Variation and Change in Postcolonial Contexts [1 ed.] 9781443884938, 9781443876117

Citation preview

Variation and Change in Postcolonial Contexts

Variation and Change in Postcolonial Contexts Edited by

Rita Calabrese, J. K. Chambers and Gerhard Leitner

Variation and Change in Postcolonial Contexts Edited by Rita Calabrese, J. K. Chambers and Gerhard Leitner This book first published 2015 Cambridge Scholars Publishing Lady Stephenson Library, Newcastle upon Tyne, NE6 2PA, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2015 by Rita Calabrese, J. K. Chambers, Gerhard Leitner and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-7611-9 ISBN (13): 978-1-4438-7611-7

CONTENTS

Preface ........................................................................................................ ix Introduction ................................................................................................. 1 Post-colonial Convergence and Divergence Rita Calabrese Section I: Variation and Change in North America Convergence and Divergence in Canadian English ................................... 17 J.K. Chambers Chapter One ............................................................................................... 21 Saying “Tomato” in Postcolonial Canada J.K. Chambers Chapter Two .............................................................................................. 31 A Corpus-based Analysis of some Canadianisms of French Origin in Canadian English Mirko Casagranda Chapter Three ............................................................................................ 41 Heritage Italian in Toronto: Analysis of Verb Variation in a Spoken Corpus Maria Parascandolo Section II: Variation and Change in Africa Towards a Standard West African English ................................................ 57 Gerhard Leitner Chapter Four .............................................................................................. 61 A Pilot Study of Acoustic Features of Word-final Affricated /t/ and /ts/ in Educated Ghanaian English Thorsten Brato

vi

Contents

Chapter Five .............................................................................................. 79 De, Tone and Property Items in Nigerian Pidgin Maria Mazzoli Chapter Six .............................................................................................. 109 Postcolonial Englishes in an EFL Context Gerardo Mazzaferro Section III: Variation and Change in Asia Multilingualism and Standardization in South Asia ................................ 123 Rita Calabrese Chapter Seven.......................................................................................... 127 The Dynamics of Obligation and Necessity in New Englishes: The Case of have to in ICE Gabriela Diaconu Chapter Eight ........................................................................................... 149 Language Formation and Change in a Complex Multilingual Context Rita Calabrese Section IV: Variation and Change in Insular Areas English in the Island Worlds ................................................................... 163 Gerhard Leitner Chapter Nine............................................................................................ 165 Pluricentricity and Multiple Layers of English in Different Habitats Gerhard Leitner Chapter Ten ............................................................................................. 191 Resignifying Standard English in Marlene NourbSe Philip’s “She tries her tongue, her silence softly breaks” Roberto Masone Chapter Eleven ........................................................................................ 209 Semantic Change, Intersubjectivity and Social Knowledge in The Sydney Morning Herald Katherine E. Russo

Variation and Change in Postcolonial Contexts

vii

Section V: Variation and Change in New Media Global Variants in New Media Discourse ............................................... 229 Rita Calabrese Chapter Twelve ....................................................................................... 233 From English to Twenglish: A New Language Variety? Roberta Facchinetti and Paola-Maria Caleffi Chapter Thirteen ...................................................................................... 255 “Fulling di spies”: Promoting Bilingualism in Jamaican Newspapers and Blogs Eleonora Federici and Manuela Coppola Glossary ................................................................................................... 271 Bibliography ............................................................................................ 277 Contributors ............................................................................................. 311 Index ........................................................................................................ 317

PREFACE

The present volume addresses recent issues concerning language change and standardization as a result of language contact in postcolonial settings. The authors are experts in several areas of specialization, and they were asked to elaborate on those aspects of language variation and change that characterize the emergence of new varieties in specific geographical areas. The areas represented are North America, Africa, Asia and the insular areas of Australia, Trinidad and Tobago. The contributors adopt different, though interrelated, research procedures ranging from linguistic diagnostics and related methodologies to the most accredited interpretative theories on the evolution of New Englishes. Since new media represent some of the most frequently used channels for conveying new linguistic habits as well as new languages, we have included a specific section of the volume on the use of emerging varieties of English in new media. A special focus has also been given to those new varieties of Philippine and Nigerian English spoken in a noncanonical post-colonial context represented by the city of Turin in Italy. The result is a collection of studies that illuminate issues of language variability from different perspectives with the aim of contributing to the long debate on language contact, diversification, speciation and standardization. We have organized the book by regions, starting with North America (Section I), Africa (Section II), Asia (Section III), and Insular Areas (Section IV), with a final section for New Media (Section V). Each section begins with a short introduction and summaries of the chapters that follow. We are pleased to present these chapters as a free-ranging exploration of issues on a common and increasingly important theme, the linguistic consequences of contact in postcolonial situations. Each chapter brings to the fore local issues, and we feel that the amalgamation of overlapping methods, contexts and results provides a kind of colloquium that enriches our understanding beyond the individual parts. —The Editors

