year and he thought that this is terrible (ICE-EAfr: S1B-048K) The frequencies of APM with respect to the same 10 nouns of measurement as before are given in Table 10.2. As Table 10.2 shows, both the absolute and normalised frequencies of the APM are clearly lower in the ICE corpora on the whole than in those from the vernacular British Isles varieties. We take this to indicate that the rates of occurrence of the APM are, indeed, conditioned by the vernacular–standard continuum examined here. Although less frequent than in ICE-GB, APM is most varied in the IndE corpus, with 6 of the 10 investigated nouns showing APM: kilo (100 percent), inch (38 percent), foot (29 percent), pound (25 percent; only one instance), metre (3 percent; one instance), and year (2 percent). In the ICE-GB data, APM is restricted to only two nouns—i.e. stone (100 percent) and foot (52 percent). The different rankings of the nouns attracting APM in ICE-India, ICE-GB, and also the SED data discussed in the previous section demonstrate that this feature of educated IndE cannot Table 10.2 Plural Marking in the ICE Corpora Corpus Ø -s Total %Ø ICE-GB 25 546 571 4.4 (528,500) (0.47) (10.33) (10.80) ICE-Ind 16 473 489 3.3 (553,400) (0.29) (8.55) (8.84) ICE-Sing 1 400 401 0.2 (496,200) (0.02) (8.06) (8.08) ICE-Phil 2 391 393 0.5 (555,700) (0.04) (7.04) (7.07) ICE-EAfr 4 210 214 1.9 (261,600) (0.15) (8.03) (8.18)
< previous page
page_238
next page >
< previous page
page_239
next page >
Page 239 at least solely be ascribed to superstratal influence from BrE. As for the other ICE varieties, our results show that year and foot are the only measure nouns attracting APM in ICE-Sing, ICE-Phil, and ICE-EAfr, and the instances of even those are very few.3 Here our findings are in some contrast with some earlier works on these varieties. Platt, Weber, and Ho (1983, 1984:46–9), for example, note that lack of plural marking is a common feature in the colloquial varieties of IndE, PhiE, and SgE. According to them, it may be caused by two factors: the reduction of consonant clusters and substratum influence from (e.g. Chinese). In EAfrE, as Schmied (2004:932) points out, plural marking may be omitted as a result of a simplification of the English system in a language contact situation. Direct substratum influence is out of the question because nominal inflections are much more complex in Bantu languages than in English. On the basis of our findings, then, APM has a very limited existence in all five ICE varieties and, perhaps surprisingly, even more so in the standard or educated national varieties that have an L2 background. This does not, however, mean that it could not be more prominent in their nonstandard varieties. Indeed, according to The World Atlas of Morphosyntactic Variation in English4 (WAMVE; Kortmann and Szmrecsanyi 2004), APM comes out fifth on the list of the most frequent morphosyntactic nonstandard features in varieties of English worldwide, being attested in 37 of the 46 investigated varieties (Kortmann and Szmrecsanyi 2004:1154). This makes it the most widespread of the features included in the present study and thus a plausible candidate for at least a vernacular ‘angloversal’ if not a universal in a cross-linguistic sense—a question that we turn to next. 3.3 APM in a Cross-Linguistic Perspective Typologically, APM with nouns of measurement is a common feature in languages of the world. It is found in vernacular varieties of several Germanic languages, including Danish, Swedish, Dutch, and German (cf. Rohdenburg 2006). Lexically unrestricted use of singular or uninflected nouns after cardinal numbers is also a feature of non-Indo-European languages such as Hungarian, Estonian, and Finnish (cf. Seppänen and Seppänen, 1984), as well as Indonesian and Modern Chinese (cf. Rohdenburg 2006). Note, however, that in some of these languages (e.g. Finnish), APM is fully grammaticalised as a feature of the standard varieties. This, of course, adds another dimension to the question of the APM as a putative vernacular universal. Leaving this last question aside for a while, one can easily think of a functional explanation for the crosslinguistic generality of APM. Indeed, Jespersen (1924:208) argues that absence of plural marking with measure nouns is an “instance of wise economy”: In Magyar, there is the inverse rule that number is indicated in a secondary and not in a primary word, but only when a substantive is
< previous page
page_239
next page >
< previous page
page_240
next page >
Page 240 accompanied by a numeral. It is, then, put in the singular, as if we were to say “three house.” This is termed “illogical” by the eminent native linguist Simonyi: I should rather call it an instance of wise economy, as in this case any express indication of the plurality of the substantive would be superfluous. (Jespersen 1924:208) In a similar vein, Stebbins (1997:16) states that APM is an example of a general tendency found in many (unrelated) languages “towards marking only the necessary and sufficient amount of number morphology required to convey plurality”. Williams (1987:176), writing from an SLA perspective, finds that the omission of plural - s after numerals, quantifiers, and partitives is a case of selective production of redundant marking, which is a common pattern of learner English. Thus, APM appears to be motivated by the cross-linguistic principle of morphological economy in both L1 and L2 type of varieties or languages. To conclude this section, our research has shown that APM with measure nouns is a common feature of the vernacular British Isles varieties of English, with the SED data, which represent very traditional dialects of EngE, topping the list. By contrast, it remains at best a marginal feature of the national standard varieties represented here by the spoken components of five ICE corpora. To that extent, the expected vernacular– standard continuum with respect to this feature is borne out by our results. This is also backed up by the WAMVE data from the nonstandard varieties of IndE, PhilE, SingE, and EAfrE, as well as numerous other such varieties, which show that APM is a very widespread phenomenon in nonstandard speech. We are aware that the results from the Asian and African ICE corpora may to some extent reflect the differing contexts in which the English language is used in the corpora. Schmied (1997:149), for example, writes that “all spoken private texts [in the ICE-East Africa] are problematic [because] English is normally used in domains related to the upper part of the formal spectrum, and other languages are preferred in conversations among family members and friends”. The original ICE text category ‘private dialogue’ of 100 texts is therefore replaced in ICE-East Africa by the category ‘direct conversation’ consisting of only 30 texts (Schmied 1997:147–8; see also Hudson-Ettle and Schmied 1999). Despite possible problems with the comparability of data, the overall trend emerging from our quantitative comparisons above is clear enough: as far as English is concerned, APM is primarily a feature of vernacular speech, whether with an L1 or L2 background (with a couple of well-known exceptions such as stone and foot, which are used by speakers of standard English as well). To that extent, APM can be considered a good example of a vernacular universal. It is also a contender for a place among vernacular universals in a ‘true’ cross-linguistic sense thanks to its widespread distribution in (varieties of) languages around the world. Furthermore, one can
< previous page
page_240
next page >
< previous page
page_241
next page >
Page 241 point out a persuasive functional rationale for the cross-linguistic popularity of APM—namely, that based on the semantic redundancy of the plural marker in the presence of an overt quantifier. What remains as a complicating factor from a cross-linguistic perspective is that APM does not appear to be restricted to vernacular varieties, but is fully grammaticalised in many standard languages. We cannot embark here on a discussion on the possible factors determining why a feature such as APM should be fully grammaticalised in one language but not in some others. Pending answers to these kinds of questions, we have to be content to say that APM can at least be considered a vernacular angloversal, but with counterparts in vernacular or other varieties of a wide range of other languages. 4. NONSTANDARD USES OF THE DEFINITE ARTICLE This section is devoted to another syntactic feature which is common in the Celtic-influenced varieties, but is also attested to varying degrees in other British Isles varieties and World Englishes, viz. wider use of the definite article in certain kinds of contexts. Standard grammars of English tell us that the definite article is typically left out in British StE in connection with, e.g. nonspecific or generic plural NPs; names of social institutions such as school, church , or hospital when used with nonspecific reference; quantifying expressions such as most , both, and half when followed by a postmodifying of- phrase; names of languages, festive days, or seasons; and means of transport or communication (see e.g. Biber, Johansson, Leech, Conrad, and Finegan 1999:261–3; Quirk, Greenbaum, Leech, and Svartvik 1985:277–9). In this study, we focus on article usage with names of social institutions and quantifiers. As in the previous section, we have selected a number of the most commonly occurring names of social institutions and quantifying expressions for a detailed examination. They are: school, hospital , church , college , university; most of , half of , and both of . The following examples from our database illustrate uses of the definite article which deviate from the usages described in standard textbooks on English grammar and can in this sense be considered nonstandard: (10)AC: I’d three brothers and two sisters, and they’re all gone bar [/] bar [\] me. MB: Hmm. AC: Yes. But the most of them, well only one that did worse than eighty. (SED Man2; Ronague, Isle of Man) (11)My- my grandmother now, she- she used- she said something to me erm, Tuesday she- she went to the hospital on Tuesday. There we are, I would say, “wyt ti mynd I’r hospit-” I- I would say, if I was to say, er, “are you going to the hospital on Tuesday”,
< previous page
page_241
next page >
< previous page
page_242
next page >
Page 242 [Mm.] in Welsh, I would say er, “wyt ti mynd i’r hospital dydd Mawrth?” I would use ‘hospital’ instead of ‘ysbyty’. (WE, Llandybie: EE) (12)A man who thinks about this > filling ?> things etcetera <,> he has missed his way in the college (ICE-Ind: S2A-035) (13)B: So Kareen Chow asked me to go. Would you go A: Depending on what is going on between the both of them (ICE-Sing: S1A-061) (14)B: Apparently there are less people traveling now because of the weather A: Uh I heard in the news that uh this morning the elementary and high school uh students uh are are excused to go to the school right now <,> but in certain areas I think only (ICE-Phil: S1A-050) (15)so he told me <-/>he told me<-/> that you see in Canada people don’t care about going to the church (ICE-EAfr: S1A-016) We are aware that definite article usage—or, indeed, article usage in general—is not something that would be governed by categorical rules and that there is a certain amount of variation in its use in any variety of English, including so-called Standard English, depending on all kinds of pragmatic and other factors (cf. Sand 2003:416, 425 and her references to e.g. Algeo 1988; Wächtler 1988). In the case of school and hospital , for example, it can be difficult to draw the line between nonspecific references to these institutions (as in go to school/be in hospital ), as opposed to references to specific instances of these social institutions. In Standard BrE, the definite article is only used in the latter case (cf. Biber et al. 1999:261). However, the wider textual and situational context of each of the occurrences usually gives reliable enough clues as to which kind of reference is at issue. As in the previous sections, we begin with the data from our set of vernacular British Isles Englishes and proceed thence to the ICE varieties and on to cross-linguistic comparisons. 4.1 Nonstandard Uses of the Definite Article in Vernacular British Isles Varieties The results on the mentioned uses of the definite article in our British Isles varieties are given in Table 10.3. In the traditional EngE dialects, there appears to be no clear preference for definite article with any of the investigated items; the numbers of
< previous page
page_242
next page >
< previous page
page_243
next page >
Page 243 Table 10.3 Definite Article Usage in EngE (SED), WE and IrE Corpus The Ø Total N % The SED 11 309 320 3.4 (515,000) (0.21) (6.00) (6.21) WE 9 124 133 6.8 (61,400) (1.47) (20.20) (21.66) IrE 21 111 132 15.9 (158,000) (1.33) (7.03) (8.35) instances range from 0 to 3, more than one instance being found only with school (3), hospital (2), and most of/on (3). In the WE corpus, the highest numbers of instances are found with hospital (4), both of (3), and school (2). In addition, we counted eight instances of the Welsh (‘Welsh language’) and four instances of the English, a feature which also has a Welsh parallel. The IrE corpus, then, has clearly the highest number of occurrences of the definite article in the contexts at issue. There are more than one instance with school (3), most of (13), and half of (2). Definite article is the preferred choice for IrE speakers only with most of (13 of 20); the next highest relative proportions are found with the other quantifying expressions, although the frequencies are very low (1 of 1 for both of , 2 of 4 for half of ). Like WE, IrE often has the definite article with names of languages and especially with that of the indigenous language, Irish (12 instances of the Irish , 3 of the English in our database). In the case of both WE and IrE, it is likely that the wider use of the definite article stems from substratum influence from the respective Celtic languages. In Welsh, the definite article is used with names of social institutions, as well as with e.g. y ddau ‘the both’ and y Saesneg ‘the English (language)’ (Thorne 1993:97– 100). Irish, too, has parallels to the use of the definite article with names of social institutions, and these parallels extend, in fact, to most other nonstandard uses found in IrE (Filppula 1999 section 5.2 for a detailed discussion). 4.2 Nonstandard Uses of the Definite Article in the ICE Varieties We now turn to the data from the five ICE corpora investigated here. Table 10.4 provides the frequencies of the definite article in the same contexts as before. The figures for ICE-GB show that the use of the definite article with names of social institutions or the mentioned types of quantifying expressions is virtually nonexistent in present-day educated BrE. The only instance is the most of , whereas all the others follow the standard practice of having zero article in these contexts. The percentages of definite article use are
< previous page
page_243
next page >
< previous page
page_244
next page >
Page 244 Table 10.4 Definite Article Usage in the ICE Varieties Corpus The Ø Total % The ICE-GB 1 192 193 0.5 (528,500) (0.02) (3.63) (3.65) ICE-Ind 34 252 286 11.9 (553,400) (0.61) (4.55) (5.17) ICE-Sing 9 236 245 3.7 (496,200) (0.18) (4.76) (4.94) ICE-Phil 9 284 293 3.1 (555,700) (0.16) (5.11) (5.27) ICE-EAfr 22 265 287 7.7 (261,600) (0.84) (10.13) (10.97) clearly higher on average in the Asian and African English corpora than in ICE-GB. IndE and EAfrE stand out from the two others in producing the highest overall percentages and most varied use of this feature; the levels of use in ICE-Sing and ICE-Phil are comparable with those observed for the SED data. Overall, definite article is the favoured variant over the zero article with hospital in all corpora but ICE-Sing. Although the rates of use range from 59 percent to 100 percent in the others, it is only 18 percent in ICE-Sing. Use of the definite article with university is also relatively common, the percentages ranging from 17 percent to 100 percent. Definite article with quantifiers appears to be rare in other than the Indian and East African varieties. In IndE, there is more than one instance of definite article use also with school (19 percent), college (19 percent), half of (22 percent), and most of (3 percent). We could add to this that, as in WE and IrE (see above), there are also seven instances of definite article with English (‘English language’). Apart from hospital and university, ICE-Sing has only two instances of definite article in the investigated contexts, viz. with the quantifiers half of and both of . The items with more than one instance in PhilE are: hospital (100 percent), church (25 percent), and university (100 percent). In ICE-EAfr, we found more than one instance with most of the investigated items, including church (17 percent), school (5 percent), and most of (2 percent). It seems to us that substratum influence is not a factor explaining the previous instances of the definite article in the Asian and African varieties because many of the relevant regional substratum languages (e.g. Hindi and Chinese) do not have a definite article (see Sand 2003:418). It is also noteworthy that in SingE and EAfrE, the general tendency is towards omission of articles and other determiners, rather than ‘overuse’ of the definite article (Schmied 2004:932; Wee 2004:1061; see also Platt, Weber, and Ho 1983).
< previous page
page_244
next page >
< previous page
page_245
next page >
Page 245 Article usage varies in PhilE, too, with the definite article being used with generic plural nouns such as the Filipino heroes (in general) instead of Filipino heroes and the food instead of food . However, it is not used with names such as United States or Rizal College (Thompson 2003:52–3). Bhatt (2004) finds similar variability in article use in IndE, but, as pointed out by Kachru (1983:78), no consistent patterns seem to exist. Despite the lack of any obvious substratal impact, insertion of definite article in place of indefinite or zero articles is common in IndE (Sedlatschek 2007, cited by Sand 2003:418). In SingE, too, the definite article tends to be preferred to the indefinite one (Wee and Ansaldo 2004). As regards EAfrE, Schmied (2004:932) notes that omission of the definite article “may partly be an overgeneralization of the British usage in contexts like I am going to church/school to new ones”, and that “subrules of StE grammar are [often] neglected”. This, however, begs the question of why such ‘neglect of subrules of StE grammar’ occurs in some varieties more than in others. This brings us suitably to the possible global or cross-linguistic implications of definite article usage, which are discussed in the next section. 4.3 Global and Cross-Linguistic Perspectives on Definite Article Usage A major impetus for choosing definite article usage as one of the features to be examined in this chapter came from Andrea Sand’s (2003) article, where she responds to Filppula’s (1999) account of definite article usage in IrE. Whereas Filppula argues that certain nonstandard uses of the definite article in IrE dialects can be ascribed to Irish substratum influence, Sand (2003:428) calls this in question by pointing out “striking parallels [to the IrE usages] in several contact varieties of English”. According to Sand, substratum influence cannot therefore be considered the common denominator for all these phenomena, but other ‘universalist’ or cognitive explanations must be sought. As one such explanation, she suggests that definiteness in general is linked to what is known as the Animacy Hierarchy, which involves placing the categories ‘human’ and ‘proper noun’ on top of the hierarchy. Because human referents can be assumed to be more relevant to communication than nonhuman ones, they tend to be definite (cf. Whaley 1997, cited in Sand 2003). This, in turn, may explain the use of the definite article in the contexts under discussion: most of the nonstandard uses relate to human referents or to institutions that are central to human life and behaviour, as well as being identifiable to speakers and hearers, and therefore tend to be marked for definiteness through the use of the article (Sand 2003:430). As regards the use of the article with quantifying expressions, Sand (2003:430) rather cryptically states that it is based on “an extension of the standard English logical use of the definite article”. What, however, remains unexplained on Sand’s account is why these universal constraints do not apply with equal force to all vernacular
< previous page
page_245
next page >
< previous page
page_246
next page >
Page 246 varieties or produce the same sets of nonstandard uses in them, let alone similar ‘extensions’ of the standard norms. Another factor that seems to set IrE and the other ‘Celtic Englishes’ apart from the contact varieties considered by Sand stems from Platt, Weber, and Ho’s (1984:52–9) observation according to which article usage in the ‘New Englishes’ such as IndE displays a general shift from the definite/indefinite distinction to a specific/ nonspecific one, affecting the way articles are used. Often this entails omission of the indefinite article with nonspecific items and the use of one/this/ these/that/those/the with specific ones. These tendencies are not characteristic of the Celtic Englishes. Another work to explore possible universal dimensions of article usage is Williams (1987:166–7), who considers the article system to be an area of the English language which is ‘inherently vulnerable’ (i.e. marked, as pointed out by Sarah Thomason during the discussions at the Symposium). This is shown by the fact that article usage frequently causes difficulties for second-language learners of English, leading to variability and modification. As regards such ‘areas of vulnerability’ in language, Williams (1987) observes that they are often subject to historical change as well as processes of language acquisition. She refers here to Slobin’s (1977) and Wode’s (1984) contentions that the same kind of psycholinguistic constraints that operate in these domains of grammar also apply to phenomena of language change, the structure of language contact varieties such as pidgins and creoles, and language typology, and can in this sense be considered universal. In light of the findings reported in the WAMVE (see Kortmann and Szmrecsanyi 2004:1155), it is incontestable that what is there labelled as ‘irregular use of articles’ is common in nonstandard varieties of English: it turned out to be the twelfth most common feature worldwide out of the 76 vernacular features and was attested in 33 of the 46 varieties included in the survey. It was a top feature in the British Isles, the Caribbean, Australia, Africa, and Asia, but not in America and the Pacific (Kortmann and Szmrecsanyi 2004:1157). On the basis of its distribution and despite its attestation in the British Isles varieties as well, it appears to be mainly a feature of L2 varieties of English, quite unlike the APM discussed in the previous section (cf. Kortmann and Szmrecsanyi 2004:1158). It must be noted, however, that the feature investigated in the WAMVE was irregular use of articles in general, while the present study focuses on the use of the definite article in contexts where it does not usually occur in (British) StE. In conclusion, our results point to a slightly different kind of vernacular– standard continuum as with the absence of plural marking. This time the main division among the vernacular British Isles varieties is between the Celtic Englishes (IrE, in particular) and the traditional EngE dialects. Still, even the latter show higher rates of occurrence of definite articles than Standard BrE. Interestingly, the Asian and African standard varieties are closer to the vernacular British Isles varieties and exhibit clearly higher frequencies of the definite article than Standard BrE. The IndE and EAfrE figures, in
< previous page
page_246
next page >
< previous page
page_247
next page >
Page 247 particular, would seem to suggest that the L2 background and universals of SLA play a role here. This explanation hardly suffices on its own to account for the salience of nonstandard uses in IrE, however: the existence of close substratal parallels to most of the nonstandard uses leads us to conclude that substratum influence is most likely in the case of IrE and, despite the somewhat lower figures in our WE database, at least probable in the case of most other Celtic-influenced varieties of English. For the Asian and African Englishes, however, it seems harder to identify plausible substratal sources, which is why universal tendencies of second-language learning override substratum explanations in these settings. 5. WIDER USE OF THE PROGRESSIVE FORM WITH STATIVE VERBS As compared with the two syntactic features discussed previously, we now move on to something that even further underscores the possible role of language contacts and the ‘L2 factor’, viz. wider use of the progressive form (PF) with stative verbs. What is also intriguing about this feature is its apparently rather recent but rapid spread into the more standardised varieties of English as well. What we mean here by the use of the PF with stative verbs refers to those instances where an ‘inherently’ stative verb (i.e. a verb denoting a state rather than a process, event, or other type of activity) is being used ‘dynamically’ with the PF. It is common knowledge that such inherently stative verbs as think, wonder, and hope can, and indeed very frequently are, used in a dynamic sense with the PF (cf. Biber et al. 1999:471–2). However, it is equally well known that the English language constrains the use of the progressive form in connection with certain (uses of) stative verbs, such as think in the sense ‘to hold a belief’, see and hear in the ‘pure’ perceptual sense, have in the sense of ‘be in possession of’, and many others like know, understand, want , like, love, and so on (cf. Biber et al., ibid.; Quirk et al. 1985:205). Although use of the PF with these verbs is generally unacceptable in their stative senses in StE, instances are found especially in many spoken varieties of English. The following are but some of the examples from our databases: (16)it were my sister come [: came] across (the) field and telled me as # there were a woman in (the) yard who were wanting (to) see me about (the) bull. (SED Db4; Youlgreave, Derbyshire) (17)A: You were listening. Were you understanding it B: No not a word of course (ICE-GB: S1A-069) (18)Because <,> they have seen them <,> in that labour camp locality <,,> and they were knowing the names (ICE-Ind: S2A-063)
< previous page
page_247
next page >
< previous page
page_248
next page >
Page 248 (19)A: I am having two room mates <,,> B: And they are belonging to <,> ? A: One belongs to Raichur and the other belongs to Mumbai <,,> (ICE-Ind: S1A-061) (20)I hope the guys who are hearing this uhm doesn’t get offended on what we are going to say (ICE-Phil: S1A-063) For our study, we have selected 10 commonly occurring verbs which represent different categories of states, as distinguished by Quirk et al. (1985:203): • Intellectual states: know, think (in the sense ‘to hold a belief’), believe, understand • States of emotion or attitude: intend, want • States of perception: see, hear • Other states: depend, belong It should be noted that only those uses of the previous verbs are considered where the subject is not presented as an active agent, and the state or process is not limited in duration (cf. Biber et al. 1999:473– 4). We also rely on the descriptions of stative verbs in e.g. Huddleston and Pullum (2002:169–70), Quirk et al. (1985:202–6), and Smith (2005:94–5) in ruling out instances where the PF indicates a process of ‘waxing’ or ‘waning’ (i.e. denoting an increase or a decrease in an action or a state), deliberate action, or other circumstances which make the situation dynamic rather than static. These circumstances may be verbspecific: in the case of see and hear , for example, we have omitted instances which indicate experiencing or witnessing and included only those with a ‘pure’ perceptual meaning (‘experience/detection of the sensation’ vs. ‘production of the sensation’ in Smith’s terminology; see Smith 2005:94). Note further that all verb phrases including auxiliary verbs, whether have or modal auxiliaries, are excluded as these constructions create semantic circumstances different from the simple finite uses of stative verbs. 5.1 Use of the PF with Stative Verbs in Vernacular British Isles Varieties Use of the PF with inherently stative verbs has long been known to be characteristic of the Celtic-influenced varieties spoken in Ireland, Wales, and Scotland. The following examples, drawn from our databases and from some previous studies, illustrate some of the typical usages: (21)I think two of the lads was lost at sea during the War. They were belonging to the, them men here. (IrE; cited in Filppula 2003:162)
< previous page
page_248
next page >
< previous page
page_249
next page >
Page 249 (22)[An’ what did they look like?] They were looking like er, er, gates, you know … (WE; cited in Paulasto 2006:225) (23)No, people don’t need the weather like what they did then - they were depending on the weather. (HebE; cited in Sabban 1982:276) (24)I wasnae liking it and the lassie I was going wi wasnae liking it. (ScE; cited in Miller 1993:121) On the basis of the literature at least, this feature does not appear to be widespread in the traditional EngE dialects, but a few occurrences were found in our SED data, exemplified in examples (25) and (26): (25)I dinna ken [: know] what a fire our fellows would have a Friday. But they were wanting to blast again a Friday. (SED Nb9; Allendale, Northumberland) (26)Everybody was depending on him. (SED O2; Steeple Ashton, Oxfordshire) In this case, the extent of use of this feature is best measured in terms of frequencies per 10,000 words.5 This is done in Table 10.5. In light of the frequent citations of this feature in the literature, the figures for the Celtic Englishes are surprisingly low and the overall differences between the corpora rather small. Nevertheless, of the British Isles varieties, IrE displays the highest normalised rate of use of the PF with the mentioned verbs (0.32 instances per 10,000 words). SED comes second at 0.10 instances per 10,000 words, while there are no instances at all in the WE corpus used for this study. Want is the most common verb to occur with the PF in the two others. Of course, the small number of verbs searched Table 10.5 Use of the PF with Stative Verbs in EngE (SED), WE and IrE Corpus N N/10,000 Words SED 5 0.10 (515,000) WE 0 0.0 (61,400) IrE 5 0.32 (158,000)
< previous page
page_249
next page >
< previous page
page_250
next page >
Page 250 for this study may hide the linguistic reality to some extent. This seems to be the case with the Celtic Englishes and with WE, in particular. Indeed, a search through the more conservative SAWD data yielded further examples of statives with the PF, for instance, with have and shape , as in examples (27) and (28), and live (in the sense ‘to reside in some place’) occurs quite often with the PF even in contexts where there is no time limitation on the length of residence involved. This last-mentioned usage is typical of IrE as well, and our IrE database also contains occurrences of the PF with such rather infrequent verbs as care and flow (when referring to a permanent state or stance), illustrated in examples (29) and (30). These observations suggest a certain degree of productivity of ‘progressive syntax’ in both WE and IrE vernacular. In traditional EngE dialects, by contrast, its use appears to be lexically more restricted; of the 10 verbs searched here, it was limited to just two verbs in our SED data, viz. want (4) and depend (1). Further searches confirmed that, with the exception of a few instances of have and live (see below), there were no occurrences of progressive uses of the other stative verbs mentioned above or exemplified in examples (27)–(30) (shape, care, flow). (27)[What did the shooting star mean then?] Well she used to say that the star was visiting, like, you know, > from here over to the other side that they were having the lights then you see? (WE, SAWD: Gn 8:1) (28)(On a sharpening stone, a strick) It was shaping like this, like a carrot, but it was square. (WE, SAWD: Gn 9:3) (29)Well, of course, Semperit is a, an Austrian = firm […] They are not caring about the Irish people, they are only looking after their own interest, […] (IrE, Dublin: M.L.) (30)[They] call it the Golf Stream […] And that ’s flowing into the Atlantic. = It is flowing into the Atlantic Ocean. (IrE, Kerry: M.C.) If we are to assume some degree of Celtic substratum influence on WE and IrE in this respect, it would be based on the fact that both Welsh and Irish use a construction involving the verb ‘be’ and a so-called verbal noun in these kinds of context. More specifically, most stative verbs in Welsh require the use of the imperfective construction ‘be’ + imperfective marker yn + verbal noun, which resembles the English PF (Heinecke 1999; Thorne 1993). Isaac (2003:61–2) gives examples of what he calls Welsh ‘progressive syntax’ with epistemic verbs such as know, believe, think, feel, and understand, which all use the periphrastic form in Welsh: Wi’n gwybod ‘I know’ (lit. ‘I am knowing’), Wi’n credu ‘I believe’, Wi’n meddwl ‘I think’,
< previous page
page_250
next page >
< previous page
page_251
next page >
Page 251 Wi’n teimlo ‘I feel’, and Wi’n deall ‘I understand’. Irish, on its part, has a similar construction involving the verbal noun, but there the usage varies depending on the lexical verb in question. For instance, it would not be used “uniformly” in Irish with the epistemic verbs mentioned earlier, as Isaac (2003:62) points out (see also Filppula 2003:163). Of the Celtic languages, ‘progressive syntax’ is in fact most in evidence in Scottish Gaelic, where the originally progressive form has come to be used as the present for “nearly all verbs” (Comrie 1976:39). This is the most likely explanation for the salient use of the PF in the Hebridean English contact variety, as documented by Annette Sabban (1982). As regards WE, however, Paulasto (2006) finds that it is not so much stative verbs that are affected by the Welsh substratum, but rather expressions denoting habitual activities, which are frequently construed with the PF; witness examples (31) and (32) (for similar use of the PF, see also Penhallurick 1996). (31)… they used to have a special sieve for this, one man with a shovel lifting up to the sieve and the other one holding it, and the wind was taking all the rubbish away from it. (WE, cited in Paulasto 2006:219) (32)An’ then we learnt later what was happening. Because he was speaking (i.e. ‘he spoke’) Welsh, he was made a fuss of. (WE, Paulasto 2006:221) Although the uses illustrated in examples (31) and (32) are a robust feature of WE, use of the PF with stative verbs, Paulasto argues, belongs to the receding features of WE dialects, appearing mainly in the speech of the elderly informants in the conservative SAWD corpus. Certain stative verbs are nevertheless common in the PF: to these belongs the stance verb live, which appears in WE in temporally unrestricted contexts (e.g. We are living in a more Welsh area in North Wales; Saintess Tybie was living somewhere around the sixth century) (Paulasto 2006:224). A search through instances of live(s) versus am/are/is living plus a locative adverbial in such contexts indicates that the PF is approximately twice as frequent in the WE corpus as in SED; some 20 percent of all instances in the former as opposed to 10 percent in the latter. The figures are necessarily imprecise because the length of residence is not always clear from the context. To sum up our findings so far: the quantitative evidence we have been able to amass here admittedly remains very scanty because of the highly selective searches, but, on the back of other, additional observations on the kinds of verbs that are used with the PF in our databases it seems fair to say that, compared with traditional EngE dialects, the use of the PF with stative verbs is at least more varied in IrE and in WE, too. On the basis of Sabban’s discussion of HebE and Miller’s description of ScE, similar uses are characteristic of those varieties as well. Because we did not find a similar range of
< previous page
page_251
next page >
< previous page
page_252
next page >
Page 252 uses in the SED corpus (although it is much larger), it can be concluded that this particular feature is not characteristic of the traditional EngE dialects and its presence in the Celtic English has to have some other explanation than conservativeness. The most plausible alternatives are either explanations in terms of substratal influences or universals of SLA, or both working together. We defer our judgment on these until after the discussion of the data from the ICE varieties and from other nonstandard varieties in the next sections. 5.2 Use of the PF with Stative Verbs in the ICE Varieties Again, we begin with a quantitative survey of the same verbs as earlier in our ICE varieties (see Table 10.6). Table 10.6 reveals, first, that in Asian and African Englishes, the frequencies of use of statives with the PF are somewhat higher on average than in the British Isles vernaculars. This is particularly so in ICE-EAfr, where this feature occurs 0.57 times per 10,000 words, with ICE-India being the runner-up at 0.43 instances per 10,000 words—both showing higher rates of use than even IrE. The remaining three corpora, including ICE-GB, display somewhat lower levels of use. Yet the documentation of this feature in ICE-GB, and its slightly higher normalised frequencies than in SED, is interesting because it gives some support to studies which have reported increased frequencies of use of the PF with statives in present-day English, including its standard forms (see e.g. Mair 2006; Mair and Hundt 1995; Smith 2005). What adds to the interest of this is Beal’s (1997:373) and Miller’s (2004:55) observations on the same phenomenon even in educated written Scottish English today. Table 10.6 Use of the PF with Stative Verbs in the ICE Varieties Corpus N N/10,000 Words ICE-GB 10 0.19 (528,500) ICE-Ind 24 0.43 (553,400) ICE-Sing 6 0.11 (496,200) ICE-Phil 9 0.18 (555,700) ICE-EAfr 15 0.57 (261,600)
< previous page
page_252
next page >
< previous page
page_253
next page >
Page 253 On a more specific level, IndE exhibits the most varied use of the PF with stative verbs: there is at least one instance with all verbs but believe, and the highest frequencies are found with the verbs know (0.11/10,000), depend (0.11), see (0.05), and belong (0.05). Were the nonfinite instances (i.e. those occurring after modals) included, the frequencies would be much higher in especially this corpus. Kachru (1983:78) considers the IndE usage to be contact-induced and to represent an extension to IndE of the Hindi-Urdu pattern with stative verbs (e.g. mãi dek rahā hũ ‘I am seeing’). This is plausible in light of the fact that Hindi is an aspect-oriented language where the perfective and imperfective aspects are highly grammaticalised; furthermore, the progressive/continuous aspect is widely used and clearly distinct from the habitual and perfective ones (Bhat 1999:127, 143). In EAfrE, this feature is even more common but less varied than in IndE, the only occurrences in the corpus being with see (0.27), intend (0.19), think (0.08), and hear (0.04). Schmied (2004:930) confirms that the nonprogressive use of the PF exists in EAfrE, e.g. with stative verbs. He gives the following example: (33)It is really very toxic to the user because it produces a lot of smoke heavy smoke and it is smelling . (ICE-East Africa) Our findings also indicate that verbs of perception can be accompanied by the PF in EAfrE. In SingE, the most common verbs to occur with the PF are think (= believe) and intend . Although Platt, Weber, and Ho (1984:72–3) mention SingE among the varieties where this feature is attested, most other sources disagree. For example, in the WAMVE, ‘wider range of use of the progressive’ is not considered to be a typical feature of SingE, whereas Zhiming (2005:250) writes that SingE follows mainstream English in not employing the PF with stative verbs. He finds no evidence in ICE-Sing for stative verbs such as love, believe, and know being used with the V-ing form, but as we have seen previously, this is possible with some other verbs. Wee (2004:1059), in turn, points out that aspect is marked lexically in SingE, the progressive being indicated by still , which may combine with the PF (see also Ansaldo 2004; Fong 2004). Finally, ICE-Phil contains more than one instance of a stative with the PF only with the verb think (‘believe that …’). Although this feature is not particularly salient in PhilE in light of our data, Thompson (2003:52–3) has found evidence of ‘overuse’ of the perfect and the progressive in PhilE, the latter especially in contexts denoting habitual action (as in He is going to school regularly). The fact that this feature occurs in PhilE and the other Asian and African varieties in far from negligible numbers clearly calls for an explanation, which cannot simply be in terms of just one explanatory factor, but may involve global or universal considerations. They are the topic of the next section.
< previous page
page_253
next page >
< previous page
page_254
next page >
Page 254 5.3 Global and Cross-Linguistic Perspectives on the Use of the PF with Stative Verbs Let us first examine the evidence from nonstandard varieties of English worldwide. As before, the WAMVE serves a useful purpose here. In that survey, use of the PF with stative verbs is labelled as ‘a wider range of uses of the Progressive’. The examples in the feature list of the survey (Kortmann and Szmrecsanyi 2004:1146) show that the two terms are close enough in meaning for meaningful comparisons. According to the WAMVE, a wider range of uses of the PF is one of the top features in American, Asian, and African Englishes, found in 27 of the 46 varieties (Kortmann and Szmrecsanyi 2004:1179; see also Mesthrie 2004; Platt, Weber and Ho 1984; Schneider 2004:1105). It is also attested in New Zealand English and Australian English (Burridge 2004:1117), which makes it a feature of both L1 and L2 varieties of English (Burridge 2004:1186). However, in the British Isles, it is prominent only in Orkney and Shetland English and IrE and somewhat less common in ScE and Northern English. As far as the L2 varieties are concerned, the WAMVE data are confirmed e.g. by Mesthrie (2004:1134), who states that the extension of the PF to stative contexts is “a striking and almost universal characteristic among L2 varieties in Africa-Asia”. One possible explanation for this lies in language contact. Thus, Gachelin (1997:33) points out the primacy of aspect over tense in many of the world’s languages, whether in the Celtic lands, Africa, or Asia, and states that this difference may have an impact on how tense-aspect-mood constructions are interpreted in contact-influenced varieties of English. In his words, [a]spect is more important than tense in many substrata that may influence varieties of English, ranging from the Celtic fringe of Western Europe […] to African and Eastern languages like Hausa, Swahili, IndoAryan, Chinese and Malay playing their roles in WAE [West-African English], EAE [East African English], IE [Indian English] and SIE [Singapore English]. (Gachelin 1997:33) We have above tried to show that, despite the modest rates of incidence in the IrE and especially WE corpora, it is reasonable to assume that the Celtic substratum has affected the usage in the Irish and Welsh varieties, and that on the basis of the facts about IndE and its sociohistorical background, the impact of Hindi and Urdu on IndE is more than likely. However, substratum explanations hardly suffice to account for the widespread nature of this particular feature in African and Asian varieties or, indeed, provide the sole explanation even in those settings where an identifiable substrate language or languages exist. We must therefore look for other explanations.