INTRODUCTION POST-COLONIAL CONVERGENCE AND DIVERGENCE RITA CALABRESE

Terminological issues: Why “postcolonial”? The term ‘postcolonial’ has originally been adopted in literature concerning variation primarily referring to those varieties of English which developed as a consequence of the colonial expansion of the British Empire. In the present book, the term applies to both “canonical” and “emerging” varieties of English which have resulted from new contact situations including those determined by globalization as well (see Section V in this volume). Therefore, the prefix “post” here is intended not only in its historical meaning, but more broadly as a term implying “change”, “differentiation” and “divergence” from a given language input along with “convergence” on a given set of linguistic features characterizing new linguistic habits by a group of non-native English speakers. This perspective also welcomes an essentially Labovian view since it assumes that “language change entails young speakers innovating or at least adopting new features” (Kerswill 2010, 230). The broader view also encompasses the so-called “postcolonial varieties” or “New Englishes” from a second language acquisition perspective and as such studies them as “contact varieties”. The number of labels used to refer to these varieties requires a brief survey of the several classifications adopted in the literature on varieties of English as in the next section.

Classification of varieties of English Varieties of English can be classified according to different parameters/variables ranging from historical, geographical, political and language acquisition criteria.

Introduction

2

Historical criteria look at the expansion of English beyond its borders in terms of subsequent diasporas chronologically definable. Jenkins (2010, 6) gives a comprehensive account of the different migratory movements firstly towards the New World and secondly to Africa and Asia starting from early 17th century as follows: x The first diaspora (from early 17th century) involved the Puritans’ migrations from England and the deportation of slaves from Africa to North America, Australia, New Zealand and determined early new mother-tongue or L1 varieties of English, also labelled as “new Englishes”1 x The second diaspora (from late 18th century) resulted in the stable colonialisation of both South Asia and Africa which had started from early 17th century and 15th century respectively and in the emergence of L2 varieties of English, known as “New Englishes”. Following the geographical path traced by people’s movements throughout different territories, it is possible to derive models like those proposed by Kachru (1988) and Algeo (2001) and presented again by Siemund (2013, 9) which combine spatial localization with linguistic characterization: 1. 2. 3. 4. 5. 6.

English in England English in the originally Celtic-speaking lands The English of North America The ‘settler’ Englishes of Australia, New Zealand and South Africa The Englishes (largely non-native) of South and Southeast Asia The Creole Englishes of Africa, the Caribbean, and the Pacific.

The non-native Englishes of South and Southeast Asia and the Creoles of Africa, the Caribbean, and the Pacific have a special status in the above classification because of their typological implications determined by the degree of contact between the substrate and the superstrate languages which will be discussed further. The delimitation of geographical areas sometimes overlaps with political borders which conveniently help identify regional varieties below the national level (e.g. Somerset English, Scottish English, Toronto English) as well as at the national level (e.g. British 1

The widespread use of the plural form “Englishes” has become a recognized convention which stresses the diversity to be found in today’s English with “no single base of authority, prestige and normativity” (Mesthrie & Bhatt 2008, 3).

Post-colonial Convergence and Divergence

3

English, Canadian English, Australian English, Nigerian English, Indian English discussed in the present volume). In turn, both political and geographical criteria tell us something on the status of the varieties characterized by specific linguistic features and different contexts of acquisition which have definitely shaped the nature of English as first, second or foreign language. From this perspective the concept of national identity loses its weight as a driving force responsible for the development of new linguistic features /dialects. Rather, that force is to be found in natural, mutual accommodation, hence the sense of identity relies upon it and is chronologically subsequent to it (Trudgill 2010, 188). As matter of fact, in a general Communication Accommodation Theory new varieties emerge initially through individual acts of linguistic adjustment performed by speakers in a specific social context where degree of contact, social group boundaries, linguistic ideologies and social identity formation play a subsequent crucial role (Kerswill 2010, 232). According to Trudgill (2009) the major typological split among varieties of English lies in fact between high-contact and low-contact varieties of English. Among the former, characterized by processes of simplification, he mentions the non-standard urban varieties in the British Isles and colonial varieties of North America, Australasia and South Africa along with shift varieties like Irish English and non-native L2 varieties like Indian English, standard Englishes and creoles. As low-contact varieties characterized by processes of complexification, Trudgill identifies the traditional dialects of English located in the British Isles and North America. His claim is based on overall coding strategies consisting in inflectional coding of grammatical information as a complexifying strategy and analytic or zero marking as a simplifying strategy. In this context differential language-acquisition abilities and processing strategies of young learners as opposed to adults play a crucial role in structural changes of languages and their varieties: simplification will occur in short-term language-contact situations involving adults, whereas complexification will develop in long-term co-territorial contact situations involving childhood bilingualism (Trudgill 2010, 23). A recently published collection of studies (Szmrecsanyi and Kortmann 2012) on linguistic complexity in the field of World Englishes and Second Language Acquisition (SLA) research has re-opened the debate concerning the traditionally conceived dichotomy of creole/simplicity and standard/complexity. Over the years, two main approaches have prevailed to objectively assess linguistic complexity, namely the absolute approach which defines complexity in terms of the number of parts in a system and their mutual connections and the relative approach which explains complexity in relation to language users and their processing and learning .

4

Introd duction

efforts (Szm mrecsanyi andd Kortmann 2012, 2 10). Acccording to th he former approach, a grammar is juudged to be more m complexx if it has morre marked phonemes, m more syntacticc and morpho ophonemic rulles, more casees of suppletion and allomorphy. The latter con nsiders that sppeakers often n simplify their native language for specific sociial purposes inn contact situ uations as well as aduult language leearners regulaarly adopt sim mplifying straategies to avoid irreguularities and to increase trransparency. Starting from m the assumption thhat the understanding of comparatively c y simple variaance (i.e. language-intternal compleexity variation n) is a prelim minary step tow wards the analysis of comparativelly complex variance v (i.e. cross-linguisstic complexity variiation) in varrieties of En nglish, Szmreccsanyi and Kortmann K (2004 and 2012) have developed d a research proggram which combines c methodologies and interppretational ap pproaches adoopted in the study of large-scale ccross-linguistiic variation with w the analyssis of language internal and syntactic variation variation. Inn their global synopsis s of morphological m in English, Kortmann annd Szmrecsan nyi (2011) deescribe 46 vaarieties of 1 and indigeniized L2 varietties along English (varriably distribuuted among L1 with Englishh-based pidginns and creoless) in terms of ppresence or ab bsence of 76 non-stanndard morphoosyntactic feaatures. Using Principal Co omponent Analysis, thhey assign each e variety of English a coordinate in twodimensionall principal coomponent spaace (Figure1, dotted boxess indicate group membberships).

Fig. 1. Visuallization of princcipal componen nts of variance iin the 76 × 46 database. d

Post-colonial Convergence and Divergence

5

The results of the satellite-view statistical analysis show that the 46 varieties cluster very nicely according to whether they are L1 varieties (represented by squares), L2 varieties (represented by triangles), or English-based pidgins and creoles (represented by circles) –and indeed better than geographically. Thus variety type turns out to be the better predictor of overall similarity or distance between individual varieties than the world region where they are spoken.

The vertical axis (Component 2) represents a given variety’s degree of analyticity resulting in a wider use of invariable periphrastic structures, whereas the horizontal axis (Component 1) indicates increased morphosyntactic complexity (Szmrecsanyi and Kortmann 2012, 16). However, from these results it must not be concluded that intensive language contact necessarily leads to simplification processes (ib.19). As a matter of fact, complexification processes are observable among contact languages as well as contact varieties.

Variation and change in postcolonial contexts: a review of research literature The study of linguistic variation has a long tradition that finds its turning point in the comparative studies carried out by the Neogrammarians and early work/research on contact languages dating back to 19th century. Other research fields based on philological constructs emerged motivated by the need to mediate between different languages and cultures. In his preface to An Anglo-Indian Dictionary2 published in 1885, George Clifford Whitworth wrote that “the term ‘Anglo-Indian’ would properly designate something which, originally Indian, has been specially modified by something English”--further specifying “but popularly it is applied to English persons residing in India and to things pertaining to them” (ib.vii). The Dictionary therefore “should contain all those words which English people in their relations with India have found it necessary or convenient to add to their own vernacular, and should give also any special significations which pure English words have acquired in India” (Whitworth 1885, vii). Whitworth’s words anticipate the main focus of later studies on contact varieties, namely the socio-cultural setting in which such varieties function. As stated by Ferguson (1959) 2

The dictionary had the meaningful subtitle A glossary of Indian terms used in English, and of such English or other non-Indian terms as have obtained special meanings in India.

6

Introduction Descriptivists usually prefer detailed descriptions of ‘pure’ dialects or standard languages rather than the careful study of the mixed, intermediate forms often in wider use. Study of such matters as diglossia is of clear value in understanding processes of linguistic change and presents challenges to some of the assumptions of synchronic linguistics.