< previous page
page_254
next page >
< previous page
page_255
next page >
Page 255 Gachelin (1997) suggests that the high incidence of the PF in the Asian and African varieties, as well as in some English-based creoles, has semantically changed it into a ‘general imperfective’, less and less sensitive to contextual constraints. According to him, this development will eventually constitute one of the major characteristics of future ‘World English’ (Gachelin 1997:43–4). In a similar vein, Mesthrie (2004:1134) notes that the use of the PF with stative verbs is so widespread in the L2 varieties of English in Africa and Asia that it can be considered an almost universal feature. Platt, Weber, and Ho (1984:73) suggest a combination of factors arising from the nature of second-language teaching and learning environments. According to them, it is possible that the wider use of the PF is partly due to the ‘overteaching’ of this construction by teachers of English: it may be that the PF has been stressed by teachers as an essential part of the English verbal system to the extent that learners do not anymore differentiate the verbs with which they use it. This can then lead to an overgeneralisation of this feature to even stative contexts. As another explanation, the same authors suggest that the use of the PF with stative verbs may have spread through analogy from such expressions as I’m having a meal , She’s having a good time, where the verb is stative but the situation is dynamic (Platt, Weber, and Ho 1984). Yet another explanation, largely similar to the kind of overextension discussed earlier, is regularisation, which, as will be remembered, is one of the items in Chambers’ list of putative vernacular universals and is also discussed in several other chapters in this volume. According to Andersen (1984:79, cited in Williams 1987:170), regularisation is an important factor in second-language learning situations and “guides the learner in constructing an internally consistent interlanguage system”. In the case of the PF, regularisation increases the scope of the construction, thus reducing the complexity of the grammatical system of English. In SingE, for example, substratum influence cannot explain nonstandard uses of the PF, but it is plausible to assume that processes of overextension and regularisation have played a role in shaping up this variety. But even in those settings where substratum influence is a factor to be reckoned with, there is no principled reason to consider these two possible sources of influence—substratum and regularisation—mutually exclusive (cf. Thomason, Chapter 15, this volume). There remains the question of what lies behind the expanding use of the PF in standard varieties of English, including even written standard BrE. This is not a recent innovation: several studies have documented a general increase in the rates of use of the PF in many ‘mainstream’ varieties of English from as early as the late Middle English period onwards. At the same time, some of the semantic constraints on its use have been relaxed, including those relating to stative verbs (see e.g. Elsness 1994; Leech and Smith 2006; Mair 2006; Mair and Hundt 1995; Smith 2005). In AmE, this tendency appears to be even more pronounced than in BrE (see e.g. Śmiecińska 2002–2003). However, it has turned out to be hard to find any general explanation for
< previous page
page_255
next page >
< previous page
page_256
next page >
Page 256 this phenomenon. Mair and Hundt (1995) explain it as one manifestation of the general ‘colloquialisation’ of StE (i.e. gradual acceptance of spoken English usages in the written standard as well). As regards the extended use of the PF with stative verbs, in particular, Leech and Smith (2006) conclude that their investigation of the LOB and FLOB data, covering a time span of some 40 years, does not reveal any general explanation for this phenomenon. In perhaps the most thorough study of the increasing use of the progressive in BrE so far, Smith (2005) discusses four possible factors, which are: (1) grammaticalisation (i.e. gradual generalising of the progressive from its original, concrete ‘in progress’ meaning to more abstract meanings); (2) colloquialisation; (3) influence of AmE on BrE and other varieties; and (4) influence of northern EngE and what he terms ‘Celtic dialects’. Interestingly, Smith finds some quantitative support for the last mentioned in data drawn from the BNC; as a further factor, he mentions the possible linguistic effect of the continuous large-scale migrations from the Celtic lands to major cities in England and to the United States, which may, according to him, have acted as a springboard for the spread of this feature into BrE and AmE, too. However, he points out that dialect contacts alone can hardly explain the dramatic rise in the frequencies of the progressive in the latter half of the twentieth century, which have not witnessed migrations from Ireland and the other Celtic areas on the same scale as the nineteenth and early twentieth centuries (Smith 2005:234–8). The foregoing discussion has shown that a variety of factors seem to be involved in the expanding use of the PF with stative verbs. Although there are grounds for arguing that in the Celtic Englishes substratal influences from the relevant Celtic languages are a contributory factor, this phenomenon is by no means unique to these dialects, and even in them may reflect more general tendencies of language contact. Among the most important of these are the primacy of aspect over tense in many of the world’s languages and the kinds of overextension, regularisation, and simplification processes described earlier. What also suggests a clear role for these general tendencies is the widespread nature of this feature in Asian and African varieties, as well as in other ‘colonial’ varieties like AmE, AusE, and NZE, all of which exhibit varying levels of language contact. For some of these varieties, as we have seen, we can identify plausible substratal parallels, which make them comparable to the Celtic Englishes. For others, the general and universal aspects have to be given priority. The rapidly increasing use of the PF with stative verbs in standard BrE may also form one chapter in these general developments, possibly sparked off by the large-scale migrations from the Celtic lands in the last couple of centuries. Finally, one cannot overlook language-internal factors either, which is here most clearly evidenced by the kind of grammaticalisation path outlined in Smith (2005). As compared with the two features discussed earlier, the use of the PF with stative verbs presents a similar picture to the second one, viz. wider use of the definite article, in that as far as the vernacular varieties are concerned,
< previous page
page_256
next page >
< previous page
page_257
next page >
Page 257 the main division is between the Celtic varieties (especially IrE) and the traditional EngE dialects, in which this feature appears to be very rare. Another similarity is the prominent role played by the L2 background and universal aspects of language contacts. The greatest difference between this feature and the two others concerns the behaviour of standard BrE, which in this case appears to partake to a greater extent in the general trend shared by colonial Englishes. What awaits further investigation is the cross-linguistic aspect and especially the extent of the developments described here amongst the world’s languages. It would seem that, at least in comparison with our first feature—absence of plural marking with measurement nouns—the wider use of the PF has a more restricted distribution because of constraints imposed by the specifics of the tense and aspect systems of each language, one of the most obvious ones being the lack of such a construction in many languages. That said, one can find examples of languages or groups of languages in typological literature where original progressives have extended their meanings to either present or imperfective meanings. For example, Dahl (1985:93) mentions Punjabi and Hindi-Urdu as cases where progressives have turned into general imperfectives. Comrie’s (1976) discussion of Scottish Gaelic was already mentioned; Comrie (1976:101) also mentions Yoruba as another such case. Bybee, Perkins, and Pagliuca (1994) also note the same kind of development in several languages and add that the extension of the ‘progressive grams’ either to a present or an imperfective involves as a major step an extension to habitual meaning (Bybee, Perkins, and Pagliuca: 141). As we have seen, this is exactly what has happened in the Celtic Englishes. 6. CONCLUSIONS In this chapter, we have sought to contribute to the ongoing debate on vernacular universals and language/dialect contacts by presenting the results of our quantitative and qualitative comparisons between vernacular varieties, on one hand, and national standard varieties, on the other. In addition to the vernacular–standard continuum, the varieties were chosen so as to enable study of the effect of the sociohistorical background of each variety. Comparisons were also made with data available from other surveys with at least some form of quantitative input; particularly important in this respect was the WAMVE, which opened a window to a large number of different nonstandard varieties of one language, viz. English. Although many variationist studies are primarily concerned with phonological features, we have examined here three morphosyntactic features, all of which had as their common denominator attestation in one or more of the so-called Celtic Englishes, and hence a possible language contact background. Unsurprisingly, our findings give no direct or straightforward answers to the questions of the existence and locus of vernacular universals and their
< previous page
page_257
next page >
< previous page
page_258
next page >
Page 258 relationship with language contact phenomena. Yet it can be stated with some degree of confidence that of the three features investigated here absence of plural marking with measure nouns comes closest to being a vernacular universal in a demanding sense: it is widely attested in vernacular varieties of English in the British Isles and elsewhere, including the most conservative EngE dialects; it also has a wide distribution in genetically unrelated languages all round the world; and it is functionally well motivated. Use of the definite article with names of institutions and quantifiers and the use of the PF with stative verbs present a more variegated picture. Our results show the L2 factor to be relevant to both, and in the case of the definite article usage, in particular, this factor is also in evidence in the colonial standard varieties. In the vernacular EngE dialects, by contrast, these two features have a low profile, which further speaks for the importance of general tendencies of SLA and language contacts in the other settings, except of course Standard BrE. The fact that stative verbs are increasingly being used with the PF even in Standard BrE can be interpreted as being due to influence from other varieties, as well as a continuation of long-term language-internal processes of grammaticalisation and colloquialisation. NOTES 1. This research was conducted under the auspices of the research project on “Vernacular Universals vs. Contact-Induced Language Change” (2005–2008), funded by the Humanities and Social Sciences Section of the Academy of Finland (Grant No. 210702). 2. In Middle Welsh, however, a plural noun was also possible, although the noun was generally in the singular (Evans 1964:47). 3. Marianne Hundt (2007) suggests that the ICE-India corpus may contain more colloquial spoken language than the other ICE corpora. If true, this could account for at least part of the observed differences between IndE, PhilE, SingE, and EAfrE. 4. The World Atlas refers to the survey conducted in connection with the compilation of The Handbook of Varieties of English (Kortmann et al. 2004) and published as a CD-ROM. 5. This is because of the well-known difficulties involved in determining whether a given instance of a progressive form of a verb could have been expressed by the simple form of the verb, whilst retaining exactly the same meaning. REFERENCES Algeo, J. 1988. British and American grammatical differences. International Journal of Lexicography 1: 1–31. Andersen, R. 1984. The one-to-one principle of interlanguage construction. Language Learning 34: 77–95. Ansaldo, U. 2004. The evolution of Singapore English: Finding the matrix. In L. Lim, ed. 2004. 127–49. Beal, J. 1997. Syntax and morphology. In The Edinburgh History of the Scots Language, edited by C. Jones, 335–77. Edinburgh: Edinburgh University Press.
< previous page
page_258
next page >
< previous page
page_259
next page >
Page 259 Bhat, D. N. S. 1999. The Prominence of Tense, Aspect and Mood . Amsterdam and Philadelphia: John Benjamins. Bhatt, R. M. 2004. Indian English: Syntax. In Kortmann et al., eds. 2004. 1016–30. Biber, D., S. Johansson, G. Leech, S. Conrad, and E. Finegan. 1999. Longman Grammar of Spoken and Written English. London: Longman. Burridge, K. 2004. Synopsis: Morphological and syntactic variation in the Pacific and Australasia. In Kortmann et al., eds. 2004. 1116–31. Bybee, J., R. Perkins, and W. Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago and London: The University of Chicago Press. Chambers, J.K. 2004. Dynamic typology and vernacular universals. In Dialectology meets Typology, edited by B. Kortmann, 127–45. Berlin and New York: Mouton de Gruyter. Comrie, B. 1976. Aspect . Cambridge: Cambridge University Press. Dahl, Ö. 1985. Tense and Aspect Systems. Oxford: Blackwell. Elsness, J. 1994. On the progression of the progressive in Early Modern English. ICAME Journal 18: 5–25. Edwards, V., and B. Weltens. 1985. Research on non-standard dialects of British English: Progress and prospects (1). In Focus on: England and Wales (Varieties of English Around the World 4), edited by W. Viereck, 97–139. Amsterdam and Philadelphia: John Benjamins. Evans, D.S. 1964. A Grammar of Middle Welsh . Dublin: Dublin Institute for Advanced Studies. Filppula, M. 1999. The Grammar of Irish English: Language in Hibernian Style. London: Routledge. ———. 2003. More on the English progressive and the Celtic connection. In H. L. C. Tristram, ed. 2003. 150– 68. Fong, V. 2004. The verbal cluster. In L. Lim, ed. 2004. 75–104. Gachelin, J.-M. 1997. The progressive and habitual aspects in non-standard Englishes. In Englishes Around the World, Vol. 1: General Studies, British Isles, North America, edited by E.W. Schneider, 33–46. Amsterdam and Philadelphia: John Benjamins. Gonzales, A. 1997. Philippine English: A variety in search of legitimation. In Schneider ed. 1997. 205–12. Heinecke, J. 1999. Temporal Deixis in Welsh and Breton . Heidelberg: Winter. Huddleston, R., and G. Pullum, eds. 2002. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Hudson-Ettle, D.M., and J. Schmied. 1999. Manual to accompany the East African component of The International Corpus of English, ICE-EA: Background information, coding conventions and lists of source texts. Department of English, Chemnitz University of Technology. Hundt, M. 2007. Global feature – local norms? A case study on the progressive passive. A paper given at the 13th Annual conference of IAWE, Regensburg, Germany, 4–6 October, 2007. Isaac, G. 2003. Diagnosing the symptoms of contact: some Celtic-English case histories.” In Tristram, ed. 2003. 46–64. Jespersen, O. 1924. The Philosophy of Grammar . London: Allen & Unwin. Joyce, P.W. 1910/1988. English as We Speak It in Ireland. 3rd edition. Dublin: Wolfhound Press. Kachru, B. 1983. The Indianization of English: The English Language in India. Delhi: Oxford University Press. ———. 1996. English as lingua franca. In Contact Linguistics, edited by H. Goebl, P. Nelde, and Z. Stary, 906–13. Berlin and New York: Mouton de Gruyter. Klemola, J. Forthcoming. The Corpus of SED Tape-recordings.
< previous page
page_259
next page >
< previous page
page_260
next page >
Page 260 Kortmann, B., E. Schneider, K. Burridge, R. Mesthrie, and C. Upton. 2004. A Handbook of Varieties of English. Volume 2: Morphology and Syntax . Berlin and New York: Mouton de Gruyter. Kortmann, B., and B. Szmrecsanyi. 2004. Global synopsis: Morphological and syntactic variation in English. In Kortmann et al., eds. 2004. 1142–202. Leech, G., and N. Smith. 2006. Recent grammatical change in written English 1961–1992: Some preliminary findings of a comparison of American with British English. In The Changing Face of Corpus Linguistics, edited by A. Renouf and A. Kehoe, 185–204. Amsterdam and New York: Rodopi. Lim, L., ed. 2004. Singapore English: A Grammatical Description . Amsterdam and Philadelphia: John Benjamins. ———, and J.A. Foley. 2004. English in Singapore and Singapore English: Background and methodology. In Lim, ed. 2004. 1–18. Macaulay, R.K.S. 1991. Locating Dialect in Discourse: The Language of Honest Men and Bonnie Lassies in Ayr . New York and Oxford: Oxford University Press. Mair, C. 2006. Twentieth-century English. Cambridge: Cambridge University Press. ———, and M. Hundt. 1995. Why is the progressive becoming more frequent in English? Zeitschrift für Anglistik und Amerikanistik XLIII: 111–22. Mesthrie, R. 2004. Synopsis: Morphological and syntactic variation in Africa and South and Southeast Asia. In Kortmann et al., eds. 2004. 1132–41. Miller, J. 1993. The grammar of Scottish English. In Real English: The Grammar of English Dialects in the British Isles , edited by J. Milroy and L. Milroy, 99–138. London: Longman. ———. 2004. Scottish English: Morphology and syntax. In Kortmann et al., eds. 2004. 47–72. Ó Siadhail, M. 1989. Modern Irish: Grammatical Structure and Dialectal Variation. Cambridge: Cambridge University Press. Parry, D. 1999. A Grammar and Glossary of Conservative Anglo-Welsh Dialects of Rural Wales. NATCECT. Occasional Publications, No. 8. Sheffield: University of Sheffield. Paulasto, H. 2006. Welsh English Syntax: Contact and Variation (Publications in the Humanities 43). Joensuu: Joensuu University Press. Penhallurick, R. 1996. The grammar of northern Welsh English: Progressive verb phrases. In Speech Past and Present. Studies in English Dialectology in Memory of Ossi Ihalainen (University of Bamberg Studies in English Linguistics, Vol. 38), edited by J. Klemola, M. Kytö, and M. Rissanen, 308–42. Frankfurt am Main: Peter Lang. Platt, J., H. Weber, and M.L. Ho. 1983. Singapore and Malaysia. Amsterdam and Philadelphia: John Benjamins. ———. 1984. The New Englishes . London, Boston, Melbourne, and Henley: Routledge and Kegan Paul. Quirk, R., S. Greenbaum, G. Leech, and J. Svartvik. 1985. A Comprehensive Grammar of the English Language. London: Longman. Rohdenburg, G. 2006. Variable plural marking with measure nouns in non-standard English and low German dialects. NOWELE 48: 111–30. Sabban, A. 1982. Gälisch-englischer Sprachkontakt . Heidelberg: Julius Groos. Sand, A. 2003. The definite article in Irish English and other contact varieties of English. In Tristram, ed. 2003. 413–30. Schmied, J. 1997. Beyond recipes, beyond Maks, beyond Africa: Texts, text-types, text collections and African realities. In Schneider, ed. 1997. 141–58. ———. 2004. East African English (Kenya, Uganda, Tanzania): Morphology and syntax. In Kortmann et al., eds. 2004. 929–47.
< previous page
page_260
next page >
< previous page
page_261
next page >
Page 261 Schneider, E.W., ed. 1997. Englishes Around the World, Vol. 2: Caribbean, Africa, Asia, Australasia. Studies in Honour of Manfred Görlach. Amsterdam and Philadelphia: John Benjamins. Schneider, E. W. 2004. Synopsis: Morphological and syntactic variation in the Americas and the Caribbean. In Kortmann et al., eds. 2004. 1104–15. Sedlatschek, A. 2007. Variation and Change in Contemporary Indian English: A Corpus-Based Description . Unpublished PhD thesis. Freiburg i.Br. Seppänen, R., and A. Seppänen. 1984. ‘Two dozen’, ‘several hundred’: An English construction and its nonEnglish parallels. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 37: 46–58. Sibayan, B. P. 1998. Reflections on some dilemmas posed by literacy in English, Filipino, and the Philippine vernaculars. In Pagtanaw: Essays on Language in Honor of Teodoro A. Llamzon, edited by M.L. Bautista, 149–57. Manila: Linguistic Society of the Philippines. Slobin, D. 1977. Language change in childhood and history. In Language Thought and Language Learning , edited by J. Macnamara, 185–214. New York: Academic Press. Śmiecińska, J. 2002–2003. Stative verbs and the progressive aspect in English. Poznań Studies in Contemporary Linguistics 38: 187–95. Smith, N. 2005. A Corpus-based Investigation of Recent Change in the Use of the Progressive in British English. Unpublished PhD thesis, University of Lancaster. Stebbins, T. 1997. Asymmetrical nominal number marking: A functional account. Sprachtypologie und Universalienforschung 50: 5–47. Thompson, R. M. 2003. Filipino English and Taglish: Language Switching from Multiple Perspectives . Amsterdam and Philadelphia: John Benjamins. Thorne, D. A. 1993. A Comprehensive Welsh Grammar . Oxford: Blackwell. Tristram, H.L.C., ed. 2003. Celtic Englishes III . Heidelberg: Winter. Wächtler, K. 1988. Beyond the rules of grammar: The case of (the) articles in English. In Essays on the English Language and Applied Linguistics on the Occasion of Gerhard Nickel’s 60th Birthday, edited by J. Klegraf and D. Nehls, 147–55. Heidelberg: Julius Groos. Wee, L. 2004. Singapore English: Morphology and syntax. In Kortmann et al., eds. 2004. 1058–72. ———, and U. Ansaldo. 2004. Nouns and noun phrases. In Lim, ed. 2004. 57–74. Whaley, L. J. 1997. An Introduction to Typology: The Unity and Diversity of Language. Thousand Oaks, London, New Delhi: Sage Publications. Williams, J. 1987. Non-native varieties of English: A special case of language acquisition. English World-Wide 8(2): 161–99. Wode, H. 1984. Some theoretical implications of L2 acquisition research and the grammar of interlanguage. In Interlanguage, edited by A. Davies, C. Criper, and A. Howatt, 162–84. Edinburgh: Edinburgh University Press. Wright, J. 1905. The English Dialect Grammar . Oxford: Henry Frowde. Zhiming, B. 2005. The aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics 41: 237–67.
< previous page
page_261
next page >
< previous page
page_262
next page >
page_262
next page >
Page 262 This page intentionally left blank.
< previous page
< previous page
page_263
next page >
Page 263 IV Methodological and Theoretical Perspectives
< previous page
page_263
next page >
< previous page
page_264
next page >
page_264
next page >
Page 264 This page intentionally left blank.
< previous page
< previous page
page_265
next page >
Page 265 11 Methods and Inferences in the Study of Substrate Influence Terence Odlin 1. INTRODUCTION The topic of substrate influence has often been viewed as a controversial area, and some specific controversies seem likely to continue for a long time to come. Even so, any careful look at the studies in the last three decades or so will affirm both massive evidence of such influence and an increasing sophistication in methodology. Indeed, the enhanced understanding of what it takes to document substrate influence has helped greatly in providing solid evidence and thus in reducing the range of controversies. This chapter focuses on the methodological advances, but a useful byproduct of looking at the advances will be some specific inferences about second language acquisition (SLA) and language contact that are warranted by the methods. The chapter will compare three important methods (as well as some related ones) and then explore implications of some of the findings obtained through such methods. However, two preliminaries are necessary: first to look at terminology and then to reflect a little on cross-linguistic comparisons. The term substrate influence will be used more than two terms widely found in research on SLA: language transfer and cross-linguistic influence . Although these latter terms are usually employed with care in SLA investigations, their uses in that research are understood to mean a particular direction of the influence, most typically, the influence of a native language on a target language. Influence is also possible in the other direction, of course, and has been the object of much study (e.g. Jarvis and Pavlenko 2008; Thomason and Kaufman 1988). Because the term substrate influence is less likely to seem ambiguous, it will be used in this chapter. For readers seeking a more detailed discussion of terminology, Odlin (1989, 2003) compares several related terms, including interference, code-switching , and positive and negative transfer. It might go without saying that cross-linguistic comparisons are a necessary condition for the study of substrate influence. However, such comparisons are not always a sufficient condition, as was seen about three decades ago in SLA studies of the acquisition of negation (surveyed by Odlin 1989:104–10). For anyone familiar with the similarities and differences between
< previous page
page_265
next page >
< previous page
page_266
next page >
Page 266 Spanish and English, the influence of the former on the latter might seem self-evident when a Spaniard using English says I no understand, both because of the morphological identity of Spanish and English no and because of its preverbal position in the translation equivalent in Spanish Yo no entiendo (albeit that the pronoun yo is not obligatory in the way that I is in the English sentence). However convincing a simple appeal to substrate influence may seem in this case, the acquisition of negation turns out to be a more complex issue because speakers of German and Japanese have likewise been observed to use the preverbal pattern although there is no closely corresponding syntactic pattern in the substrate languages. In German, for instance, negators normally follow verbs in independent clauses. As the discussion of methods will show, even such problematic cases can be sorted out. Nevertheless, it is also true that other factors besides substrate influence can and do affect the acquisition of a second language, and so the need for good methods to distinguish substrate influence from other factors is all the greater. 2. THREE METHODS A variety of methods can be used to establish the likelihood of substrate influence. However, three allow for especially plausible inferences. One method compares how a structure is used in the native language, the target language, and the interlanguage, the latter a term coined by Selinker (1972) and one owing much to this method (Selinker 1969). Although Selinker’s approach involves just one native language group, the second approach compares the interlanguage performances of learners with two or more distinct native languages, as will be illustrated in studies by Mesthrie and Dunne (1990) and Ringbom (1987, 2007). The third method, propounded by Jarvis (2000), combines the first two methods. The value of the method used by Selinker is seen in his study of interlanguage (IL) word order of native speakers of Hebrew learning English. Interviews with school-age learners showed frequent examples of adverb phrases placed immediately after verbs as in I like very much movies instead of I like movies very much . From interviews with native speakers of English and native speakers of Hebrew, Selinker was able to demonstrate a statistically significant relation between the IL and Hebrew data, with the IL patterning more like the native than like the target language. His results seem all the more plausible because he took measures to provide comparable discourse contexts for the three groups. The second method involves comparisons of how learners having different native languages perform on the same structure. Any careful use of this method entails having at least two groups of learners who are similar except in terms of their native language. In the study of Mesthrie and Dunne (1990), such comparability arose, sadly, from the segregation enforced during the apartheid era in South Africa. The people interviewed spoke South African
< previous page
page_266
next page >
< previous page
page_267
next page >
Page 267 Indian English, with either an Indic or Dravidian language being the mother tongue of these individuals. Relativization patterns were the focus of the investigation, as this domain shows striking contrasts in the languages of the two families. Where Indic languages employ correlative constructions, Dravidian languages employ structures where the head noun is premodified by the relative clause. The contrast is evident in the following illustrations from Gujarati (Indic) and Tamil (Dravidian) given by Mesthrie and Dunne (1990): (1) Gujarati Je veparii marii sathe avyo, te CORRELATIVE businessman me with came that veparii Harilal ka bhaaii che businessman Harilal of brother is ‘The businessman who came with me is Harilal’s brother.’ (Literally, ‘Which businessman came with me, that businessman is Harilal’s brother.’) (2) Tamil Taccaan aticca vannaan cenneki carpenter.nom beat.past.rel part washerman.nom Madras.dat poonaan go.past.3sg.masc ‘The washerman whom the carpenter beat went to Madras.’ (Literally, ‘The carpenter-beat(en) washerman went to Madras.’) In the Gujarati example, the correlatives je … te somewhat resemble English forms either … or and neither … nor (which do not, however, play any role in relative clauses). The closest parallel in Standard English to the Tamil case is the sort of reduced clause that can occur before the modified noun as in a-never-to-beforgotten-experience . However, basilect speakers of South African Indian English studied by Mesthrie and Dunne produce two structurally distinct patterns of relativization as seen below (in the Dravidian example, the square brackets are insertions to help interpret the utterance): (3)Indic But now, which-one principal came here, she’s just cheeky like the other one. ‘The principal who arrived recently is just as stern as the previous one.’ (4)Dravidian People [who got working-here-for-them] sons, like, for them nice they can stay. ‘It is nice for people who have sons [who are] working for the company, since they are allowed to stay on in the barracks.’
< previous page
page_267
next page >
< previous page
page_268
next page >
Page 268 The likelihood of substrate influence in such cases seems high because of the clear structural contrast between the Indic and Dravidian substrate languages and between the ILs, along with the structural resemblances between the IL structures and those in the substrate languages. Although the first and second methods obviously differ in terms of the number of native languages involved, both succeed in making a plausible case for an important type of substrate influence: divergences between the IL and the target language. However, the second method shows an additional strength in that it can also help to argue for cases of convergence between the native and target languages. In his work on the acquisition of English in Finland, Ringbom (1987, 2007) emphasized the importance of the facilitating effects that cross-linguistic similarities can promote. In Finland, the teaching of English does not vary a great deal for speakers of Swedish or speakers of Finnish (and Ringbom considers the cultural and socioeconomic differences to be minimal), and yet the former group generally finds learning English much easier. Ringbom (1987) documents the difference in learning outcomes in a chapter comparing results on a variety of tests of comprehension and production, including grammar as well as vocabulary. Because Swedish and English are Indo-European (and Germanic) languages, and Finnish is non-Indo-European, the relative success of the ‘Swedes’ (to use Ringbom’s term) clearly suggests convergence. Indeed, Ringbom (1987:109) concludes, “[t]he effect of cross-linguistic influence is overwhelmingly facilitative since the difference is clearly in the Swedes’ favour”. Like Ringbom, Jarvis (1998, 2000) compared the performance of Finns and Swedes learning English in Finland, but he added a methodological refinement reminiscent of Selinker’s approach in that he also collected data from native speakers of the substrate languages as well as of the target language. By thus having five distinct data sets instead of three, Jarvis has been able to make finer-grained inferences about substrate influence. Although he employed a variety of tasks to compare the different groups, the one that will be discussed here was a written narrative of part of a film, the Charlie Chaplin classic Modern Times, which co-starred Paulette Goddard. One scene in the film shows a collision in the street between Chaplin and Goddard, and students writing in Swedish often described what they saw with one of two different verbs, krocka (‘crash’) and springa på (‘run on’). In contrast, the students writing in Finnish overwhelmingly chose törmätä (‘crash’) and almost never juosta (‘run’). In the English narratives of the Swedes, several pupils used sentences such as She run on Charlie Chaplin , which suggests that springa på was their model. The Finns never employed this construction, and therefore its frequent appearance in the English of the Swedes strongly argues for substrate influence. Further evidence of such influence comes from a study of spatial constructions (Jarvis and Odlin 2000) which identified comparable differences in certain prepositional choices in the same film descriptions. For instance, in a scene with Chaplin
< previous page
page_268
next page >
< previous page
page_269
next page >
Page 269 and Goddard on a lawn, the IL phrase sat to the grass was used by several Finns, but not by any Swedes. The third method can also provide fine-grained characterizations of convergences. The results of an investigation by Jarvis (2002) show that Swedes have a major advantage over Finns in the accurate use of definite and indefinite articles in English. His results are certainly similar to those of Ringom (1987), who found Swedes to use articles more often than Finns. However, Ringbom’s study could not provide a detailed characterization of the facilitating effects of cross-linguistic similarity because he had no production data in Swedish or Finnish. In contrast, Jarvis was able to show that the Swedes had a clearer sense of the discourse contexts requiring English articles, and he found the grammatical correspondences in the Swedish essays to be close to those of the Swedes writing in English, correspondences not evident in the Finnish essays.1 2.1 Other Approaches The three methods detailed here do not exhaust the ways that one might try to demonstrate substrate influence. Some other approaches have proven valuable and may sometimes be the only viable way to verify such influence. For instance, Filppula (1986, 1999) has considered the regional variation seen in the frequency of cleft and absolute constructions, the logic of his analysis being that in places where the Irish substrate disappeared earlier (Dublin and Wicklow), the effects of substrate influence should be less robust than in regions such as Clare and Kerry, where Irish still survives somewhat as a community language. Where Filppula considered stratification by region, Sabban (1982) considered stratification by age; in her analysis of the effects of Gaelic on Hebridean English, the older speakers tended to produce more structures showing substrate influence. Comparable evidence comes from a detailed investigation by Paulasto (2006) which found an especially frequent use of focus fronting constructions among older bilingual speakers of Welsh English. Other studies of social variation by class and register have likewise made convincing cases for native language effects on a second language (e.g., Schmidt 1987), and other approaches seem viable as well. Indeed, a fairly wide range of useful methodologies has been discussed elsewhere (e.g., Jarvis and Pavlenko 2008; McWhorter 1996). However, all three of the methods focused on in this chapter make inferences about substrate influence that are especially hard to refute. In the negation example discussed previously, the use of preverbal no by speakers of German, Japanese, and Spanish might seem to argue solely for developmental factors in the acquisition of negation. However, a closer look at the evidence shows that speakers of German eventually use postverbal negators, something not found much in the English of Spanish speakers (Odlin 1989; Wode 1983). Furthermore, Japanese speakers do not rely so heavily on the negator no , but instead also opt for
< previous page
page_269
next page >
< previous page
page_270
next page >
Page 270 other negators such as not more frequently than do Spanish speakers (Odlin 1989; Stauble 1984). Such IL differences, which emerge in comparisons akin to the second method, show that developmental and substrate influences can and do coexist in SLA. 2.2 Some Cautions The three methods do not come cost-free, however. In the first method, it is necessary to collect data from three different populations; otherwise, the statistical comparisons that Selinker considers crucial to this method become impossible. The second method likewise requires three data sets, although in practice this requirement is sometimes relaxed with regard to data from native speakers of the target language, especially when structural descriptions are readily available and when the results of such data collection might appear to be self-evident (although this is obviously a risky assumption in some cases). Without a doubt, the third method requires the greatest effort because five data sets are involved. Although Jarvis has convincingly demonstrated the value of this approach, the collection effort makes it unlikely that there will be an abundance of such studies anytime soon. As already noted, the first method presupposes statistical comparisons, and a common expectation with the second and third methods is likewise that a statistically significant difference between groups can be determined. However, some instances of substrate influence seem quite convincing even in the absence of any possible statistical comparisons or of the explicit comparisons involved in the three methods. An uncontroversial case of substrate influence is evident in the so-called after perfect still found in Ireland and some parts of Scotland. Although problems involving the onset of such influence have been considered (e.g. Filppula 1999; McCafferty 2004), there is a wide consensus about the Celtic substrate. Sabban (1982:155) offers a useful example: I’m after forgetting all that lot now, which is synonymous with I have forgotten all that now. The interviewee was an elderly native speaker of Scottish Gaelic, which has a very similar perfect construction with either an déidh or air (‘after’) and with analogous forms used in Ireland, tar éis and i ndiaidh . What makes the after perfect uncontroversial is its absence in dialects of English outside of the regions where Celtic English speakers have been prominent. Evidence of this uniqueness can be found in the recent Handbook of Varieties of English (Kortmann, Schneider, Burridge, Mesthrie, and Upton 2004).2 Cases such as the after perfect indicate a uniqueness due to substrate influence, and other such cases no doubt exist. For instance, Jarvis and Pavlenko (2008, Chapter 3) note the following sentence from a Finnish speaker: Charlie Chaplin get up and say shelle [to her]: “ Remember my—and bread .” The form shelle shows a Finnish allative case inflection used with the pronoun she. Such cases are rare in the data collected by Jarvis, but they do show that in earlier stages of proficiency some learners consider it possible to
< previous page
page_270
next page >
< previous page
page_271
next page >
Page 271 use Finnish case forms in English. This belief seems to disappear fairly early, although other types of influence from the case system persist for much longer (Jarvis and Odlin 2000). The after perfect and the shelle examples show that plausible instances of substrate influence sometimes exist in the absence of statistical comparisons. Even so, the value of the three methods remains great because they can help resolve what might not seem to be clear-cut cases of substrate influence, and they allow for a wider range of inferences. 3. INFERENCES ABOUT ARTICLES All three of the methods discussed in the preceding sections allow for inferences about substrate influence even in cases that might otherwise be more controversial. Each method has its distinct merits (and also limitations), but in any event, most SLA studies that make a good case rely on the second method, comparing the IL productions of two or more groups having different substrate languages (usually their native languages). Although the third method, pioneered by Jarvis, is different, the results obtained with that method allow for many of the same types of inferences warranted by the second method, and so findings from that approach will be included in the following discussion. Of research that employs the second and third methods, the largest collection of studies on the same grammatical domain focuses on articles. As such, this evidence can shed new light on analyses by Siegel (2000, 2005) of substrate influence in the development of Hawaiian Creole English. Before examining Siegel’s position, it is necessary to consider the claims of Bickerton (1981, 1984) regarding HCE because Siegel’s research challenges some key positions of Bickerton. As many language contact researchers know, Bickerton has viewed HCE as an especially clear case of what his hypothesized language bioprogram will lead to when certain conditions of creolization are met, especially an impoverished input from any pidgin languages and a population of younger speakers who will develop a creole largely through the workings of the bioprogram. For Bickerton, certain structural features distinguish HCE from any of the pidgins also found in the archipelago. For example, he has claimed that the system of tense, modality, and aspect is unlike the pidgins, and the same is supposedly true for serial verbs, complementizers, articles, and other structures (Siegel [2000] lists a dozen or so features discussed by Bickerton in Roots of Language [1981] and in another analysis published in Behavioral and Brain Sciences [1984]). It is beyond the scope of this chapter to discuss Bickerton’s entire analysis or the entire range of empirical challenges that Siegel has made, but the discussions of articles and serial verbs can be broached profitably in light of the work in SLA on these areas. Bickerton does not, it should be emphasized, deny that substrate influence can affect the development of pidgins. Indeed, some of his own work
< previous page
page_271
next page >
< previous page
page_272
next page >
Page 272 (e.g., Bickerton and Givón 1976) employs the second method discussed in this chapter to contrast the wordorder patterns of speakers of Ilocano and Japanese in their Hawaiian Pidgin English (HPE). However, he draws a firm line between such influence in the development of pidgins and in creoles, claiming that only the former will show real substrate effects, at least in cases where the social conditions are conducive to creole formation. The role of pidgins in true creolization is minimal for Bickerton, his argument being that the complexity of structures in the creoles exceeds any input available from the pidgins. Thus, in the case of articles, Bickerton (1981:22) sees the lack of success of Japanese speakers of HPE as evidence of the allegedly impoverished input that HPE would provide for the developing creole, and he makes a similar argument for Filipino speakers of HPE. Interestingly, he does not dwell on the contrasting behavior of the Japanese and Filipinos, the former whom he characterizes as underusing articles and the latter as overusing them. The difference between the groups could suggest some kind of substrate influence in the pidgin, but such influence Bickerton might consider to be generally unhelpful in the formation of HCE, especially if neither group shows much accuracy. His position in his 1984 analysis remains essentially unchanged with regard to articles (Bickerton 1981:176). One of the problems that Siegel finds in Bickerton’s analysis is that speakers of Portuguese and Chinese were in Hawaii in sufficient numbers to be considered major players in the formation of the creole, with their presence in large numbers predating that of speakers of Japanese or Filipino languages. Siegel (2000) details this argument with considerable demographic evidence, and he proceeds to attempt to identify cases of likely substrate influence from Chinese and Portuguese. In the regard to articles, he looks only at the possibility of influence from Chinese, mainly on the indefinite article in HCE, but perhaps also, he suggests, on the definite article (2000:115–17; 2005:68–70). Whatever merit Siegel’s analysis of Chinese influence in this regard, his substratist explanation seems incomplete in ignoring, as it does, the likely influence of Portuguese, which has definite and indefinite articles. As described earlier, Jarvis (2002) found Swedes having a major advantage over Finns in the accurate use of definite and indefinite articles in English, and his results corroborate those of Ringbom (1987). Moreover, the convergence that the Swedes were able to achieve is found in other research as well as studies discussed by Odlin (1989:34; 2003:461; 2006:24–5) indicate. If second-language learners regularly take advantage of such correspondences, it naturally follows that speakers of Portuguese would have likewise been able to construct an IL whose article system would resemble the target system. The likelihood of substrate influence in this grammatical domain seems all the greater because there were also Spanish speakers in Hawaii around the same time, some coming from Puerto Rico and some from Spain, and the total population of these three groups outnumbered Chinese speakers by the year 1910 (Siegel 2000:200).
< previous page
page_272
next page >
< previous page
page_273
next page >
Page 273 To my knowledge, every study that has compared speakers of substrate languages that have articles with those whose languages do not use articles shows an advantage for the former group over the latter. It thus seems reasonable to predict that this is the norm in language contact, including controversial cases such as Hawaiian Creole English. Such a prediction does not rule out the possibility—indeed, the likelihood—that children in a multilingual community will develop new patterns of article use. However, the SLA studies show that a key assumption Bickerton made about articles is unwarranted. That is, the IL data resulting from certain convergences between English and the substrates of Portuguese and Spanish would be sufficient input for children in a creole community to get a sense of how article systems work. Perhaps something like a bioprogram would be conducive to some kinds of restructuring over others, but children in Hawaii would have access to data from bilinguals using an article system. The article studies I have cited warrant the inferences in the preceding paragraph. However, the Jarvis (2002) study is especially significant because its methodology helps to counter a longstanding criticism by Bickerton of substratist analyses. Bickerton has insisted that there be a point-by-point correspondence between a particular substrate language and the creole if such explanations are to be taken seriously. The Jarvis study raises serious problems for anyone insisting on a point-by-point correspondence. On the one hand, the Swedes show an overall advantage in using articles, and the correspondences with Swedish are often close. Nevertheless, there are cases where the native language does not have the predicted effect, where, for example, Swedes omit a definite article in their English narratives even while Swedish usage in the same context shows the definite article. Some uses of zero articles appear in Table 5 of his study (Jarvis 2002:402), and such omissions are sometimes errors from the standpoint of the English norm. Jarvis (p.c.) notes that “26 Swedes omitted the definite article where it should have been used”. Such omissions represent departures from the point-by-point correspondence demanded by Bickerton, but such instances are not anywhere frequent enough to blur the overall advantage that the Swedes show in comparison with the Finns in their use of definite articles.3 3.1 Inferences About Serial Verbs As stated earlier, the large number of SLA studies of IL article use allows for especially strong inferences about other language contact situations. Even so, research using the second method allows for inferences in other structural domains as well. In one case, such research corroborates Siegel’s analysis of Chinese influence on Hawaiian Creole English, namely in the area of serial verbs. Helms-Park (2001, 2003) investigated verbs of causation in the IL of speakers of Vietnamese, on the one hand, and speakers of Hindi and Urdu, on the other hand. Her findings on motion predicates will illustrate the contrasts found. Her 2003 investigation shows that the
< previous page
page_273
next page >
< previous page
page_274
next page >
Page 274 Vietnamese speakers frequently used periphrastic causative expressions involving motion (e.g. She has managed to rise the kite fly over the tallest building and The man dropped the can of paint fell ). This finding is consistent with the fact that Vietnamese relies heavily on serial verbs. The likelihood of substrate influence is high not only because of the character of the substrate, but also because of the performance of the speakers of Hindi and Urdu, who never produced such constructions (a behavior consistent with the absence of serial verbs in Hindi and Urdu). Like Vietnamese, Chinese relies heavily on serial verbs, and Siegel posits substrate influence in this area in the Hawaiian Pidgin English of early Chinese immigrants. These individuals would thus be able to model serial verbs for the succeeding generation that fashioned HCE. This transmission would thus be similar to the development of articles, except that Siegel’s claims about Chinese influence seem stronger with serial verbs because there is little in Portuguese or Spanish that could result in substrate influence in such HCE verbs. The likelihood of Chinese and Vietnamese influence in the development of serial verbs in ILs obviously supports similar conclusions in other work on pidgins and creoles and other cases of language contact (e.g., Ho and Platt 1993; Migge 2003). 3.2 Inferences About Formal and Functional Properties The methodology of SLA research on articles and serial verbs helps to assess the likelihood of substrate influence in language contact situations where the close comparisons possible in some SLA investigations may not be feasible. However, articles and serial verbs are specific structures, and so it is natural to wonder whether any wider inferences may be possible. A few examples to be discussed will suggest that the comparative methodology can indeed support broader inferences. As detailed previously, the correlative strategy in South African Indian English is evident in sentences such as which-one principal came here, she’s just cheeky like the other one. Such cases prompted Odlin (2001) to speculate whether correlation as a formal property is transferable in structural domains besides relativization. Other evidence does in fact indicate that the formal patterning in correlative structures is susceptible to substrate influence in different language contact settings. Ringbom (1987) notes the following example written by a native speaker of Finnish: Weather moves quickly from the other kind to the other kind (= ‘Weather changes from one kind to another quickly,’ and he sees this as modeled on Finnish toinen … toinen, ‘other … other’) (Ringbom 1987:125). He considers this an error characteristic of native speakers of Finnish and not of speakers of Swedish. Some strikingly similar cases of correlation appear in a description by Mesthrie (2004:969) of Black South African English: Others are for the proposal , others are against it (= ‘Some are for … others are …’). Although Mesthrie does not discuss any particular African language to account for these
< previous page
page_274
next page >
< previous page
page_275
next page >
Page 275 structures, Buthelezi (1995:248) cites Zulu as one likely source because it has a correlative construction omunye … nomunye and offers both a literal translation (‘one … and one’) and a more idiomatic one (‘the other and the other’). Apart from correlation, other formal properties may prove to be commonly transferable, but whatever the range of formal properties, functional characteristics also matter. Several cases of substrate influence involving focus constructions illustrate the importance of meaning, and once again, the comparison of second-language varieties proves helpful in establishing such influence. In their study of Swedish influence, Odlin and Jarvis (2004) identified a certain pattern of relativization, and the same pattern occasionally appeared in cleft sentences. The cleft pattern, which was used by L1 speakers of Swedish, involved a distinctive relative pronoun: He [Chaplin] say it were he som take the bred. Odlin and Jarvis found such uses of som only among native speakers of Swedish, not among any native speakers of Finnish. The distinctiveness of som in such cases argues clearly for substrate influence and constitutes one of a number of cases of focus constructions showing semantic or pragmatic influences from the native language (Odlin 2008). Formal similarity and functional motivation need not operate separately, of course, and the case of som clearly indicates an interlingual identification made by second-language users on the basis of formal as well as functional similarity. The studies of articles cited earlier in this chapter likewise indicate that substrate influence combines formal patterning (the co-occurrence of a determiner with a noun) and functional motivation (the expression of referential information such as definiteness). Yet another implication of the work on the transfer of article systems seems to be the significance of obligatory categories. Boas (1938:132) stressed the importance of cross-linguistic grammatical variation with respect to what categories are obligatory and contrasted the importance of categories such as definiteness and tense in English with evidentiality in other languages. If the common occurrence of (generally helpful) native-language influence seen in articles reflects the importance of obligatory systems, it is inviting to surmise that other obligatory categories will manifest themselves at least sometimes in substrate influence.4 The category of evidentiality that Boas cited proves useful, in fact, for verifying such a surmise, as seen in two recent studies of verb tense (Guler 2005; Odlin and Alonso-Vázquez 2006). In the same discourse contexts, speakers of different languages show different uses of the past perfect. Native speakers of Turkish used the past simple and past perfect to mark tense and evidentiality contrasts in English, whereas speakers of Castilian Spanish simply marked tense and not evidentiality. Guler’s results on Turkish learners of English are consistent with other studies of language contact, where obligatory evidentiality marking in the substrate language leads to a special IL structure (e.g., Bunte and Kendall 1981; Klee and Ocampo 1995).
< previous page
page_275
next page >
< previous page
page_276
next page >
Page 276 4. SUMMARY This chapter has focused on three methods used in second-language research to demonstrate the likelihood of substrate influence. Two of the methods require comparisons of the native language, the target language, and the IL of learners, but the methods differ. One of them requires only one group of second-language learners, all having the same native language, whereas the other method requires at least two groups having different native languages. (Another difference between these methods is that the first requires actual data produced in the native language to compare with the IL data, whereas the other method often simply relies on structural descriptions of the native language.) The third method requires five data sets: two distinct sets of IL productions and also two sets of native language productions in the same discourse context, along with a data set of native speaker productions in the same context. Although the last of these methods is the most powerful, it also requires the greatest collection effort. Most SLA research that has offered persuasive evidence of substrate influence has relied on the second method, requiring two sets of IL productions that reflect distinct patterns in the native language. Differences in IL productions in articles and serial verbs provide a chance to evaluate just how likely substrate influence was in the formation of Hawaiian Creole English. The SLA research is quite consistent in studies of articles, showing that speakers of a language with an article system find acquiring an article system in a new language to be much easier than do speakers of a language having no articles. Because Portuguese is a language having articles, it follows that Portuguese speakers would have had such an advantage in acquiring a superstrate with articles (be the superstrate considered HCE or some contact variety of nonstandard English). Although there is less evidence comparing two ILs in the same discourse context, speakers of a language with serial verbs seem much more likely to attempt constructing serial patterns in their IL, as seen in the research of Helms-Park (2003). This finding supports an inference that Chinese speakers in Hawaii were responsible, as Siegel (2000) suggests, for the formation of serial verbs in the creole. The insistence of Bickerton (1981, 1984) that there be a point-by-point correspondence in real cases of substrate influence proves to be unwarranted in light of the research of Jarvis (2002) on articles which show that native-language influence contributes much to the greater success of one L1 group in comparison with another, although neither group shows an invariable point-by-point correspondence between the native-language and the IL productions. Thus, in the case of Portuguese (or, for that matter, Chinese) influence on HCE, any insistence on a point-by-point correspondence is also unwarranted. Wider inferences seem possible from the SLA methods discussed. Syntactic correlation seems to be a transferable formal property because it appears in a number of different language contact settings. Focus constructions
< previous page
page_276
next page >
< previous page
page_277
next page >
Page 277 likewise recur in many settings, although functional factors also seem to motivate the influence. Furthermore, when categories such as definiteness and evidentiality are obligatory in the native language, they seem especially likely to influence IL productions. NOTES 1. Finnish has a pronoun se (with various allomorphs marking different cases) that can serve functions similar to those of articles. Jarvis (2002:405) details the uses of these forms by native speakers of Finnish in their narratives of the Chaplin film, but the figures indicate that their use is much less frequent than is the use of articles by native speakers of English and native speakers of Swedish. Accordingly, if se has any effect at all on the use of English articles by speakers of Finnish, the influence is marginal. 2. Readers who wish to verify this should be warned that some information in the Handbook suggests that the structure is also found in Cameroonian English. However, there is nothing in the Handbook chapter on Cameroonian English (Mbangwana 2004) about the after perfect. It is sometimes suggested that Irish missionaries may have modeled the after perfect in Cameroon or elsewhere. In such a case, superstrate influence of a Celtic English instead of substrate influence would be directly involved. Bobda (2006) discusses some likely cases of superstrate influence from Irish English, but he does not mention the after perfect, nor does another study of English in Cameroon (Wolf 2001). 3. Bickerton (1981:56) sees the functions of articles in creoles as restricted, where definite articles are typically used for presupposed and specific reference, indefinite articles for asserted and specific reference, and zero articles for nonspecific (including generic) reference. Regardless of whether the use of articles is so uniform from one creole to the next (as Bickerton claims), the fact remains that in some language contact situations there are speakers whose native language has articles, whereas in other situations the native languages do not have articles. To expect there to be no influence in such language contact settings in contrast to other second-language settings would require an explanation which neither Bickerton nor anyone else seems to have offered. 4. Caution is advisable in making predictions even about obligatory categories. Among the reasons for caution is the fact that grammatical categories can code sociolinguistic information that is highly languagespecific. Thus, for instance, verb forms in Japanese and Persian play a part in the honorific systems of those languages, and it is probably true that most speakers would not try to find cross-linguistic correspondences for the same social categories in the verb systems of languages such as English. REFERENCES Bickerton, D. 1981. Roots of Language. Ann Arbor: Karoma. ———. 1984. The language bioprogram hypothesis. Behavioral and Brain Sciences 7: 173–221. ———, and T. Givón. 1976. Pidginization and syntactic change: From SXV and VSX to SVX. In Papers from the Parasession on Diachronic Syntax , edited by S. Stever, C. Walker, and S. Mufwene, 9–39. Chicago: Chicago Linguistic Society. Boas, F. 1938. Language. In General Anthropology, edited by F. Boas, 124–45. Boston: Heath.
< previous page
page_277
next page >
< previous page
page_278
next page >
Page 278 Bobda, A.S. 2006. Irish presence in colonial Cameroon and its linguistic legacy. In The Celtic Englishes IV , edited by H. Tristram, 217–33. Potsdam: Potsdam University Press. Bunte, P., and M. Kendall. 1981. When is an error not an error? Notes on language contact and the question of interference. Anthropological Linguistics 23: 1–7. Buthelezi, Q. 1995. South African Black English: Lexical and syntactic characteristics. In Language and Social History: Studies in South African Sociolinguistics , edited by R. Mesthrie, 242–50. Cape Town: David Philip. Filppula, M. 1986. Some Aspects of Hiberno-English in a Functional Sentence Perspective . Joensuu: University of Joensuu. ———. 1999. The Grammar of Irish English: Language in Hibernian Style. London: Routledge. Guler, N. 2005. Transfer of Turkish Evidentiality . MA report, Ohio University. Helms-Park, R. 2001. Evidence of lexical transfer in learner syntax: The acquisition of English causatives by speakers of Hindi-Urdu and Vietnamese. Studies in Second Language Acquisition 23: 71–102. ———. 2003. Transfer in SLA and creoles: The implications of causative serial verbs in the interlanguage of Vietnamese ESL learners. Studies in Second Language Acquisition 25: 211–44. Ho, M., and J. Platt. 1993. Dynamics of a Contact Continuum: Singaporean English. Oxford: Oxford University Press. Jarvis, S. 1998. Conceptual Transfer in the Interlanguage Lexicon . Bloomington: Indiana University Linguistics Club. ———. 2000. Methodological rigor in the study of transfer: Identifying L1 influence in the interlanguage lexicon. Language Learning 50: 245–309. ———. 2002. Topic continuity in L2 English article use. Studies in Second Language Acquisition 24: 387–418. ———, and T. Odlin. 2000. Morphological type, spatial reference, and language transfer. Studies in Second Language Acquisition 22: 535–56. ———, and A. Pavlenko. 2008. Cross-linguistic Influence in Language and Cognition. New York: Routledge. Klee, C., and A. Ocampo. 1995. The expression of past reference in Spanish narratives of Spanish-Quechua bilingual speakers. In Spanish in Four Continents, edited by C. Silva-Corvalán, 52–70. Washington, DC: Georgetown University Press. Kortmann, B., E. Schneider, K. Burridge, R. Mesthrie, and C. Upton, eds. 2004. A Handbook of Varieties of English. Volume 2: Morphology and Syntax . Berlin: Mouton de Gruyter. Mbangwana, P. 2004. Cameroon English: Morphology and syntax. In Kortmann et al., eds. 2004. 898–908. McCafferty, K. 2004. Innovation in language contact. Diachronica 21: 113–60. McWhorter, J. 1996. A deep breath and a second wind: The substrate hypothesis reassessed. Anthropological Linguistics 38: 461–94. Mesthrie, R. 2004. Black South African English: Morphology and syntax. In Kortmann et al., eds. 2004. 974– 92. ———, and T. Dunne. 1990. Syntactic variation in language shift: The relative clause in South African Indian English. Language Variation and Change 2: 31–56. Migge, B. 2003. Creole Formation as Language Contact . Amsterdam: John Benjamins. Odlin, T. 1989. Language Transfer. Cambridge: Cambridge University Press. ———. 2001. Language transfer and substrate influence. In Concise Encyclopedia of Sociolinguistics , edited by R. Mesthrie, 499–503. Amsterdam: Elsevier. ———. 2003. Cross-linguistic influence. In Handbook of Second Language Acquisition, edited by C. Doughty and M. Long, 436–86. Oxford: Blackwell.
< previous page
page_278
next page >
< previous page
page_279
next page >
Page 279 ———. 2006. Could a contrastive analysis ever be complete? In Cross-Linguistic Influence in the Second Language Lexicon , edited by J. Arabski, 22–35. Clevedon, UK: Multilingual Matters. ———. 2008. Focus constructions and language transfer. In Morphosyntactic Issues in Second Language Acquisition Studies , edited by D. Gabryś-Barker, 3–28. Clevedon, UK: Multilingual Matters. ———, and C. Alonso-Vázquez. 2006. Meanings in search of the perfect form: A look at interlanguage verb phrases. Rivista di Psicolinguistica Applicata 6: 53–63. ———, and S. Jarvis. 2004. Same source, different outcomes: A study of Swedish influence on the acquisition of English in Finland. The International Journal of Multilingualism 1: 123–40. Paulasto, H. 2006. Welsh English Syntax: Contact and Variation. Joensuu: Joensuu University Press. Ringbom, H. 1987. The Role of the First Language in Foreign Language Learning . Clevedon, UK: Multilingual Matters. ———. 2007. Cross-Linguistic Similarity in Foreign Language Learning . Clevedon, UK: Multilingual Matters. Sabban, A. 1982. Gälisch-Englischer Sprachkontakt . Heidelberg: Julius Groos. Schmidt, R. 1987. Sociolinguistic variation and language transfer in phonology. In Interlanguage Phonology, edited by G. Ioup and S. Weinberger, 365–77. Rowley, MA: Newbury House. Selinker, L. 1969. Language transfer. General Linguistics 9: 67–92. ———. 1972. Interlanguage. International Review of Applied Linguistics 10: 209–31. Siegel, J. 2000. Substrate influence in Hawai’i Creole English. Language in Society 29: 197–236. ———. 2005. Recent evidence against the language bioprogram hypothesis. Studies in Language 30: 51–88. Stauble, A. 1984. A comparison of a Spanish-English and a Japanese-English second language continuum. In Second Languages: A Cross-Linguistic Perspective , edited by R. Andersen, 323–53. Rowley, MA: Newbury House. Thomason, S., and T. Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Wode, H. 1983. On the systematicity of L1 transfer in L2 acquisition. In Papers on Language Acquisition, Language Learning and Language Teaching, edited by H. Wode, 144–49. Heidelberg: J. Groos. Wolf, H. 2001. English in Cameroon. Berlin: Mouton de Gruyter.
< previous page
page_279
next page >
< previous page
page_280
next page >
Page 280 12 Some Offspring of Colonial English Are Creole Salikoko S. Mufwene 1. INTRODUCTION Linguists have cherished the following myth in discussions of the evolution of English and its speciation into several varieties, viz., those spoken by descendants of Europeans, wherever they evolved, are English dialects, whereas most of the nonstandard vernaculars that have evolved among populations of nonEuropean descent are creoles and separate languages altogether. Holm (1988, 2004) and some other creolists have attempted to accommodate the ‘cline’ in the evolution of colonial English (Schneider 1990) and other European colonial languages by positing an intermediate category of semi-creoles, which include varieties such as African American Vernacular English (AAVE), but not Old Amish English. Although AAVE and the like have diverged extensively enough to be considered separate languages by some (as is evident from the literature on ‘Ebonics’, like on Afrikaans), they have not developed enough of those features (so far controversial) that would make them ‘creole’. However, as is evident from Mühlhäusler (1985) and Mufwene (1988), the sentiments of the (native) speakers of these varieties, which, according to the received doctrine, are what should matter the most, have generally been ignored. Like speakers of AAVE, those of both Gullah and Kanaka English claim they speak English. Linguists have disfranchised them by simply stipulating that these varieties are not mutually intelligible with other English vernaculars (Mufwene 2001). The same disenfranchising treatment has also been extended to new English varieties that have emerged in former British exploitation colonies, which have been characterized as indigenized or non-native Englishes.1 That mutual intelligibility is not a reliable criterion has mattered very little in this case. It is also as if indigenization as a process by which an exogenous phenomenon is adapted to a particular territory and autonomization as the process by which a language variety develops a separate norm (Chaudenson 1992, 2001) did not apply in the transformations that changed English in the settlement colonies of North America, Australia, and New Zealand in particular.
< previous page
page_280
next page >
< previous page
page_281
next page >
Page 281 The prior myth has its foundation in some questionable assumptions in linguistics, including the following which are disputed in Mufwene (1998, 2001, 2005a): 1) External factors, including language contact, play a negligible role in determining what is assumed to be normal and ordinary language change. If they do, as in the case of creoles, then the change is no longer ordinary. Creoles are thus claimed to have developed by nonordinary and unusual processes (Hock and Joseph 1996), which make them children out of wedlock, so to speak (cf. Mufwene 2001). Thus, they can be disfranchised from the family of descendants of the English language, consistent with Thomason’s (2001, 2002) and Thomason and Kaufman’s (1988) position that creoles cannot be classed genetically with their lexifiers, as they allegedly inherited only most of their vocabularies from the latter, but selected their grammars from other sources (cf. Mufwene 2003, 2005a).2 2) There is a creolization process that can, at any point in human history and anywhere, transform any language into a creole, despite increasing evidence (including Mufwene 2001, 2005a; Trudgill 1986, 2004) which suggests or argues that there cannot be such a global language-restructuring process (pace Trudgill’s 2004 conclusion against his own data).3 3) Although there are no structural criteria for distinguishing dialects from languages, we can be certain that English creoles (in the present case) are not English dialects because speakers of ‘native Englishes’ who are not familiar with these vernaculars claim they do not understand them (see e.g. Holm 1988; Romaine 1994). The documented fact that there are so many dialects of the same language which are not mutually intelligible has simply been ignored, as much as the fact that the ecologies of the development of creoles and other ex-colonial varieties of English have not been identical (Schneider 2007). 4) With the exception of indigenized Englishes, the development of new non-creole varieties of English outside Europe reflects internal language evolution, i.e. involving only negligible contact with other languages, consistent with the way that the role of Old Norse and Norman French can apparently be ignored in the development of precolonial English in England (see e.g. Thomason and Kaufman 1988). Why we have favored a model of historical linguistics that downplays the role of social history is intriguing. I question different aspects of the prior myth and assumptions briefly in this chapter. I argue that there is nothing wrong with treating creoles as dialects of their lexifiers, consistent with the sentiments of several of their speakers, despite the traditional overemphasis on the role of contact in their development at the expense of natural
< previous page
page_281
next page >
< previous page
page_282
next page >
Page 282 processes of language evolution so evident in them (Mufwene 1998, 2001, 2005a). In the process, I also question common accounts of the origins of dialect differences such as the following: 1) North American English varieties are different from British varieties because their speakers are separated by the Atlantic Ocean; or 2) speakers of two dialects separated by a mountain or a river have different structural features because the mountain or river prevents them from communicating with each other on a regular basis. Such observations do not explain why such diverging dialects have usually also preserved several of the features they shared before the geographical split, nor how they stopped sharing some of those that now distinguish them from each other. These questions underscore why it is useful to invoke ecology (both internal and external), which in theories of evolution is said to roll the dice, determining the fate of any species. Ultimately, we must address the question of whether such speciation processes traditionally claimed to be internally motivated are different from those which produced creoles and are said to be externally motivated. Comparisons of Mufwene’s (1996a, 2001, 2002, 2005a) competition-andselection approach to the development of creoles with Trudgill’s (1986, 2004) account of the development of new English dialects in North America, Australia, New Zealand, and the Falkland Islands suggest that the restructuring processes are of the same kind and subject to the same constraints (Mufwene 2006). One could also argue that their speakers are justified in claiming that they speak dialects of their lexifiers (Mufwene 2005a). A related question here is whether contact was not involved in the development of colonial dialects. The data in Trudgill (1986, 2004) support Mufwene’s position that it was, subject to varying ecological factors invoked later. However, it will be informative to start with the evolution of precolonial English. 2. HOW DID ENGLISH EVOLVE? Bailey and Maroldt’s (1977) position that Middle English evolved by ‘creolization’ has justly been rejected, notably by Thomason and Kaufman (1988—but see also Mufwene 2005a for a complementary discussion). Unfortunately, the question of the origins of English itself, which could be just as informative about whether the emergence of creoles is peculiar, has generally also been dodged. How did English develop out of the languages brought over by the Jutes, the Angles, and the Saxons to England? How did Modern English evolve to be different typologically from modern Dutch and modern German, for instance, regarding the postposing of auxiliary verbs in clause-final position in subordinate clauses? Assuming that the contribution of Celtic languages to the structures of Old English is negligible—because the Celts were as marginalized from the invaders as the Native Americans were centuries later by the European colonists (Mufwene
< previous page
page_282
next page >
< previous page
page_283
next page >
Page 283 2001)4—what led to the emergence of Old English and the concurrent disappearance of the languages that the Jutes, Angles, and Saxons had brought with them to England? May we simply assume that in the first place these West Germanic languages were probably not structurally identical, perhaps not mutually intelligible, and that contacts among them alone would have led to the development of Old English? May the emergence of Old English as a process have been a precursor of that of the koinéization that produced nonindigenized and non-creole colonial English varieties, especially those such as on the Falkland Islands which evolved almost exclusively from the contact of metropolitan English vernaculars? Other interesting questions arise, of course. For instance, was there competition among the Germanic languages brought over to England? Which one prevailed, albeit in a modified form influenced by the others? How were its features selected? From a language evolution perspective, is there a genetic connection between the language of the Angles (who had settled in Mercia and Northumbria) and Old English, which may account for the etymological connection between the names Angle , English, England? As explained in Mufwene (2000), names for languages and people have traditionally been derived from the same base morphemes. This is true all over the world. As a non-Anglicist, I am not aware that these questions have often been addressed in the literature. Although I have no direct linguistic evidence, history suggests a developmental scenario of Old English that may not be so different from that of the colonial English varieties spoken by descendants of Europeans, out of the contact of significantly diverse metropolitan dialects of English. In both cases, we need not assume a priori that the varieties that came in contact with each other were (fully) intelligible. Note that, because of parochialism among the different European nations until the nineteenth century, the history of North America suggests that American English started essentially from the koinéization of metropolitan dialects and the resulting American koiné would be influenced by continental European languages only after the Anglo economic system prevailed and the other Europeans started functioning in it and shifting to English as their vernacular. These shifts were of course also concomitants of the gradual ethnic integration of populations of European descent, along with redefinitions of the White race (Mufwene, to appear). Increased immigrations from continental Europe, which would give White demographic majority to continental Europeans all the way into the twentieth century, also increased the possibility of adstrate influence from continental European languages. Only the critical mass achieved by the Anglo populations early during the colonial period and the fact that the overall population increased incrementally prevented a drastic divergence of post-Revolution American Englishes from colonial American Englishes. The fact that a large proportion of locally born and immigrant children could acquire the target language natively, favoring the Anglos’ features over those of the non-Anglos was also a significant ecological factor perhaps more important than the incremental immigration
< previous page
page_283
next page >
< previous page
page_284
next page >
Page 284 of the adult populations. The children generally avoided the stigmatizing xenolectal features produced by adult non-native speakers, thereby selecting most, although not all, of these out of the targeted American English, as is happening today with immigrants (Mufwene 2005a). In the same vein, we must also factor in the fact that Old English was not homogeneous. We must then ask whether its different varieties were mutually intelligible and, more importantly, to what extent they were typologically similar. These questions are difficult to answer, but the literature on the subject matter indicates that there was lexical and structural variation in Old English (Crystal 1995; Hogg 1992). However, the same literature also reports that speakers of these different varieties formed alliances to defend themselves against common enemies. Therefore, populations moved around and the varieties that they spoke came in contact with each other, which led to restructuring. This change has been associated with the beginning of the simplification of Old English morphosyntax, although one may wonder whether this subsystem was equally complex in all varieties—for instance, whether all varieties were completely fusional or whether at least some of them contained a substantial share of periphrastic strategies that contributed to the development of predominantly analytical morphosyntax in today’s English. These considerations don’t of course entail that we can ignore the presence of Latin and Norman French in medieval England and the impact that they must have exerted indirectly on English varieties of the upper class, perhaps primarily in favoring grammatical strategies that were consistent with these foreign systems. A good example of this evolution by congruence is the pied-piping of the preposition with the WHconstituent in relative clauses, as in This is the question with which I am concerned. Lüdtke (1995) helps us put in perspective Thomason and Kaufman’s (1988) mistaken position that contact with Norman French is not the reason that Old English changed into Middle English. How can we ignore the socioeconomic significance of ‘Norman English’ spoken by the Norman aristocracy after Norman French lost its prestige to Parisian French in the twelfth and thirteenth centuries and the Norman aristocrats decided to anglicize? Don’t the seeds of modern Standard English lie partly in these social dynamics? Doesn’t the fact that preposition stranding is rather the norm in colloquial and nonstandard English5 suggest that Latin and French, which were also used by the intellectual elite, would have favored the native pied-piping option only in the then emergent Standard English? Lack of a native prestigious English model in a polity where foreign languages, Latin and Norman French, had served the upper class favored the acceptance of the French-influenced variety of English as a prestigious one. From that time onwards, there have been a few areas of divergence between, on the one hand, the proletarian, nonstandard vernaculars and, on the other hand, the standard, prestigious varieties of English. For instance, it is doubtful that the nonstandard vernaculars have been as much affected by the
< previous page
page_284
next page >
< previous page
page_285
next page >
Page 285 Great Vowel Shift as the standard variety. WH-relatives are not common in nonstandard varieties, which also entail the absence of pied-piping in them. In short, contact has played an important role in the bifurcated evolution of English in England and thus of its speciation into diverse social and regional dialects. The role of contact has of course been too obvious to be denied in the development of Irish and Scottish English. The growing literature on Irish English (including Hickey 2005; Kallen 1997) speaks for itself. In this case, the lower, rather than the upper, class is the one that has been the most associated with substrate influence from Gaelic. This appears to be plausible since because English did not spread in Ireland until the introduction of the potato farms and the investment of the Irish migrant workers in this colonial language, although it had been introduced centuries earlier. The migrant workers succeeded in making English utilitarian where the school system had failed (Odlin 2003). The presence of the Irish in the extra-European colonies has often also been invoked to account partially for differences between the relevant colonial varieties and British English. However, does the relevant literature tell the full story? 3. CONTACT AND THE DEVELOPMENT OF NORTH AMERICAN VARIETIES OF ENGLISH The positions I question in this section have been more suggested than explicitly stated in the literature. Nonetheless, it helps to discuss them in the interest of understanding language evolution and speciation. As noted by Mufwene and Pargman (2003), the literature suggests that only the development of AfricanAmerican English varieties was influenced by language contact; those spoken by North American Whites would putatively be outcomes of (primarily) internal-language evolution. Little attention has been paid to differences in the ecologies of their developments—for instance, the contact patterns of various metropolitan dialects in the colonies and when significant contact with continental European languages took place, most likely after rather than before the American Revolution. Such a historical perspective accounts not only for the koinéization that produced new varieties even in colonies where the settlers were almost exclusively from England (at least before the twentieth century), but also for the fact that North American White varieties of English are less divergent from the metropolitan varieties than the African-American varieties. The reason for the greater divergence of African-American varieties lies not only in the nature of substrate influence, but also in the fact that continental Europeans shifted quite late to English as their vernacular and pressures for conformity to a colonial English that had already emerged without their agency must have worked on them in the same ways as on later immigrants. Accordingly, continental European children, both immigrants and American-born,
< previous page
page_285
next page >
< previous page
page_286
next page >
Page 286 acquired the local English natively while their parents typically died with their xenolectal features, affecting the target language only minimally or marginally (Mufwene 2005a). In any case, if the role of contact is not factored into accounts of the evolution of English in colonial populations, one cannot explain why no single North American English variety matches one single British dialect. Even Appalachian and Ozark varieties of English, which have been associated with strong Gaelic or Irish English influence, do not exactly match Irish English! Nor is New England’s English, which is considered the American variety the closest to British English, a faithful reproduction of East Anglian English.6 One must also wonder whether it is true that, as the literature often claims, North American varieties of English are apparently more conservative because they did not participate in changes that have affected the British Isles since the eighteenth century. If they are, how and to what extent do they reflect (pre-)eighteenth-century British English? Didn’t later emigrations from the United Kingdom to North America bear on linguistic developments in the latter territory? Or is dialect contact, combined with the founder effect, the primary reason that North American Englishes appear to be conservative? I express all these concerns because the socioeconomic history of the Western world suggests that the restructuring of European languages into new colonial varieties was conditioned by similar ecological factors earlier in Europe and that the trigger role of contact cannot be dismissed casually from accounts of their putatively internally motivated evolution (Mufwene 2007). The same history also suggests that some of the same restructuring processes attested outside Europe started in Europe in cities such as London, Bristol, and Liverpool as contact settings from which British subjects emigrated, as proposed by Chaudenson (1992, 2001 for colonial French) and Buccini (1995) for colonial Dutch. Differences in the outcomes of these developments follow more from variation in the individual ecologies (internal and external) of language evolution than in the nature of the restructuring processes involved (Mufwene 2001, 2005a). English in England underwent changes during the same time it was colonizing North America and other parts of the world. The routes of population movements and language contact which produced changes in British English were concurrent with the population movements that brought several English and Irish, largely destitutes from the low class and speakers of nonstandard varieties of English, to North American colonies and elsewhere. According to Bailyn (1986), the emigrations to North American colonies were largely spillovers from the population movements in England. Although the specifics of the various contact settings were not identical, basically similar populations relocated and came in contact both within and outside the British Isles, notwithstanding the fact that they came in contact with other populations in most colonies. Although the early colonial European population in New England was relatively homogeneous in its origin from primarily East Anglia, that of the
< previous page
page_286
next page >
< previous page
page_287
next page >
Page 287 Chesapeake area (Virginia, Maryland, Delaware, and Pennsylvania) and of the Southeastern colonies (especially South Carolina and Georgia) was more heterogeneous. Aside from a significant proportion of Irish indentured servants, it contained exiles and indentured laborers from Germany, France, and other places in Europe. In some cases, the English took over colonies that had belonged to other European nations, such as the New Netherland (New York and New Jersey) from the Dutch and Nova Scotia and Maine from the French. Although, as allegedly in England several centuries earlier, the invaders did not apparently mingle much with the Natives, they came in contact with other varieties of their own languages, as well as with other languages. By normal processes of accommodation, which entails selection of subsets of features out of larger pools where they competed with each other, new varieties arose (Mufwene 1996a, 1996b; Trudgill 1986, 2004). It is especially informative that although studies such as Tagliamonte (1996), Smith and Tagliamonte (1998), and Shackleton (2005) continue to show continuities of nonstandard British English features in North America, none of the American varieties in which the features are attested matches exactly any particular metropolitan dialect. The previous studies and others, especially Clarke (1997) and Trudgill (1986, 2004), which focus on varieties unlikely to have been influenced by African languages, cast doubt on the position that the vernaculars spoken by descendants of Africans outside Europe developed by different restructuring processes from those spoken by descendants of Europeans. The ecologies in which vernaculars spoken by descendants of Africans developed differed from those of the others by the additional presence of African substrate features in the pool of competing grammatical strategies. The influence that these languages exerted in determining which features of the lexifier would be selected into the emergent vernaculars must have become more obvious since the eighteenth century, when the plantation industry started to thrive, when the African slave labor became dominant in the New World and continued to grow rapidly thanks more to population replacement than to birth, and when on the large rice and sugar cane plantations the European and African populations were typically segregated (Mufwene 1996a). However, on the small farm holdings and plantations, social relations between the African slaves and the European indentured servants seem to have remained relatively close.7 Consequently, both groups developed similar varieties of English, as is obvious from AAVE and American White southern English varieties, both of which are related to small farms and to tobacco and cotton plantations up to the nineteenth century.8 As much as the Black-nanny myth has been cherished to explain why American White Southerners speak like African Americans, one must remember that the vast majority of those Whites could not afford a nanny in the first place. The explanation for similarities in their speech ways lies more in similarities of their colonial social status (indentured laborers and slaves) and of (European) linguistic inputs (nonstandard English varieties),
< previous page
page_287
next page >
< previous page
page_288
next page >
Page 288 as well as in the relatively intimate relations which obtained between (descendants of) Europeans and (descendants of) Africans up to the institutionalization of segregation with the Jim Crow laws in the late nineteenth century. These considerations make the role of contact in the development of new varieties of English as significant as some other ecological factors which I discuss briefly later—more extensively in Mufwene (1996a, 2001, 2005a, 2007)—as responsible for cross-group and/or cross-setting similarities and differences. Such factors include the following: 1) the nature of the varieties brought in contact (standard vs. nonstandard), 2) the proportion of speakers of the different varieties (White majority vs. non-White majority), 3) the nature and frequency of interactions among speakers of different socioeconomic and/or ethnic backgrounds (depending on whether the societies were integrated or segregated), and 4) the Founder Principle/Effect and patterns of population growth. The varieties whose inputs were mostly nonstandard, including creole varieties, share a lot of nonstandard features. For instance, creoles and White nonstandard vernaculars are alike in not having pied-piping and favoring preposition stranding (e.g. the pen with which the President signed the bill vs. the pen [that] the President signed the bill with ). Likewise, both kinds of vernaculars form the nominal plural with them (at least as an alternative), as in them folks—although the constraints are not exactly the same— for instance, the fact that, in creoles, dem in such a function does not have a (strong) distal force and can be used in associative plural constructions, as in Felicia (an) dem ‘Felicia and her associates’. Such divergence need not shock us; after all, the substrate languages had a role to play in the restructuring process. Sometimes it drove evolution in a more divergent path from the metropolitan varieties, favoring innovative constructions that seem to have no models in the lexifier.9 Nor need we be shocked by the fact that AAVE is not as different from White nonstandard varieties of English as the creoles are. As explained above, where AAVE developed, segregation was institutionalized much later—in the late nineteenth century —than in the settings where creoles developed, where it was institutionalized in the early eighteenth century.10 AAVE also developed where the proportion of Africans was hardly higher, if at all, than that of Europeans, which made a lot of allowance for children of the African-born slaves to acquire English directly not only from Black Creoles (the locally born), but also from White ones (Mufwene 1999). As a matter of fact, the nature of interaction across class and/or ethnic boundaries appears to be a more significant factor than the demographic disproportion, on which genetic creolistics has capitalized to date. Although one may claim that it differs from American Southern English mostly in the statistical distribution of structural features, AAVE has developed in settings where (descendants of) Africans have been minorities. Indirect evidence for my hypothesis comes from varieties such as Old Amish, Ozark, and Appalachian Englishes (e.g. Keiser 2003; Wolfram, Christian, and Dube
< previous page
page_288
next page >
< previous page
page_289
next page >
Page 289 1988), whose speakers are descendants of settlers who were not native English speakers either during the colonial period or during immigration waves or migrations westward (from the original 13 United States) in the nineteenth century. Having evolved in social and/or geographic isolation from the majority populations’ varieties, they are significantly different from what is considered mainstream American middle-class English, in fact not at all closer to it than AAVE is. The other part of the evidence comes from the divergence hypothesis as formulated especially by Bailey and Maynor (1987) and Bailey and Thomas (1998). Segregation has increased the divergence of White and African-American vernaculars in the United States (see also Labov and Harris 1986), and these are becoming obvious even in the American South. This is evident even in the Northern and Southern cities vowel shifts (Labov 2001), which show that White and African-American populations socialize primarily within their ethnic boundaries (Mufwene 2005b). Note also that, although the late nineteenth century marks the passage of the Jim Crow laws and the very beginnings of the Great Migration of Southern Blacks to segregated North American cities, it is also the beginning of the integration of White Americans. It started the gradual end of parochialism among European immigrants along national lines. The integration fostered the gradual disappearance in the twentieth century of varieties thitherto identified as Italian and German Englishes, for instance. The ethnic dialectal distinctions that still survive in the United States today do not reflect different kinds of language restructuring processes; they reflect the varying impact of divergent social ecologies on language evolution (Mufwene 2001, 2005a). The way ethnic dialect differences have almost vanished among Euro-Americans since the twentieth century is similar to what happened much earlier among the slaves in all the colonies. Assuming that the slaves were able to tell, based on the structural properties of a person’s speech (especially their prosodic features), the part of Africa their bozal cohort was from, those differences were obliterated by the focusing which produced among the locally born slaves the sociolects now identified as AAVE and Gullah. Some regional differences have emerged within these sociolects, but they reflect variation in the ecologies in which they have evolved. Unlike the descendants of Asian contract laborers in Hawaii, African Americans and other descendants of slaves in the New World have lost the ethnic distinctions that their ancestors had brought from Africa. There is also a sense in which African Americans are more American than just descendants of African slaves. This observation is made more evident by a superficial comparison of dialect complexes in the United States and the Caribbean. Despite differences that are obvious to the locals, White and Black Caribbeans sound more like each other than like European and African Americans, respectively. Likewise, European and African Americans sound more like each other than like White and Black Caribbeans,
< previous page
page_289
next page >
< previous page
page_290
next page >
Page 290 respectively. Even race segregation in both regions could not prevent these convergences between the languages varieties of populations of European and African descents in their respective colonies. In the case of North America, another factor must also be considered. It can be argued that, although present-day AAVE and Gullah had probably stabilized by the end of the nineteenth century, such is not the case for today’s White American varieties of English. Indeed, Bailey (1997) argues that today’s American White Southern English varieties reflect socioeconomic changes that affected the American South in the nineteenth century. The abolition of slavery and the collapse of the plantation industry put an end to importation of African labor. Yet the North American population continued to grow significantly through massive immigrations from Europe. A large proportion of these immigrants were less proletarian than the founder population (consisting largely of indentured servants and farmers); according to Hirsch (1991), they were more integrated (although it is all a matter of degree) into the extant White American population than were the African Americans. On the one hand, because the different waves of immigration increased the overall population only by installments, gradual integration of the immigrants within the host population prevented any major language restructuring of the kind that had already produced significant divergence between North American and British Englishes. This is all consistent with the Founder Principle, according to which characteristics of today’s language varieties were largely, although not exclusively, determined by selections made by the founder population (or those who came earlier; Mufwene 1996a, 2001). On the other hand, the greater proportion of less destitute or more affluent and better educated immigrants in the nineteenth century favored the emergence of middle-class varieties that were less conservative. However, one cannot ignore the ecological significance of population structure. Where there was not as much integration, the immigrants produced new divergent vernaculars too, such as in the Midwest, where the Scandinavian immigrants produced the variety stereotyped in the movie Fargo . Gradual integration within the extant Anglo populations have of course eroded such ethnic sociolects, although these have left their marks on present-day varieties, such as the often invoked construction come/bring NP with , which reflects the German(ic) construction kommen/bringen NP mit . Although Bailey (1997) is correct in noting migrations of Americans within and into the South and Southwest, one cannot ignore the more general fact that most of today’s United States developed out of migrations westward from the original 13 colonies that became independent in the late eighteenth century. The migrations westward entailed contacts even among the earliest colonists, and these entailed relatively more restructuring, however minor this may have been compared with the earlier restructuring that produced differences between North American varieties and British dialects. As noted earlier, African Americans participated in these migrations too. The Black Exodus of 1879 and the Great Migration at the turn of the century brought
< previous page
page_290
next page >
< previous page
page_291
next page >
Page 291 several of them to Northern, Midwestern, and Western cities. However, most of them have lived segregated in ghettos, where they have socialized primarily among themselves and less with the White residents of the same cities, participating marginally, if at all, in the restructuring of White American vernaculars. They have continued to speak AAVE varieties that differ little grammatically from the varieties that the migrants brought from Southern states, where 90 percent of the African-American population had lived up to the end of the nineteenth century.11 That is, today’s African Americans have restructured their ethnic variety minimally, which lends some support to the typical assumption among its students that it is relatively homogeneous across the United States.12 Overall, contact and variation in the sociohistorical ecologies of language evolution continue to be the explanation for the speciation of English into several new varieties in North America and the Caribbean. Here I argue that the same kind of evolution took place elsewhere, differing only in the specific ways in which internal and external ecologies determined the outcomes. From a structural point of view, there is nothing which justifies assuming that, because they arose from language contact and mixing, English creoles and AAVE are not dialects of English, or that they descend less legitimately from it than any other new variety. The other varieties also appear to have developed under contact conditions. In the case of English varieties spoken by Euro-Americans, suffice it to note that fewer than half of Euro-Americans today descend from immigrants from the United Kingdom; English was therefore in contact with other European languages, whose structural features are different from its own in some respects today and then. Therefore, those continental European languages must have influenced the evolution of English among Euro-Americans, although apparently less extensively. As explained earlier (see also Mufwene 2005a, 2006), this has to do with how late (i.e. in the nineteenth century, rather than earlier) the continental European immigrants shifted to English as their vernacular after the essence of today’s White American English(es) had already developed. The impact of the late shifters can be compared to that of recent immigrants in the twentieth century, who also joined the extant population incrementally, with their children acquiring the extant varieties natively and the parents eventually dying with their xenolectal accents. That is, with the populations relatively integrated, children were the primary agents of selection, favoring native over nonnative features and minimizing the impact of xenolectal grammatical features on the extant varieties. 4. INDIGENIZED ENGLISHES AND THE EVOLUTION OF ENGLISH The starting question here is: does the same approach apply to places where indigenized varieties of English have developed? As in Mufwene (2001,
< previous page
page_291
next page >
< previous page
page_292
next page >
Page 292 2005a, 2007), I answer this question affirmatively, keeping in mind that, although the algebraic equation for language restructuring in contact conditions is apparently the same, the ecologies in which the formula applies are not identical. The latter assign different values to several variables and yield differing outputs. In fact, places such as South Africa highlight forcefully the need to invoke varying ethnographic ecologies even within the same polity to account for differing outputs—just like in North America in some ways, but still differently in some other respects.13 First, we must note the need to distinguish between, on the one hand, communities where English has been appropriated as a vernacular (such as among African Americans and among Indians and some colored people in South Africa) and, on the other hand, those where English has been appropriated primarily as a lingua franca and is used as an umpteenth language variety (such as among the Bantu populations and Afrikaners in South Africa, among Black Africans in former British colonies, and among Indians in India). Where English functions as a lingua franca, it is associated primarily with the school system and the elite, notwithstanding the less educated who have produced a more restructured variety of their own (Kachru 1983, 2005), often showing more influence from the substrate languages. Where English functions as a vernacular, the history is more complex. The less prestigious segments of the populations typically speak varieties that originated in some nonstandard native varieties, and an adequate assessment of the extent of divergence of their systems requires comparison with native nonstandard varieties, not the standard one, contrary to the common tradition in studies of AAVE and English creoles. However, assessing the extent of divergence of the indigenized lingua-franca varieties requires comparison with the scholastic models from which they developed.14 These facts underscore the assumption that variation in the characteristics of the new English vernaculars or lingua francas is partly correlated with variation in the varieties targeted by those who produced them. Another part of the explanation lies in the languages previously spoken by (descendants of the) new speakers, what is known as substrate influence in creolistics. How many kinds of Englishes evolve in a particular polity has generally been determined by its contact history. Places like Nigeria are an interesting case where English was actually introduced twice, at different periods and in different ways. Nigerian Pidgin English evolved from the contact of nonstandard English with African languages when the English developed and maintained trade colonies in the region and had sporadic interactions with those who traded with them in the region. The trade language was then nonstandard English, which was approximated through naturalistic learning and was significantly modified by its new speakers. Nigerian English, the indigenized variety, evolved from scholastic English, introduced in the nineteenth century, during the exploitation colonization of the territory. The British administrators needed colonial auxiliaries (especially bureaucrats and teachers) who would interface between the colonizers and the indigenous
< previous page
page_292
next page >
< previous page
page_293
next page >