Over the past thirty years the interest of linguists in diatopic variation has been polarized around specific issues within well defined research areas: due to the work carried out by Braj Kachru the study of New Englishes has become part of mainstream Sociolinguistics, focusing on the characterization of the spread of English “viewed in terms of three concentric circles representing the types of spread, the patterns of acquisition and the functional domains in which English is used across cultures and languages” (Kachru 1985, 12). Major studies first carried out on individual creoles have provided a unifying framework in the field of Contact Linguistics when post-creole continua were identified e.g. in Jamaica (De Camp 1971) and Hawaii (Carr 1972) giving rise to important assumptions on the common origin and development of these varieties and their classification as proper varieties of the superstrate language. In line with Sridhar & Sridhar’s (1986) plea for an integrated approach, more recently Mesthrie & Bhatt (2008) have called for extensive studies aiming to bridge the paradigm gap between studies of Second Language Acquisition (SLA generally focused on individuals) and New Englishes. For a long time these two lines of linguistic research have been treated as two different and unrelated areas mainly because of the “linguistic taboos … to establish these New Englishes as full-fledged varieties with the potential to develop endonormative and local standards and norms [not to] be conflated, with the error-focused description and analysis of foreign language learners’ output as a deviation from an exonormative norm” (Mukherjee & Hundt 2011, 1f). However, both branches have to do with four related questions concerning codification, innovation, de-Englishization and non-native creativity (Kachru 1985, 17). As a matter of fact, from a variationist perspective, those concepts acquire a functional-specific meaning that leads to a cline of “acceptability” and “creativity” depending on their specific context and function of usage. Taking into account cognitive factors underlying contact language phenomena such as code-switching, codemixing, and isomorphism (one form-one meaning to maximize transparency) enables us to find the shared basis of underlying acquisition phenomena which effectively relate those SLA studies to research on New Englishes. However, the fundamental difference between the two groups of studies is that the former is based on individual learners’ interlanguage as the developing competence in a second or foreign language aiming at a

Post-colonial Convergence and Divergence

7

target-like language proficiency/acquisition. The latter instead pertains to “aggregates of people” (Mesthrie & Bhatt 2008, 157) who would use their interlanguages in specific domains and communicative contexts through the creation of new structural, lexical and pragmatic norms. Therefore, in postcolonial contexts, substrate influence is not evidence of a failure to learn target language constructions, but rather the choice from a “pool of variants” on the ground of harmony with L1 constructions (ib.). The following account of research literature on language variation and change restates the dynamic framework proposed by Chambers (2004, 128) who in turn referred to Croft’s dynamic paradigm (Croft 1990, 258f). This model unifies all three types of variation defined as cross-linguistic (mainly concerned with typological issues), intralinguistic (mainly pertaining sociolinguistic and language acquisition processes), and diachronic (concerning historical linguistics matters).

Language contact, typology and universal markedness: Internal and external factors in variety formation Past literature looking at contact “tended to assume uncritically that contact was always the source of new features registered in particular languages” (Hickey 2010, 1). Language contact has been so far invoked especially to account for irregularities in the correspondences revealed by the comparative method but not as the cause of regular changes that lead to language speciation. More precisely, Thomason (2001, 62) distinguishes two major categories of change. The first type includes changes that bring no structural modification to the source language, and the second one affects more profoundly the latter leading either to typical interference features or to a complete shift to the new language. As matter of fact, changes are not necessarily replacive. Depending on the effects they have on language structuring, changes can be classified as 1. additive when they introduce new rules or units to the current system, 2. subtractive when they involve loss of variants and, 3. reinforcing or weakening current variants. It is also worth noting that a cline of contact is often observable due to underlying differences in the degree to which languages in contact influence each other. More recent research has however tested current hypotheses on the interrelationship between language universals and language variation and given rise to new challenging theories on contact varieties. The notion of “vernacular universals” (Chambers 2004) seems to limit the supposed tendency towards the absolute creativity of these varieties relying on the identification of universally shared features across varieties of English

8

Introduction

around the world3. In particular, this notion postulates the direct origin/source of such features in universals of language development, specifically in the context of new dialect formation. Therefore, they can be “identified partly in terms of their social patterning, in so far as there are regularities in the way they are socially embedded” (Chambers 2004, 128). Along with the concept of universality of individual features implied in this theory a similar hypothesis has been proposed supporting the idea of “the universality of conspiracies of morphosyntactic features” and strategies for coding certain types of grammatical information which can be identified for individual types of varieties in any language” (Kortmann & Szmrecsanyi 2011, 265). This hypothesis would contribute to identify types that rather than being characterized by individual features “all conspire and jointly instantiate an overall coding strategy which is constitutive of the variety type” (ib.). This assumption would also converge on Labov’s idea of “clusters of properties” and their “intermediate combinations in terms of discreteness, abstractness, grammatical conditioning, and social conditioning” which combine together to bring about linguistic change (Labov 2010, 542)4. The challenging search for general principles underlying language complexity is achieved by these theoretical models with varying degrees of strength and certainty and in relation to the properties of human physiology and psychology in a measurable way (ib.600f). A further challenging issue would be to show that both internal and external motivations for language change are needed in any full account of language history and of synchronic variation (Hickey 2010, 31). Further approaches to language contact are also emerging based on a view of language as the practice of communicative interactions and of grammatical categories as triggers of language processing tasks (Matras 2010, 66). The resulting koiné would emerge from the loss of distinctive features in favour of features with a high degree of mutual intelligibility and high prestige in everyday communication (Noonan 2010, 58). As a matter of fact, spoken language proves to be the area of language where the major changes emerge and then stabilize through exposure to and imitation of firstly model speakers and secondly, model written texts with the consequent emergence of new language standards (Deumert & Vandenbussche 2003, 456). Contact between speech and written language would lead to the emergence of new spoken norms and a new written standard which combines structural and lexical elements of two different linguistic 3

Recurring features are, for instance, levelling of irregular verb forms, multiple negation, or copula deletion. 4 In Labov (2010) such description is used to explain the regularity of sound change.