Page 293 populations and they taught them English. The actual interactions in English with the colonial auxiliaries were limited to the work place because, as a rule, the colonizers did not socialize with the colonized. However, English also evolved to function as a lingua franca within the elite, especially when they did not share an indigenous vernacular. It has also come to function as a socioeconomic class marker. This particular pattern of xenolectal usage of the colonial language fostered some divergence, structural and pragmatic, that is characterized as the indigenization of English. On more or less the model of creole speech continua, indigenized English communities are also internally variable (see also Schneider 2007). The extent to which a particular speaker’s idiolect diverges from the metropolitan native educated norm depends on his or her level of education (Kachru 1983, 1996, 2005), the kinds of teachers he or she has had (Bambgose 1992), his or her first vernacular (Bambgose 1998), and whether he or she has interacted significantly with native speakers, notwithstanding his or her language learning skills. Needless to say, the settings in which indigenized Englishes have emerged are in some ways reminiscent of those in which creoles have emerged and diverged from their lexifiers. One of the reasons for the divergence lies in the segregation that led the new speakers to use their colonial vernacular for communication more among themselves than with European native speakers. Segregation fosters divergence regardless of the nature of the input, which is obviously different between English creoles and indigenized varieties. Differences among indigenized varieties are of course also largely due to variation in the structural characteristics of the substrate languages. This factor accounts for differences between creoles too. For instance, the substrate languages were not the same in the Caribbean and Hawaii, although one must also factor in the fact that the English taken to Hawaii was mostly late eighteenth- and early nineteenth-century American middle-class English and not seventeenth- and eighteenth-century British nonstandard English. However, even in the Caribbean, the demographic strengths of the different African languages were not identical from one colony to another, especially at different times during the growth of the colonies; therefore, the dynamics of competition and selection in the feature pools were not identical either. Aside from differences in the natures of their inputs, indigenized Englishes differ from creoles in that the latter function as vernaculars because they have typically evolved in settlement colonies in which their creators did not have the option of continuing to speak their ancestral languages as their vernaculars. In this respect, South Africa is especially interesting because it represents both types of worlds: a settlement colony for non-Blacks and an exploitation colony of some sort for the Blacks. With the exception of Afrikaners, it is vernacular English varieties that have developed among the immigrants, including South African Indians, whereas indigenized varieties have developed among the Blacks. The Afrikaners have developed indigenized, second-language varieties for reasons that are obvious from their
< previous page
page_293
next page >
< previous page
page_294
next page >
Page 294 social history since the nineteenth century: their determination to foster their own non-European and Afrikaner/African identity and their resistance to the adoption of English as a vernacular. (They have promoted education in Afrikaans for themselves.)15 These ecological subsystems are of course in contact with each other, and little is known to date on how the emergent English varieties have influenced each other. In any case, these observations underscore once more the role which language contact and the nature of social interaction have played in the evolution of English within ever-changing ethnographic ecologies. They highlight the significance of the kind of language variety (standard vs. nonstandard) to which a population has been exposed and the primary mode of acquisition: interactive oral mode for vernacular varieties versus the school (through primarily the written medium) for the indigenized, lingua-franca varieties. In both cases, language varieties previously acquired tend to influence the structure of the emergent variety. There are in fact some parallelisms between, on the one hand, the polities where English has evolved into indigenized varieties and, on the other hand, those where it has restructured into both basilectal-creole and the acrolectal varieties. Although creoles have been lexified by nonstandard varieties and developed through strictly oral modes of language transmission, the true acrolects are school-based, just like the prestigious varieties of indigenized Englishes. On the other hand, indigenized Englishes owe the features that make them the most peculiar to oral transmission in the daily interactions where less competent learners introduce features from their native languages. Some of these spread to become communal peculiarities. 5. CONCLUSIONS Going back to the title of this chapter, English creoles are outcomes of the same restructuring processes that have affected the evolution of English everywhere and its speciation into different varieties under the influence of language and/or dialect contact. Creoles’ structural features have been partly determined by the settings of their emergence, viz., plantation settlement colonies in which those who developed them needed a new vernacular for communication even among themselves in the colonial communities. The structural distance between these vernaculars and the standard varieties of their lexifiers, with which they have typically been compared—misguidedly—is determined in part by the structural features of their nonstandard lexifiers and by those of the substrate languages with which the latter came in contact. Research and the debate on how the features were selected into the new vernaculars continue, and more and more interesting questions arise which we must answer (see Mufwene 1996a, 2002). Basically, the same restructuring equation has applied everywhere, only the values of the variables in this algebraic equation have varied and thus produced the spectrum
< previous page
page_294
next page >
< previous page
page_295
next page >
Page 295 of varieties of English we now know (Mufwene 1996b, 2005a). Some of the members of this spectrum have been disfranchised as indigenized varieties and some others are creoles. The reasons are not evidently linguistic (DeGraff 2003, 2005; Mufwene 2001, 2005a, to appear). The traditional way in which indigenized Englishes have been opposed to native Englishes is not particularly informative about differential language evolution. The fact that the native varieties are spoken mostly by native speakers does not change the fact that they too are outcomes of contact among dialects of English and with other languages of the Western European stock in conditions of relative absorption of the then non-English-speaking populations by the native-speaking populations. This particular sociolinguistic evolution is a function of how the socioeconomic structures of settlement colonies have evolved, imposing one economic model in which everybody must operate to earn their living and only one language to compete, within the system, for job opportunities and for social acceptance. The way the term native has been used to distinguish language varieties rather than idiolects is misleading. It suggests that non-native speakers have contributed more in kind to the restructuring of indigenized Englishes than to that of native Englishes. It overlooks the fact that English creoles, which are also native vernaculars, can very well be considered as nonstandard English varieties in plantation settlement colonies, just like Old Amish English, for example. To the extent that non-native speakers played an important role in the divergence of creoles from other nonstandard vernaculars that evolved concurrently in the colonies, the distinction between native and non-native Englishes seems to be related to whether the non-European appropriators of the colonial language constituted the overwhelming majority during the critical stage of its transformation. Sociologically, indigenized Englishes are thus at the stage where the emergent creoles were before they were spoken predominantly by native speakers. However, this particular demographic consideration highlights only an ecological factor that bears differentially on language evolution. It does not lead to the conclusion that the actual mechanisms of language restructuring were different from one kind of setting to another. I also observed earlier that the reason that most North American White English varieties are believed not to be as divergent from metropolitan varieties as AAVE and Gullah are has to do with how late continental European immigrants shifted to English. To maintain the traditional opposition between, on the one hand, native Englishes and, on the other hand, indigenized and creole Englishes, one may want to argue that the proportion of native speakers should maintain a substantial critical mass during the formative stages of a new variety. That is, the indigenized and creole varieties that emerged in exploitation and plantation settlement colonies are not considered (fully) English because they evolved in settings where the native speakers failed to maintain this critical mass. However, what has this ecological factor got to do with the way in which a language changes, as
< previous page
page_295
next page >
< previous page
page_296
next page >
Page 296 opposed to the extent to which it has diverged from the model targeted by its appropriators? The traditional practice is not without precedents in the scholarship on the history of Indo-European languages. On the one hand, the evolution of Latin into the Romance languages during its appropriation by the Celts, who adopted it as their vernacular after the Romans had abandoned their Western European colonies, produced new languages considered mutations from Latin. On the other hand, Old English (a local development from continental Germanic languages brought to England by Germanic colonists) evolved only into Middle English and later Modern English while it was appropriated by the insular Celts. That is, it has putatively remained the same language, although the transformations are apparently as significant as those that produced the Romance languages. An important reason for this different interpretation of the evolutionary processes seems to lie in the fact that (descendants of) the Germanic colonists never left England and maintained a critical mass among the Anglicizing Celts. On the other hand, the continental Celts adopted a vernacular whose original speakers had left and could hardly control its fate among them. Thus, the identification of English creoles as new languages seems to be analogous to that of the Romance languages. What seems strange, however, is that English creoles have been claimed not to be genetically related to English. The argument appears to be the same for indigenized Englishes, as they have evolved in contact settings in which the appropriators of the colonial language were not integrated with the European colonists. They have consolidated mostly after the independence of the different nations and the adoption of English as the official language and the lingua franca of the elite. Kachru and Nelson (2006) may be right in discussing AAVE in their book since the latter too has diverged from American White Southern English mostly over the past century because of the passage of the Jim Crow laws. As an ecological factor, segregation also accounts for the emergence of varieties, such as Italian and German Englishes, which would disappear gradually only after the Euro-Americans became relatively more integrated economically first and then socially. European populations such as the Old Amish, who remain economically and socially isolated, maintain their distinctly divergent varieties to date. Interestingly, the factor also accounts for why Gullah, on the coast of South Carolina and Georgia, emerged earlier, as a divergent variety, than AAVE. European and African-American populations in coastal South Carolina were segregated much earlier, in 1720 (Wood 1974), when the overall African population in South Carolina was double that of the European populations and the disproportion was much higher on the coastal plantations (9 to 1, according to Turner 1949). Although the race of speakers is a factor favoring the current distinction between creole and non-creole vernaculars, there is nothing in the whole scenario that suggests that English underwent different processes in restructuring into Gullah than into AAVE,
< previous page
page_296
next page >
< previous page
page_297
next page >
Page 297 only differences in the ecologies of language contact account for variation in the outcomes of the changes. There is nothing in all this that would make it a mistake to conclude that Gullah and other English creoles are also new natural offspring of English. We can certainly conclude this chapter with the wording of its title: “Some offspring of colonial English are creole”. Approaching the evolution of English as I have so far in this chapter, we can learn more about the variable ways in which the ecologies of social interactions affected the outcomes of the evolution of English in different colonies. However, we will not learn much by assuming a priori that creoles developed in their own unique, nonordinary ways and are not dialects nor some other kinds of offspring of English. Nor can we continue to assume that only creoles and indigenized Englishes were affected by contact. We should focus more on contact as a trigger of change, perhaps as a factor that could prompt us to reexamine the traditional distinction between internally and externally motivated change which underlies arguments for the genetic distinction between native English, indigenized Englishes, and English creoles (Mufwene 2001, 2005a; Pargman 2002). I have said almost nothing about (expanded) pidgins, but it is easy to extend the above discussion to them. Like creoles, they have been determined by an oral mode of ‘transmission’ and they have been lexified by nonstandard varieties. An important difference is that, like the indigenized Englishes, they have functioned more as lingua francas than as vernaculars. It is especially in dense urban communities in which the new speakers have interacted regularly across traditional ethnolinguistic lines that the new varieties have acquired vernacular functions. These particular ecologies account for the expansion of their systems from the original pidgins, for more focusing (in the sense of Le Page and Tabouret-Keller 1985), and thus for the emergence of their own separate norms (Chaudenson 1992, 2001). They have become indistinguishable from creoles with respect to the complexity of their structures, and, like creoles and indigenized Englishes, they show influence from the substrate languages. Whatever typology we may need to understand the evolution of English and its speciation into diverse varieties—some of which are not mutually intelligible—must involve more dimensions than what Thomason and Kaufman (1988) cover in their discussion of outcomes of language contact, certainly many more than the traditional approach to the history of English has suggested. They all seem to have to do with variation in the ecologies of language contact, including the variety to which the new appropriators of the language were exposed, the modes of transmission, patterns of social interaction with the original population of native speakers, and the settings in which they continue to be spoken. NOTES 1. Kachru and Nelson (2006:9) describe this as the perception by linguists that English has spread into basically “two diasporas”, with the first including
< previous page
page_297
next page >
< previous page
page_298
next page >
Page 298 “Great Britain, Australia, North America, and New Zealand”, whereas the second applies to former British exploitation colonies in Africa and Asia, where it is spoken by minorities of national populations. 2. In this respect, indigenized Englishes have fallen between the cracks because they are not (entirely) nonstandard and do not typically function as vernaculars. What makes them curious to Westerners is that they have structural features that mark them off from the native varieties. 3. The indigenization process seems to have been conceived of in the same spirit as the putative process of creolization, except that it has received less attention among students of the history of English who do not work on indigenized varieties. 4. This position happens to be a myth, which is being forcefully debunked by the growing literature on Celtic Englishes (e.g. Filppula, Klemola, and Pitkänen 2002; Tristram 2006). This shows that the Celtic languages have made a more significant contribution to the history of English, since its early stages, than has traditionally been recognized. Mufwene’s (2001) conjecture that the Germanic colonists in England drove the Celts westwards in the same way the European colonists in North America drove Native Americans westwards and into reservations should be interpreted just as an interesting research lead. 5. Preposition stranding is incidentally the only option selected into creoles, which were of course lexified by nonstandard English vernaculars. 6. According to Bailyn (1986) and Fischer (1989), the founder population of the New England colony originated predominantly in East Anglia, a peculiarity that may have subjected their dialect to less restructuring, especially if one agrees that continental Europeans shifted to English too late to impact its evolution in a way that would be consistent with the fact that they are now more numerous in North America than the descendants of the English. 7. This is not to say that the African and European laborers were considered legally equal. Discrimination based on race, which is different from segregation, existed. It was the basis for the enslavement of Black Africans after a brief, initial colonial period (nearly half a century) during which they were treated as indentured servants, at least in Virginia (Tate 1965). With the exception of coastal South Carolina, where the Africans quickly became the overwhelming majority (by 1720; see Wood 1974), racial segregation took much longer before being institutionalized in the American South, with the Jim Crow laws first passed in 1877. (See also Schneider 1995 for similar considerations on the impact of late segregation on language evolution in the American South.) 8. Bailey and Thomas (1998) and Wolfram and Thomas (2002), who also doubt the creole origins of AAVE, think that AAVE and American White Southern English varieties started diverging in the late nineteenth century. Since 1988, Poplack and Tagliamonte (1991, 1994) have jointly highlighted several structural similarities between White nonstandard vernaculars in North America and African-Diaspora nonstandard, non-creole vernaculars, suggesting that they all developed in similar ways, selecting similar features and similar constraints, although minor differences obtain naturally here and there. Such findings have typically been characterized in the literature as quantitative rather than qualitative. Bailey (1989) argues that some of the changes (although perhaps not many of them) which have happened since the late nineteenth century reveal systemic, not just quantitative, differences. See also Kautzsch (2002). 9. As explained in Mufwene (1996a, 2001, 2005a, 2005b), substrate influence may be interpreted as the role played by the substrate languages in determining what particular structural options would be selected from among the
< previous page
page_298
next page >
< previous page
page_299
next page >
Page 299 competing variants in the lexifier. As in the case of the associative plural, it could have regularized a construction that had been statistically marginal in the lexifier. 10. Post-emancipation population structures have in fact evolved in opposite directions in the Caribbean and North America. Being a minority in the economically stratified Caribbean and governed from the United Kingdom, Caribbean poor Whites allied themselves with the Black majority, although the emergent upper socioeconomic class would consist predominantly of Whites. Now politically and economically independent from the United Kingdom, Whites in the United States were also the general majority, and the more affluent of them imposed alone the current world order. Having emerged to promote the interests of the poor Whites, especially in the South, the White supremacist movement helped pass the Jim Crow laws in the late nineteenth century, promoting the segregation of races and discrimination against non-Whites, particularly against descendants of Africans. This fostered the divergence of AAVE from the Southern White nonstandard vernacular, with which it shares origins. (See also Schneider 1995 on the subject matter.) 11. Studies such as by Poplack and Tagliamonte suggest that much of present-day AAVE must have stabilized already by the nineteenth century, although, as pointed our by Bailey and Maynor (1989) and Bailey and Thomas (1998), a few features (phonological and grammatical) have changed since the last quarter of the nineteenth century, consistent with their version of the divergence hypothesis of AAVE and White American nonstandard vernaculars. 12. Labov (1972) states this view and Labov (1998) reiterates it, invoking support from Baugh (1983). There is definitely no significant regional variation in structures, especially from one city to another. After all, during the Great Migration, African Americans, who had been predominantly rural by the end of the nineteenth century, moved largely to cities. Coming from typically the same southern regions, they relocated according to the same urban plan, marked by residential segregation. Their vernaculars would be restructured similarly by some sort of koinéization. Nonetheless, native speakers report of differences not only between urban and rural varieties, but also between regional ones. African Americans in Chicago can normally tell when they are speaking with an African American from New York or from the South. Charles DeBose (p.c., January 1995) agreed with me that African Americans in New Orleans sound different from those in the Midwest. Regardless of whether these differences are only or primarily prosodic deserves investigation, which Troike (1973) wished had been undertaken already in the 1970s. 13. This is still new territory to me. I base my considerations here tentatively on some of the contributions to de Klerk (1996). 14. Here again one must remember that some, if not most, of the teachers are non-native speakers using norms that often differ from those of British and North American Englishes, as we are reminded by, for instance, Kachru (1983) and Gupta (1991). 15. Things have been changing in interesting ways since the change of political regimes in 1994. More and more affluent South Africans of all races have been raising their children as native English speakers, although the overall population is still far from integrated. REFERENCES Bailey, C.-J.N., and K. Maroldt. 1977. The French lineage of English. In Pidgins – Creoles–Languages in Contact , edited by J. M. Meisel, 21–53. Tübingen: Narr.
< previous page
page_299
next page >
< previous page
page_300
next page >
Page 300 Bailey, G. 1997. When did Southern American English begin? In Schneider, ed. 1997. 255–75. Bailey, G., and N. Maynor. 1987. Decreolization? Language in Society 16: 449–73. ———. 1989. The divergence controversy. American Speech 64: 12–39. Bailey, G., and E. Thomas. 1998. Some aspects of African-American vernacular English phonology. In S.S. Mufwene et al. 1998. 85–109. Bailyn, B. 1986. The Peopling of British North America: An Introduction . New York: Random House. Bambgose, A. 1992. Standard English in Nigeria: Issues of identification. In The Other Tongues: English Across Cultures , edited by B. Kachru, 140–61. Urbana: University of Illinois Press. ———. 1998. Torn between the norms: Innovations in world Englishes. World Englishes 17: 1–14. Baugh, J. 1983. Black Street Speech: Its History, Structure, and Survival . Austin: University of Texas Press. Buccini, A.F. 1995. The dialectal origins of New Netherland Dutch. In The Berkeley Conference on Dutch Linguistics 1993, edited by T. F. Shannon and J. Snapper, 211–63. Lanham, MD: University Press of America. Chaudenson, R. 1992. Des îles, des hommes, des langues: essais sur la créolisation linguistique et culturelle . Paris: L’Harmattan. ———. 2001. Creolization of Language and Culture . London: Routledge. Clarke, S. 1997. On establishing historical relationships between New and Old World varieties: Habitual aspect and Newfoundland vernacular English. In Schneider, ed. 1997. 277–93. Crystal, D. 1995. The Cambridge Encyclopedia of the English Language. Cambridge: Cambridge University Press. DeGraff, M. 2003. Against creole exceptionalism. Discussion note. Language 79: 391–410. ———. 2005. Linguists’ most dangerous myth: The fallacy of creole exceptionalism. Language in Society 34: 533–91. de Klerk, V., ed. 1996. Focus on South Africa. Amsterdam: John Benjamins. Filppula, M., J. Klemola, and H. Pitkänen, eds. 2002. The Celtic Roots of English. Joensuu: University of Joensuu. Fischer, D. H. 1989. Albion’s Seed: Four British Folkways in America. New York and Oxford: Oxford University Press. Fisiak, J. 1995. Linguistic Change under Contact Conditions. Berlin: Mouton de Gruyter. Gupta, A.F. 1991. Almost a creole: Singapore colloquial English. California Linguistic Notes 23: 9–21. Hickey, R. 2005. Dublin English: Evolution and Change. Amsterdam: John Benjamins. Hirsch, A.R. 1991. Black ghettos. In The Reader’s Companion to American History, edited by E. Foner and J. A. Garraty, 109–13. Boston: Houghton Mifflin. Hock, H.H., and B.D. Joseph. 1996. Language History, Language Change, and Language Relationship. Berlin: Mouton de Gruyter. Hogg, R.M., ed. 1992. The Cambridge History of the English Language, Vol. 1: The Beginnings to 1066. Cambridge: Cambridge University Press. Holm, J. 1988. Pidgins and Creoles, Vol. 1: Theory and Structure. Cambridge: Cambridge University Press. ———. 2004. Languages in Contact: The Partial Restructuring of Vernaculars. Cambridge: Cambridge University Press. Joseph, B.D., D. Destefano, N.G. Jacobs, and I. Lehiste, eds. 2003. When Languages Collide: Perspectives on Language Conflict, Language Competition, and Language Coexistence. Columbus: Ohio State University Press.
< previous page
page_300
next page >
< previous page
page_301
next page >
Page 301 Kachru, B. 1983. South Asian English. In English as a World Language, edited by R. W. Bailey and M. Görlach, 353–83. Ann Arbor: University of Michigan Press. ———. 1996. English in South Asia. In The Cambridge History of the English Language, Vol. 5: English in Britain and Overseas: Origins and Development , edited by R. Burchfield, 497–553. Cambridge: Cambridge University Press. ———. 2005. Asian Englishes: Beyond the Canon. Hong Kong: Hong Kong University Press. Kachru, Y., and C.L. Nelson. 2006. World Englishes in Asian Contexts. Hong Kong: Hong Kong University Press. Kallen, J., ed. 1997. Focus on Ireland. Amsterdam: John Benjamins. Kautzsch, A. 2002. The Historical Evolution of Earlier African American English: An Empirical Comparison of Early Sources . Berlin: Mouton de Gruyter. Keiser, S.H. 2003. Pennsylvania German and the “Lunch Trail Treat”: Language shift and cultural maintenance in two Amish communities. In B. D. Joseph et al., eds. 2003. 3–20. Labov, W. 1972. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. ———. 1998. Co-existent systems in African-American vernacular English. In Mufwene et al., eds. 1998. 110–53. ———. 2001. Principles of Linguistic Change: Social Factors. Malden, MA: Blackwell. ———, and W.A. Harris. 1986. De facto segregation of Black and White vernaculars. In Diversity and Diachrony , edited by D. Sankoff, 1–24. Amsterdam: John Benjamins. Le Page, R., and A. Tabouret-Keller. 1985. Acts of Identity: Creole-Based Approaches to Language and Identity . Cambridge: Cambridge University Press. Lüdtke, H. 1995. On the origin and Modern and Middle English. In Fisiak, ed. 1995. 51–53. Mufwene, S.S. 1988. Why study pidgins and creoles? Column. Journal of Pidgin and Creole Languages 3: 265–76. ———. 1996a. The Founder Principle in creole genesis. Diachronica 13: 83–134. ———. 1996b. The development of American Englishes: Some questions from a creole genesis perspective. In Varieties of English around the World: Focus on the USA , edited by E. W. Schneider, 231–64. Amsterdam: John Benjamins. ———. 1998. What research on creole genesis can contribute to historical linguistics. In Historical Linguistics 1997, edited by M. Schmid, J. Austin, and D. Stein, 315–38. Amsterdam: John Benjamins. ———. 1999. Some sociohistorical inferences about the development of African-American English. In The English History of African American English, edited by S. Poplack, 233–63. Oxford: Blackwell. ———. 2000. Population contacts and the evolution of English. The European English Messenger 9: 9–15. ———. 2001. The Ecology of Language Evolution . Cambridge: Cambridge University Press. ———. 2002. Competition and selection in language evolution. Selection 3: 45–56. ———. 2003. Genetic linguistics and genetic creolistics. Short note. Journal of Pidgin and Creole Languages 18: 273–88. ———. 2005a. Créoles, écologie sociale, évolution linguistique. Paris: L’Harmattan. ———. 2005b. Language evolution: The population genetics way. In Gene, Sprachen, und ihre Evolution , edited by G. Hauska, 30–52. Universitaetsverlag Regensburg. ———. 2007. Population movements and contacts: Competition, selection, and language evolution. Journal of Language Contact 1: 63-91.
< previous page
page_301
next page >
< previous page
page_302
next page >
Page 302 ———. To appear. Race, racialism, and the study of language evolution in America. In Language Variety in the South: Historical and Contemporary Perspectives , edited by M. Picone and K. Davis. Tuscaloosa: University of Alabama Press. ———, and Pargman, S. 2003. Competition and selection in the development of American Englishes. World Englishes 22: 201–22. Mufwene, S. S., J. R. Rickford, G. Bailey, and J. Baugh, eds. 1998. African-American English: Structure, History, and Use. London: Routledge. Mühlhäusler, P. 1985. The number of pidgin Englishes in the Pacific. Papers in Pidgin and Creole Linguistics, No.1 Pacific Linguistics, A-72: 25–51. Odlin, T. 2003. Language ecology and the Columbian exchange. In B. D. Joseph et al., eds. 2003. 71–94. Pargman, S. 2002. Internal and External Factors in Language Change . PhD dissertation, The University of Chicago. Poplack, S., and S. Tagliamonte. 1991. African American English in the Diaspora: Evidence from old line Nova Scotians. Language Variation and Change 3: 301–39. ———. 1994. -S or nothing: Marking the plural in the African American Diaspora. American Speech 69: 227– 59. Romaine, S. 1994. Language in Society: An Introduction to Sociolinguistics . Oxford: Oxford University Press. Schneider, E. W. 1990. The cline of creoleness in English-oriented creoles and semi-creoles of the Caribbean. English World-Wide 11: 79–113. ———. 1995. Verbal -s inflection in “early” American Black English. In Fisiak, ed. 1995. 315–26. ———, ed. 1997. Englishes Around the World 1. General studies, British Isles, North America: Studies in Honor of Manfred Görlach. Amsterdam: John Benjamins. ———. 2007. Post-Colonial Englishes: The Dynamics of Language Diffusion. Cambridge: Cambridge University Press. Shackleton, R. 2005. English-American speech relationships: A quantitative approach. Journal of English Linguistics 33: 99–160. Smith, J., and S. Tagliamonte. 1998. We were all thegither … I think we was all thegither: Was regularisation in Buckie English. World Englishes 17: 105–26. Tagliamonte, S. 1996. Has it ever been “perfect”: Uncovering the grammar of Early Black English. York Papers in Linguistics 17: 351–96. Tagliamonte, S., and S. Poplack. 1988. Tense and aspect in Samaná English. Language in Society 17: 513– 33. ———. 1993. The zero-marked verb: Testing the creole hypothesis. Journal of Pidgin and Creole Languages 8: 171–206. Tate, T.W. 1965. The Negro in Eighteenth-century Williamsburg . Williamsburg, VA: The Colonial Williamsburg Foundation. Thomason, S.G. 2001. Language Contact: An Introduction . Washington, DC: Georgetown University Press. ———. 2002. Creoles and genetic relationship. Column. Journal of Pidgin and Creole Languages 17: 101–9. ———, and Kaufman, T. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Tristram, H. L. C., ed. 2006. Celtic Englishes IV: The Interface between English and the Celtic Languages. Potsdam: Potsdam University Press. Troike, R. 1973. On social, regional, and age variation in Black English. The Florida Foreign Language Reporter, Spring/Fall, 7–8. Trudgill, P. 1986. Dialects in Contact . New York: Basil Blackwell. ———. 2004. New-dialect Formation: The Inevitability of Colonial Englishes . Oxford: Oxford University Press. Turner, L. D. 1949. Africanisms in the Gullah Dialect. Chicago: University of Chicago Press.
< previous page
page_302
next page >
< previous page
page_303
next page >
Page 303 Wolfram, W., D. Christian, and N. Dube. 1988. Variation and Change in Geographically Isolated Communities: Appalachian English and Ozark English. Tuscaloosa: University of Alabama Press. Wolfram, W., and E. R. Thomas. 2002. The Development of African American English. Malden, MA: Blackwell. Wood, P.H. 1974. Black Majority: Negroes in Colonial South Carolina From 1670 Through the Stono Rebellion . New York: Norton.
< previous page
page_303
next page >
< previous page
page_304
next page >
Page 304 13 Vernacular Universals and the Sociolinguistic Typology of English Dialects Peter Trudgill 1. INTRODUCTION It is clear that recent typological work on varieties of English by Chambers, beginning with Chambers (2001), has stimulated some important and valuable research. His suggestion that different types of variety of English might exhibit different types of linguistic characteristics has been shown to be correct, and we now have a solid picture of what a number of these characteristics are and their origins. For example, although Trudgill and Hannah (1982) pointed out informally that varieties of English with some degree of English-as-a-second-language (ESL) input seemed to have some features in common (such as undifferentiated tag questions, such as You’re leaving, is it ?, and loss of the mass/count distinction, e.g. chalks, informations), we now have a structured and fuller, richer, and more differentiated understanding of such phenomena. Since Chambers (2001), it has become clear from work such as Kortmann and Schneider (2004) and Schneider and Kortmann (2004) that a taxonomy of varieties of English based on their linguistic characteristics has to include, amongst others, a number of significantly different typological categories. These include—to make an important distinction not made by Trudgill and Hannah— second-language varieties and shift varieties (Mesthrie 1992) of English. The former include varieties such as Indian and Nigerian English, which for the most part do not have native speakers, but which are spoken and written in polities where they have a very well-established and important role. The latter are varieties which have a history of being second-language varieties, but which are now, as a result of language shift, only or mainly native-speaker varieties. This category includes Welsh English and Irish English. However, in this chapter, I want to argue that one particular attempt to find a typological distinction between types of English, also inspired by Chambers’ work, has not been successful. This is the search for vernacular universals of English. I suggest that this attempt to establish a typological distinction between vernacular and nonvernacular varieties, by showing that vernacular varieties have a number of features in common which are
< previous page
page_304
next page >
< previous page
page_305
next page >
Page 305 absent from nonvernaculars, has ultimately been in vain. The truth is that we have not found any vernacular universals—or at least not enough to make the concept fruitful. I also argue that the reason that we have not found any vernacular universals is because we have been looking in the wrong place. The term vernacular has been used in many different ways in sociolinguistic work, and this has on occasion led to confusion. One obvious confusion is between vernacular meaning ‘informal/unmonitored’ (i.e. where the term is applied to varieties which are actually different styles) and vernacular meaning ‘non-standard’ (i.e. where the term is applied to dialects/sociolects) (see Trudgill 1999a). In this chapter, I use the term vernacular in its dialect/sociolect sense, as Chambers for the most part does also. I argue that we have not found what we have been looking for because we have been seeking generalisations about ‘vernacular’, i.e. nonstandard varieties of English, as opposed to Standard English, where the vernaculars have been conceived of as consisting of the koineised nonstandard colonial varieties of North America, Australasia, and South Africa, together with the koineised nonstandard urban varieties of English in the British Isles. We have been concentrating on a typological distinction between Standard English and such ‘vernaculars’, when in fact these two types of variety lie along a typological continuum of native-speaker varieties. I argue that the reason we have found no vernacular universals is that Standard English and the vernaculars, so conceived, are not typologically significantly different, and that the important typological distinction lies elsewhere. 2. NONVERNACULAR AND/OR NONUNIVERSAL? The fact is, I suggest, that the proposed vernacular universals do not, in the end, stand up to the closer examination that the several years since 2001 have permitted. Let us consider the list of vernacular universals of English proposed by Chambers (2004): 2.1 Phonological 1. (ng) or Alveolar Substitution, as in walkin’ It is true that this phenomenon occurs in all vernacular varieties, but the problem is that it also occurs in informal styles of nonvernaculars—the various forms of pronunciation associated with Standard English (see Wells 1982:262). ‘Substitution’ may also not be the appropriate term here from a diachronic perspective because this feature mostly derives from a merger of Old English gerund and participle forms (Wells 1982:262).
< previous page
page_305
next page >
< previous page
page_306
next page >
Page 306 2. Morpheme-Final CC Simplification The examples given by Chambers— pos’office, han’ful —collapse two different phenomena. Absence of /d/ in handful is usual for all native English speakers, including Standard English speakers, except in the most careful of styles, as is morpheme-final CCC > CC simplification generally. What does seem to be vernacular is CCV > CV simplification, as in pos’office, but it is not clear how widespread and/or vernacular this is— although Schneider (2004:1126) gives a brief summary. Wells (1982) cites only AAVE, Newfoundland, and the U.S. South vernacular varieties as having this feature. 3. Final Obstruent Devoicing, as in hunret , cubbert This does not appear to be sufficiently widespread in accents of English around the world to qualify. The two examples that Chambers (2004) cites both involve word-final unstressed syllables ending in /d/. It is true that the Traditional-Dialect of northern Norfolk (Trudgill 2003) has this pattern, but the number of relevant lexical items is rather small: wicked , naked , David, and hundred (past-tense verb forms are not affected). The number of accents in which even this limited form of final /d/ devoicing occurs does not seem to be large enough to qualify. It is not mentioned by Wells (1982). 2.2 Grammatical 4. Conjugation Regularization, as in John seen it This is a phenomenon which has certainly occurred in all vernacular varieties of English. But the fact that it has also occurred in Standard English, although to a lesser extent (see below), and the fact that vernacular varieties have not undergone total regularisation, make it difficult to argue persuasively in favour of this feature as a vernacular universal. 5. Default Singulars as in we was The fact that very large numbers of vernaculars (in large areas of the north of England, for instance; see Beal 2004:122) have regularised in favour of were rather than was , as in I were , he were , suggests that the correct generalisation here is not ‘default singular’, but simply regularisation, especially because the regularisation also applies to present-tense be in both the positive and negative (e.g. I be, he be; I ain’t , he ain’t ) (see below). This would make point 5 into a special case of 4. 5a. There Was + Plural This is cited by Chambers as coming into the same category, but seems best analysed as a separate variable to be distinguished from 4 and 5 because it is
< previous page
page_306
next page >
< previous page
page_307
next page >
Page 307 entirely paralleled in, and subsidiary to, the present-tense usages of there is and there’s , as well as where’s as in where’s my keys ? What we appear to be witnessing here is a development triggered initially by the grammaticalisation of there’s as a single item paralleled by, for example, French il y a. Another reason for distinguishing it from 4 and 5 is that, although there’s + plural is certainly ‘universal’, it is hardly vernacular because it also occurs in the speech of Standard English speakers, as indicated by Chambers. 6. Multiple Negation This is certainly only vernacular, but it is difficult to argue for it as a vernacular universal when in fact it is confined to the vernaculars simply because it has been lost in Standard English—because of a linguistic change which took place in (pre-)Standard English. It is rather the absence of multiple negation which is a feature of non vernaculars in English (see below). 7. Copula Absence This is not widespread enough to qualify (see Kortmann and Szmrecsanyi 2004:1154) and is especially associated with a particular typological category of English varieties, namely creoles. Other possible candidates for the label of vernacular universal might be two other features mentioned in Trudgill (1999a, see below)—the reintroduction of plural forms of you and reflexive pronoun regularisation— but we would need more evidence before we could assert this with any confidence. 3. PECULIARITIES OF STANDARD ENGLISH The question then is: why might this interesting and intriguing idea, in the end, have proved a blind alley? Here is one possibility. Trudgill (1999a) points out that, compared with all other dialects of English, Standard English is somewhat peculiar. The paper cites a list of the grammatical idiosyncrasies of Standard English—features which set it apart from all or many other dialects. Interestingly, Hope (2000), in a paper that must have been written before Trudgill (1999a) was published, also points to certain unusual features of Standard English. It is therefore very interesting to see that Hope and I agreed to a considerable extent about what these features were. Also interestingly, Chambers (2001) introduced the term ‘vernacular universals’ to refer to linguistic features which are absent from Standard English, but which recur in many different nonstandard varieties of English around the world. Following on from this and other earlier writings, Chambers (2004) cites amongst his ‘best candidates’ for the label of grammatical vernacular universals some of the features mentioned by Hope and/or by me.
< previous page
page_307
next page >
< previous page
page_308
next page >
Page 308 Here are those features cited by Trudgill and also listed by Hope and/or Chambers: 1. Third-Person Marking “Standard English has an unusual and irregular present-tense verb morphology in that only the third-person singular receives morphological marking: he goes versus I go. Many other dialects use either zero for all persons or - s for all persons” (Trudgill 1999a: 124). Hope (2000:52) cites as an oddity of Standard English the “retention of -s inflection in the third-person present tense, typologically the least likely slot to be morphologically marked”. 2. Multiple Negation, as cited above “Standard English lacks multiple negation, so that no choice is available between I don’t want none, which is not possible, and I don’t want any. Most nonstandard dialects of English around the world permit multiple negation” (Trudgill 1999a). Hope (2000:53) cites as an oddity of Standard English its “failure to allow double negatives”. Chambers (2004:129) also cites as a putative vernacular English universal “multiple negation, or negative concord, as in He didn’t see nothing”. 3. Second-Person Pronouns, as cited above “Standard English fails to distinguish between second-person singular and second-person plural pronouns, having you in both cases. Many nonstandard dialects maintain the older English distinction between thou and you, or have developed newer distinctions such as you versus youse” (Trudgill 1999a). Hope (2000:52) cites as an oddity of Standard English “the typologically unusual five-person pronoun system”. 4. Forms of be “Standard English has irregular forms of the verb to be both in the present tense (am, is, are ) and in the past (was, were ). Many nonstandard dialects have the same form for all persons, such as I be, you be, he be, we be, they be, and I were, you were, he were, we were, they were ” (Trudgill 1999a). Chambers (2004:129) lists as a possible vernacular universal a feature which he calls, as cited above, “default singulars, or subject–verb nonconcord, as in They was the last ones”. 5. Irregular Verbs “In the case of many irregular verbs, Standard English redundantly distinguishes between preterite and perfect verb forms both by the use of the
< previous page
page_308
next page >
< previous page
page_309
next page >
Page 309 auxiliary have and by the use of distinct preterite and past participle forms: I have seen versus I saw . Most other dialects have I have seen versus I seen ” (Trudgill 1999a). Chambers (2004:129) also lists as a vernacular universal, as cited above, “conjugation regularization, or the levelling of irregular verb forms, as in Yesterday John seen the eclipse and Mary heared the good news”. This would all seem to be a case of great minds thinking alike and at the same time, but it is apparent that there is a difference of perspectives here—and this may be where the problem lies. For Hope and me, it was not the presence of parallels between the different nonstandard varieties that was significant. Rather, it was the oddness of Standard English compared with all other varieties which was of importance. That is, we did not regard the features we cited so much as possible universals of vernacular English; rather, we saw them as peculiarities of nonvernacular (i.e. Standard English). These are features where Standard English is the odd one out—and it is worth remembering that, amongst the dialects of English spoken by native speakers, Standard English is very much a minority dialect. In the United Kingdom, for example, only about 12 percent of the population are native speakers of Standard English (Trudgill 1999a). My suggestion is thus that, for at least certain features of the English language that are found in nonstandard varieties but not in Standard English, this is perhaps best regarded as being due to unusual or idiosyncratic developments in the minority Standard English, not to some special qualities of the majority vernaculars. This is not just playing with nomenclature. This perspective means that we do not have to explain the presence of these particular features in nonstandard English. For Hope and me, it is rather the peculiarities of Standard English that have to be examined and analysed. In fact, this analysis is at a superficial level rather banal and easily arrived at. The main and really rather obvious explanation for the presence of idiosyncrasies in Standard English has quite simply got to do with the fact that Standard English is standardised—that it is a standard. 4. STANDARDISATION The link with standardisation is the following. According to Milroy (2000:14), “standardisation inhibits linguistic change”. Milroy says indeed that the aim of standardisation is to “fix and ‘embalm’ (Samuel Johnson’s term) the structural properties of the language in a uniform state and prevent all structural change” [Milroy’s italics]. This is especially true of written standards, so that a growing gap between written and spoken language is a well-known phenomenon—one could cite modern Czech or modern French as good examples. We must therefore expect standards to be in many respects more conservative than nonstandard varieties.
< previous page
page_309
next page >
< previous page
page_310
next page >
Page 310 Now, most of the features mentioned earlier as cited by Hope, Chambers, and/or myself can readily be accounted for along these lines by noting that the nonstandard dialects are more regular than the standard. This can be easily illustrated by means of another nonstandard feature listed in Trudgill (1999a), but not mentioned by either Hope or Chambers—namely, reflexive pronouns. Standard English has an irregular formation of reflexive pronouns, with some forms being based on the possessive pronouns—myself , yourself, ourselves, and yourselves—and others being based on the objective pronouns— himself and themselves. Nearly all nonstandard dialects have a regular system employing possessive forms throughout (e.g. hisself , theirselves). Then we can recall that regularisation of grammatical irregularities over time is a very well-known feature of linguistic change. In fact, we can say that is a diachronic universal. It is well known in historical linguistics that, although regular sound change leads to grammatical irregularity, it is then very often the case that these irregularities will at some later stage—possibly a much later stage—be regularised through analogy: “the drastic effects which sound change can have on the morphology of a language are often alleviated through analogy” (Murray 1997:335). We can therefore suppose that, with respect to some of the features just cited, Standard English is somewhat more irregular than the nonstandard dialects simply as a consequence of the fact that this particular very usual type of linguistic change has not taken place in the standard in certain cases because of its conservatism. 5. FAILURE TO REGULARISE IN STANDARD ENGLISH The following features represent instances of greater regularity (i.e. greater conservatism) in nonstandard than Standard English: 1. Third-Person Marking As noted earlier, Standard English is irregular in that it treats the third-person singular of present-tense verbs differently from the other five persons. The nonstandard dialects, in contrast, are perfectly regular in that they have undergone innovations which have the consequence that these dialects treat all persons of the verb the same. 2. Forms of be Similarly, Standard English has irregular forms of the verb to be both in the present tense (am , is, are ) and in the past (was , were ). Most nonstandard dialects have regularised the past-tense irregularity by having the same form for all persons and both numbers. As Chambers (2004:131) points out,
< previous page
page_310
next page >
< previous page
page_311
next page >
Page 311 his term ‘default singulars’ highlights only one form of this regularisation, although he mentions only one possibility out of the other two. Different nonstandard dialects either have were in all persons and both numbers—we were , he were , which is extremely widespread in dialects spoken by millions of people in large areas of northern England and elsewhere (Beal 2004:122; Shorrocks 1999:168), or they have was in all persons and both numbers, as Chambers has noted—we was , he was ; or they have was in all persons and both numbers in the affirmative, but were in all persons and both numbers in the negative— I was , you was , I weren’t , and you weren’t . ‘Default singular’ does not seem to work, however, even as a term to refer to those dialects which have was throughout. Was is the plural verb form as well as the singular verb form in these dialects, and to refer to they was as singular seems to be using unfortunate Standard English-centred terminology. ‘Non-concord’ is also unfortunate in the same way. In these dialects, we cannot say that the verb does not agree with the subject—what we have to say is that, for example, the third-person plural verb form is was . Interestingly, we cannot even be at all sure that the term ‘default singular’ is historically justified either—in the sense that we do not know for sure that an original singular form was generalised to the plural. In Old English, Middle English, and Early Modern English, were was also the second-person singular form (Lass 1992, 2000). We have to bear in mind that the subjunctive, where were was the singular as well as the plural form, may well also have been involved, and that in many dialects it is not entirely clear that either was or were has been generalised especially because these verb forms are so often unstressed. The Traditional-Dialects of East Anglia, for example, had a form which was usually written wor for all persons, but in a nonrhotic dialect we cannot be absolutely sure that this is, as it were, a variant of was and not were . The modern dialect of Norwich actually has an affirmative form for all persons which is pronounced /wùz/, which might be some kind of hybrid (Trudgill 2003). The suggestion, then, is that singular versus plural is not really the point. What we are seeing here is the sorting out, the simplification, in different ways in different Germanic languages and dialects, of a morphophonological irregularity which was indeed caused by a regular sound change, and one that happened a very long time ago indeed—Verner’s Law (Prokosch 1939:221; Trudgill 2008). In English, there are also many nonstandard dialects which have totally regularised the present-tense paradigm as well by having be for all persons: I be, you be, he be, we be, and they be. In other nonstandard dialects, other forms of regularisation have occurred, with is or am occurring for all persons: I is, you am , etc. Of course, the negative form ain’t also occurs for all persons. So to pick out we was , you was , and they was as in any way special seems to overlook the big picture—the much more widespread process of regularisation of forms of BE that has taken place throughout.