Post-colonial Convergence and Divergence

9

systems (Haugen 1972, 57). One of the main issues to be addressed will therefore be the clarification of the historical interactions and language contact phenomena occurring between pre-existing language standards and the emerging standard language which makes the process of standardization a special type of language change within a more general theory of language contact (ib.). In recent years an effective theoretical approach to language variability and change is emerging which assumes as a primary source of language diversification around the world population movements and contacts (Mufwene 2007). Under this view all causes of change even the so-called internally-motivated changes (ib.65) in any language result from mutual accommodations among speakers in a continual process of competition and selection which changes patterns of variation in a speech community. The hypothesis on the deep interrelation between variability and social dynamics or language variability and social structure has a fairly long history even though “until the advent of sociolinguistics, there were no concentrated attempts at discovering the social significance of linguistic variation” (Chambers 2013, 6). In 1981 Romaine (5-7) pointed out the close interrelation between social dynamics and domains of grammar: Syntax is the marker of cohesion; therefore, individuals try to eliminate alternatives in syntax. In contrast to syntax, however, vocabulary is a marker of divisions in a society (cf. for example, Bright and Ramanujan1964); and we may find individuals actively cultivating alternatives in order to make more subtle social distinctions. […] According to Labov, social and stylistic variation presuppose the option of saying the same thing in different ways, i.e. the variants of the variables have the same referential meaning, but are somehow different with respect to their social and stylistic significance.

In Mufewene’s (2001, 2005) theory of language ecology and evolution the emergence of contact varieties can be regarded as a series of selective actions from “a pool of linguistic variants” from different language backgrounds and experiences available to speakers in a contact setting (Schneider 2007, 22). The “natural” selection of specific features as stable elements of the new variety depends on both extralinguistic and intralinguistic factors such as the nature of the linguistic input elements, typological similarities between the languages involved in the process of competition and selection as well as the influence of social determinants (Schneider 2007, 23). However, assuming that the external influence is the only cause of change in which every specific ecology of language use determines its local evolutionary trajectory (Hruschka 2009) gives only a partial perspective of the matter. A new, though controversial, theory on

Introduction

10

the formation of colonial Englishes has been proposed by Trudgill (2004, 1-3) in a deterministic view of language change and gradual variety differentiation. In particular, he supposes the combination of six factors to explain the differences between colonial Englishes and British English, which can be briefly outlined as follows: 1. 2. 3. 4. 5. 6.

Adaptation to a new physical environment Linguistic changes in the mother country Linguistic changes in the colony Language contact with indigenous languages Language contact with other European languages Dialect contact determined by the different geographical and social origins of the settlers

In relation to factor 2 in particular, Mukherjee and Gries (2009) note that “the input variety of British English is a diachronically changing reference point” (ib.34) so that “for any comparison of a New English variety with its historical input variety a diachronic corpus of the native variety is needed” (ib.35). Earlier studies have in fact shown that in some varieties of English features typically associated with them can be traced back to colonial English spoken by masters to their slaves (Poplack & Tagliamonte 2001). The emphasis on the continuation of vernacular traits at new locations is the basis of the so-called “retentionist hypothesis” (Harris 1984). Such retentions may cause problems in establishing genetic relationships according to the family tree model and isolating the effects of contact. This is an important aspect generally neglected in research on second language varieties but is emerging in recent approaches as an important factor (value system of a society and its indigenous culture) in the development of ESL as a variety not meant as entirely identical with the target language. A recent approach to the evolution of Post-colonial Englishes has been proposed by Schneider’s Dynamic Model (2003, 2007, 2011) which builds upon the assumption that such emerging varieties have typically followed “an underlying, fundamentally uniform evolutionary process caused by the social dynamics between the two parties involved in a colonization process” (Schneider 2011, 113). The evolution is characterized by language negotiation and accommodation phenomena from the perspective of group identity formation. The model assumes five developmental stages with specifically social and linguistic implications: 1. Foundation 2. Exonormative stabilization

Post-colonial Convergence and Divergence

11

3. Nativization 4. Endonormative stabilization 5. Differentiation Leitner (2010, 27) says, “These phases define a path from the transplantation of (a range of) dialects (by large or small numbers of speakers) into novel environments where they mix as a result of interaction (= dialect contact) and are in contact with indigenous and or other migrant languages”. The Dynamic Model can be compared to Trudgill’s new-dialect formation theory in the first two stages in which feature selection and diffusion occur, later followed by minority variants leveling and consequent norm focusing and stability (Trudgill 2004, 83-89). The stabilized form gradually turns into a local, non prestigious form which then will undergo a process of endonormative stabilization as well as final differentiation. Schneider (2007, 92) argues that it is not primarily the occasional occurrences of “well-known” distinctive features “that attribute its uniqueness to a variety; it is the subconscious set of conventions regulating the norm level of speech habits, of what is normally done and uttered, the ‘way things are said’ in a community”. In this view, innovations as well may be derived occasionally from the speech of individuals because of internal conditions, but then spread to the entire community for external reasons (ib. 99). The label linguists give to a variety often correlates with the function of that variety in the community in which it emerges and the extent to which users of the variety tend to converge on a series of norms (Trousdale 2010, 93). The formation of a new variety involves therefore not only changes in norms, but a steady process from relative absence of norms followed by focusing, namely the reduction in the number of variants (especially phonemes and morphemes) with consequent simplification of phonological and morphophonemic complextity or koineization to the re-emergence of norms (Kerswill 2010, 230f). The initially competing and then co-occurring forces which characterize the pidginization and standardization processes respectively are outlined in Table 1. PIDGINIZATION 1. Reduction in function 2. Cross-linguistic influence or contact-induced changes 3. Simplification as regularization 4. Increase in transparency

STANDARDIZATION 1. Convergence 2. Reduction of linguistic variants 3. Syntactic elaboration/expansion 4. Social symbolism of standard norms

Table 1 Strategies of the pidginization and standardization processes

12

Introduction

The overall process has been effectively explained and synthesized by Leitner (2012, 251): Having adapted to new environments, native varieties […] need to integrate new ethnic varieties that develop after first generation settlements [and ] to create a new balance between speech and public writing which narrows the gap […] and makes writing open to expressions of speech. Second language English (in the traditional former colonies) is more concerned with the conflict between the heritage and the demands of globalization.[…] Some of the varieties concerned develop stable rules of their own, which show marked differences from any of the native varieties. Some of them develop away from another and as they do that, they create powerful new epi-centres in pluricentric English.

Internal and external factors in variety formation- A number of language/cognitive phenomena represented in Table 1 can be seen as governed by “universal laws of onto-genetic second-language acquisition and phylogenetic language shift” (Schneider 2007, 89). Linguists generally tend to distinguish between intralinguistic and extralinguistic factors when discussing principles underlying change and evolution. Conditions which are purely internal to language depend on speech production, perception, and processing, whereas extralinguistic conditions include factors determined by language use in historical and sociolinguistic settings. Some of them can be explained by speakers’ simplification strategies aiming to increase the economy of speech production in language contact situations which leads to the omission of inflectional endings, copulas, articles or overgeneralization such as the extension of plural endings to mass nouns, the progressive marker to state verbs and invariant tags to tag questions. The role and nature of the superstrate along with the substrate pressure constitute an important issue in the study of new variety formation. In particular, from a diachronic perspective it is of utmost importance for the characterization of the superstrate to recognize its intermediaries. In fact, the British arriving to colonies came from different parts of Britain and different educational backgrounds. Moreover, the local British community is seen to be expanding to what Schneider (2007, 37) defines as “British plus”: a community of speakers “seasoned with the additional flavor of the colonial experience which those who stayed ‘home’ do not share”. As a result, “new locally born generations of British children develop a hybrid cultural identity” (ib.). The cross-cultural language contacts trigger important changes in the linguistic system of English as used by both communities, the native settler community and the indigenous one, firstly on the lexical level and later in the syntactic and morphological systems mo-

Post-colonial Convergence and Divergence

13

tivated by the need to refer to things belonging to the local environment along with a process of language convergence necessary and functional to “mutual negotiation” and intelligibility between two communities whose shared variety is a second language for some and a first language for others (ib.45). This assumption points out the so-called “paradigm gap” between SLA theories, mainstream linguistics and research on New Englishes discussed in the next paragraph. The integration of these apparently different paradigms into a more comprehensive theoretical framework would have the effect of creating a new field of research for testing current linguistic hypotheses on the interrelationship between language universals, language variation and creativity and SLA theories. The purpose of this unifying research procedure would have therefore the immediate effect to reveal the underlying functioning of universal processes of LA which interact with both universal sets of linguistic principles and language-specific sets of parameters. Starting from the generative assumption that an “individual’s internal (I) language is a set of parameters acquired by the interaction of UG [Universal Grammar] and the visible data (External (E) language tokens)” (Matras 2010, 66f). Changes may occur in the parameter settings of different stages of L as well as differences in/between dialects/varieties of the same language across time, geographical and social space. As a consequence, the speakers of Creoles use parametric values of their own grammar in assigning a value to the parameters of the language they are creating/ being exposed to. Under this view, contact phenomena would be considered as arranged on a continuum from the most automatic (like interference/transfers/errors) to the most conscious ones produced for stylistic and creative purposes. All are, however, functional in the sense that they are the product of language processing in goal-oriented communicative interaction (Hickey 2010, 7ff.). Therefore change in certain structural categories is associated with the task-oriented function of these categories, i.e. “with the way they support language processing in discourse [… and] in this respect contact phenomena are seen as enabling rather than interfering with communicative activity” (ib.). The crucial question concerning transfer of items from one system to another leads to two important considerations concerning: 1. The basic distinction between systemic and non-systemic elements; 2. How borrowings of linguistic items made on an individual level spread throughout an entire community. Systemic elements can be explained by the distinction between grammatical category and its exponents, i.e. some languages borrow grammatical categories but not the manner of expressing them in the source language (ib.). Non-systemic elements are single