< previous page
page_311
next page >
< previous page
page_312
next page >
Page 312 3. Irregular Verbs The vast majority of verbs in English have preterite and past participle forms which are identical: I loved and I have loved. For many irregular verbs, however, Standard English has distinct preterite and past participle forms: I have seen versus I saw . This is due to the 6,000-year-old feature of Indo-European, the “morphological process of vowel change referred to as ablaut” (Hock and Joseph 1996:510). Most nonstandard dialects of English have regularised this irregularity for most verbs and have I have seen versus I seen , I went versus I should’ve went , and so on. It is, then, not difficult to come up with examples of cases where Standard English has not undergone the regularisation that the nonstandard dialects have experienced. 6. DIALECT MIXTURE However, we now come to something which at first sight seems a little mysterious. Not only is regularisation a diachronic universal, it is also very much a process which is especially associated with dialect contact and language contact (Trudgill 1986). Contact is well known to lead to simplification (Mühlhäusler 1977)—and regularisation is one form of simplification. (But it is only one form of simplification, as we shall see below.) Therefore, we can suppose that the regularising changes that have occurred in nonstandard English occurred at least in part as a consequence of contact. It seems, then, that we are not dealing with the issue of vernacular universals versus contact. There is no ‘versus’ about it. Rather, what we are focussing on are contact-induced nonstandard features. If this is so, however, then we have a puzzle on our hands because Standard English and its precursors have been just as much subject to dialect contact as other dialects and probably much more so than most of them. Modern Standard English is the outcome of very considerable dialect contact indeed, and it was subject to the process of dialect mixture over a period which lasted for several centuries. This fact is rather well known, and there is considerable evidence of it. For example, Wyld (1936:50) describes the language of the Proclamation of Henry III, a text written in 1258 in the precursor of modern Standard English, as demonstrating a “characteristic blending of Midland elements with those which are typically southern”. He then writes of similar texts from the period 1300– 1325that the “Southern element is still considerable but the Midland element is larger” than before (1936:51), and he then goes on to say (1936:52), of the fourteenth-century standardised English of the time of Chaucer, that “London English had by this time settled down into a definite blending of the various dialectal elements”. This mixture, however, was in a state of flux for many generations. For example, “Chaucer’s poetry
< previous page
page_312
next page >
< previous page
page_313
next page >
Page 313 contains more purely Southern and South-eastern elements than his prose” (1936:52), and “the precise regional dialect constituents of London English were not finally fixed in their present proportion and distribution during the fifteenth century, nor indeed for some time after the beginning of the following century” (1936:97). The language of Wyclif, for example, is very Midland in origin; it has hardly any purely Southern features, but it does have some East Midland features, together with “some indication of Northern influence” (1936:59). Hope (2000:50), too, says of the London speech which became standard that “it is tempting to ask what dialects were not present in this Londonish– East Midlandish–Northenish–Southernish ‘single’ ancestor” of Standard English”. Strang (1970:165) extends Wyld’s period of flux a good deal further when she writes that “in London circles during the 16c and 17c many different usages … were jostled together”. Lass (2000:91) adds that “the modern standard has a heterogeneous dialect base”. It is also interesting to note, although this is not the concern of this chapter, that at the level of phonology, Samuels (1972:147) speaks of ‘coexisting subsystems’; and Wells writes of ‘rival pronunciations’ in the seventeenth and eighteenth centuries. This picture of long-term mixture and flux is not at all surprising when we note from Keene (2000:109) that “most adult Londoners were born outside the city” and that “in the 18th century the outsiders may have been as many as two-thirds of the total”. 7. REGULARISATION IN STANDARD ENGLISH So, if Standard English is so much the result of dialect contact, why does it not demonstrate regularisation? In actual fact, of course, Standard English has undergone very considerable regularisation. Milroy says that when it comes to standardisation, “the intention is to prevent change”, but that in fact “the effect is to inhibit it”. It would not, therefore, prevent contact-induced regularisation; it would simply slow it down relative to nonstandard varieties. Standard English is relatively more irregular than the nonstandard dialects because certain forms of regularisation have not yet taken place. This is in turn the result of the fact that standardisation has a conservative, retarding effect on linguistic change. Standard English has undergone regularisation, but it has progressed less far along the expected regularisation path. But it has progressed along it. Examples of diachronic regularisation in Standard English are very easy to come by. For example, very many formerly irregular verbs have become regular not only in modern nonstandard dialects, but in Standard English also (e.g. the preterite of help is now helped rather than the earlier irregular holp). Indeed, Standard American English, which has been exposed to considerable more dialect and language contact that Standard English English, has regularised even more verbs, such as burn, learn, and spell.
< previous page
page_313
next page >
< previous page
page_314
next page >
Page 314 8. LOSS OF REDUNDANCY IN STANDARD ENGLISH It is also the case, importantly, that regularisation is only one aspect of simplification. (Note than simplification is being used here as a technical linguistic term which has no implications with implications for the expressive possibilities or cognitive consequences of any language structures.) In addition to the regularisation of irregularities, dialect contact-induced simplification can also take the form of the loss of redundancy (Trudgill 1986). I want to argue that Standard English has undergone this form of simplification also, in some cases to a greater extent that the nonstandard dialects. Redundancy, and therefore loss of redundancy, takes two major forms. The first type of loss involves the loss of morphological categories . Sometimes this loss is compensated for by the use of more analytical structures, as in usage of prepositions instead of case. Sometimes it is just straightforward loss, as with the loss of grammatical gender or the loss of the subjunctive in English. We can speak of loss of redundancy ‘with or without repair’. The second occurs in the form of reduction in repetition of information. Repetition of information is illustrated (e.g. in grammatical agreement, where there is more than one signal that, say, a noun phrase is feminine; or in obligatory tense marking, e.g. when all verbs in a past-tense narrative are marked for past tense). Here reduction in redundancy will take the form of reduction in the number of repetitions, as in the loss of agreement. We can now note that many of the features we discussed earlier demonstrate loss of redundancy as well as regularisation. For example, for forms of be, dialects no longer distinguish redundantly between persons of the verb by both subject pronoun or noun and verb form. For irregular verbs, many dialects no longer distinguish between preterites and past participles redundantly both by the presence or absence of auxiliary have and a difference of form. Of course, the verb system of modern English generally is considerably less redundant morphologically than that of Middle English. This recognition that loss of redundancy is also involved in dialect contact-induced simplification now leads us to recognise that, in fact, in some respects, Standard English is actually more simplified than the nonstandard dialects. For example, the loss of the second-person pronoun forms thou/thee/thy/ thine represents the clear loss of a morphological category—the distinction singular/plural is no longer operative in the second person. As we saw earlier, this development occurred in very many nonstandard dialects also, but most of them have restored the distinction and thus the category. Another example is provided by the phenomenon of multiple negation. Another label for multiple negation is negative concord or negative agreement. Standard English, as is well known, has lost this form of agreement. The loss of multiple negation is often ascribed to the successful efforts of eighteenth-century prescriptivists exercising normative pressures based on peculiar ideas about logic. However, Auer and Gonzalez-Diaz (2005) cite
< previous page
page_314
next page >
< previous page
page_315
next page >
Page 315 empirical evidence from the history of English which shows that there is a “danger of overestimating the explanatory potential of prescriptivism”. For instance, in the case of the loss of double comparatives such as more nicer from Standard English, they show that prescriptivism “can only be considered a mere reinforcing factor of a process that was set in motion earlier”. So it is very relevant that Rissanen (2000) has shown that the decline of multiple negation in written English started well before the period in which prescriptivism became operative—it began at the latest in the fifteenth and sixteenth centuries (see also Nevalainen 2006; Nevalainen and Raumolin-Brunberg 2003). So it is probable that what set this process ‘in motion’ was the considerable degree of dialect contact to which pre-Standard English had been subjected. I do not, then, agree with Hope when he says, with reference to single negation in Standard English, that “as speakers make the choices that will result in standardisation, they unconsciously tend towards more complex structures, because of their sense of the prestige and difference of formal written language. Standard English would then be a ‘deliberately’ difficult language” with features “which would not survive in a ‘natural’ linguistic environment”. I see no evidence of speakers ‘choosing’ here, and surely absence of multiple negation makes for simplicity, not complexity. So in fact Standard English is not so very peculiar, and Standard English and the nonstandard dialects are not so very different. They all demonstrate the effects of simplification—both regularisation and loss of redundancy—even if simplification has taken place at different speeds and, in the case of different linguistic phenomena, in the two types of variety. We can say that, as far as producing simplification is concerned, the natural conservatism of the standard variety has been to a certain extent balanced against the greater degree of dialect contact which it has experienced. This commonality of simplification, then, may well be why we have not succeeded in finding any genuinely vernacular universals. 9. THE TRUE TYPOLOGICAL SPLIT There is a sense in which the kind of simplification—defined as regularisation, reduction in repetition of information, and reduction in morphological categories—as outlined earlier, is not especially challenging to students of language change because it is really rather predictable. There is no total consensus about adolescent and adult language-learning abilities in the literature, but the view from sociolinguistics (see e.g., Labov 1972) is that children acquire new dialects and languages more or less perfectly up to the age of about 8, whereas there is very little or no chance of them learning a language variety perfectly after the age of about 14. Lenneberg’s (1967) critical period hypothesis, or developmental threshold, shows us that simplification will occur in contact situations to the extent that adult-only second-language
< previous page
page_315
next page >
< previous page
page_316
next page >
Page 316 or dialect learning is concerned. Speech communities which have frequent contacts of this type with other societies are therefore relatively more likely to produce languages and dialects which demonstrate simplification, the most extreme cases typologically being pidgins and creoles. Given that this is so, then perhaps a more challenging question is: to what extent we can expect to find the reverse type of linguist processes in the reverse type of situation. To what extent can we expect to see the development of irregularity and redundancy in low -contact varieties? To what extent can we expect to see the development of morphological categories in low-contact varieties? I think the answer is reasonably clear. We are forced to assume that it is precisely in low-contact situations that we are most likely to find the development of such features—otherwise none of the world’s languages would have them. From the point of view of this volume, then, the most interesting point is not the difference between Standard English and the mainstream vernacular varieties. Both of them have undergone the relatively predictable changes that we have labelled simplification—the loss of special third-person marking, the loss of multiple negation, or the loss of irregular verbs. The most challenging issue lies in the difference between different types of vernacular. In Wells’ (1982:3) classification of varieties of English, he establishes a category of varieties which he calls General English. General English consists of both Standard English and the different varieties of nonstandard General English. However, opposed to General English is another category of varieties which he labels Traditional-Dialect. What we have been discussing so far are differences in degree and kind of regularisation between different varieties of high-contact General English, both standard and nonstandard. I have already discussed the history of contact in (pre-)Standard English, but it is clear that the dialects of nonstandard General English that we are most familiar with are also for the most part urban and/or colonial varieties which also have a considerable history of contact. Basically, in looking at both Standard English and nonstandard varieties of General English, we have simply been noting the different effects of the simplification that accompanies the sort of contact situations in which they have been involved. What is much more difficult to explain, and therefore perhaps more interesting, is the reverse type of linguistic change, which, I am supposing, typically occurs in low-contact situations (see Trudgill 2002). As far as English is concerned, we can explore this type of more opaque linguistic change by examining, not nonstandard varieties of General English with their widespread and in many case worldwide simplifications, but those other nonstandard varieties, the Traditional-Dialects which, as Wells says, are increasingly hard to find and which occur only in parts of England, especially those further away from London, and in southern and eastern Scotland and parts of Northern Ireland, with the addition of Newfoundland and perhaps parts of the Appalachians.
< previous page
page_316
next page >
< previous page
page_317
next page >
Page 317 The relevance of the Traditional-Dialects is that we can check to see if the relatively low-contact linguistic situations in which the Traditional-Dialects grew up really have led to the development of features which represent the reverse of the regularisation which has led to some of our General English vernacular universals—namely, expansion in irregularity, increase in repetition of information, and the growth of morphological categories. Traditional-Dialects are especially important, I would suggest, because it is the sociolinguistic situations in which these varieties developed which most closely resemble the situations in which most language varieties have developed over the past tens of thousands of years of human linguistic history and where the difficult questions lie. Here are just a small but maybe suggestive number of English Traditional-Dialect features which do not occur in the high-contact Standard English or those high-contact nonstandard mainstream varieties which seem to have most vernacular universals, and which illustrate, if only briefly, the growth of complexity which, I suggest, may depend for its genesis on low-contact linguistic environments. 9.1 Increase in Redundancy: The Growth of New Morphological Categories 1. Transitive versus Intransitive Infinitives A few Traditional-Dialects in a small area of the southwest of England saw the development of a new and fascinating marking of the difference between transitive and intransitive infinitives. Intransitive infinitives (and objectless transitives) in these dialects were marked by the word-final morpheme - y, whereas transitive infinitives were unmarked. So in Dorset we find (Trudgill 1999b: 103): (1) Can you zew up thease zeam? ‘Can you sew up this seam?’ vs. The cat vell zick an’ woulden mousy. ‘The cat fell sick and wouldn’t catch mice’. This is unparalleled anywhere else in the English-speaking world and is quite possibly unparalleled anywhere else at all. My enquiry on the LingTyp list asking for examples of other languages which have morphological marking for intransitive but not transitive infinitives received three answers, none of them producing a precise parallel. The Australian Aboriginal language Warrgamay has two sets of inflections—those that can attach to intransitive verbs and those that cannot attach to intransitive verbs—but this applies to all verb forms and not just infinitives. A similar situation applies in Ulwa, a Misumalpan language of Nicaragua. The closest we can come to it is in the
< previous page
page_317
next page >
< previous page
page_318
next page >
Page 318 Anywa language, a Nilo-Saharan language of Ethiopia and Sudan, where special marking does occur on intransitive infinitives, but only on those which are derived from transitives, which are then marked relative to their transitive counterparts: ‘weave [tr]’ vs. ‘to do weaving’ (Reh 1996:187). 2. A number of Traditional-Dialects in the southwest of England developed a distinction between two categories which, although common enough in languages of the world, is unknown in any of the General English varieties. This is the phenomenon described by Ihalainen (1976), in which there is a category distinction between habitual verb forms such as: (2) I do/did go there every day vs. I goes tomorrow/I went last week. 3. Also in the English southwest, we see the development, also unparalleled elsewhere in the Englishspeaking world, of the expression of a pronominal category difference between count and mass nouns, such that inanimate count nouns are pronominalised with he but mass nouns with it : (3) Pass the loaf—he’s over there vs. I likes this bread—it be very tasty. 9.2 Increase in Redundancy: Introduction of Repetition of Information One example of this can be found in the Traditional-Dialect of Norfolk, where double tense marking or ‘pasttense infinitives’ can be found. For example: (4) Have the fox left? No that ain’t, do Bailey would’ve let them went. ‘No it hasn’t, or Bailey would’ve let them [the hounds] go.’ (Trudgill 1995) Precisely the same phenomenon occurs in the highly isolated variety of English spoken on Tristan da Cunha, where constructions such as We didn’t want to went also occur (Schreier 2003). 9.3 Irregularisation It is not always the case that grammatical change leads to greater regularity. The reverse can be the case, and it is therefore worth noting that it appears to be much more common in Traditional-Dialects than in General varieties. In the Traditional-Dialect of Norfolk, for example, we find a number of
< previous page
page_318
next page >
< previous page
page_319
next page >
Page 319 irregular preterites which occur in cases where Standard English has regular forms. These include (Forby 1830/1970; Trudgill, 2003): hoe hew mow mew owe ewe row rew save seft sew /sou/ sew /su:/ (as of stitches) show shew snow snew sow sew /su:/ (as of seed) thow (thaw) thew shriek shruck wave weft wrap wrop This is worth some further investigation. In fact, a few of these verbs had strong or irregular preterites on their first attested appearance in Old English or Middle English. This is true of mow and row, whereas sew had both strong and weak forms in Old English. In these cases, then, regularisation has taken place in Standard English and in the General English nonstandard dialects, whereas the Traditional Norfolk Dialect has preserved the original irregularity. The other irregular forms, however, are not historical. Show, for example, was a weak verb in Old English, with the preterite sceawed, as was thaw ; and the other verbs were also regular from their earliest appearance. The only exception is owe , which goes back to Old English ägan , with the early Old English preterite ahte, which later of course gave rise to ought, and was superseded as preterite by regular awede > owed . So the Norfolk forms ewe, hew, sew (of sew), shew, snew, thew , shruck, seft, weft, and wrop are all innovations which involved irregularisation. The changes to ewe, hew, sew , shew, sew , and thew are obviously the result of analogy with the blow class, but the origins of shruck, seft, weft, and wrop are less clear. It is often said that some forms of American English, including Standard English, have replaced the regular preterite of dive , namely dived , with an irregular preterite dove—clearly by analogy with the class of irregular verbs such as ride/rode. However, this is not a case of the development of irregularity in Standard English. According to the Old English Dictionary, dove is also a form which occurs in dialectal British English, and, more importantly, the verb was a strong verb in Old English, with the past participle dofen. 10. CONCLUSION The most important outcome of the work inspired by Chambers (2001, 2003, 2004), Kortmann et al. (2004), and Schneider and Kortmann (2004)
< previous page
page_319
next page >
< previous page
page_320
next page >
Page 320 has been the establishment of a typology of English varieties on the basis of their linguistic characteristics (as opposed to their functions and/or histories). We have briefly noted earlier two of the important dimensions of typological differentiation which this research has produced: creole versus non-creole varieties, and second-language versus shift versus long-established mother tongue varieties. The main thrust of this chapter, however, has been to attempt to establish a third such dimension. We have observed that, in terms of linguistic change, we can explain simplification in high-contact dialects and languages partly in terms of imperfect learning by adults. We have also seen that, in terms of the discussion of vernacular universals versus contact, many of the proposed but failed vernacular universals of English are at least partly due to contact, and that many of the nonvernacular features of Standard English are also partly due to contact—hence, the failure. The more difficult linguistic change question is: how are we to explain developments that occur in lowcontact languages and dialects? The truly important typological dimension of differentiation we have established in this chapter, on the basis of linguistic characteristics, is not between the standard and nonstandard forms of General English, which are both characterised by (different forms and different degrees of) simplification, but between General English and the low-contact Traditional-Dialects which are more likely to produce complexification. REFERENCES Auer, A., and V. Gonzalez-Diaz. 2005. Eighteenth-century prescriptivism in English: A re-evaluation of its effects on actual language usage. Multilingua 24(4): 317–41. Beal, J. 2004. English dialects in the North of England: Morphology and syntax. In Kortmann et al., eds. 2004. Vol. 2. 114–41. Chambers, J.K. 2001. Vernacular universals. In ICLaVE 1: Proceedings of the First International Conference on Language Variation in Europe , edited by J.M. Fontana, L. McNally, M.T. Turell, and E. Vallduví. Barcelona: Universitat Pompeu Fabra. ———. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Implications . Oxford: Blackwell. ———. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective , edited by B. Kortmann, 127–45. Berlin: Mouton de Gruyter. Forby, R. [1830] 1970. The Vocabulary of East Anglia. New York: Augustus Kelley. Hock, H.H., and B. Joseph. 1996. Language History, Language Change and Language Relationship. Berlin: Mouton de Gruyter. Hope, J. 2000. Rats, bats, sparrows and dogs: biology, linguistics and the nature of Standard English. In Wright, ed. 2000. 49–56. Ihalainen, O. 1976. Periphrastic do in affirmative sentences in the dialect of east Somerset. Neuphilologische Mitteilungen 77: 608–22. Keene, D. 2000. Metropolitan values: Migration, mobility and cultural norms, London 1100–1700. In Wright, ed. 2000. 93–114.
< previous page
page_320
next page >
< previous page
page_321
next page >
Page 321 Kortmann, B., E. Schneider, K. Burridge, R. Mesthrie, and C. Upton, eds. 2004. A Handbook of Varieties of English, Vol. 2. Berlin: Mouton de Gruyter. Kortmann, B., and B. Szmrecsanyi. 2004. Global synopsis: Morphological and syntactic variation in English. In Kortmann et al., eds. 2004. Vol. 2. 1142–202. Labov, W. 1972. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Lass, R. 1992. Phonology and morphology. In Cambridge History of the English Language, Vol. 2, edited by N. Blake, 23–155. Cambridge: Cambridge University Press. ———. 2000. Phonology and morphology. In Cambridge History of the English Language, Vol. 3, edited by R. Lass, 56–186. Cambridge: Cambridge University Press. Lenneberg, E. 1967. Biological Foundations of Language. New York: Wiley. Mesthrie, R. 1992. English in Language Shift. Johannesburg: Witswatersrand University Press. Milroy, J. 2000. Historical description and the ideology of the standard language. In Wright, ed. 2000. 11–28. Mühlhäusler, P. 1977. Pidginisation and Simplification of Language. Canberra: Pacific Linguistics. Murray, R. 1997. Historical linguistics. In Contemporary Linguistics, edited by W. O’Grady et al., 313–71. London: Longman. Nevalainen, T. 2006. An Introduction to Early Modern English. Edinburgh: Edinburgh University Press. ———, and H. Raumolin-Brunberg. 2003. Historical Sociolinguistics . London: Longman. Prokosch, E. 1939. A Comparative Germanic Grammar . New Haven: Yale University. Reh, M. 1996. The Anywa Language. Cologne: Ruediger Koeppe Verlag. Rissanen, M. 2000. Standardisation and the language of early statutes. In Wright, ed. 2000. 117–30. Samuels, M. 1972. Linguistic Evolution: With Special Reference to English. Cambridge: Cambridge University Press. Schneider, E. 2004. Global synopsis: Phonetic and phonological variation in English world-wide. In A Handbook of Varieties of English, Vol. 1, edited by E. Schneider, B. Kortmann, K. Burridge, R. Mesthrie, and C. Upton, 1111–37. Berlin: Mouton de Gruyter. Schneider, E.W., B. Kortmann, K. Burridge, R. Meshtrie, and C. Upton, eds. 2004. A Handbook of Varieties of English, Vol. 1. Berlin: Mouton de Gruyter. Schreier, D. 2003. Isolation and Language Change: Contemporary and Sociohistorical Evidence from Tristan da Cunha English. Basingstoke: Palgrave Macmillan. Shorrocks, G. 1999. A Grammar of the Dialect of the Bolton Area . Frankfurt: Lang. Strang, B. 1970. A History of English. London: Methuen. Trudgill, P. 1986. Dialects in Contact . Oxford: Blackwell. ———. 1995. Grammaticalisation and social structure: Nonstandard conjunction-formation in East Anglian English. In Grammar and Semantics: Papers in Honour of John Lyon, edited by F.R. Palmer, 136–47. Cambridge: Cambridge University Press. ———. 1999a. Standard English: What it isn’t’. In Standard English: The widening debate , edited by T. Bex and R.J. Watts, 117–28. London: Routledge. ———.1999b. The Dialects of England. Oxford: Blackwell. ———. 2002. Linguistic and social typology. In Handbook of Linguistic Variation and Change , edited by J.K. Chambers, N. Schilling-Estes, and P. Trudgill, 707–28. Oxford: Blackwell.
< previous page
page_321
next page >
< previous page
page_322
next page >
Page 322 ———. 2003. The Norfolk Dialect. Cromer: Poppyland. ———. 2008. English dialect “default singulars,” was versus were , Verner’s Law, and Germanic dialects. Journal of English Linguistics 36(4). ———, and Hannah, J. 1982. International English: A Guide to Varieties of Standard English. London: Edward Arnold. Wells, J.C. 1982. Accents of English. Cambridge: Cambridge University Press. Wright, L. 2000. The Development of Standard English 1300–1800 . Cambridge: Cambridge University Press. Wyld, H.C. 1936. History of Modern Colloquial English. 3rd edition. Oxford: Blackwell. ACKNOWLEDGMENTS Very many thanks indeed to Jan Terje Faarlund, Wolfgang Schulze, Michael Noonan, Andrew KoontzGarboden, Bernd Kortmann, and the conference participants, especially Terttu Nevalainen and Benedikt Szmrecsanyi, for their help with this chapter.
< previous page
page_322
next page >
< previous page
page_323
next page >
Page 323 14 Linguistic Universals and Vernacular Data1 Peter Siemund 1. INTRODUCTION Modern linguistics not only wants to describe, it aims at explanations and predictions based on linguistic theories. The extension of linguistic theorising to vernacular data is a contemporary enterprise much in vogue. Vernacular data are a valuable testing ground for established theories of language and an indispensable corrective for our assumptions about universal properties of languages. The study of vernaculars profits from this enterprise by gaining access to new methodologies, hypotheses, motivations, and explanations. The main objective of the present contribution is to explore the aforementioned relationship between vernacular data and linguistic universals, as they have resulted from modern linguistic theorizing. I will be asking whether and how research into linguistic universals can help us in the explanation of vernacular phenomena and what the methodological problems posed by such an approach are. The focus of this study will be on vernacular data of English and mainly comprises data from different regional varieties. The methodological backbone is furnished by functional typology, which has extensively contributed to our knowledge of the patterns and limits of variation and has made ample suggestions for providing adequate and convincing explanations for them. The subsequent text is structured as follows. Because it is my goal here to discuss vernacular data against the background of linguistic universals, I will first of all provide an assessment of the major theoretical frameworks in which universals have been discussed, identify the domains of language for which universals have been proposed, and try to characterise the basis as well as the motivations underlying the universals proposed. In addition, I will be pointing out the relevance of universals research for the study of English varieties. In the second part of my chapter, I will propose a typology of English varieties, or different types of vernacular data, and correlate these with the various linguistic domains for which universals have been proposed. This discussion will show that varieties of English are a highly heterogeneous field for which universal concepts, including their motivations, from various
< previous page
page_323
next page >
< previous page
page_324
next page >
Page 324 domains (change, contact, acquisition, discourse) are relevant. I will argue that there is no reason to believe in a set of universals valid for vernacular data alone, at least not for English. Either we discuss universals that apply to language in general or we focus on specific linguistic domains. In the third part of my chapter, I will discuss three grammatical domains of English (pronominal gender, reflexives and reflexive marking, and tense and aspect), based on vernacular data from different domains, which are accessible to an interpretation in terms of linguistic universals. I will propose that the notions of semantic maps and conceptual space, as developed in functional typology, can be fruitfully exploited for the analysis of vernacular data. Observable grammatical variation across different varieties of English is largely congruent with semantic maps resulting from cross-linguistic typological work. 2. THE MAJOR PARAMETERS OF LINGUISTIC UNIVERSALS The idea of a linguistic universal is quite simple at first sight: a restriction on the set of logically possible properties of language. It is intuitively plausible because not everything that is in principle feasible is attested. Linguistic universals do not govern language particulars, but capture fundamental and general properties of language and thus by definition apply to all languages. On closer scrutiny, however, the notion of a linguistic universal is quite problematic. To begin with, linguistic universals are highly theory-dependent. A universal formulated in the generative paradigm has little to do with the universals embraced in functional typology. This applies, in particular, to the explanations for universals put forward in these paradigms. Another problem is that linguistic universals are quite domain-specific even though they are meant to cover universal properties of language. Universals can be conceived of as restrictions on the architecture (grammar) of language, as universal constraints on language change (grammaticalisation), as restrictions on language contact processes, and also as constraints in terms of language-acquisition processes. There also seem to be universal constraints on the structure of discourse. Another crucial problem concerns the basis of linguistic universals (i.e. their motivation and explanation). It is clearly unsatisfactory merely to formulate constraints and generalisations of language structure without trying to account for them. The explanation of linguistic universals is again highly theory-dependent. Different theoretical frameworks give radically different explanations for universals. Discussing vernacular data against the background of linguistic universals research thus offers numerous trapdoors, which necessarily arise as a consequence of this multifarious notion. Depending on the theoretical
< previous page
page_324
next page >
< previous page
page_325
next page >
Page 325 framework chosen, we can expect different research objectives as well as answers. Representative of the generative paradigm are the studies by Benincá (1989) and Henry (1995), whereas Kortmann and Szmrecsanyi (2004) and Siemund (2008a) analyse vernacular data in the functionalist-inductive paradigm. An important question is whether to regard vernacular data as a linguistic domain in its own right complementing the study of language architecture, change, contact, acquisition, discourse, and perhaps others. Doing so would allow us to postulate linguistic universals that in principle could be relevant to vernacular data alone, as done in Chambers (2001, 2003, 2004) and Sand (2003, 2004). Alternatively, one could hold that the universals discussed in the context of these other domains are also—at least partially—relevant to the analysis of vernacular data. I will adopt this latter position here because varieties of English (i.e. English vernaculars) cannot be understood without taking into account change, contact, acquisition, and perhaps also discourse. Varieties of English, or particular properties found in these varieties, can be analysed as the result of grammaticalisation processes, language contact, and (mainly L2) acquisition. Eventually, there is the question of the basis of linguistic universals (i.e. their proper motivation). Again, different theoretical frameworks give different answers, but even within functionalist paradigms, a host of competing motivations can be found side by side. None of these, I think, is specifically geared towards vernacular data. 3. THEORETICAL FRAMEWORKS, UNIVERSALS, AND EXPLANATION In drawing a distinction between functionalist-inductive and formalist-deductive paradigms, I am clearly taking a somewhat simplifying position here, but it appears that many or most current approaches to the study and explanation of language can be grouped with either of these two major paradigms (Newmeyer 2000, 2005). These two major paradigms have developed totally different—and often incompatible— conceptions of universals and explanation, and hence this is the first important parameter that needs to be investigated here. 3.1 Functionalist-Inductive Paradigms The primary concern of all work within functionalist-inductive approaches such as functional typology has been the identification of cross-linguistic generalisations and properties that can be shown to hold for most, if not even all, languages. Cross-linguistic work in this paradigm works bottom-up (or inductively), trying to derive generalisations from many, in the ideal
< previous page
page_325
next page >
< previous page
page_326
next page >
Page 326 case genetically unrelated, languages and intends to identify universally valid constraints on the architecture of language. Linguistic universals developed within this paradigm can be conceived of as restrictions that apply to all languages. Such cross-linguistic generalisations have been found in the domains of phonology, morphology, and syntax, although most of them are statistical in nature and usually do not hold for all languages. Because absolute universals are either trivial (e.g. ‘all languages have consonants and vowels‘) or rare, functional typology has, to a greater extent, been interested in implicational connections where the existence of property X in a language is tantamount to the existence of property Y. Great value is attached to implicational connections of this type because they say something about the mutual dependencies of properties of language and make clear predictions about possible languages. Moreover, they can be combined to form chains of implicational connections known as implicational hierarchies. Well-known examples are the Animacy Hierarchy and the Accessibility Hierarchy, as shown in examples (1) and (2). (1) 1, 2 < 3 < proper names < human < animates < inanimates (2) subject < direct object < indirect object < comparison Such hierarchies have proved relevant for various grammatical phenomena (number marking, case marking, gender marking, relativisation, etc.). For example, if number distinctions are marked on inanimate nouns, the animacy hierarchy in example (1) predicts that such number distinctions will also be drawn for all noun classes further to the left (Corbett 2000; Croft 2004:128). In the same way, the accessibility hierarchy in example (2) makes us expect that if indirect objects can be relativised from, the same should apply to direct objects and subject constituents (Croft 2004:147; Keenan and Comrie 1977). More recent theorising in functional typology has put great emphasis on the concepts of ‘semantic maps’ and ‘typological space’ (Croft 2004; Haspelmath 1997). The idea here is to identify contiguous semantic or functional areas which in principle can be covered by linguistic forms. Such contiguous functional domains are stable cross-linguistically and in this sense universal. The model predicts that languages can project grammatical categories on different, yet contiguous areas in conceptual space and thus captures the universal and the language-specific. Many, if not all, implicational hierarchies can be reinterpreted as semantic maps. Croft (2004:133ff) proposes the term ‘conceptual space’ for the universal aspects and ‘semantic map’ for the areas in the conceptual space covered by formal markers of specific languages. The claim that grammatical categories cover contiguous areas in conceptual space is known as the Semantic Map Connectivity Hypothesis (Croft 2004:134): “any relevant language-specific
< previous page
page_326
next page >
< previous page
page_327
next page >
Page 327 and/or construction-specific category should map onto a connected region in conceptual space”. The functionalist-typological approach, including its conceptions of language-universal and languageparticular facts, is of great relevance to the study of vernacular data because it is first and foremost interested in formulating constraints on variation (Bisang 2004). Hence, this approach can—almost by definition—help us in the analysis of vernacular data. As a matter of fact, for most languages investigated by typologists, extensive written documentation does not exist, and the main source of information is vernacular data. The discussion of English, and other Western languages for that matter, within functionalist-typological approaches is usually restricted to the standard variety, although the inclusion of the regional as well as the social dimension could yield important additional insights. Because functional typology makes claims that in principle hold for all languages, including their respective varieties, vernacular data should be captured by these claims and should also provide important input for making the relevant claims more stable. Speakers of vernacular codes have the same cognitive basis as speakers of standard varieties; hence, the same or similar functional explanations should be available for vernacular data. 3.2 Formalist-Deductive Paradigms Generative grammar (also Universal Grammar [UG]) aims at the creation of a coherent theory of language built on highly abstract categories and principles which are relevant for all languages. UG is not interested in universals based on surface properties of language. All hypotheses in this framework have the status of universals (i.e. they are meant to apply to all languages) until falsified. Thus, the ultimate objective of work done within this paradigm is, just as in the functional-typological approaches, the unravelling of universal constraints on the architecture of language. Nevertheless, the methodology is quite different. Rather than finding generalisations on the basis of the formal properties of many unrelated languages (bottom–up), the formalist-deductive approach starts by carefully analysing a single language, thus deriving properties that in principle hold for all languages (top–down). The generalisations offered have the status of hypotheses that need to be tested on other languages. Surface properties of languages, in sharp contrast to functionalist-inductive paradigms, have no scientific value in this framework and are only relevant in so far as the hypotheses need to be tested on language material. Work done in this paradigm has led to important insights into the structure of language. A fundamental property that apparently holds for all human languages is structure dependence (Chomsky 1971, 1975). The principle of structure dependence says that structural operations are not carried out on surface elements, but operate on abstract categories. To give a
< previous page
page_327
next page >
< previous page
page_328
next page >
Page 328 simple example, it is well known that syntactic operations are not based on surface strings, but on constituents. Because surface strings do not contain marks of constituents, the latter must be something abstract. A prominent and convincing example of structure dependence is the verb-second property of some Germanic languages. Example (4) shows possible permutations of example (3), and all of them have the finite verb placed in the second structural position. (3) Paul hat die Suppe mit dem Löffel gegessen. Paul has the soup with the spoon eaten ‘Paul has eaten the soup with the spoon.’ (4) a. Die Suppe hat Paul mit dem Löffel gegessen. b. Mit dem Löffel hat Paul die Suppe gegessen. c. Gegessen hat Paul die Suppe mit dem Löffel. A specific operation that works on the relocation of constituents and not on individual words is movement. There is no language that would, say, move the third word in a sentence to the initial position to mark some particular function. In the same manner, there is no knowledge of a language that would exploit the reversal of all sentential elements (i.e. move the last element to first position, the last but one element to second position, etc.) for some grammatical functions. Another fundamental property of language that apparently is universal is recursion (i.e. the repeated and potentially infinite embedding of a clause under a matrix predicate) (John says that Paul knows that Stella believes …) or the stacking of adjectives. There have been numerous other proposals for allegedly universal properties of language (Binding Principles, Subjacency, etc.), most of which, however, are difficult to test on languages other than English. The analysis of vernacular data in the formalist-deductive paradigm appears more difficult than in the functionalist-inductive paradigm. This is for several reasons. First, universals like structure dependence and recursion are so general that they are probably only of little value for the analysis of vernacular data. It appears safe to expect vernacular varieties to conform to these requirements. Second, most of the other principles proposed in this framework struggle with the evidence of Standard English. The Binding Principles, to mention just one prominent case, are notoriously difficult to bring into harmony with the data from the standard varieties, let alone nonstandard Englishes (König and Siemund 2000a; Reinhart and Reuland 1993; Siemund 2003, 2004). Third, the analysis of more subtle properties crucially relies on negative evidence and clear speaker judgments, which are difficult to come by for vernacular data. Finally, most grammatical domains for which variation can be observed in English are not really accessible to a discussion in the formalist-deductive paradigm because this framework
< previous page
page_328
next page >
< previous page
page_329
next page >
Page 329 has only little to say on tense and aspect, pronominal gender, agreement, definiteness, and the like. 3.3. A Note on Explanation Functionalist-inductive paradigms, on the one hand, and formalist-deductive paradigms, on the other hand, apply widely diverging concepts of explanation. Generally speaking and again somewhat simplifying, the main point where the two approaches differ is the relevance of external (i.e. extralinguistic explanation). A fundamental belief shared by most work done within the functionalist-inductive paradigm is that language structure cannot be explained by taking recourse to itself: language cannot explain language. All plausible explanation must be external and thus rooted in physical conditions, cognition, psychology, processing, learning, and the like. Formalist-deductive paradigms, on the contrary, completely deny the relevance of external explanations. All explanation can plausibly proceed only within the confines of a clearly defined linguistic theory, and all observable structure of language should in principle be derivable from a set of assumptions made in a theoretical model. Explanation in the formalist-deductive paradigms is strictly system-internal and based on the classical deductive-nomological model (Berg 1998; Hempel 1965). Having said that, one needs to emphasise that both approaches rely on abductive reasoning for the generation of hypotheses (Peirce 1931). The crucial difference lies in the nature of the hypotheses. Functionalist-inductive approaches generate hypotheses concerning the relationship between the structure of language and external factors. Formalist-deductive approaches propose theory-internal hypotheses. 4. UNIVERSALS AND LINGUISTIC DOMAINS The foregoing sections have shown that the linguistic framework adopted crucially determines the conceptions of universals as well as the explanations put forward for them. Another important factor influencing the shape of universals is the domain of (the study of) language which is investigated. Such domains include the architecture of language (probably the most general of all), language change, language contact, language acquisition, and the structure of discourse—perhaps others as well. These domains have given rise to astonishingly different conceptions of universals. Because the architecture of language is the most general domain in which the quest for universals is pursued, conceptions of universals strongly depend on the theory chosen. The reader is therefore referred back to the discussion of theoretical frameworks in section 3. I will here also try to assess the value of these domains, including their notions of universals, for the analysis of vernacular
< previous page
page_329
next page >
< previous page
page_330
next page >
Page 330 data. It will turn out that all of these domains are relevant for varieties of English. 4.1 Language Change Universals of language change have mainly been discussed within grammaticalisation theory (Heine, Claudi, and Hünnemeyer 1991; Hopper and Traugott 2003; Lehmann 1995). The fundamental empirical observation on which this framework rests is that certain lexical elements over time develop into grammatical morphemes. This process of grammaticalisation is highly systematic and involves the gradual reduction of form and content, as well as the integration of the relevant elements into a grammatical paradigm (content words turn into function words). It has also been observed that the lexical sources drawn on for the development of new grammatical morphemes are relatively restricted, with functionally similar grammatical morphemes across different languages being based on similar lexical source elements. Prominent grammatical domains that have been treated within grammaticalisation theory include tense and aspect markers (source lexemes: verbs of motion, volition, having, standing, etc.; Dahl 1987), reflexive markers (source lexemes: body part nouns; König and Siemund 2005), indefinite articles (based on the numeral one), and many more (Heine and Kuteva 2002). Grammaticalisation theory describes pervasive cross-linguistic tendencies. Although it is not possible to predict precisely when and which grammaticalisation processes occur, the observed processes can usually be assigned to some recurrent type. The universal aspects of language change identified in this framework relate to (1) the development of function words from content words, (2) the restricted and recurrent set of source lexemes across languages, and (3) the mechanisms involved (fusion, reduction of form, semantic bleaching, metonymic extension, conventionalisation of implicatures, etc.). Insights gained from grammaticalisation theory are extremely important for the discussion of vernacular data. This is due to several reasons. First, with respect to specific grammaticalisation processes, vernacular forms can be more conservative, but also more progressive relative to the respective standard variety and can thus offer windows into the past and the future. Second, vernacular forms are often void of normative pressures and can thus illustrate how grammaticalisation processes work without such pressures, both in terms of direction and relative speed. Third, vernacular forms may illustrate grammaticalisation processes that are not attested in the respective standard variety, but which are otherwise consonant with crosslinguistic paths of development. I will confine myself to mentioning just two prominent examples here— namely, the development of a sentence-final interrogative particle (innit < isn’t it ) and the pronoun of the second person plural y’all (cf. Example [5]). These grammatical morphemes are not found in the standard varieties.2
< previous page
page_330
next page >
< previous page
page_331
next page >
Page 331 (5) a. Cake goes well with tea, innit? b. I really like y’all’s house. 4.2 Language Contact Because many vernacular forms of English are the result of language-contact processes (Irish English, Singapore English, etc.), insights from contact linguistics, in particular the universal processes identified there, seem imminently significant for the study of the relevant vernacular varieties. Work in contact linguistics can be expected to help to decide controversial cases of influence through contact and to offer new perspectives on established phenomena. Moreover, much traditional work on contact vernaculars assumes that all phenomena must be reconstructible in either of the contact varieties/ languages, thus screening out possible innovations as the result of the contact situation. The analysis of language-contact situations has unearthed important parameters governing the exchange of features between the languages involved. Moreover, it has been possible to observe recurrent processes across numerous unrelated contact situations (Curnow 2001; Siemund 2008b; Thomason 2001a, 2001b; Winford 2003). Some of these principles and processes are attributed universal status, others represent pervasive generalisations. Among the most important factors known or hypothesised to foster (or restrain) the mutual influence of languages in contact are the relative size of the communities in contact, the intensity of contact, the language-learning processes involved, the typological distance between the contact languages, as well as the status of potentially borrowable items in the overall linguistic system (compare the borrowing scale proposed in Thomason 2001b). Moreover, it has been claimed that language contact fosters complexity of the languages involved, that gaps in the morphemic inventory facilitate borrowing, that marked features are less likely to be transferred than unmarked features, and that language shift mainly affects phonology and morphosystax, but less so the lexicon.3 It is clear from this outline that the discussion of language contact has moved considerably beyond the notions of ‘transfer’, ‘calquing’, ‘mutual reinforcement’, and so on, which are still predominantly found in discussions of English vernaculars in contact with other languages. Although Thomason (2001b: 63) somewhat pessimistically concludes that “anything” can be transferred from one language to another, the factors identified earlier should allow a more qualified discussion of English contact varieties than currently available. Most challenging for language contact studies is the identification of transfer that does not involve the exchange of phonetic substance (i.e. the transfer of meaning or syntactic relations). Recent contributions to language contact emphasise the role of grammaticalisation processes such that speakers of some language A grammaticalise a construction on the model
< previous page
page_331
next page >
< previous page
page_332
next page >
Page 332 of a contact language B using material from A (Heine and Kuteva 2003, 2005). One prominent example from a contact variety of English—Irish English—is discussed in Heine and Kuteva (2005:102) and involves the socalled ‘ after -perfect’, as in example (6). This construction apparently is not just a loan translation because foregoing stages in the grammaticalisation process are attested (cf. He’s after the flu ‘He just had the flu’), although the chronological sequence is difficult to recover from the documentation available. (6) She’s after selling the boat. ‘She has just sold the boat.’ In addition, it has been shown that even the transfer of discourse preferences can invoke contact-induced change. For example, two languages in contact may share a certain word-order pattern. This specific order of elements may be a majority pattern in the one language, but a minority pattern in the other. The transfer of discourse preferences may turn a minority pattern into a frequently used pattern, thus potentially changing the basic word-order type of a language (Heine 2008). None of these more recent ideas have been systematically applied to the analysis of English contact varieties. A first attempt to analyse the medial-object perfect of Irish English as a case of contact-induced grammaticalisation can be found in Pietsch (2005). 4.3 Language Acquisition Results of language-acquisition research appear most relevant for those types of vernacular data that can be assumed to have been extensively shaped by acquisition processes, both of first- and second-language acquisition (SLA). As for vernacular varieties of English, it is pidgins and creoles as well as shift varieties that can be assumed to fulfil this criterion.4 In principle, SLA strategies appear relevant for pidgins and shift varieties, whereas first-language acquisition strategies are probably most relevant for creoles. This presupposes a view of creoles as nativised pidgins. Recurrent and, in this sense, universal phenomena of (untutored) SLA concern the errors made, the causes and motivations of the observable errors, and the order in the acquisition of grammatical morphemes and constructions (Ellis 1995). Typical errors of SLA are omissions, additions, substitutions, and word-order changes. These errors are usually motivated by transfer from the first language, overgeneralisations, analogy, ignorance of rule restrictions, incomplete application of rules, and wrong hypotheses about the target language. The existence of a natural (i.e. universal) order (Krashen 1983) in SLA is, as far as I can see, an open issue. In contrast, studies focusing on the acquisition of specific grammatical morphemes or
< previous page
page_332
next page >
< previous page
page_333
next page >
Page 333 constructions (e.g. pronouns) have been more successful because apparently universal sequences of acquisition can be identified. Most important for our present concerns is the relationship between the universals identified in SLA research and those discussed in other areas of linguistics, notably UG and language typology. The two extreme positions that one can assume in theory are that (1) all regular phenomena of SLA are the consequence of more fundamental linguistic universals, and (2) they are independent of one another. In practice, however, we find that this problem has rarely been addressed because most scholarship interested in these issues has simply been concerned with proving that typological, as well as UG-based universals, can be confirmed by SLA research, thus neglecting many important results from acquisition research. As for typological universals, what we typically find discussed in SLA studies are word-order universals and the Accessibility Hierarchy. UG-based universals in SLA mainly concern subjacency and pro-drop. This is not to deny that this line of research has produced valuable results. For example, Gass and Ard (1984) discuss a number of typological universals from phonology, word order, relativisations, as well as tense and aspect in relation to SLA. Interestingly enough, they find that cross-linguistic universals from all of these domains bear on SLA with the exception of word-order universals. To illustrate this point briefly, second-language learners can be expected to obey the Accessibility Hierarchy (Keenan and Comrie 1977), but it is apparently not the case that they obey the so-called harmonic relations between word orders (Croft 2004:63). As already mentioned, so far most work has been directed toward showing that SLA, and acquisition in general for that matter, is consonant with linguistic universals established independently. However, in the long run, it may well turn out that at least some of these universals are the result of regular principles of language acquisition. 4.4 Discourse Structure The final linguistic domain to be included here is the structure of discourse. The reason is simple: most, if not all, English vernaculars are spoken registers and work in a less syntacticised mode in the sense of Givón (1979, 1984). The study of discourse structure should in principle yield important generalisations for varieties of English shaped by SLA processes, shift varieties, and probably also for pidgins. Work on discourse structure has uncovered important principles underlying the organization of such less syntacticised modes. A good overview of the most important principles can be found in Perdue (2006). These are that (1) topical elements precede comments, (2) focused elements occur last in a sentence or discourse unit, and (3) noun phrases with a high degree of control appear before other noun phrases (Klein 2001).
< previous page
page_333
next page >
< previous page
page_334
next page >
Page 334 4.5 Summary The foregoing discussion has shown that many so-called linguistic universals are quite domain specific, meaning that work in linguistic areas like language change, language contact, language acquisition, and the structure of discourse has given rise to universals that have relatively little to do with one another. This fact as such is not particularly problematic, but when it comes to the discussion of vernacular data (here: varieties of English) against linguistic universals, we need to develop an understanding of what we have in mind when we talk about universals. In essence, the problem is that many universals are domain-specific, wheras vernacular data are a cover term for various linguistic domains. The main lesson to be learned from these considerations in my view is that universals research can only be seriously taken advantage of if the specific properties of vernacular data are taken into consideration. It clearly makes no sense to discuss universals of language contact in the context of traditional English dialects alone. Alternatively, it would be necessary to make the concept of a universal so general that it can cover all linguistic evidence. 4.6 The Basis of Linguistic Universals Within the formalist-deductive paradigm, the basis of linguistic universals is assumed to reside in the language faculty, but because this is a black box, practically no more than that can be said. Functionalistinductive approaches are interested in providing external explanations for observable universal properties of language and mainly address physical and cognitive constraints. Physical constraints mainly address limitations of the human articulatory and perceptual apparatus and explain many phonological restrictions. This includes trivial constraints such as the avoidance of consonant clusters, where the individual consonants are produced at extremely distant places of articulation (say, dentals followed by pharyngeals), but also more subtle cases (e.g. voiced obstruents entail the existence of the corresponding voiceless obstruents). Cognitive constraints capture restrictions concerning the processing and storing of linguistic data in the human cognitive apparatus and frequently address issues of economy and complexity. It is assumed that storing and processing is costly, that there is a limit to it, that there are conditions that negatively affect the resources available, and that the human cognitive system possesses certain principles that make the storing and processing of linguistic data more economical. Among the most important of these principles, we find analogy, entrenchment, frequency (e.g. Zipf’s law), harmony, iconicity, metaphor, metonymy, horror aequi, weight, as well as some others. Recent scholarship in language typology has produced the notions of
< previous page
page_334
next page >
< previous page
page_335
next page >
Page 335 semantic maps and cognitive space (Croft 2004; Haspelmath 1997), which represent powerful cognitive constraints on the structure of language. 5. LINGUISTIC UNIVERSALS AND VARIETIES OF ENGLISH As we have seen, linguistic universals are domain-specific and to a large extent disjunctive relative to these domains. In this section, I will argue that varieties of English (i.e. vernacular data) are also highly heterogeneous, and that any attempt to apply universals research to varieties of English must take these differences into consideration. The differences between and across varieties of English are frequently underestimated in the quest for their universal structural properties. I believe that it is vital to assess the sociocultural and historical background of each variety. This is mainly for three reasons: (1) different types of varieties may give rise to different types of universals, (2) different universals may be in need of different explanations, and (3) even the same phenomena in different varieties may be in need of different explanations. Varieties of English across the globe can be classified in different ways, but as far as I can see, we can distinguish at least the following major types: 1. Native Englishes — British English — North American English — English of Australia and New Zealand — South African English 2. Shift varieties — Irish English — Singapore English — Indian South African English 3. L2 varieties — Singapore English — Indian English — many African Englishes 4. Creole Englishes (Tok Pisin, Sranan, Saramaccan, etc.) 5. Various Pidgin Englishes This scheme of classification—admittedly—is relatively rough, but any such scheme is faced with the problem of enormous internal differentiation of some varieties. Indian English, to give just one prominent example, is partly a native variety, partly a shift variety, and partly an L2 variety. Singapore English can be regarded as a shift variety and as an L2 variety. I will not elaborate this point further here.
< previous page
page_335
next page >
< previous page
page_336
next page >
Page 336 Each of these varieties deserves a separate treatment when it comes to the discussion of linguistic universals. However, a proper treatment of these varieties is severely complicated by the fact that universal concepts from more than one linguistic domain are usually relevant for them. Intuitively, we expect the following correlations: — Fundamental constraints on the architecture of language should in principle be relevant for all these varieties. One of these principles would be structure dependence, and this general constraint seems to be satisfied by all varieties. — Universals arising from language change (grammaticalisation) should be relevant for understanding the native Englishes and clearly also for Creole Englishes. They should be less important for L2 varieties, pidgins, and shift varieties. However, Heine and Kuteva (2005) argue for the importance of grammaticalisation processes in language-contact situations. — Language-contact studies have given rise to many important generalisations of cross-linguistic validity. In principle, these should apply to English shift varieties and L2 varieties because these are essentially contact varieties, and perhaps also creole Englishes. They should also be relevant for dialect contact. — Universal properties of L2 acquisition must be utterly important for making sense of English L2 varieties as well as shift varieties. — Universals of discourse structure can be assumed to influence or even determine L2 varieties and pidgins. — Principles of L1 acquisition are known to be important for creole languages. The previous list illustrates a many-to-many relation between types of varieties and the linguistic domains in which universals of language have been observed. The picture is complicated by the fact that neither linguistic domains nor the varieties of English identified earlier are neatly delimited objects of study. But even if we abstract away from the problems of delimitation, it is clear that a multitude of different linguistic universals is in principle relevant for varieties of English. Specifically, the following points need to be taken into consideration: — There are different types of English varieties, and hence great care must be exercised in attempting generalisations across all of them. — Generalisations based on single dimensions (e.g. the regional dimension) may miss the special character of many varieties. — Concentrating on phenomena that all varieties have in common (so-called ‘vernacular universals’ and ‘angloversals’) may mask important details.5
< previous page
page_336
next page >
< previous page
page_337
next page >
Page 337 — Any apparent generalisation across all varieties may have different causes in different varieties. In what follows, I will provide some exemplification for the points made earlier. Chambers (2004:129) offers a list of seven grammatical phenomena which he discusses as plausible candidates of what he terms ‘vernacular universals’ (i.e. universal properties of vernacular code). Examples include phonological processes like the reduction of consonant clusters (walking walkin; post office pos office), the devoicing of final obstruents (hundred hundret), the regularisation of paradigms and patterns of concord, multiple negation, and copula deletion. According to Chambers (2004:128), these phenomena “recur in vernaculars wherever they are spoken”. The problem is that these so-called ‘vernacular universals’ are not specific to vernaculars. Phonological processes as the ones discussed by Chambers can most likely be found in all spoken registers. Regularization processes occur all the time as is witnessed by diachronic linguistics. Multiple negation is quite common cross-linguistically—not only in vernaculars, and so on and so forth. Kortmann, Schneider, Burridge, Mesthrie, and Upton (2004) outline an approach to varieties of English that mainly takes the regional dimension into consideration. In this handbook of variation, the individual varieties of English are more or less treated as separate languages and arranged, discussed, and analysed in a style that is familiar from cross-linguistic typological work. As I said earlier, focusing on the regional dimension alone may hide many other important differences between these varieties. Interestingly enough, Kortmann and Szmrecsanyi (2004), in the same volume, as well as Szmrecsanyi and Kortmann (Chapter 2, this volume) analyse varieties of English against their salient grammatical properties, thereby finding clusters that match the varieties of English identified earlier surprisingly closely. Apparently, such clusters of grammatical properties can identify different types of varieties. Much work in the study of English varieties is concerned with uncovering the source of a certain phenomenon. Because many varieties of English are contact varieties, it is tempting to analyse many nonstandard phenomena as cases of transfer from the relevant contact language, say the loss of dental fricatives. But such a position neglects that many phenomena can also be the result of language change or L2 acquisition. The same applies, mutatis mutandis, to various cases of grammaticalisation, which may be contact-induced or may have happened anyway (e.g. medial object perfects in Irish English). In Sand (2003) and Siemund (2002a, 2002b, 2008a), the distribution of definite articles and systems of pronominal gender, respectively, is investigated across different varieties of English and explained by taking recourse to the hierarchy of individuation and ultimately the concept of semantic maps. Although the hierarchy of individuation successfully restricts the
< previous page
page_337
next page >
< previous page
page_338
next page >
Page 338 scope of variation, it does not as such explain the extension (or reduction) in the use pattern of a formal device relative to Standard English. This point will be illustrated in detail in the next section. 6. THREE CASE STUDIES: PRONOMINAL GENDER, REFLEXIVITY, AND TENSE AND ASPECT I will now turn to a discussion of three case studies of grammatical variation across different varieties of English. The grammatical domains are pronominal gender, reflexivity and reflexive marking, as well as tense and aspect. They have been chosen because I have done work on these areas myself, and they illustrate nicely the points that I want to make. The main methodological point addressed in this chapter is that varieties of English (i.e. vernacular data) should either be regarded as ‘normal’ linguistic systems which are subject to universal restrictions on the general architecture of language or that varieties should be matched with linguistic domains for which universal aspects have been identified (change, contact, acquisition, and discourse). The three case studies discussed in what follows will draw on some prominent examples of universals of the architecture of language identified in functional typology. 6.1 Pronominal Gender The distribution of gender-marked pronouns like he, she, and it in Standard English is determined by animacy (he/she—it ) and the sex (he—she ) of the relevant referents. In some regional varieties of English, in contrast, we find extensive use of the animate pronouns (he/she) for objects of the inanimate world. For example, in the traditional vernaculars of Southwest England, the animate pronouns he and she are used for all objects as long as they are countable or individuated, whereas the neuter pronoun it is only used for masses and abstract referents. Similar systems of pronominal gender exist in Newfoundland English and Tasmanian Vernacular English, but also in spoken Dutch, as well as some Danish and Frisian dialects (Siemund 2002a, 2002b, 2008a; Trudgill 1990). This is shown in Figures 14.1a and 14.1b. Some exemplification from the southwest of England is provided in examples (7) through (10), taken from Siemund (2008a: 41–6): (7)[What’s the matter with your hand?] Well, th’old horse muved on, and the body of the butt valled down, and he [the hand] was a jammed in twixt the body o’ un and the sharps (bran-pollard). (8)[Of an ash tree which was leaning over a road, a man said to me] Our Frank limb un last winter, but I don’t never think he’ll never be able vor to be a-got upright.
< previous page
page_338
next page >
< previous page
page_339
next page >
Page 339
Figure 14.1a Pronominal gender in Standard English. (9) Thick there cask ’ont hold, tidn no good to put it [the liquid] in he [the cask]. (10) [In carving a joint it is very common to ask:] Do you like it [timber] green [immature] or dry? In Siemund (2002a, 2002b, 2008a), it is argued that these systems of pronominal gender comply with the well-known hierarchy of individuation, such that the animate pronouns are used for the leftmost part of the hierarchy shown in Figure 14.2, whereas neuter it is used for the area to the right. Crucially, varieties of English differ in the location of the cut-off point on the hierarchy. One way of representing the hierarchy of individuation can be found in Figure 14.2 (Sasse 1993). Discussing systems of pronominal gender in terms of the hierarchy of individuation allows us to make some clear predictions. For example, we would expect animate and inanimate pronouns to cover adjacent, noninterrupted areas on the hierarchy—in agreement with the Semantic Map Connectivity Hypothesis proposed by Croft (2004:134). We also predict that sporadic occurrence of animate pronouns with reference to something
Figure 14.1b Pronominal gender in varieties of English.
< previous page
page_339
next page >
< previous page
page_340
next page >
Page 340
Figure 14.2 The hierarchy of individuation. inanimate, as known from various spoken registers of English, occur along the hierarchy of individuation (i.e. to the right of the cut-off point). Conversely, systems of pronominal gender where animate and inanimate pronouns cover discontinues areas or where certain areas of the hierarchy remain unexpressed are not predicted and consequently ruled out. To the best of my knowledge, such unpredicted systems are not attested across varieties of English. Some predicted and unpredicted systems are shown Figures 14.3a and 14.3b. The main advantage of analysing systems of pronominal gender in terms of the hierarchy of individuation consists in the fact that this hierarchy is well established cross-linguistically and has been shown to determine the distribution of various other grammatical categories, such as number, case, and so on (Croft 2004:122ff). The main problem surrounding the hierarchy is that its conceptual status, its motivation, is not entirely clear. Although an ordering with respect to degree of individuation appears intuitively plausible, we are still in need of studies that would demonstrate its conceptual basis.
Figure 14.3a Attested systems of pronominal gender.
< previous page
page_340
next page >
< previous page
page_341
next page >
Page 341
Figure 14.3b Unattested systems of pronominal gender. The bottom line of the foregoing discussion is that the hierarchy of individuation provides us with a universal generalisation while allowing us to describe variety-specific facts. In other words, it captures the universal and the specific. Systems of pronominal gender across varieties of English crucially differ in the position of the cut-off point separating the functional domains of animate and inanimate pronouns. The hierarchy of individuation does not make any claims concerning the conditions that force different varieties to adopt different cut-off points. These conditions are perhaps recoverable from the relevant types of varieties (contact variety, L2 variety, etc.). 6.2 Reflexives and Reflexive Marking The second domain that I will briefly analyse here in terms of linguistic universals is the one of reflexive marking (i.e. the co-indexing of a sentential constituent with the subject): (11) Johni saw himselfi in the mirror. This domain of grammar has seen a lot of important innovations in the history of English, developing from a state with no reflexive markers (Old English) to a state with a complete paradigm of such markers: myself , yourself, himself, and so on. Reflexive markers of English are complex expressions consisting of a pronominal form and the morpheme self/selves. Different stages of this historical development are preserved in varieties of English. Moreover, some varieties have also produced local innovations. Across varieties of English, the domain of reflexives and reflexive marking gives rise to essentially three types of variation: (1) variation in the paradigm of reflexives (e.g. hisself ), (2) reflexives based on source lexemes other than self (body part nouns like yet ‘head’ in Tok Pisin, sinkii ‘skin’ in Saramaccan, and bodi ‘body’ in Nigerian Pidgin), and (3) variation with respect to the distribution of reflexives and pronouns. Here, I will exclusively be concerned with the third issue. The other topics are discussed in Siemund (2003).
< previous page
page_341
next page >
< previous page
page_342
next page >
Page 342 The distribution of reflexive markers (complex self -forms) and simple pronouns in the history of English is conditioned by mainly two factors (König and Siemund 2000b). On the one hand, it can be observed that complex self -forms first develop in the third person (him/herself), then spread to the second person (yourself), and finally also appear in the first person (myself ). The relevant contexts are given in example (12) on the basis of examples from Modern English. The standard varieties do not allow the use of simple pronouns for the marking of reflexive relations in these contexts. (12) a. He brews him/himself a cup of coffee. b. You brew you/yourself a cup of coffee. c. I brew me/myself a cup of coffee. On the other hand, the development of complex self -forms is influenced by the structural position in which they occur. It turns out that pronouns coreferential with the subject fuse with an adjacent self -form first in the position of direct objects. Complex self -forms then also appear in the position of indirect object and can finally also be found as obliques. This is shown in example (13). The standard varieties do not allow simple pronouns in a coreferential interpretation in examples (13a) or (13b), but require self -forms there. However, they tolerate either simple pronouns or self -forms as obliques coreferential with the subject in example (13c). (13) a. He went to bathe him/himself. b. He organised him/himself a band of traders. c. He noticed a snake near him/himself. Varieties of English demonstrate various intermediate positions along the developmental paths (paths of grammaticalisation) sketched earlier. For example, in the traditional vernaculars of England, the use of thirdperson pronouns is attested in a reflexivising function in the position of direct objects. Appalachian English allows third-person pronouns in this function as indirect objects. Many spoken registers of the standard varieties tolerate cases such as example (12c) with a simple pronoun. The crucial point in the present context is that the paths of grammaticalisation sketched out previously (person, grammatical relations) can be interpreted as consequences of well-established linguistic universals. Cross-linguistic work on reflexive marking has found that reflexive markers of the first and second person are extremely robust predictors for the existence of reflexive markers of the third person. In other words, the implicational hierarchy in example (14) applies. Moreover, there is a convincing functional explanation because pronouns give rise to referential ambiguities—between coreferential and noncoreferential interpretation—only in the third person (marginally in the second person, never in the first person).
< previous page
page_342
next page >
< previous page
page_343
next page >
Page 343 (14) 3 person > 2 person > 1 person In a similar spirit, the initial appearance of complex self -forms in direct object positions and its subsequent spread to other structural positions can be interpreted as a result of the grammatical relations hierarchy in example (15). This hierarchy is equally the result of cross-linguistic comparative work and in this sense a linguistic universal (Croft 2004:142ff). (15) direct object > indirect object > oblique Both the person hierarchy in example (14) and the grammatical relations hierarchy in example (15) can be viewed as semantic maps, in a similar way as the hierarchy of individuation discussed in the previous section. These maps have the status of linguistic universals, whereas varieties of English illustrate specific facts in harmony with these universals. As for reflexives and reflexive marking, these maps represent paths of grammaticalisation with varieties of English identifying different positions along these paths. 6.3 Tense and Aspect In what follows, I will address a third empirical domain of English varieties that is of great relevance for the discussion of linguistic universals: tense and aspect marking. Although the following remarks will be made in a more speculative vein, I hope to show that variation phenomena from this domain of grammar are also open to an analysis in terms of semantic maps. Tense and aspect marking across varieties of English mainly provides us with variation along the following dimensions, relative to Standard English: (1) the use of past and present tense instead of the present perfect, (2) the use of resultative constructions, (3) the use of special perfect constructions (e.g. the medialobject perfect), (4) the use of the progressive with state verbs, and (5) the occurrence of constructions not found in Standard English (e.g. the Irish English after -perfect). I will here exclusively be concerned with perfect constructions, mainly from Irish English. It is well known that Irish English has a set of perfect constructions quite different from Standard English (Filppula 1999). Among the most prominent of these constructions is the medial-object perfect, whose main characteristic is the occurrence of the object constituent before the perfect participle. It can only be found with transitive verbs. An illustration is shown in example (16). (16) John has the dinner eaten. ‘John has eaten the dinner. (i.e. the dinner is gone)’ Such medial-object perfects can plausibly be analysed as the intermediate stage of a prominent path of grammaticalisation, along which resultative constructions develop into perfects, as in example (17).
< previous page
page_343
next page >
< previous page
page_344
next page >
Page 344 (17) resultatives medial-object perfects perfects Medial-object perfects differ from resultatives like example (18a) in that the object referent need not necessarily be in a relationship of possession with the subject referent and not in the state described by the participle. Unlike true perfects, as in example (18b), the auxiliary and the participle do not form one constituent. (18) a. John has the enemy bound. ‘The enemy is in a bound state and John possesses the enemy in this state.’ b. John has eaten the dinner. The source of the medial-object perfect in Irish English has been an area of great controversy (Filppula 1999). It has been analysed as a retention phenomenon from some earlier English dialect, a transfer from Irish, and, more recently, as a case of contact-induced grammaticalisation (Pietsch 2005). Whatever the correct reconstruction, all of these analyses are compatible with the path of grammaticalisation in example (17). In the formation period of Irish English, the three constructions identified earlier could be found side by side, albeit not necessarily with the same speaker or in the same area. Provided that it is justified to interpret the path of grammaticalisation in example (17) as a semantic map, we make some clear predictions concerning the distribution of these constructions at a given point in time. More precisely, we would expect speakers to have either one of these constructions, all of them concurrently, or the beginning or end point of the path of grammaticalisation in combination with the intermediate stage. We exclude the case that speakers have resultatives and true perfects at the same time without allowing medial-object perfects. Interpreted in this way, perfect constructions in Irish English, and perhaps other varieties of English as well, open up a principled space of variation that is constrained by the semantic map derived from example (17). Individual speakers select a well-defined segment from this space of variation. Again, this model allows us to unite language universal and variety-specific facts. 7. SUMMARY AND CONCLUSION The point of departure for the current contribution was my intention to apply methods and results from research into linguistic universals for the analysis of varieties of English—or vernacular data in general. In doing so, I have tried to pinpoint a couple of, in my view, important methodological problems that need to be taken care of before any such endeavour can be successfully pursued. Moreover, I have provided an analysis of
< previous page
page_344
next page >
< previous page
page_345
next page >
Page 345 three grammatical phenomena across different varieties of English within the framework of functional typology, a representative of the functionalist-inductive paradigm, addressing universal and specific aspects of variation. According to my line of argumentation, there are mainly two problems for an analysis of vernacular data in terms of linguistic universals. On the one hand, the notion of linguistic universals crucially depends on the theoretical framework chosen, as well as the domain of language investigated (change, contact, acquisition, discourse, and perhaps others). Furthermore, linguistic theories determine the explanatory apparatus that can be made use of. In sum, then, linguistic universals are a highly heterogeneous species. On the other hand, vernacular data are not a particularly well-defined object of study, as is witnessed by the enormous differentiation of English varieties. We find native varieties, shift varieties, L2 varieties, creoles and pidgins, as well as a lot of overlap among them. Although this list is not conclusive, it shows that lumping these varieties together for analytical purposes carries the danger of simplification. To be sure, I do not wish to imply that vernacular data and the study of linguistic universals cannot be brought together, only that we have to be careful. As a heuristic, I proposed that we can mainly proceed in two ways. First, we can make use of known constraints on the architecture of language and evaluate vernacular data against them. After all, what holds of language in general should also apply to vernacular data. Second, we can evaluate vernacular data in the context of domain-specific universals (i.e. universals of language change, contact. etc.). The latter approach, however, requires vernacular data to be compatible with these domains. For English at least, this is no problem, although it may be difficult to assign specific varieties to these domains. In the last part of this chapter, I tried to illustrate the analysis of data from different varieties of English, taken from three grammatical domains, in terms of fundamental constraints on the architecture of language, specifically the hierarchy of individuation, the person hierarchy, and the grammatical relations hierarchy and paths of grammaticalisation. Building on proposals made within functional typology, I argued that these hierarchies and paths of grammaticalisation can be viewed as semantic maps, or restrictions on conceptual space, and that the variation observable across varieties of English is ordered and constrained by these semantic maps. The advantage of this approach is that it captures both universal and specific facts. NOTES 1. The research work reported in this chapter has been conducted within the Sonderforschungsbereich 538 Multilingualism at the University of Hamburg. Funding by the Deutsche Forschungsgemeinschaft is gratefully acknowledged. I would like to express my gratitude to all participants of the international
< previous page
page_345
next page >
< previous page
page_346
next page >
Page 346 symposium on World Englishes held at the University of Joensuu Research Station in Mekrijärvi, North Karelia, Finland, 1–3 September 2006, and specifically to Markku Filppula and Juhani Klemola for organising this magnificent scientific event. 2. There are numerous other highly interesting grammatical phenomena attested across different varieties of English. Some of them can be analysed as the result of cross-linguistically pervasive processes of grammaticalisation (Siemund 2004; Trudgill 1995, 1997). 3. It should be noted that some of the statements made in the literature contradict each other. For instance, although it has been claimed that language contact may increase complexity, we also find many statements saying it leads to simplification. Moreover, in language-contact situations, languages may not only replicate structures they lack, but also those they already possess. 4. The term ‘shift variety’ will be used here for those varieties that are the result of a non-English-speaking community giving up their first language in favour of English (Irish English, Singapore English). It is based on the notion of contact-induced language shift as discussed in Thomason (2001b: 74–5). 5. In saying that some of the so-called ‘angloversals’ may be the result of language acquisition strategies, Mair (2003) explicitly addresses this problem. REFERENCES Benincá, P., ed. 1989. Dialect Variation and the Theory of Grammar . Dordrecht: Foris. Berg, T. 1998. Linguistic Structure and Change: An Explanation from Language Processing. Oxford: Clarendon Press. Bisang, W. 2004. Dialectology and typology—An integrated perspective. In Kortmann, ed. 2004. 11–45. Chambers, J.K. 2001. Vernacular universals. In ICLaVE 1: Proceedings of the First International Conference on Language Variation in Europe , edited by J.M. Fontana, L. McNally, M.T. Turell, and E. Vallduví, 52–60. Barcelona: Universitat Pompeu Fabra. ———. 2003. Sociolinguistic Theory: Linguistic Variation and Its Social Implications . Oxford, UK/Malden, US: Blackwell. ———. 2004. Dynamic typology and vernacular universals. In Kortmann, ed. 2004. 127–45. Chomsky, N. 1971. Problems of Knowledge and Freedom . New York: Pantheon. ———. 1975. Reflections on Language. New York: Pantheon. Corbett, G.G. 2000. Number . Cambridge: Cambridge University Press. Croft, W. 2004. Typology and Universals. Cambridge: Cambridge University Press. Curnow, T.J. 2001. What language features can be ‘borrowed’? In Areal Diffusion and Genetic Inheritance: Case Studies in Language Change, edited by R.M.W. Dixon and A.Y. Aikhenvald. Oxford: Oxford University Press. 412–36. Dahl, Ö. 1987. Tense and Aspect Systems. Oxford: Blackwell. Ellis, R. 1995. The Study of Second Language Acquisition. Oxford: Oxford University Press. Filppula, M. 1999. The Grammar of Irish English. London: Routledge. Gass, S., and J. Ard. 1984. Second language acquisition and the ontology of language universals. In Rutherford, ed. 1984. 33–68. Givón, T. 1979. On Understanding Grammar . New York: Academic Press. ———. 1984. Universals of discourse structure and second language acquisition. In Rutherford, ed. 1984. 109–36. Haspelmath, M. 1997. Indefinite Pronouns . Oxford: Oxford University Press.
< previous page
page_346
next page >
< previous page
page_347
next page >
Page 347 Heine, B., U. Claudi, and F. Hünnemeyer. 1991. Grammaticalization: A Conceptual Framework . Chicago: University of Chicago Press. Heine, B., and T. Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. ———. 2003. On contact-induced grammaticalization. Studies in Language 27: 529–72. ———. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. ———. 2008. Contact-induced word order change without word order change. In Siemund and Kintana, eds. 2008. Hempel, C.G. 1965. Aspects of Scientific Explanation and Other Essays in the Philosophy of Science . New York: Free Press. Henry, A. 1995. Belfast English and Standard English: Dialect Variation and Parameter Setting. Oxford: Oxford University Press. Hopper, P.J., and E.C. Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press. Keenan, E., and B. Comrie. 1977. Noun phrase accessibility and universal grammar. Linguistic Inquiry 8: 63– 99. Klein, W. 2001. Elementary forms of linguistic organisation. In New Essays on the Origin of Language, edited by S. Ward and J. Trabant, 81–102. Berlin: Mouton de Gruyter. König, E., and P. Siemund. 2000a. Locally free self-forms, logophoricity and intensification in English. English Language and Linguistics 4(2): 183–204. ———. 2000b. The development of complex reflexives and intensifiers in English. Diachronica XVII: 39–84. ———. 2005. Intensifiers and reflexive pronouns. In World Atlas of Language Structures, edited by M. Dryer, M. Haspelmath, D. Gil, and B. Comrie, 194–7. Oxford: Oxford University Press. Kortmann, B., ed. 2004. Dialectology Meets Typology. Berlin and New York: Mouton de Gruyter. Kortmann, B., E. Schneider, K. Burridge, R. Mesthrie, and C. Upton eds. 2004. A Handbook of Varieties of English. 2 Vols . Berlin: Mouton de Gruyter. Kortmann, B., and B. Szmrecsanyi. 2004. Global synopsis: Morphological and syntactic variation in English. In Kortmann et al., eds. 2004. Vol. 2. 1142–1202. Krashen, S. 1983. Second Language Acquisition and Second Language Learning . Oxford: Oxford University Press. Lehmann, C. 1995. Thoughts on Grammaticalization. München: Lincom. Mair, C. 2003. Kreolismen und verbales Identitätsmanagement im geschriebenen jamaikanischen Englisch. In Zwischen Ausgrenzung und Hybridisierung, edited by E. Vogel, A. Napp, and W. Lutterer, 79–96. Würzburg: Ergon. Newmeyer, F.J. 2000. Language Form and Language Function . Cambridge, MA: MIT Press. ———. 2005. Possible and Probable Languages: A Generative Perspective on Linguistic Typology. Oxford: Oxford University Press. Peirce, C.S. 1931. The Collected Papers: Pragmatism and Pragmaticism. Cambridge, MA: Harvard University Press, Vol. V, §4, 171. Perdue, C. 2006. “Creating language anew”: Some remarks on an idea of Bernard Comrie’s. Linguistics 44(4): 853–71. Pietsch, L. 2005. Re-inventing the ‘perfect’ wheel: Grammaticalisation and the Hiberno-English medial-object perfects. Arbeiten zur Mehrsprachigkeit 65:2005. Universität Hamburg. Reinhart, T., and E. Reuland. 1993. Reflexivity. Linguistic Inquiry 12: 657–720. Rutherford, W.E., ed. 1984. Language Universals and Second Language Acquisition. Amsterdam: John Benjamins.
< previous page
page_347
next page >
< previous page
page_348
next page >
Page 348 Sand, A. 2003. The definite article in Irish English and other contact varieties of English. In The Celtic Englishes III , edited by H.L.C. Tristram, 413–30. Heidelberg: Winter. ———. 2004. Shared morpho-syntactic features of contact varieties: Article use. World Englishe s 23(2): 281– 98. Sasse, H.-J. 1993. Syntactic categories and subcategories. In Syntax. An International Handbook of Contemporary Research, edited by J. Jacobs, A. von Stechow, W. Sternefeld, and T. Vennemann, 646–86. Berlin: de Gruyter. Siemund, P. 2002a. Mass versus count: Pronominal gender in regional varieties of Germanic languages. Zeitschrift für Sprachtypologie und Universalienforschung (STUF) 55(3): 213–33. ———. 2002b. Animate pronouns for inanimate objects. Pronominal gender in English regional varieties. In Anglistentag 2001 Vienna, edited by D. Kastovsky, 19–34. Trier: Wissenschaftlicher Verlag. ———. 2003. Varieties of English from a cross-linguistic perspective: Intensifiers and reflexives. In Determinants of Grammatical Variation, edited by B. Mondorf and G. Rohdenburg, 479–506. Berlin: Mouton de Gruyter. ———. 2004. English. In Variationstypologie , edited by T. Roelcke, 1–29. Berlin: Mouton de Gruyter. ———. 2008a. Pronominal Gender in English. A Study of English Varieties from a Cross-linguistic Perspective . London: Routledge. ———. 2008b. Language contact: Constraints and common paths of contact induced change. In Siemund and Kintana, eds. 2008. Siemund, P., and N. Kintana, eds. 2008. Language Contact and Contact Languages. Amsterdam: John Benjamins. Thomason, S. 2001a. Contact-induced typological change. In Language Typology and Language Universals. An International Handbook, edited by M. Haspelmath, E. König, W. Oestereicher, and W. Raible, 1640–48. Berlin: Mouton de Gruyter. ———. 2001b. Language Contact. An Introduction . Edinburgh: Edinburgh University Press. Trudgill, P. 1990. The Dialects of England. Cambridge, MA: Blackwell. ———. 1995. Grammaticalisation and social structure: Nonstandard conjunction-formation in East Anglian English. In Grammar and Semantics: Papers in Honour of John Lyons, edited by F.R. Palmer, 136–47. Cambridge: Cambridge University Press. ———. 1997. British vernacular dialects in the formation of American English: The case of East Anglian do. In Linguistic History and Linguistic Modelling: A Festschrift for Jacek Fisiak on his 60th Birthday, edited by R. Hickey and S. Puppel, 749–58. Berlin: Mouton de Gruyter. Winford, D. 2003. An Introduction to Contact Linguistics. Oxford: Blackwell.