14

Introduction

words and phrases, pragmatic markers and sentence adverbials that do not necessarily require integration into the system of the substrate language and can be picked up by adults in a contact situation and therefore accommodated without any degree of restructuring. The term accommodation acquires a special role here since it has been recognized not only as peculiarly characteristic of any individual, but also as the main triggering force inducing spread of linguistic change from an individual to an entire community (see above examples from Trudgill, Labov, Chambers). Language structure is a largely mechanical system, out of the reach of conscious recognition or adjustment by its users […]. Therefore our efforts to change language consciously must be confined to higher-level stylistic options: the selection of words, and the construction of phrases and sentences within a narrowly limited set of choices. […]. A broader approach will carry us into the ways in which language structure fits into our social needs (Labov 2010, 604-5).

These different approaches seem to converge on a model of linguistic change involving diverse perspectives that can be termed as typological, variationist (both synchronic and diachronic), cognitive and populationbased. In conclusion, two major predictors seem to emerge from the discussion on contact-induced change which become also relevant to internal explanations of language change: universal markedness, and degree of integration within a linguistic system. Rita Calabrese

SECTION I NORTH AMERICA

CONVERGENCE AND DIVERGENCE IN CANADIAN ENGLISH J.K. CHAMBERS

In North America, English has taken varying forms depending on different factors which include regional, social, and ethnic backgrounds of its speakers. In this section, we will take a closer look at Canadian English with particular reference to the speech communities in Toronto. The linguistic situation characterizing Canadian English (CanEng) makes it emblematic of the different directions a language may take with respect to its “root-language” as a consequence of different historical events that shape its structural properties despite the historically attested common origin.1 Over the last decades, there has been a growing interest in the description of Canadian English especially focusing on phonological variants (Canadian Raising, Canadian Shift) and other typical features characterizing speech (discourse tag eh?). However, in almost all descriptive approaches to CanEng the synchronic perspective of such descriptions predominates with respect to insightful diachronic analyses of evolution of CanEng. In particular, the main critical issues have been concerned with the composite nature of CanEng which has frequently been depicted as a blending of British and American English speech patterns and variously described as an endonormative variety even though not yet stabilized and autonomous. It is also worth noting that most evidence of Americanization or diffusion of American norms in most literature on CanEng is based on isolated phonological and lexical items retrieved from questionnaire surveys rather than systematic investigation of the inherent variability of natural speech data, without considering the linguistic constraints and social meanings associated with variant usage in the Canadian context.

1

Leitner (2012, 133) considers Canadian English as a “third path” along with British English and American English.

18

Section I

When looking back at past literature in the field, Lighthall (1889) represents one of the earliest attempts to identify different regional and social varieties of English within the country. The identification of distinct regional dialect regions has not, however, been prominent in the study of CanE, apart from two general exceptions, i.e. the island of Newfoundland (along with Labrador) and Quebec English, which constitute separate dialect areas along with more distinct dialect enclaves. A recent national survey carried out by Boberg (2010) has highlighted the regional fragmentation of CanEng characterized by phonetic and lexical isoglosses tracing specific dialect areas (Newfoundland, Quebec, Ontario and West). As such “subtle differences” can be heard in Ottawa, Toronto, Calgary and Vancouver, they present a challenge to the conventional view that CanEng is geographically homogeneous over the vast territory extending from Vancouver to Ottawa (Boberg 2008, 150) as well as to the Loyalist Base theory (e.g. Chambers 2009, 71-3) which interprets the apparent homogeneity of central and western CanE in terms of westward expansion (Dollinger and Clarke 2012, 460). According to this view, Ontarians (themselves Loyalists) were among the first to settle western communities and set the speech patterns, in accordance with the founder principle “those who come and settle first have linguistically more input in the koinèization process” (Dollinger 2012, 460). But the biggest problem of many existing studies of CanE is their middle-class basis which hardly considers the rural/urban split (ib.). Another potential source of heterogeneity in CanE is the influence of L2 speakers and the development of ethnic varieties. For instance, data from Montreal reported in Dollinger and Clarke (2012, 460) would place Canada in Phase 5 of Schneider’s Postcolonial English model (2007), and contribute to classify CanEng as a variety characterized by high linguistic diversification. As a matter of fact, well-established communities, such as Italians, have developed features of their own, but it seems that outside of Montreal other communities continue to assimilate features of general Canadian speech patterns. CanEng can be said to be originated from different waves of migration (see Chambers in this volume) that far from being easily identifiable from a linguistic perspective, make it even more difficult to determine with any certainty what specific features the varieties of English spoken by these groups would have displayed. It is, however, easier to characterize the speech of the immigrants who came in the first half of the nineteenth century from the British Isles. They would have spoken regional varieties from all over Great Britain and Ireland with northern and western (Irish) varieties better represented than the