< previous page
page_348
next page >
< previous page
page_349
next page >
Page 349 15 Why Universals VERSUS Contact-Induced Change? Sarah G. Thomason 1. INTRODUCTION This chapter explores several theoretical issues that arise in the study of universals of language change, from the perspective of both processes of dialect divergence and language contact. A major conclusion is that drawing a dichotomy between proposed ‘vernacular universals’ and contact-induced change is not a good idea because many linguistic changes involve both kinds of process—that is, various processes of contact-induced change and also universal tendencies of various kinds. Historical linguists traditionally appeal to three ultimate causes of language change: drift, which refers to structural tendencies inherent in a given language (resulting from what is often called pattern pressures or structural imbalances); dialect borrowing; and foreign interference. The last two are of course not separable in any precise way, for two reasons. First, the spread of every linguistic change is due to contacts among speakers; and second, dialect borrowing and foreign interference are points on a continuum—it is impossible to draw a neat line between situations in which dialects influence each other and situations in which separate languages influence each other, because the overall process by which sister dialects become sister languages is gradual. Still, different methods have been developed for the study of dialect borrowing, i.e. interference between systems that are lexically and structurally very similar, and foreign interference, primarily the study of interference between systems that are not close lexically and/or structurally (Thomason 2003). Drift as a cause of change subsumes pattern pressures specific to a particular language as well as universal structural tendencies, especially those driven by markedness. Underlying the concept of drift is an assumption that a prominent (although by no means the only) driving force behind internally-motivated language change is ease of learning, which includes both ease of perception and ease of production. Because ease of learning also informs many or most types of contact-induced change, it is hardly surprising to find that the same types of change, and often the very same changes, result from drift and interference. For this reason, anyone seeking the best explanation
< previous page
page_349
next page >
< previous page
page_350
next page >
Page 350 for a given linguistic change must consider potential internal motivations and also potential external motivations. The relevance of this point for the study of vernacular universals in general and World Englishes in particular is that, because different causes can have similar effects, a universal feature of ‘nonstandard’ English-lexicon varieties (where ‘standard’ is taken to mean conforming to the prescriptive norms of the historically most prominent English-dominant nations, i.e. the United Kingdom, the United States, Canada, Australia, and New Zealand) might not arise from the same source(s) in every variety. So, for instance, the typologically rare interdental fricatives might be absent in some nonstandard English dialects of England as a result of drift, but their absence in an English-lexifier creole might be due instead to structural contribution from the creole creators’ original native languages—and also, quite possibly, to drift. It is therefore not sufficient, in arguing for or against a contact explanation for a particular change, to show that the same change has occurred elsewhere under different circumstances; and the possibility of multiple causation, which is likely very common in developments that have led to World Englishes and English-lexifier pidgins and creoles around the globe, must not be overlooked. (See section 4 for a discussion of how to decide whether language contact has played a role in motivating a particular change.) To the three traditional causes of linguistic change, we must now add a fourth: deliberate change by groups of speakers (Thomason 1997, 2007). Such changes, which frequently but by no means always involve lexical innovations, are found primarily in the languages of small speech communities, but they also occur in large speech communities, usually (maybe only) through formal language-planning activities. I will not consider deliberate changes in this chapter, but they lurk in the background, ready to throw a monkey wrench into any simplistic theory of what kinds of linguistic change are possible. The empirical focus of this chapter is twofold: first, on changes that have occurred under well-established contact conditions (including, to a limited extent, pidgins and creoles), and second, on changes that have occurred in circumstances that suggest that drift was the major factor. Changes in both standard and nonstandard dialects will be examined. Although the topic of vernacular universals has been most extensively explored in studies of World Englishes (see e.g. Chambers 2004), my main examples will come from other languages because the concept of vernacular universals is intended to be general, not specific to English. Because of the frequent emphasis on simplification as a (or the) major component of processes of change in such cases, I will concentrate especially on changes that are implausible as simplifications (section 2). I will argue that, despite the ease-of-learning view of much internally-motivated change, neither internallymotivated linguistic change nor contact-induced linguistic change is predictably either simplifying or complicating (section 3). In other words, some internally-motivated changes simplify a language’s structure
< previous page
page_350
next page >
< previous page
page_351
next page >
Page 351 whereas some complicate it, and some externally motivated changes complicate a language’s structure whereas others simplify it. And many changes are neutral: they have neither effect. Finally, the chapter concludes with a consideration of the possibility (or not) of distinguishing changes due to universal tendencies from changes due to language contact (section 4). 2. ARE STANDARD DIALECTS MORE COMPLEX THAN NONSTANDARD VARIETIES? Chambers (2006) claims that “[t]he theory of Vernacular Roots begins with the obvious but hitherto unexploited observation that dialects become more complex as they become more standard or literary”. But is this observation in fact accurate for English and/or for standard dialects in general? Certainly the opposite claim has been made for English, as in the following passage from Thomason and Kaufman (1988:329): Standard languages tend to be simpler (at least at the time of codification) than many of the vernacular dialects on which they are based, partly because they must accommodate the production habits of speakers on the low end of the range of structural complexity within a network of dialects. Thomason and Kaufman, however, distinguish between two general types of standardization processes— evolutionary and revolutionary (1988:209–10)—and the tendency towards simplicity is likely to be true only of standard dialects of the first type. Evolutionary processes arise through dialect mixture when a particular region or city becomes dominant for sociopolitical reasons; its speech becomes a model for official and social purposes, speakers of other dialects move to the dominant locale and adopt its dialect, with modifications through dialect mixture or even koinéization, and so the standard language emerges. Standard English and many other standard languages of Europe arose in this way. In the other general type of standardization, the revolutionary (or artificial) process, there is little admixture (at least at first) because one region or city’s dialect is elevated in toto (and rather abruptly) to standard status. This process is common around the world; one European example is the former Standard SerboCroatian, together with its newly emerging derivatives Croatian, Bosnian, and so on. Examples are more common outside Europe (e.g. Standard Indonesian). The relevance of this distinction for the question of the simplicity of standard dialects is that the second type—standard varieties that arise by fiat rather than through relatively gradual development— are unlikely to fit into a model that assumes that all standard varieties are more complex than all nonstandard dialects: Standard SerboCroatian, for instance, is (or was) just another dialect that happened to get elevated to standard status, so it
< previous page
page_351
next page >
< previous page
page_352
next page >
Page 352 is arguably no more complex, and no simpler, than any of the nonstandard SerboCroatian dialects. (I use the term ‘SerboCroatian’ to refer to the dialects of the majority language of the former Yugoslavia before the breakup of that country into several nations. The specific developments discussed in this chapter all occurred prior to the recent breakup of the old Yugoslavia.) Whether Standard English (or any other gradual, ‘evolutionary’ standard dialect) is simpler or more complex than nonstandard dialects is an empirical question that cannot be solved by mere assertion, by Chambers or Thomason and Kaufman or anyone else. In one sense, admittedly, a standard dialect that becomes the vehicle of administration of a modern nationstate is almost certain to be more complex than any nonstandard dialect: if the dialect that has no history of national usage comes to be used for running a country, new vocabulary will certainly be needed, especially administrative lexicon. This increase in technical terminology is all too likely to be offset by a sharp decrease in such lexical domains as ethnobiological terminology and other nonurban semantic fields, if the new standard dialect happens to come from a previously rural dialect (but usually it doesn’t). (An increase in literary lexicon is not a necessary concomitant, though, especially if the culture previously had a robust oral literature.) For the topic of vernacular universals, however, structural changes are more interesting. Here’s one example in which the development of a grammatical dimension in the standard dialect turned out to be simpler than a partly parallel development in certain nonstandard dialects. The language is Lithuanian, and the grammatical dimension is case inflection in nouns. During the sixteenth and seventeenth centuries, as a result of interference from the case-rich language(s) of shifting Finnic speakers, Lithuanian acquired three new cases: an illative, an allative, and an adessive. All three cases were widely used in the sixteenth and seventeenth centuries, but the allative and the adessive have now vanished from Standard Lithuanian. In certain isolated southeastern dialects, however, all three of these new cases are still in regular use, along with the inherited cases that are shared by the standard and nonstandard dialects (see Thomason and Kaufman 1988:242–3 for discussion of this example, which comes ultimately from Senn 1966:92; cf. also Fairbanks 1977:117). Other examples of standard-dialect structure that is arguably simpler than nonstandard-dialect structure can be found easily in newly standardized languages. Here’s an example from Montana Salish, a gravely endangered language spoken in northwestern Montana (United States). The elaborate Montana Salish phonemic inventory has 38 consonants, including 4 pharyngeal resonant consonants: plain, glottalized, labialized, and glottalized labialized. Efforts to preserve and revitalize the language have resulted in an increase in written materials, which, although this is not a dialect that is ever likely to be used for official governmental purposes, involves a process of standardization: choices are being made about which symbols should be
< previous page
page_352
next page >
< previous page
page_353
next page >
Page 353 included in the orthography. Symbol choice does not in itself present major problems because the writing system is recent (so that it fits the phonemic structure closely) and is based on an Americanist version of the International Phonetic Alphabet. But the four pharyngeal phonemes are problematic because most of the fluent speakers, and all of the younger tribal members who wish to learn their heritage language, either don’t hear the pharyngeals at all or perceive them as vowel length. If they don’t hear them, they can’t write them. So although the few remaining elders who speak the language fluently have regularly occurring pharyngeals in numerous words, it is difficult to justify incorporating these phonemes systematically into the writing system—it would guarantee consistent misspellings. The fluent elders, who don’t generally attempt to write their native language, will continue to pronounce the four pharyngeals when they speak, but no one else will either pronounce or write them; and in this respect, the newly standard dialect will be simpler than the ‘nonstandard’ dialects spoken by these elders. Finally, although pidgins are frequently cited (incorrectly, in my opinion, for historical reasons) as simplified versions of the vocabulary-base language, even pidgins often have complex features that are lacking in the standard variety of the lexifier language. One example is shared by the English-lexicon pidgins American Indian Pidgin English (AIPE, now extinct) and Tok Pisin of Papua New Guinea: this is the marking of transitive verbs by a transitive suffix—the same suffix, in fact, although these suffixes must represent independent developments, not some old historical link between these two pidgins. The suffix is –um (AIPE)/–im (Tok Pisin), both ultimately from English him. Their use is clear in sentences like AIPE Me see-um chief ‘I see/saw the chief’ (literally ‘1sg. see.transitive chief’) and Tok Pisin Na em i pain-im wanpela man ‘Then s/he found a man’ (lit. ‘then s/he predicate. marker find-transitive one man’). So far, my examples of structural complication in nonstandard dialects (and pidgins), as opposed to standard dialects, have come from language-contact situations: shift-induced interference in the case of the new Lithuanian cases, borrowing from English (perhaps mediated by literacy in English) in the case of the shaky Montana Salish pharyngeal phonemes, and language creation in multilingual contexts in the case of the AIPE and Tok Pisin transitive suffixes. Examples that complicate nonstandard dialect structure are easy to find in internally motivated change, too. One instance is found in the history of the oblique plural cases in masculine o- /C-stem noun inflection in several SerboCroatian dialects (see Thomason 1977 for the development of SerboCroatian C-stem masculine nouns from older o- stem masculine nouns). SerboCroatian is traditionally divided into four primary dialect groups: Štokavian, Čakavian, Kajkavian, and Torlak, although Torlak is sometimes seen as a subdivision of Štokavian (see Ivić 1958 for an authoritative discussion of the various dialect groups). The process of drift has led to a variety of configurations of the seven inherited noun cases in SerboCroatian dialects. In the three oblique cases of
< previous page
page_353
next page >
< previous page
page_354
next page >
Page 354 Table 15.1 Masculine o- Stem Oblique Plural Cases in SerboCroatian and Its Immediate Ancestors Language/Dialect Inst.pl Dat.pl Loc.pl Proto-Slavic: -ï -omŭ -ĕxŭ 14th-c. Štok. SC: -i -omĭ -exĭ Standard SC: -ima -ima -ima Resava SC (Štok.): -i -ima -a/-i/-ima Novî SC (Čak.): -i -on -ih Virje SC (Kajk.): -i -om -e N. Timok SC (Torlak): 0 0 0 (no oblique plural cases) the plural—the instrumental, dative, and locative—many dialects, including Standard SerboCroatian, have undergone partial or total syncretism; other dialects have not. In Table 15.1, the starting point for these three cases is represented by late Proto-Slavic and echoed, after regular sound changes, by fourteenthcentury Štokavian SerboCroatian (SC) (Svane 1958). Following these two languages are the relevant case suffixes for modern Standard SerboCroatian (a Štokavian dialect) and one representative nonstandard dialect from each of the four major dialect groups: Resava (Štokavian; Ivković 1927), Novî (Čakavian; Belić 1909– 1910), Virje (Kajkavian; Fancev 1907), and North Timok (Torlak; Stanojević 1911). Of the five modern dialects, only Northern Timok, which has lost all oblique plural cases, is simpler than Standard SC, in which the three oblique plural cases have merged morphologically. The other Štokavian dialect, Resava, has undergone partial syncretism in the instrumental and locative cases—one variant of the locative is identical to the instrumental suffix –i, and another locative variant, –ima, is identical to the dative suffix –ima—while the other two modern nonstandard dialects, Novî and Virje, have kept all three cases separate. There are other dialects besides Standard SC that have merged all three of these cases and others besides Resava that have merged two of them. The point, however, is that Standard SC is simpler than many nonstandard dialects in this feature. In other features the standard dialect is more complex than at least some nonstandard dialects, and in still other features the standard dialect equals at least some nonstandard dialects in complexity. No valid generalization can be made, therefore, about the overall complexity of Standard SC by comparison to nonstandard dialects. The same conclusion can be drawn from a set of changes in Slavic territory as a whole—namely, the history of the noun-class category of animacy
< previous page
page_354
next page >
< previous page
page_355
next page >
Page 355 in Slavic languages. The changes in the individual languages are, as far as can be determined, internally motivated (i.e. due to drift). In some languages and dialects, the result is a simpler system than in ProtoSlavic; in others, the result is a more complex system; in still others, the Proto-Slavic morphological pattern is preserved intact. There is no correlation between the morphological results and the status of a given language/dialect as standard or nonstandard. The morphological distinction between animate and inanimate nouns was clearly very new in late ProtoSlavic, and its origin has been a subject of lively controversy for decades (see e.g. Thomason and Kaufman 1988:249–50). In the earliest Slavic texts, its morphological and semantic range was quite limited: the distinction was present only in the accusative singular of masculine o- stem nouns, and in this noun class, only free, adult, human males were grammatically animate. Specifically, in an animate masculine o- stem noun, the accusative singular suffix was identical to the genitive singular suffix, whereas in inanimate o- stem nouns the accusative singular matched the nominative singular. In all the daughter languages of Proto-Slavic the animacy category has expanded semantically since ProtoSlavic times, usually to include all mammals and often to include other creatures as well. Morphologically, as noted earlier, the results vary considerably between and often within languages. Table 15.2 shows the patterns in late Proto-Slavic and three modern daughter-language systems: Standard SerboCroatian, Novî SerboCroatian (Belić 1909–1910), and Standard Russian. Standard SerboCroatian has the exact same morphological pattern for the animacy category as in late ProtoSlavic: the only grammatically animate nouns are masculine o- stems, and the distinction appears only in the accusative singular: accusative = genitive singular in animate nouns and accusative = nominative singular in inanimate nouns. In Novî SerboCroatian, the distinction has remained in the singular of masculine o- /Cstems, but it has Table 15.2 Animacy in Slavic Languages Masc. o-Stem Nouns Language/Dialect Inanimate Animate Proto-Slavic: sg.acc = sg.nom sg.acc = sg.gen Standard SC: sg.acc = sg.nom sg.acc = sg.gen Novî SC: acc = nom acc = gen Standard Russian: sg.acc = sg.nom sg.acc = sg.gen St. Russ., all noun classes: pl.acc = pl.nom pl.acc = pl.gen
< previous page
page_355
next page >
< previous page
page_356
next page >
Page 356 also spread to the plural. The result—formally, removal of the ‘singular’ specification required for ProtoSlavic and Standard SC—is a simpler system in terms of rule complexity: the Novî rule is more general. It isn’t clear whether the Standard Russian system is simpler or more complicated than the Proto-Slavic and Standard SC system. The singular pattern is identical to both of those languages: animacy is grammatically relevant only in masculine o- stem nouns and only in the accusative singular. But in the plural, all noun classes—not just the masculine o- /C-stems, but also the a- stems and the old i- stems (modern feminine Cstems)—have an animate/ inanimate distinction, via extension of the singular masculine o- stem pattern to the o- stem plural and then generalization of the plural pattern: accusative = genitive in plural animate nouns and accusative = nominative in plural inanimate nouns. Unlike the Novî SC pattern, the Standard Russian pattern may be considered more complicated than the older system because two rules are needed, one for singular and one for plural nouns; these two rules cannot be conflated. Nevertheless, both Standard Russian and Novî SC have generalized analogically from the late Proto-Slavic pattern. The difference is that Russian has taken the process further, with a resulting asymmetry between singular and plural in the grammar of animacy. Finally, Standard Czech has generalized the animate/inanimate distinction in a different way, moving toward a system in which animate masculine nouns are distinguished from inanimate nouns throughout the paradigm. The animate/inanimate distinction has spread to most cases in both the singular and the plural of both the o- stem masculine and the a- stem masculine nouns. The result is an extremely messy system indeed, far more complicated than the late Proto-Slavic pattern, in part because the generalization hasn’t gone all the way—in a few cases, animate and inanimate masculine nouns still have identical suffixes. At least some nonstandard Czech dialects show the same developmental tendencies as the standard dialect in the animacy category, but their declensional systems now seem to be retreating from the earlier tendency to increase the grammatical scope of the distinction (Thomason 1976). The point of this rather elaborate description of the morphological development of the animacy category in several Slavic languages is that no safe predictions can be made about changes in standard versus nonstandard dialects. Not all the languages have changed the original morphological pattern in any way. In those languages that have changed, all the developments are variations on a theme—the morphological expansion, through analogic changes, of the animate/inanimate distinction. Some of these expansions simplify the overall system; some complicate it. There is no evidence that standard dialects are more likely to develop greater complexity than nonstandard dialects, in spite of the fact that the greatest increase in complication, in this small sample, is found in a standard dialect (and even that fact could be due simply to the relative difficulty of locating grammars of nonstandard dialects without traveling to the various countries).
< previous page
page_356
next page >
< previous page
page_357
next page >
Page 357 3. EASE OF LEARNING IN EXTERNALLY- AND INTERNALLY-MOTIVATED CHANGE The investigation of proposed vernacular universals leads directly to the issue of markedness, or ease of learning, because this would be a (or the) fundamental building block of a theory of universal tendencies of development. But ease of learning, important as it is believed to be in processes of language change, is a slippery concept. There is experimental evidence to support some claims of (for instance) perceptual saliency in phonetics, but for most areas of language change no experimental evidence is available. Markedness is crucial here, but of course markedness remains a controversial matter. Nevertheless, the two most widely used tests of markedness—range of occurrence in the world’s languages and age of learning by firstlanguage acquirers—remain useful as rough indicators of marked (harder to learn) versus unmarked (easier to learn) linguistic structures. Consider, for example, the English interdental fricatives: they are quite rare in the world’s languages, and they tend to be learned later than other fricatives in first-language acquistion of English. It is therefore reasonable to claim them as universally marked segments. By contrast, nearly every language in the world has an /n/ phoneme, and children learn it early: [n] is therefore considered to be a universally unmarked segment. In morphology, agglutinative morphology, being more transparent in its easily segmentable affix boundaries, is unmarked relative to flexional morphology, in which two or more morphemes are often bundled together into a single indivisible affix. In this section I will talk about ease of learning and markedness as if they were straightforward, but readers should keep in mind that many of the most important questions about these concepts have not yet been answered satisfactorily. The discussion must therefore be considered tentative. It is not true that all types of language change are linked to ease of learning, but probably the majority of nondeliberate changes do have something to do with it. For internally-motivated sound change, there is a now rich body of research on the diachronic role of phonetic context in both perception and production; in internal morphosyntactic change, analogic processes generalize or extend patterns in ways that often make learning the system easier. For contact-induced change, the overall picture is more complicated (see below). Ease of learning plays a role in proposals of vernacular universals because the simplifications that are predicted for nonstandard dialects are thought to make learning the dialects easier. Only some of the changes discussed in section 2 are clearly connected to ease of learning. The Montana Salish pharyngeals—which are universally marked—are losing ground with speakers and especially writers because they’re hard to hear, at least for speakers who are now English-dominant (i.e. all the remaining speakers of the language, including the most fluent native speakers in their daily lives). Most of the changes in the SerboCroatian oblique plural cases are syncretic, a process that arguably reduces the
< previous page
page_357
next page >
< previous page
page_358
next page >
Page 358 learner’s burden by reducing the number of morphological distinctions that must be learned; and the development of the Slavic animacy category in Novî SC is clearly simplificatory. The contact-induced change that introduced three new cases in Lithuanian, however, complicated the language’s case system, the AIPE and Tok Pisin transitive suffix cannot be viewed as a simplification of English verbal structure, and animacy developments in some Slavic languages—most strikingly in Standard Czech—are definite complications of the system. Ease of learning most obviously influences contact-induced change in cases of shift-induced interference, where imperfect learning plays a major role in the process of interference. It does not appear to be as important, and it is often quite irrelevant, in borrowing in the narrow sense of Thomason and Kaufman 1988, namely, under conditions of full bilingualism, where imperfect learning plays no role in the process (the discussion below and in section 4 is based in part on material in Thomason 2001, Chapter 4). Shift-induced interference comprises at least two, and often three, different types of process. First, secondlanguage speakers of a target language who are not fully bilingual in the target language (TL) may incorporate features of their native language, their L1, in speaking the TL. Changes in a TL that are carried over from a shifting group’s original L1 cannot be predicted to result in simplification of the TL if they become a permanent part of it. Sometimes, as in the example of the new Lithuanian cases that arose through interference from shifting Finnic speakers, the innovations in the TL complicate its system; in other instances, shift-induced interference results in simplification of particular structure points, as in the loss of grammatical gender in a Latvian dialect due to carryover from Livonian, a Finnic language which (like other Uralic languages) lacks gender and whose speakers shifted to the Latvian dialect (Comrie 1981:147). Second, imperfect learning by a group of speakers may involve a failure (or refusal) to learn certain TL features—especially marked features, namely, those that are harder to learn. Changes of this type do tend to simplify particular structure points in the TL, although a favorite and generally valid truism in historical linguistics must always be borne in mind: a change that simplifies the system in one place may well complicate it in another. So, for instance, recent learners of Montana Salish, all of whom have English as their L1, have failed to acquire the pharyngeal phonemes, which belong to a universally marked phonetic category and are in fact demonstrably hard to hear in this language. This failure is leading to the loss of all four pharyngeal phonemes, a clear simplification in the consonant inventory; but it is doubtful that this loss constitutes an overall simplification in the system, because the phonological effects of the pharyngeals on neighboring segments remain, especially vowel lengthening and vowel lowering, possibly adding one or more new vowel phonemes to the phonemic inventory. Most claims of overall simplification through imperfect group second-language acquisition (or, for that matter, through any other process of language change) must
< previous page
page_358
next page >
< previous page
page_359
next page >
Page 359 therefore be hedged so as to confine the claim to a particular area of the receiving language’s structure. The third type of process that is often, although not always, operative in shift-induced interference is adoption by the original TL community of a subset of the innovations introduced by the shifting speaker group. This typically happens when the shifting group merges with (part of) the original TL group to form a single speech community. This new merged community, at first, will feature both what might be called TL1, the language of original TL speakers, and TL2, the shifting group’s version of the TL, modified by carryover from their original L1 and failure to learn certain features of the TL. As part of the merger of the two originally separate speech communities, TL1 speakers may borrow some of the innovations that are present in TL2, thus forming a third, merged version of the TL, TL3. As with the learners’ failure (or refusal) to learn certain TL1 features, the innovations incorporated into TL3 are likely to be the least marked innovations, in this case because the original TL1 speakers are likely to have only passive knowledge of TL2 and may not be fully familiar with all the marked innovations present in TL2. This is by no means a hard-and-fast prediction, of course; language change is never wholly predictable. Other social, demographic, and linguistic factors (such as the relative sizes of the two speaker groups) also play a role in determining the final shape of TL3 (and indeed in the formation of TL2 as well). But in the formation of TL3, as in the failure-to-learn aspect of the formation of TL2, ease of learning is likely to be a significant factor. In borrowing in my narrow sense—that is, incorporation of features from one language the innovator knows well into another language that the innovator knows well—ease of learning is much less likely to play a major role. The reason is that the agents of change, the people who introduce the innovations into the receiving language, are fluent speakers of both the source language and the receiving language. Imperfect learning plays no role because there is no imperfect learning involved in the process. There are therefore no intrinsic barriers to adopting marked, harder-to-learn features from the source language, and such features are easy to find in borrowing situations—and much more common than in cases of shift-induced interference. This does not of course mean that all or even most structural features transferred in borrowing contexts are marked: unmarked, easy-to-learn features are also very common as interference features in these situations, because they are certainly very common in the source language (as in all languages). In addition, it may be that only some speakers of the receiving language also speak the source language fluently, and in such a case the chances for general adoption of marked structural features borrowed by fluent bilinguals will be sharply reduced. In any case, both marked and unmarked structures are common interference features in contact situations intense enough to permit structural borrowing. A fairly typical set of changes in intense borrowing situations is found in interference in Asia Minor Greek dialects from Turkish (see
< previous page
page_359
next page >
< previous page
page_360
next page >
Page 360 Thomason and Kaufman 1988:217–22). Some of the borrowed features are (locally, at least) simplifications, e.g. the replacement of the inherited interdental fricatives by stops, the general (although not absolutely complete) loss of noun–adjective agreement, and the general (though again incomplete) loss of grammatical gender. Other borrowings from Turkish arguably complicate Asia Minor Greek structure, among them the development of vowel harmony in some of the most-affected dialects. Still others neither clearly complicate nor clearly simplify Asia Minor Greek, e.g. the change from SVO to SOV word order and related word-order changes. Overall, then, great caution must be exercised in making predictions of simplification in contact-induced change. Ease of learning can safely be expected to play a significant role only in shift-induced interference, and even then only when shifting speakers fail (or refuse) to learn certain marked TL features and/or when original TL speakers fail (or refuse) to adopt some of the marked innovations from TL2 when the two speaker groups form a single speech community. This of course does not mean that the majority of contactinduced changes are simplificatory. It may well be the case that they are, at least as far as local simplification is concerned, i.e. simplification at a particular structure point as opposed to the receiving language’s grammar as a whole. But it may also be the case that most contact-induced changes complicate the grammar or are neutral in their effects, neither simplifying nor complicating the grammar—especially if we consider the grammar as a whole, not just one single structure point at a time. Similarly, internally-motivated change, even when it is clearly linked to ease of learning, is not predictably simplificatory, in large part because of the simplify-here-but-complicate-there truism mentioned earlier. Examples abound. One familiar case is the merger of the high front unrounded vowel phoneme /i/ and the high back unrounded vowel phoneme /ï/ in some Uralic and Mongolian languages. This merger complicates the vowel harmony pattern because stems with original /ï/ take back-vowel suffix harmony, while stems with original /i/ take front-vowel suffix harmony—an obvious complication for learners. As we saw in section 2, the internally-motivated developments of the animacy category in some Slavic languages, most strikingly in Czech, proceeded by perfectly ordinary analogic changes, but resulted ultimately in a much more complicated system than in late Proto-Slavic. 4. CONCLUSIONS: CAN CHANGE DUE TO UNIVERSAL TENDENCIES BE DISTINGUISHED FROM CONTACT-INDUCED CHANGE? The answer to this question is: yes, sometimes. The hedge is vital for two reasons. First, for the vast majority of linguistic changes, we cannot find a proximate or ultimate cause. The reason is that, given enough time, later changes will accumulate in large enough numbers to destroy the evidence
< previous page
page_360
next page >
< previous page
page_361
next page >
Page 361 that might have permitted us to pin down a specific cause for a particular change. And second, many changes have more than one cause, and one or more of the causes of a particular change may be internal while another, or others, may lie in language contact. The best we can do, therefore, is to lay out criteria for assessing the likelihood that language contact played a role in triggering a specific change. Here are the criteria that are (I believe) both necessary and sufficient to motivate a claim of contact-induced change (see Thomason 2001:91–5 for more detailed discussion). A precondition is that the receiving language must be considered as a whole, because structural interference will not be isolated in a system. (I focus on structural interference here because identifying lexical transfer is, by comparison, child’s play.) First, a (group of) source language(s) must be identified, and a case must be made for contact intense enough to justify a claim of structural interference. Second, it must be shown that the proposed source language and the proposed receiving language share several structural features—not absolutely identical features, because in contact-induced change there is very often alteration in transferred features, but features that are similar enough to be plausible candidates for a historical relationship through interference. Third, we must be able to show that the proposed receiving language did not always have the shared features; that is, we have to prove that the language has changed. Fourth, conversely, we must prove that the proposed source language has not changed—that is, that it had the shared features before it came into contact with the other language. Fifth and finally, we need to look for internal motivations as well, given the very real possibility of multiple causation. In other words, a given change, or set of changes, might well be due to universal tendencies and also to contact-induced change. If we are successful in satisfying criteria 1–4, not just for a single innovation in the receiving language but for a nontrivial number of independent innovations (i.e. innovations in different grammatical subsystems), then it is reasonable to claim that contact played a role in motivating the changes. If not, then not. Admittedly, there may be cases where one or more of the four main criteria can be omitted without weakening the argument hopelessly, but it will be harder to make a convincing case if there are gaps in the evidence. For instance, ancient contacts between Hittite, an Indo-European language, and Hattian, which had no known relatives and which was demonstrably in close contact with Hittite in the ancient Near East, are strongly suggestive of Hittite interference in Hattian (Goedegebuure 2007). Hattian was a verb-initial language, but Hittite was a consistent SOV language, as were other languages of the region. The appearance of certain SOV-related morphosyntactic features in Hattian, together with a paucity of Hittite loanwords, suggests that some Hittites shifted to the prestigious Hattian language, bringing along some easy-to-transfer features of their original L1. The case is not complete, however, because we have no historical evidence about the original word-order status of Hattian and no related languages to compare its
< previous page
page_361
next page >
< previous page
page_362
next page >
Page 362 structure to: we therefore cannot prove that Hattian has changed—that it didn’t always have the SOV-ish word-order features. Still, given the extreme difference in its basic ordering schema and related features, the case for Hittite interference is rather strong. The basic lesson, in any case, is that a role for language contact in motivating a change or set of changes can be firmly established. Establishing a role for drift, paradoxically, is often more difficult; but in many instances this too is possible, in a growing number of grammatical subsystems. What we find is that the very same changes, especially changes from relatively marked features to relatively unmarked features at particular structure points, can and do arise sometimes from internal causes and sometimes from external causes. This means that an investigation of proposed vernacular universals has only just begun if one can identify shared features in nonstandard (or, for that matter, standard) dialects, because many, most, or all proposed vernacular universals are likely to be universally unmarked. Therefore, if at least some of the dialects are or were in close contact with other languages, both internal and external causes are more likely than not; and often both together may account for the same change. Identical innovations may arise from universal structural tendencies, from pattern pressures internal to a particular language, from shift-induced interference, from borrowing, or from some combination of these. This is why I believe that it is unwise to ask whether universal tendencies or contact-induced changes are responsible for vernacular universals and other similar or identical innovations in closely related dialects/languages. The question embodies several presuppositions that I reject on the basis of much empirical counterevidence: that similar or identical changes can safely be assumed to arise from the same cause; that there is usually or always a single motivation for a given change; and that the processes underlying universal structural tendencies differ significantly from the processes underlying many contactinduced changes—specifically, as I have argued in this chapter, processes involving ease of learning are common to both. But I also reject the recurrent claims that contact-induced change, at least in cases of shift-induced interference, predictably simplifies structure. Peter Trudgill, for instance, seems to espouse this position when he says that simplification in ‘high-contact dialects’ is predictable in part because of the influence of second-language learners of a TL (e.g. Chapter 13, this volume; in Trudgill 2004, he presents a more nuanced theory, but there too his predictions are not valid in any general sense—see e.g. the commentaries in the same journal issue, most notably Hajek 2004 and Rice 2004). As noted in section 3, however, only two of the three processes involved in shift-induced interference are more likely than not to introduce simpler features, and even those features may simplify target-language structure only locally, not throughout the grammar. Moreover, borrowing in my narrow sense is no more likely to simplify the receiving language’s grammar than it is to complicate it.
< previous page
page_362
next page >
< previous page
page_363
next page >
Page 363 In the end, then, a one-size-fits-all theory is unlikely to succeed in accounting for the rich, and richly varying, body of data from language contact situations around the world. Linguists’ knowledge and understanding of contact phenomena are increasing rapidly, and with that increase comes a deeper understanding of the enormous complexity in this as in other areas of human linguistic behavior. REFERENCES Belić, A. 1909–1910. Zametki po čakavskim govoram. Izvestija Otdelenija Russkogo Jazyka i Slovesnosti Imperatorskoj Akademii Nauk 14(2): 181–266. Chambers, J.K. 2004. Dynamic typology and vernacular universals. In Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective , edited by B. Kortmann, 127–45. Berlin: Mouton de Gruyter. ———. 2006. Linguistic continuum from vernacular to standard . Paper presented at the conference World Englishes: Vernacular Universals vs. Contact-Induced Change, Mekrijärvi, Finland, 1–3 September 2006. Comrie, B. 1981. The Languages of the Soviet Union. Cambridge: Cambridge University Press. Fairbanks, G. 1977. Case inflections in Indo-European. Journal of Indo-European Studies 5: 101–31. Fancev, F. 1907. Beiträge zur serbokroatischen Dialektologie: der kaj-Dialekt von Virje, mit Berücksichtigung der Dialekte Podravina’s. Archiv für slavische Philologie 29: 305–89. Goedegebuure, P. 2007. Hittite-Hattian contacts . Paper presented at HistLing, University of Michigan, 23 March 2007. Hajek, J. 2004. Consonant inventories as an areal feature of the New Guinea-Pacific region: Testing Trudgill’s hypotheses. Linguistic Typology 8: 343–50. Ivić, P. 1958. Die serbokroatischen Dialekte: ihre Struktur und Entwicklung, Vol. 1: Allgemeines und die štokavische Dialektgruppe . ’S-Gravenhage: Mouton. Ivković, M. 1927. [A collection of folk tales from Resava.] In Srpske narodne pripovetke. Srpski Etnografski Zbornik 41, edited by V. Čajkanović. Rice, K. 2004. Language contact, phonemic inventories, and the Athapaskan language family. Linguistic Typology 8: 321–43. Senn, A. 1966. Handbuch der litauischen Sprache. Heidelberg: Carl Winter Universitätsverlag. Stanojević, M. 1911. Severno-Timočki Dijalekat. Srpski Dijalektološki Zbornik 2: 360–463. Svane, G.O. 1958. Die Flexionen in štokavischen Texten aus dem Zeitraum 1350–1400 . Aarhus: Universitetsforlaget. Thomason, S.G. 1976. Analogic change as grammar complication. In Current Progress in Historical Linguistics (ICHL 2), edited by W. Christie, 401–9. Amsterdam: North Holland. ———. 1977. A fragment of Serbocroatian declensional history. Folia Slavica 1(1): 124–55. ———. 1997. On mechanisms of interference. In Language and Its Ecology: Essays in Memory of Einar Haugen, edited by S. Eliasson and E.H. Jahr, 181–207. Berlin: Mouton de Gruyter. ———. 2001. Language Contact: An Introduction . Edinburgh and Washington, DC: Edinburgh University Press and Georgetown University Press. ———. 2003. Contact as a source of language change. In A Handbook of Historical Linguistics, edited by R.D. Janda and B.D. Joseph, 687–712. Oxford: Blackwell.
< previous page
page_363
next page >
< previous page
page_364
next page >
Page 364 ———. 2007. Language contact and deliberate change. Journal of Language Contact 1(1): 41–62. ———, and T. Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley and Los Angeles: University of California Press. Trudgill, P. 2004. Linguistic and social typology: The Austronesian migrations and phonemic inventories. Linguistic Typology 8: 305–20.
< previous page
page_364
next page >
< previous page
page_365
next page >
Page 365 Contributors David Britain is Senior Lecturer in Linguistics at Essex University in England. He came to Essex in 1993 after having spent two years as a Postdoctoral Fellow at Victoria University of Wellington in New Zealand. His research has mainly focussed on the linguistic consequences of dialect contact, much of it examining the new dialect that emerged in the British Fens following seventeenth-century reclamation, but he has also investigated new dialect formation in New Zealand English and the extent to which the Southern Chain Shift can account for diphthong change in British and Southern Hemisphere Englishes. He is editor of Language in the British Isles (CUP, 2007), is co-editor of Social Dialectology (Benjamins, 2003, with Jenny Cheshire), and is currently an Associate Editor of Journal of Sociolinguistics . Other work underway includes research on geographical aspects of sociolinguistics (such as the inadequacy of innovation diffusion models in accounting for the spatial spread of linguistic changes), as well as on a project led by Dr Kazuko Matsumoto investigating the birth and death of a colonial dialect of Japanese spoken in the Republic of Palau in the Western Pacific. J.K. Chambers is Professor of Linguistics at the University of Toronto. He is the author of Sociolinguistic Theory: Linguistic Variation and Its Social Significance (second edition, 2003), co-editor with Peter Trudgill and Natalie Schilling-Estes of The Handbook of Language Variation and Change (2002), and co-author with Peter Trudgill of Dialectology (second edition, 1998), as well as other books and scores of articles. General research includes studies of dialect acquisition, dialect topography, and linguistic variation. He works extensively as a forensic consultant and maintains a parallel vocation in jazz criticism, including the prizewinning biography Milestones: The Music and Times of Miles Davis (1998). Homepage www.chass.utoronto.ca/~chambers. Karen P. Corrigan has held lectureships at University College, Dublin, and the Universities of Edinburgh and York (U.K.). She was Principal Investigator on the ‘Newcastle Electronic Corpus of Tyneside English’ project (http://www.ncl.ac.uk/necte/) and is currently Professor of Linguistics
< previous page
page_365
next page >
< previous page
page_366
next page >
Page 366 and English Language at Newcastle University. She has recently published Syntax and Variation (Benjamins, 2005; with Leonie Cornips of the Meertens Institute, Amsterdam) and Creating and Digitizing Language Corpora, Volumes 1 & 2 (Palgrave, 2007; with Joan Beal and Hermann Moisl). She was awarded a Leverhulme Trust Research Fellowship (2000–2002) during which some of the research for her chapter in the present volume was undertaken and she is currently completing an Arts and Humanities Research Council-funded project that will culminate in Irish English, Volume 1: The North of Ireland to be published by Edinburgh University Press. Markku Filppula was awarded his PhD by the National University of Ireland (Dublin) in 1986. He is Professor of English at the University of Joensuu and Docent in English Philology at the University of Helsinki. He is also a member of the Finnish Academy of Science and Letters and currently Director of LANGNET, the Finnish Graduate School in Language Studies and of NordLing , the Nordic Network of Graduate Schools in Language studies. He has published widely on Hiberno-English, other ‘Celtic Englishes’, and language contacts in general. He is the author of Some Aspects of Hiberno-English in a Functional Sentence Perspective (1986) and The Grammar of Irish English: Language in Hibernian Style (1999), coauthor of English and Celtic in Contact (2008), and co-editor of The Celtic Roots of English (2002) and Dialects Across Borders (2005). He wrote the chapter on Irish English morphology and syntax for A Handbook of Varieties of English, Vol. 2: Morphology and Syntax (ed. by B. Kortmann, E. Schneider et al., Mouton de Gruyter, 2004) and on the emergence of the ‘Celtic Englishes’ for The Handbook of the History of English (ed. by A. van Kemenade and B. Los, Blackwell, 2006). Sue Fox was awarded her PhD in 2007 by the University of Essex. Her research reflects her interest in language variation and change and the social processes that bring about language change, particularly in a multi-cultural urban environment such as London. Her PhD work explores the social interaction among different ethnic groups using an ethnographic approach which combines Communities of Practice and Social Network Analysis frameworks. Sue is currently employed at Queen Mary, University of London where she has worked as a researcher in two projects funded by the Economic and Social Research Council (2004– 2007 and 2007–2010), both of which focus on London English. Elaine Gold teaches Linguistics at the University of Toronto and at Queen’s University in Ontario, Canada. She has been very active in the field of Canadian English: she teaches an annual course on Canadian English, co-organized the first conference on Canadian English in 2005, is co-editor of a special issue of the Canadian Journal of Linguistics on Canadian English, co-edits an annual volume of student Working Papers on Canadian English, and is directing the development of a research Web site on
< previous page
page_366
next page >
< previous page
page_367
next page >
Page 367 Canadian English. Dr. Gold has presented papers and published on the topics of Canadian English and Language Contact, including papers on the function and frequency of the discourse particle eh?, the history of Canadian English orthography, borrowings from Yiddish into Canadian English, and topics in YiddishHebrew language contact. She is currently researching the Bungi English-Cree dialect of Manitoba and is writing a book chapter overview of sociolinguistic research in Canada. Juhani Klemola is Professor of English Philology at the University of Tampere and Visiting Research Fellow at the University of Leeds. He studied English Philology, General Linguistics, and Philosophy at the University of Helsinki, where he obtained his BA in 1986 and MA in 1988. He received his PhD in Language and Linguistics from the University of Essex in 1996. Klemola has held academic positions at the University of Helsinki (1988–1989), the University of Joensuu (1984–1994), the University of Leeds (1994–1998), the University of Tampere (1998), and the University of Helsinki/Academy of Finland (1998–2002). In 2002, he was appointed Professor of Modern English Language at the University of Vaasa, and in 2003, he was appointed to his present position at the University of Tampere. His main research interests lie in the areas of dialect syntax, contact linguistics, and historical dialectology. Klemola is co-author of English and Celtic in Contact (2008) and editor/co-editor of Types of Variation (2006), Dialects Across Borders (2005), The Celtic Roots of English (2002), and Speech Past and Present (1996). He has published widely in the areas of English dialectology, Celtic Englishes, and historical linguistics. Bernd Kortmann is Full Professor of English Language and Linguistics at the University of Freiburg, Germany. His publications include four monographs, three edited volumes, a two-volume handbook-cumCD-ROM on the phonology and morphosyntax of the varieties of English around the world (2004), and about 70 articles and reviews in journals and collective volumes. His research interests include semantics, semantic change, grammaticalization, typology, history of linguistics, and English grammar. His main research interest over the last years has been the grammar of nonstandard varieties of English, especially from a typological perspective. As a result of his research efforts, three volumes on syntactic variation in English and Germanic dialects have been published in 2004 and 2005, among them the multimedia reference work, A Handbook of Varieties of English. Since 1996, he has been the editor of the Mouton de Gruyter series Topics in English Linguistics (together with Elizabeth Traugott, Prof. em. at Stanford University). He is the director of the graduate programme Master of European Linguistics, Chairman of the Board of the Language Teaching Centre at the University of Freiburg, and member of the Board of the Hermann-Paul Centre of Linguistics.
< previous page
page_367
next page >
< previous page
page_368
next page >
Page 368 Salikoko S. Mufwene is the Frank J. McLoraine Distinguished Service Professor of Linguistics and the College at the University of Chicago, where he also serves on the faculty of the Committee on Evolutionary Biology and as an associate of the Department of Comparative Human Development. He has taught principally at the University of the West Indies, Mona campus (1980 and 1981), at the University of Georgia (1981–1991), and at the University of Chicago (January 1992 to date). His current research is on language evolution, including language birth and death, all approached from the perspectives of population genetics and macroecology. Globalization has also featured prominently in his most recent work. He is the author of The Ecology of Language Evolution (Cambridge UP, 2001), Créoles, écologie sociale, évolution linguistique (l’Harmattan, 2005) and Language Evolution: Contact, Competition and Change (Continuum, 2008), editor and co-translator of Robert Chaudenson’s Creolization of Language and Culture (Routledge 2001), and lead editor of AfricanAmerican English (Routledge, 1998) and Polymorphous linguistics: Jim McCawley’s legacy (MIT Press, 2005), among others. He is also the series editor of the Cambridge Approaches to Language Contact. Terttu Nevalainen is Professor of English Philology at the University of Helsinki and the Director of the Research Unit for Variation, Contacts and Change in English (VARIENG). Her research focuses on historical sociolinguistics, language change, and Modern English before 1900. Dr. Nevalainen’s publications include “Early Modern English Lexis and Semantics”, in Roger Lass (ed.) The Cambridge History of the English Language, Vol. 3 (CUP 1999); Historical Sociolinguistics (Longman 2003; with Helena Raumolin-Brunberg); and An Introduction to Early Modern English (EUP 2006). She is a co-editor of a number of publications, including History of Englishes (Mouton de Gruyter 1992), Sociolinguistics and Language History (Rodopi 1996), Placing Middle English in Context (Mouton de Gruyter 2000), Gender in Grammar and Cognition (Mouton de Gruyter 2000), Types of Variation: Diachronic, Dialectal and Typological Interfaces (Benjamins 2006) and The Dynamics of Linguistic Variation: Corpus Evidence on English Past and Present (Benjamins 2008). She is one of the original compilers of the Helsinki Corpus of English Texts (HC) and the director of the project “Sociolinguistics and Language History”, which has produced the Corpus of Early English Correspondence (CEEC), and in collaboration with Ann Taylor and Anthony Warner (University of York), its tagged and parsed version (PCEEC), released in 2006. Terence Odlin is Associate Professor of English at Ohio State University. His publications include one book, Language Transfer, an edited volume, Perspectives on Pedagogical Grammar , two co-edited volumes, Language Contact, Variation, and Change and Studies of Fossilization in Second Language Acquisition, as well as several articles, book chapters, and
< previous page
page_368
next page >
< previous page
page_369
next page >
Page 369 reviews. His current research interests focus on semantic and pragmatic factors in cross-linguistic influence. Heli Paulasto (née Pitkänen) is Researcher in English Linguistics at the University of Joensuu, Finland, where she was awarded her MA degree in 1998 and her PhD in 2006. She was appointed as a student researcher at the Finnish Graduate School for Language Studies, LANGNET (1999–2002), and as a researcher in two research projects funded by the Academy of Finland (2000–2003 and 2005–2008). In her research, she combines variationist, historical, and contact linguistics: Paulasto’s doctoral dissertation Welsh English Syntax: Contact and Variation (2006) focuses on the outcomes of language contact and regional and diachronic patterns attested in the Welsh varieties of English. Dr Paulasto is co-author of English and Celtic in Contact (2008) and co-editor of The Celtic Roots of English (2002), both together with Markku Filppula and Juhani Klemola. Daniel Schreier has taught and lectured in Switzerland, New Zealand, Ger-many, and the United States, and he is now Associate Professor at the University of Zurich, Switzerland, where he holds a Chair of English Linguistics. His research interests include the sociolinguistics of English as a World Language, language variation and change, contact linguistics and new-dialect formation (based on fieldwork on the South Atlantic Islands of Tristan da Cunha and St Helena), the formation and development of New Zealand English in the early twentieth century, effects of geophysical isolation on language change, as well as synchronic and diachronic aspects of consonant change in English. He is the author of Isolation and Language Change (2003) and Consonant Change in English Worldwide: Synchrony Meets Diachrony (2005; both published with Palgrave Macmillan), St Helenian English: Origins, Evolution and Variation (2008, Benjamins), and co-author (with Karen Lavarello-Schreier) of Tristan da Cunha: History People Language (2003; London: Battlebridge). Peter Siemund is Professor and Chair of English Linguistics at the University of Hamburg. He earned his PhD at the Free University of Berlin in 1997 with a study on intensive and reflexive self -forms in English (Routledge). In 2002, he successfully defended his postdoctoral dissertation (Habilitation) on Pronominal Gender in Varieties of English (Routledge) at the Free University of Berlin. Peter Siemund has worked and published on reflexives and reflexive marking, systems of pronominal gender, tense and aspect, argument structure, interrogative constructions, speech acts and sentence types, as well as some minor topics. His work is firmly rooted in functional typology and includes synchronic and diachronic approaches as well as linguistic variation. Since 2003, he has been conducting a project on Hiberno-English: Variation and Universals in Contact-Induced Language Change within the Sonderforschungsbereich 538: Multilingualism , at the University of Hamburg.
< previous page
page_369
next page >
< previous page
page_370
next page >
Page 370 Benedikt Szmrecsanyi is Assistant Professor of English Language and Linguistics at the University of Freiburg, Germany. He received his PhD from the University of Freiburg with a thesis on morphosyntactic persistence in spoken English. His research interests include corpus linguistics, language variation and change, psycholinguistics, and dialectology and dialectometry in a typological perspective. Sali A. Tagliamonte is a Professor in the Linguistics Department at the University of Toronto, Canada. She is author of Analysing Sociolinguistic Variation (CUP) and co-author of African American English in the Diaspora: Tense and Aspect (Blackwell). She has published on African-American English, British, Irish and Canadian dialects, teen language, and television. Her ongoing research focuses on morphosyntactic and discourse-pragmatic features using cross-variety and apparent time comparisons in synchronic corpora to explore linguistic change. Sarah G. Thomason. After receiving her PhD in 1968, Sarah Thomason taught Slavic linguistics at Yale (1968–1971) and then general linguistics at the University of Pittsburgh (1972–1998); since January 1999, she has been at the University of Michigan, where she is now the William J. Gedney Collegiate Professor of Linguistics. Thomason has worked with the Salish & Pend d’Oreille Culture Committee in St. Ignatius, Montana, since 1981, compiling a dictionary and other materials for the tribe’s Salish language program. Her current research focuses on contact-induced language change and Salishan linguistics. Among her major publications are ‘Chinook Jargon in areal and historical context’ (1983), ‘Genetic relationship and the case of Ma’a (Mbugu)’ (1983), ‘Before the Lingua Franca: Pidgin Arabic in the eleventh century A.D’ (with Alaa Elgibali, 1986), Language Contact, Creolization, and Genetic Linguistics (with Terrence Kaufman, University of California Press, 1988), Contact Languages: A Wider Perspective (edited volume, John Benjamins, 1997), Language Contact: An Introduction (Edinburgh University Press and Georgetown University Press, 2001), ‘Pronoun borrowing’ (with Daniel Everett, 2005), and ‘Language contact and deliberate change’ (2007). Peter Trudgill has carried out research on dialects of English, Norwegian, Greek, Albanian, and Spanish and has published more than 30 books on sociolinguistics and dialectology. He was Professor of Linguistics at the University of Reading, England; Professor of Sociolinguistics at the University of Essex, England; Professor of English Linguistics at the University of Lausanne, Switzerland; and Professor of English Linguistics at the University of Fribourg, Switzerland. He is currently Emeritus Professor of English Linguistics at Fribourg University; Honorary Professor of Sociolinguistics at the University of East Anglia, Norwich, England; Adjunct Professor of Sociolinguistics, Agder University, Kristiansand, Norway; and Adjunct Professor in the Research Centre for Linguistic Typology at La Trobe University, Melbourne, Australia. He is a Fellow of
< previous page
page_370
next page >
< previous page
page_371
next page >
Page 371 the British Academy and has honorary doctorates from the University of Uppsala, Sweden; the University of East Anglia; and La Trobe University. He is writing a book to be entitled Sociolinguistic Typology: Language in Contact and Isolation. Donald Winford is professor of Linguistics at the Ohio state University. He did his undergraduate degree at King’s College, University of London, 1965–1968, graduating with First Class Honors in English. He completed his D. Phil. (Linguistics) at the University of York, England in 1972 with a dissertation titled “A Sociolinguistic Description of Two Communities in Trinidad.” His teaching and research interests are in creole linguistics, variationist sociolinguistics, contact linguistics, and African-American English, and he has published widely in those areas. His current research projects include a study of Gbe influence on the Surinamese creoles, and a sociolinguistic description of the African American speech community of Columbus, Ohio. His other main interest is the integration of linguistic and psycholinguistic approaches to contact phenomena, using Van Coetsem’s model of language contact as a basic theoretical framework. He is the author of Predication in Caribbean English Creoles (1993) and An Introduction to Contact Linguistics (2003). He served as president of the Society for Caribbean Linguistics from 1998 to 2000 and has been editor of the Journal of Pidgin and Creole Languages since August 2001.
< previous page
page_371
next page >
< previous page
page_372
next page >
page_372
next page >
Page 372 This page intentionally left blank.
< previous page
< previous page
page_373
next page >
Page 373 Name Index A Adger, D., 133, 139 Ainsworth, H., 185 Aitchison, J., 143 Algeo, J., 242 Allerton, D., 200n8, 200n17 Andersen, R.W., 226, 255 Anderson, J., 184 Anderwald, L., 40, 43, 69, 120 Ansaldo, U., 221, 224, 233, 245, 253 Ard, J., 333 Arnold, D., 157n4 Ash, S., 184 Auer, P., 314 B Bailey, C.-J.N., 282 Bailey, G., 67, 289–290, 298n8, 299n11 Bailyn, B., 286, 298n6 Bambgose, A., 293 Bao, Z., 221–222 Barbiers, S., 38 Bardovi-Harlig, K., 221, 224–225 Barnes, W., 219 Bauer, L., 180, 185 Baugh, J., 299n12 Bayley, R., 59, 63, 68 Beal, J.C., 133, 140, 142, 154, 157n11, 158n19, 199n4, 306, 311 Belić, A., 354–355 Bell, A., 59, 68, 200n6 Benincá, P., 325 Benton, R., 185 Berg, T., 329 Bhat, D.N.S., 253 Bhatt, R.M., 233, 245 Biber, D., 43, 52n5, 82, 98, 100n8, 241–242, 247–248 Bickerton, D., 4, 5, 271, 272, 273, 276, 277n3 Bisang, T., 327 Blain, E., 163–167, 169–171, 175 Blevins, J., 59 Bliss, A.J., 212, 214 Boas, F., 275 Bobda, A.S., 277n2 Borsley, R.D., 157n4 Bowerman, S., 184 Breivik, L.E., 83–84 Britain, D., 43, 69–70, 72, 104, 111, 120, 122, 124, 178–179, 182, 185, 194–195, 200n9, 201n18 Broadbent, J., 197 Brorström, S., 171–172 Brown, A., 221 Brunner, K., 68 Buccini, A.F., 286 Bunte, P., 275 Burridge, K., 47, 254, 337 Burrowes, A., 215, 217 Buthelezi, Q., 275 Bybee, J., 221, 257 Byrd, D., 184 C Chambers, J.K., 1–3, 10, 14, 16, 19–22, 31, 34, 36–37, 43, 57–59, 68, 74, 80, 98, 103–105, 109–111, 113– 114, 120, 163–165, 172, 177–178, 199, 208, 221, 231, 255, 304–311, 319, 325, 337, 350–351 Chaudenson, R., 280, 286, 297 Cheshire, J., 69, 72, 104, 120, 200n14 Childs, B., 4, 59–60 Chomsky, N., 5, 136, 327 Christian, D., 59, 66, 68–71, 104, 288 Clark, H., 199n3 Clarke, S., 163, 287 Claudi, U., 330
< previous page
page_374
next page >
Page 374 Claxton, A., 182 Comrie, B., 144, 146–147, 156, 157n6, 157n10, 157n11, 158n20, 251, 257, 326, 333, 358 Cook, V., 144 Corbett, G.G., 96, 99, 100n2, 326 Cornips, L., 156, 158n18 Cox, H., 183 Craig, H., 183 Crawford, W., 82, 97, 100n3 Croft, W., 6, 316, 333, 335, 339–340, 343 Cruttendon, A., 179 Crystal, D., 186, 284 Curme, G., 68, 105 Curnow, T.J., 331 Cutler, C., 185 Cysouw, M., 41, 52n4 D Dahl, Ö., 257, 330 De Graff, M., 48, 178, 295 de Klerk, V., 299n13 Dekeyser, X., 138 Denison, D., 100n7 Deterding, D., 186, 221 Doherty, C., 134 Doron, E., 25–26 Doyle, A.C., 57 Dreyfuss, G., 146 Dube, N., 70, 104, 288 Duffield, N., 133, 157n8 E Edwards, V., 120, 141, 234–235 Eisikovits, E., 69, 71 Ellegård, A., 217 Ellis, A.J., 109 Ellis, R., 332 Elsness, J., 255 Elworthy, F.T., 216–217, 219 Epsteins, S.D., 207 Evans, D.S., 258n2 F Fabb, N., 135, 157n4 Fairbanks, G., 352 Fancev, F., 354 Fasold, R., 59–60, 183 Feagin, C., 69, 71, 104 Filppula, M., 114, 134, 137, 157n11, 172–173, 212–214, 224, 232, 243, 245, 248, 251, 269–270, 298n4, 343–344 Finegan, E., 82, 87, 241 Finlay, C., 134 Fischer, D.H., 298n6 Fischer, O., 84, 89, 96, 136–138, 157n5 Fisher, J., 180, 183 Flanigan, B.O., 174 Flemming, E., 184 Flynn, S., 207 Fodor, J., 156 Foley, J.A., 233 Foley, K., 183 Foley, M., 69 Fong, V., 253 Forby, R., 319 Forsström, G., 69, 105 Fosler-Lussier, E., 200n6 Foulkes, P., 180, 184 Fox, B., 156, 157n5 Fox, S., 183, 192–195, 200n13, 200n14 Fox Tree, J., 199n3 Fries, C.C., 103 G
Gachelin, J.-M., 168, 171, 254–255 Garofolo, J., 200n7 Gaskell, M., 183 Gass, S., 333 Geisler, C., 134 Gick, P., 197, 200n9 Girand, C., 200n6 Givón, T., 272, 333 Gnanadesikan, A., 197 Goedegebuure, P., 361 Gonzales, A., 234 Gonzalez-Diaz, V., 314 Goodluck, H., 133, 139, 157n8, 157n13 Gordon, E., 64, 74 Greenbaum, S., 89, 135, 241 Greenberg, J.H., 39 Grieve, H., 183 Guilfoyle, E. , 133 Guler, N., 275 Gupta, A.F., 299n14 Guy, G., 59–60, 63 Guzzo, S., 194–195 H Haegeman, L., 135, 156 Hagège, C., 48 Hajek, J., 362 Halle, M., 180 Harrington, S. , 133 Harris, J., 134, 179, 206, 211–213, 228
< previous page
page_374
next page >
< previous page
page_375
next page >
Page 375 Harris, W.A., 289 Haspelmath, M., 6–7, 37, 326, 335 Haugland, K.E., 81, 87, 89 Hay, J., 74, 82, 180, 185 Hazen, K., 68, 70, 74 Healy, A., 180, 183 Heine, B., 330, 332, 336 Heinecke, J., 250 Helms-Park, R., 273, 276 Hempel, C.G., 329 Hendrick, R., 28–29 Henry, A., 134, 325 Henry, P.L., 211 Herrmann, T., 43, 133–135, 137, 142, 146, 149, 155–156, 157n5, 157n11, 157n14 Herzog, M., 104 Heselwood, B., 179–180, 197, 199 Hickey, R., 285 Hirsch, A.R., 290 Ho, M.L., 167, 206, 220–221, 226, 230, 233, 239, 244, 253–255, 274 Hock, H.H., 281 Hogg, R.M., 284 Holm, J., 280–281 Holmes, J., 59, 68, 185 Hope, J., 157n5, 307–310, 313, 315, 320 Hopper, P., 180, 330 Horvath, B.M., 105, 109–110, 116–117, 126–127 Horvath, R.J., 105, 109–110, 116–117, 126–127 Hosali, P., 48 Huddleston, R., 82, 98, 248 Hudson-Ettle, D.M., 240 Hundt, M., 97, 100n8, 252, 255–256, 258n3 Hünnemeyer, F., 330 I Idsardi, W., 180 Ihalainen, O., 114, 141, 216, 227n1, 318 Isaac, G., 250–251 Ivić, P., 353 Ivković, M., 354 J Jago, F., 217 Jarvis, S., 265–266, 268–273, 275–276, 277n1 Jespersen, O., 105, 137, 239–240 Johansson, S., 82, 241 Johnson, W., 200n9, 201n18 Johnston, P., 135 Jones, M., 199n4 Joseph, B.D., 281, 312 Joyce, P.W., 237 Jurafsky, D., 200n6 K Kachru, B.B., 67, 233, 245, 253, 292–293, 296, 297n1, 299n14 Kallen, J., 211, 285 Kaufman, T., 4, 11, 265, 281–282, 297, 351–352, 355, 358, 360 Kautzsch, A., 298n8 Keating, P., 184 Keenan, E.L., 144, 146–147, 157n6, 326, 333 Keene, D., 313 Keesing, R., 225 Keiser, S.H., 288 Kendall, M., 275 Khan, F., 59, 67–68 Kim, J.O., 44 Kiparsky, P., 208 Kirk, J.M., 134 Klee, C., 275 Klein, W., 48, 224, 333 Klemola, J., 15, 114, 217, 219, 227n1, 232, 298n4 König, E., 328, 330, 342 Kortmann, B., 3, 33–36, 38, 44, 47, 49, 80–82, 98, 104, 239, 246, 254, 258n4, 270, 304, 307, 319, 325, 337
Kós-Dienes, D., 217 Krashen, S., 332 Kroch, A., 20 Kurath, H., 185 Kuteva, T., 156, 157n10, 158n20, 330, 332, 336 L Labov, W., 24, 52n1, 59–60, 67–69, 104, 138, 153, 158n17, 183, 185, 289, 299n12, 315 Laitinen, M., 84 Lamont, G., 171 Lass, R., 59, 183–185, 311, 313 Lee, H., 67 Leech, G., 82, 89, 135, 241, 255–256 Lefebvre, C., 218, 226 Lehmann, C., 330 Lenneberg, E., 315
< previous page
page_375
next page >
< previous page
page_376
next page >
Page 376 Le Page, R.B., 214–215, 297 Lim, L., 233 Ling, L.E., 221 Lodge, K., 182 Lombardi, L., 7, 196 Lüdtke, H., 284 Luick, K., 68 M Macaulay, R.K.S., 237 Mair, C., 2, 252, 255–256, 246n5 Maroldt, K., 282 Martínez-Insua, A.E., 80, 84, 92, 100n6 Martohardjono, G., 207 Matthews, P.H., 80 Maxwell, D., 156 Maynor, N., 289, 299n11 Mbangwana, P., 277n2 McArthur, T., 75 McCafferty, K., 270 McCarthy, J., 180, 197 McCloskey, J., 133, 139, 147, 157n8 McDavid, R., 185 McIntosh, C., 86 McMahon, A., 180 McWhorter, J., 45, 47–48, 269 Meechan, M., 69 Mesthrie, R., 53, 133, 137–138, 140–141, 146–147, 157n5, 157n11, 167, 199n1, 254–255, 266–267, 270, 274, 304, 337 Meurman-Solin, A., 138 Michael, I., 86–87 Migge, B., 219, 274 Miller, J., 146, 149, 158n16, 249, 251–252 Milroy, J., 309, 313 Milroy, L., 74 Minkova, D., 137 Montgomery, M., 70–71 Morris, R., 83 Mueller, C.W., 44 Mufwene, S.S., 48, 104, 183, 217, 280–291, 294–295, 297, 298n9 Mühlhäusler, P., 4, 47, 280, 312 Mulligan, D.A., 163–166, 168–169 Murray, J.A.H., 70, 114 Murray, L., 89–90 Murray, R., 310 Myers-Scotton, C., 227 Myhill, J., 184 N Neilson, W., 133, 139 Nelson, C.L., 296, 297n1 Neu, H., 62 Nevalainen, T., 80, 84–87, 94, 96, 99, 100n4, 104, 170, 315 Newmeyer, F.J., 325 Newton, C., 197–198 Niles, N.A., 214, 218 O Ó Siadhail, M., 133, 139, 157n8, 236 O’Brien, R., 183 Ocampo, A., 275 Odlin, T., 206, 265–266, 269–272, 274–275, 285 O’Grady, W., 23 Ojanen, A., 182 Orgun, C., 180 Orton, H., 71 P Page-Verhoeff, J., 199n4 Pagliuca, W., 221, 257 Pargman, S., 285, 297 Parry, P., 237
Patrick, P., 59, 61, 63, 65, 68 Paulasto, H., 173, 232, 251, 269 Payne, D.L., 28, 31n2 Peirce, C.S., 329 Peitsara, K., 182 Penhallurick, R., 251 Perdue, C., 48, 333 Pérez-Guerra, J., 80, 84, 92, 100n6 Perkins, R., 221, 257 Peterson, P., 157n4 Pietsch, L., 332, 344 Pinker, S., 19–20 Platt, J., 167, 206, 220, 226, 233, 239, 244, 246, 253–255, 274 Poletto, C., 158n18 Policansky, L., 134, 141 Poplack, S., 63, 106, 111, 163, 298n8, 299n11 Postlethwaite, R., 89 Potter, S., 183 Poussa, P., 154, 157n10 Prince, A., 19–20, 197 Prokosch, E., 311 Pullum, G., 82, 98, 248 Pyles, T., 105 Q Quirk, R., 68, 89–91, 96, 135, 241, 247–248
< previous page
page_376
next page >
< previous page
page_377
Page 377 R Ramchand, G., 133, 139 Rampton, B., 185 Raumolin-Brunberg, H., 85, 94, 100n4, 315 Raymond, W., 180, 183, 200n6 Reaser, J., 47 Reh, M., 318 Reinhart, T., 328 Reuland, E., 328 Reynolds, D., 221 Rice, K., 362 Ringbom, H., 266, 268–269, 272, 274 Rissanen, M., 315 Ritchie, W.C., 220 Rohdenburg, G., 239 Romaine, S., 137–138, 141, 144–147, 149, 156, 157n5, 157n7, 157n13, 158n16, 281 Rupp, L., 199n4 Rydén, M., 171–172 S Sabban, A., 249, 251, 269–270 Samuels, M., 313 Sand, A., 126, 242, 244–246, 325, 337 Sanderson, S., 71 Santa Ana, O., 63, 68 Sasse, H.-J., 339 Schilling-Estes, N., 68, 71–72, 104 Schlauch, M., 211 Schmidt, R., 269 Schmied, J., 234, 239–240, 244–245, 253 Schneider, E., 57, 80–82, 98, 254, 270, 280–281, 293, 298n7, 299n10, 304, 306, 319, 337 Schreier, D., 21, 59, 63–67, 69–70, 73–74, 82, 104, 123–124, 178, 318 Schuchardt, H., 48 Schumann, J.H., 145 Schwartz, B.D., 208 Schütze, C.T., 158n18 Scott, S., 139, 157n8, 163–164, 166, 168–169 Sebregts, K., 180 Sedlatschek, A., 245 Selinker, L., 266, 268, 270 Senn, A., 352 Seppänen, A., 137, 157n11, 157n14, 158n16, 239 Seppänen, R., 239 Shackleton, R., 287 Sherrod, N., 183 Shorrocks, G., 182–183, 311 Sibayan, B.P., 234 Siegel, J., 225–226, 271–272, 274, 276 Siemund, P., 211, 213, 325, 328, 330–331, 337–339, 341–342, 346n2 Sigley, R., 146, 150 Singler, J., 223 Sivertsen, E., 188 Slobin, D., 246 Śmiecińska, J., 255 Smith, J., 69–71, 106, 111, 120, 133, 287 Smith, N., 248, 252, 255–256 Sorace, A., 144 Sprouse, R.A., 208 Stanojević, M., 354 Stassen, L., 37 Stauble, A., 270 Stebbins, T., 240 Stene, A., 178, 181, 186, 200n17 Stobie, M., 163, 166–167, 170–171 Stockwell, R.P., 137 Strang, B., 313 Sundby, B., 81, 87, 89 Svane, G.O., 354 Svartvik, J., 89, 135, 241 Szmrecsanyi, B., 33–36, 47, 49, 239, 246, 254, 307, 325, 337
next page >
T Tabouret-Keller, A., 297 Tagliamonte, S., 63, 69–72, 98, 106, 111, 120, 123, 127, 133–134, 142, 147–149, 154–156, 164, 287, 298n8, 299n11 Tallerman, M., 156, 158n15 Tarallo, F., 146 Tate, T.W., 298n7 Tay, M., 185 Taylor, B., 183 Temperley, D., 133 Temple, R., 63 Thomas, A., 2 Thomas, E., 60–61, 63, 67, 179, 185, 289, 298n8, 299n11 Thomason, S.G., 4, 11, 173, 255, 265, 281–282, 284, 297, 331, 346n4, 348–353, 355–356, 358, 360–361 Thompson, C., 183 Thompson, C., 183 Thompson, R.M., 234, 245, 253
< previous page
page_377
next page >
< previous page
page_378
next page >
Page 378 Thompson, S., 156, 157n5 Thorne, D.A., 237, 243, 250 Tieken-Boon van Ostade, I., 81, 84–86 Todaka, Y., 184 Torbert, B., 4, 47, 59, 61 Traugott, E.C., 105, 180, 218, 330 Tristram, H.L.C., 298n4 Troike, R., 299n12 Trudgill, P., 80, 185, 281–282, 287, 304–312, 314, 316–319, 338, 346n2, 362 Turner, L.D., 296 U Uffmann, C., 180, 196 Upton, C., 270, 337 V Vennemann, T., 180 Viereck, W., 199n4 Vincent, L., 45 Visser, F., 68, 105 W Wächtler, K., 242 Wade, R., 168 Wagner, S., 183 Walters, F., 163–169 Warren, P., 185 Washington, J., 183 Watermeyer, S., 183, 185 Weber, H., 167, 220, 233, 239, 244, 246, 253–255 Wee, L., 244–245, 253 Weinreich, U., 104 Wells, B., 197–198 Wells, J., 179, 305–306, 313, 316 Weltens, B., 234–235 Whaley, L.J., 245 Widdowson, J., 71 Williams, J., 215, 240, 246, 255 Williams, J.P., 215, 240, 246, 255 Winford, D., 215, 219–220, 223, 331 Wode, H., 246, 269 Wolf, H., 277n2 Wolfram, W., 4, 59, 60–63, 66–72, 75, 104, 183, 288, 298n8, Wood, P.H., 296, 298n7 Wrenn, C.L., 68 Wright, J., 182, 234 Wyld, H.C., 312–313 Z Zhiming, B., 233, 253 Zribi-Hertz, A., 136, 143
< previous page
page_378
next page >
< previous page
page_379
next page >
Page 379 Subject Index A Accessibility Hierarchy 13, 138–147, 153, 156, 157n14, 326, 333 acrolect 21–22, 29–31, 104, 114, 294 agreement, see also concord agreement mismatch 81, 84–85, 97–98 default singulars 1, 7, 13, 20, 24–25, 29–30, 36, 38, 57, 59, 68, 80–81, 84, 87–88, 91–100, 103–105, 109– 116, 119–127, 165, 170, 306, 308, 311 look-ahead agreement 24–25 notional plural 81, 87–88, 95–97, 100 number agreement 15, 20, 80–100 subject–verb agreement 20–24, 27–31, 80–89, 92, 100, 104 subject–verb nonagreement 12, 23–24, 28, 87, 95, 99, 165, 170 Agreement Hierarchy 99 allative case 270, 352 allomorphy 180–187, 189, 191, 195, 197–198, 200 analyticity 7, 45, 48–49 Angloversals 2, 33, 35–36, 39, 41–42, 49, 239, 241, 336, 346n5 animacy 171, 338 Animacy Hierarchy 245, 326 animacy in Slavic languages 354–356, 358, 360 Arabic 23, 25–26 ARCHER 84 areoversal 2, 33, 35, 37, 42 articles definite article 9–10, 180, 182–184, 188–190, 192–193, 195, 198, 199n4, 231, 241–246, 256, 258, 272–273, 277n3 indefinite article 180, 182–183, 186–189, 192, 199n4, 200n16, 272, 277n3, 330, 337 aspect 211, 212, 216, 221, 253–254, 256, 257, 271 B Barbadian Creole 5, 11, 207 basilect 20–21, 29, 104, 111, 114, 216, 219, 226, 267, 294 Biblical Hebrew 23, 25–26, 30 biconditional implication 39–41, 49 bilingualism 11, 64, 133–134, 269, 273, 358–359 borrowing 137, 331, 349, 353, 359–360, 362 with full bilingualism 358 British National Corpus (BNC) 97–98, 256 Buckie 69, 107–108, 110, 112–123 Bungi 10, 15, 163–175 C Celtic languages 231, 243, 251, 256, 282, 298n4 child language 1, 104, 137, 177, 180, 197, 208 clefting 24, 269, 275 complexity 5–7, 272, 297, 317, 331, 334, 346n3, 351–356 grammatical complexity 156, 255 morphosyntactic complexity 45, 47, 49 syntactic complexity 144, 157n7 concord, 100, 111, 337; see also agreement false concord 87, 89, 99 mixed concord 100n8 notional concord 89, 100 person-number concord 68
< previous page
page_379
next page >
< previous page
page_380
next page >
Page 380 plural concord 96, 98 singular concord 12, 88, 92, 96, 99 subject-verb concord 38, 68, 208, 211 subject-verb nonconcord 1, 10, 36, 38, 163, 165–166, 170, 172, 308, 311 consonant cluster reduction (CCR) 4, 12, 57–68, 74–75, 337 consonant cluster simplification 1, 10, 163–165, 170, 177, 208 constraint constraint hierarchy 71, 105, 109,–111, 121, 126, 165 constraint ranking 111, 119, 121, 126 existential constraint 111, 114–115 external constraint 122, 133, 142, 149–150, 153, 255, 334 internal constraint 12–13, 52n1, 60–61, 64–71, 75, 133–134, 147–149, 156, 180, universal constraint 12–13, 66, 109–110, 157n6, 226, 245, 324, 326, 327 Contrast Analysis 117–119, 121, 123, 126–127 co-occurrence pattern 49 copula 12, 21–22, 81–82, 89, 211–212, 225–226 copula absence 2, 36–37, 52n1, 208, 226, 307 copula deletion 2, 36, 52n1, 208, 337 Corpus of Early English Correspondence (CEEC) 84 Corpus of Early English Correspondence Extension (CEECE) 12, 81, 85, 91–92, 100n4 correlative structures 274 Cree 10, 163, 170–174 creolization 57, 64–65, 68, 75 cross-linguistic influence 265, 268 Cullybackey 107, 110, 112–113, 115–123 Cumnock 107–108, 110, 112–123 D Danish 239, 338 default singulars, see agreement definite article, see article determinative (construction) 96, 98, 100 dialect dialect contact 3, 9–10, 15, 58, 60, 64, 231, 256–257, 286, 294, 312–315, 336 dialect mixture 64, 312, 351, 313 dialect variation 30 dialectology 34, 62, 177 drift 12, 349–350, 357–360, 362 Dutch 239, 282, 286, 338 E ease of learning 8–9, 11, 349–350, 357–360, 362 ecology 104, 210, 233 of language evolution 282 sociolinguistic 225 English Early Modern English (EmodE) 83–84, 99–100, 135, 171, 210–216, 311, 137, 140, 146, 156 Middle English (ME) 68, 82–84, 89, 137–138, 181, 282, 284, 296, 311, 314, 319 Late Middle English 80, 84, 95, 255 Late Modern English 84, 90 Old English (OE) 20, 68, 83, 136–138, 180, 236, 282–284, 296, 305, 311, 319, 341 colonial varieties of English 256, 281, 285–286, 305, 316 precolonial English 281–282 postcolonial English 72 indigenized Englishes 9, 15, 206, 209, 280–281, 291–297, 298n2 varieties of English: African-American English 59, 63–64, 67, 220, 285 African American Vernacular English (AAVE) 24, 52n1, 183, 280, 287–292, 295–296, 298n8, 298n10, 298n11, 306 American English 43, 59, 61, 63–64, 82, 87, 98, 136, 200n6, 218, 282–284, 286, 291, 299n14, 313, 319, 335 American Southern English 288 Appalachian English 59, 68–70, 218, 286, 288, 342 West Virginian Appalachian English 15, 74 Barbadian English (Bajan) 212, 214–215
< previous page
page_380
next page >
< previous page
page_381
next page >
Page 381 British English 13, 58, 71–72, 177–201, 215, 232–233, 285–287, 290, 319, 335 Cameroonian English 277n2 Canadian English 10, 111, 163, 165 Cockney 182 East African English (EAfrE) 232, 244, 254, 233–234, 239–240, 244–246, 253, 258n3 Fenland English 69–70, 72, 182 Hawaiian Creole English 220, 225, 271, 273, 276 Hawaiian Pidgin English (HPE) 272, 274 Hebridean English 172, 249, 251, 269 Hiberno-English 214 Indian English (IndE) 12, 15, 59, 67–68, 74, 126, 133, 232–233, 238–240, 244–246, 253–254, 258n3, 267, 304, 335 Irish English (IrE) 5, 11, 40, 114, 134, 157n11, 206–207, 209–215, 223, 225, 227, 232, 235–237, 243–252, 254, 257, 277n2, 285–286, 304, 331–332, 335, 337, 343–344, 346n4 Jamaican Creole English 59, 65, 68, 215 Jamaican English 63, 126 Kanaka English 280 London English 72, 312–313 New Englishes 2, 13, 40, 43, 167, 172–173, 206, 227, 246, 280, 282 New Zealand English (NZE) 15, 40, 58–59, 63–65, 68, 74, 76, 82, 122, 124, 150, 184–185, 198, 254, 280, 282, 335 Nigerian Pidgin English 292, 341 Northern English 84, 121, 170, 254 Philippines English (PhilE) 167, 232–234, 240, 244–245, 253, 258n3 Scottish English 40, 70, 146, 163, 168–169, 170–171, 173,–174, 252, 285 Singapore English (SingE) 5, 11, 67, 126, 167, 183, 185–186, 198, 206–207, 209–210, 219–227, 232–234, 240, 244–245, 253–255, 258n3, 331, 335, 346n4 South African English 184–185, 335 South African Indian English (SAIE) 133, 167, 267, 274 South Armagh English 4, 13, 133–159 Southwest English dialects 120, 211–212, 214–219, 225, 227n1, 317–318, 338 St Helenian English 10, 15, 63, 65–66, 74 Trinidadian English 220–221 Tristan da Cunha English (TdCE) 13, 29, 70, 73–74, 122–124, 177–178, 318 Welsh English 2, 40, 232, 236–237, 242–244, 247, 249–251, 254, 269, 304, Estonian 239 evidentiality 277, 275 existential constraint, see constraint existential construction 70, 74, 80–102, 306–307 expletives 22–25, 29–30, 31n1 external constraint, see constraint F final devoicing 166, 170, 172–173 Finnish 38, 239, 268–271, 274–275, 277n1 formalist-deductive paradigms 327–329, 334 Frisian 338 functional typology, see typology 14, 323–327, 338, 345 functionalist-inductive paradigms 325, 327–329, 345 G Gaelic 170–174, 269, 285–286 Irish Gaelic 236 Scottish Gaelic 251, 257, 270 gender gender marking 22, 31, 326 gender variation 93–94, 150–156, 190–194 grammatical gender 314, 358, 360 pronominal gender 14, 49–50, 174, 324, 337–341 German 239, 266, 269, 282 glide 179, 195–197 glottal stop 7, 183–185, 189–190, 194–198, 199n4, 200n17
< previous page
page_381
next page >
< previous page
page_382
next page >
Page 382 grammaticalization 49, 84, 98, 100n7, 180, 256, 258, 307, 324–325, 330–332, 336–337, 342–345, 346n2 contact-induced grammaticalization 332, 344 grammatical person 13, 69, 73, 111–114, 116, 126 Gullah 280, 289–290, 295–297 Guyanese creole 215, 220 Guysborough Enclave 106, 108–110, 112–123 Guysborough Village 106, 108, 110, 112–113, 115–123 H habitual do 40, 42, 44, 50, 52n3, 212 Helsinki Corpus 84, 137 hiatus hiatus breaking 7, 178, 180–186, 194, 197–199 hiatus resolution 9, 177–199 hierarchy of individuation 337, 339–341, 343, 345 Hungarian 38, 239 I ICE Corpora 233–235, 237–240, 243 ICE East Africa 233, 237–239, 242, 244, 252–253 ICE Great Britain 233, 237–239, 244, 247, 252 ICE India 233, 237–239, 242, 244, 247–248, 252, 258n3 ICE Philippines 233, 237–239, 242, 244, 248, 252 ICE Singapore 233, 237–239, 242, 244, 252–253 imperfect learning 320, 358–359 implicational hierarchies 98, 326, 342 implicational tendency 34, 39, 49 indefinite article, see article interference foreign 349 shift-induced 353, 358–360, 362 structural 349, 352, 361 interlanguage (IL) 1, 11, 36, 48, 104, 177, 208, 224–225, 255, 266, 268–273, 275–277 internal constraint, see constraint intrusive /r/ 179, 180–181, 184–185, 187, 195–197, 200nn15–17 Irish 13, 25, 28, 133–134, 138–140, 142–143, 146–147, 156, 157n8, 157n13, 210–214, 232, 233, 236, 243, 245, 250–251, 269, 344 irregularisation 318–319 L L1 variety 10, 34–35, 39–44, 46–48, 64, 146, 208, 224, 231, 233–234, 239–240, 246–247, 254–257, 335– 336, 341, 345 L2 variety 10, 14, 33–35, 39–42, 44, 46–48, 64, 146, 208, 224, 231, 233–234, 239–240, 246–247, 254–257, 335–336, 341, 345 language acquisition 5, 14–15, 59, 184, 197, 207–209, 224, 246, 324, 329, 332–334 child language acquisition 197 first-language acquisition 5, 8, 144–145, 207, 220, 332, 336 second-language acquisition (SLA) 3, 144–145, 178, 206–207, 232, 265, 332, 337, 358 Language Biogprogram Hypothesis 1, 3, 4, 7, 11, 15, 37–38, 105, 177, 207–208, 223, 225, 271, 273 language/linguistic change 3, 12, 14, 68, 73–75, 85–86, 209, 246, 281, 295, 307, 309–310, 313, 316, 320, 329–330, 334, 336–337, 349–350, 357–360 contact-induced change 8–9, 11, 57, 206, 210, 227, 332, 349, 357–358, 360–362 externally-motivated 9, 282, 297, 351, 362 internally-motivated 9, 282, 286, 350, 353, 355, 357, 360, 362 simplifying changes 350, 360 language evolution 5, 281–283, 285–286, 289, 291, 295, 298n7 language transfer 134, 138, 265 Latin 284, 296 linking /r/ 179–180, 184–185, 187, 189–190, 195–197, 200nn15–17 local constraint, see constraint look-ahead mechanisms 4, 7, 22–31 M markedness 5–8, 11–12, 104, 196–197, 349, 357
< previous page
page_382
next page >
< previous page
page_383
next page >
Page 383 marked 5–8, 11, 19, 26–27, 29, 39, 58–59, 67, 104, 138, 143, 145, 150, 196, 199, 246, 308, 331, 357–360, 362 unmarked 3, 5–8, 11, 40–41, 51, 178, 196–197, 199, 201n18, 331, 357, 359, 362 mass/count distinction 304, 318, 339 Manitoba 10, 163, 170 Maryport 107–110, 112–123 mesolect 29, 59, 63, 65, 68 multiple causation 173, 350, 361 multiple negation, see negation Multivariate Analysis 13, 126 N naturalness 3–5, 7, 11–12 negation 13, 35, 43, 51, 119–121, 265–266, 269 multiple negation 2, 36–37, 42, 44, 47, 51, 57, 177, 208, 307–308, 314–316, 337 negative concord 2, 36, 42, 51, 308, 314 nondistinction he/she 166–167, 171–172 Norman French 281, 284 normative grammar 81, 85, 87, 90, 94, 96–97, 99, 100 North Preston 106, 108, 110, 112–123 Northern British 133–134, 155–156 Northern Irish 147–148, 156 Northern Subject Rule 13, 51, 82, 114, 117 nouns of measurement absence of plural marking in 10, 15, 231, 234–241 number-transparent noun 98 O obligatory categories 275, 277n4 Old Norse 281 Orkney English 171 P perfect after -perfect 50, 227, 270–271, 277n2, 332, 343 ‘be’ perfects 10, 42, 44, 50, 164, 168–172, 174 completive 50, 218–219, 221, 223 ‘continuative’/’extended’ 213 experiential 221–222 hot news 212–214 medial-object 332, 337, 343–344 resultative 212, 214, 218–219 phyloversal 2, 33 pidginisation 57, 64, 75 postmodification 135, 143, 156 prescriptive grammar, see normative grammar prevocalic context / position 9, 68, 75, 180–193, 195, 199n3, 199n5, 199n9 progressive progressive aspect 211–212, 214, 216, 218–219, 221, 223 progressive constructions 10, 164, 167–169, 171–174 progressive form 10, 50, 231, 247–257, 258n5, 343 R Red River Settlement 163 reflexive markers 14, 324, 330, 338, 341–343 reflexivity 338–343 regularisation 7, 9, 11, 68, 70–72, 255–256, 306–307, 310–317, 337 of hiatus resolution system 177–199 conjugation regularization 1, 19–20, 36–38, 306, 309 relative clauses ancillary relativization strategy 135–138, 140, 143, 146–147 nonrestrictive relative 51, 135, 137–138, 154 relative clause formation 143–144 relativization 35, 42–43, 51–52, 157n13, 267, 274 restrictive relative 135, 137–138, 154 restructuring (processes) 9, 11, 64–65, 67, 133, 225, 227, 273, 281–282, 284, 286–290, 292, 294–296, 299n6 resultatives 9, 343–344 resumptive pronoun 33, 52, 136 S Samaná 13, 69, 106, 111, 113–114, 117–118, 120–123, 126
semantic maps 17, 324, 326, 335, 337, 339, 343–345 shift varieties 14, 304, 332–333, 335–336, 345, 346n4
< previous page
page_383
next page >
< previous page
page_384
next page >
Page 384 simplification 9, 11, 66, 81, 170, 194, 224–225, 239, 256, 284, 311–312, 314–316, 320, 345, 346n3, 350, 357–358, 360, 362 sociocultural contact 9, 178, 185, 194, 196, 198 source language 12, 135, 138–140, 359, 361 Spanish 23, 25–26, 30, 126, 266, 269–270, 272–274 standardization 125, 155, 351, 352, 309–310, 313, 315 subject type hierarchy 98–99 subject–verb inversion 83 subordinating and 137–138, 140, 142, 146 substrate/substratum 13, 68, 133, 140, 156, 227 Celtic 243, 250, 254, 256 Gaelic 171, 173, 236 Irish 210, 245 Welsh 2, 251 substratal input 209, 225–226, 245 substrate influence 11–12, 67, 206, 216, 219–221, 226, 237, 265–279, 285, 298n9, 252, 223, 225, 227, 239, 244, 247, 255 substrate languages 11, 209–210, 217, 221, 225–226, 233–234, 244, 257, 266, 268, 271, 273, 275, 288, 292–294, 297, 298n9 substrate transfer 65, 143 superstrate 209–210, 227, 276 superstrate influence 4, 277n2 superstrate input 210, 225–226 Survey of Anglo-Welsh Dialects (SAWD) 237, 250–251 Survey of English Dialects (SED) 218, 232, 235–238, 240–241, 243–244, 247, 249–252 T target language 135–136, 225, 265–266, 268, 270, 276, 283, 286, 332, 358, 362 tense future 6, 211, 214–216, 219, 221, 223 past 210, 214, 223 present 210, 214, 223 tense-aspect systems 11, 14, 35, 39, 40, 42, 44, 50, 171, 209–210, 214–215, 219–221, 223–225, 227, 324, 329, 330, 333, 343 there construction, see existential construction Tiverton 107, 110, 112–123 TMA system 207, 209, 215, 224 typology 104, 177, 323 cross-linguistic typology 43, 337 functional typology 14, 34, 323–327, 338, 345 language typology 246, 333–334 sociolinguistic typology 20, 304–320 typoversal 2, 38, 52n2 U universal constraint, see constraint Universal Grammar (UG) 5, 7–8, 207–208, 223, 327, 333 unrestricted versal 49 V VARBRUL 99 variable grammar 94–95 varieties of English, see English varioversal 2, 33, 35, 40–41, 49 verb irregular verbs 19, 30, 38, 40, 44, 57, 308, 312–314, 319 levelling of irregular verb forms 1, 36, 41–42, 50–51, 309 serial verbs 52, 226, 271, 273–374, 276 stative verbs 10, 217, 231, 247–256, 258 transitive/intrasitive infinitives 317–318 VSO languages 28–30, 31n2 W was/were levelling 15, 57, 59, 68–76, 84, 93, 96, 100, 110, 120–121 Welsh 2, 28, 158n15, 232–233, 237, 243, 250–251 Middle Welsh 258n2 West African languages 206, 210, 224, 226 Wheatley Hill 107–113, 115–123 Wincanton 107, 110, 112–113
< previous page
page_384
next page >
< previous page
page_385
Page 385 word order 6, 29, 30, 35, 45, 52, 158n15, 213, 266, 272, 332–333, 360–362 World Atlas of Language Structures 37 World Atlas of Morphosyntactic Variation in English (WAMVE) 239–240, 246, 253–254, 257, 258n4 Y York 21, 29, 63, 69, 72, 106–113, 115–125, 178
< previous page
page_385