Convergence and Divergence in Canadian English

19

southeastern varieties on which modern Standard British English is based (Boberg 2010, 244). It seems most likely that the formative period of CanEng, during the late eighteenth and the early nineteenth centuries, was characterized by a mixture and levelling of regional varieties of English from Ireland, Scotland and England as well as northeastern American colonies. The historical outline presented so far inevitably crosses/matches with theoretical assumptions necessary to explain the evolutionary path of CanEng might have followed towards its linguistic differentiation and identity (Chambers 2012). The relation of settlement history and linguistic variation, the connection between ethnic differences and national identity, and the processes of convergence vs divergence all have vital relevance for Canada (Görlach 2003) and should be therefore further investigated with renewed methodologies. The historical perspective is in fact the approach adopted by Chambers in his contribution demonstrating how both past and recent changes are moving inexorably toward completion on the same timeline in the historical framework in which his analysis of Canadian variants can be set. Casagranda’s study shows how the French language has contributed to language change within the Canadian variety and investigates how lexical variation can be considered as a marker of identity and one of the strategies adopted by speakers to convey the plurality of languages and cultures of Canada. Parascandolo presents the preliminary results from a case study of variation in the verbal system of Italian speakers who live in the Greater Toronto Area. The contributions in the present section can variably be read as addressing the two main issues that have characterized research on Canadian English so far, namely autonomy and homogeneity of the Canadian variety.

CHAPTER ONE SAYING “TOMATO” IN POSTCOLONIAL CANADA J.K. CHAMBERS

In North America Canadian English, once highly conservative, began undergoing several changes in the 1940s quietly at first but later with accelerating urgency. The changes included every structural level from pronunciation to grammar. Some variants were in flux for centuries and others came seemingly out of nowhere, but all of them are now moving inexorably toward completion on the same timeline. Changes that take place in languages often seem arbitrary, the result of changing fashions and random fluctuations. However, several of the changes in Canadian English have a clear social motive— they represent the elimination of Victorian British variants, and they provide the most concrete evidence for the dissolution of Canada’s colonial tie to Great Britain. The linguistic changes are extensive, as I have said, and often fairly complicated. It is the motive behind them that is most interesting for understanding postcolonial Canada, and the underlying motive is embedded in two centuries of Canadian history. In order to make it manageable, I will look closely at one fairly straightforward pronunciation variable and then extrapolate from it to show the larger picture. The variable is the pronunciation of the word tomato. 1. Does the main vowel in toMATo sound like maht or mat or mate? For about 150 years, Canadians had three possible pronunciations of the word tomato based on the stressed vowel in the second syllable. The three pronunciations were these: 1.

[thԥɎm‫ܤ‬: tho] with the low-back open vowel /‫ܤ‬/

2.

[thԥɎmætho] with the front lax vowel /æ/

3.

[thԥɎmejdo] with the tense diphtong /ej/

22

Chapter One

These pronunciations coexisted for almost two centuries.1 Some Canadian said tomato so that the main vowel sounded like maht (as in 1), and some Canadians said tomato so that the main vowel sounded like mate (as in 3), and a few Canadians said it a third way, as if the main vowel sounded like mat (as in 2). Naturally, some Canadians switched between two of them (1-2 or 3) depending upon the situation. The coexistence of variants like these indicates a kind of linguistic tolerance was common in Canada for years and years. It is a phenomenon known as “the uniquely Canadian double standard.” It has its origins in Canada’s historical status as (what was once called) “British North America.” While the southern half of the North American continent tossed off its colonial ties to Britain in the American Revolution to become the United States of America, the northern half that would become Canada retained those colonial ties. When Canada’s ties to Britain grew weaker and eventually disappeared in the second half of the twentieth century, it was the result of evolution, not revolution. But for two centuries before that, Canada absorbed influences from both Britain and the United States. Linguistically, the consequences were easy to see— or, literally, easy to hear. Whenever British usage differed from American usage, Canadians typically used both. The pronunciation of tomato is an instructive example: pronunciation (1) is British, and pronunciation (3) is American. Tomatoes were discovered in Mexico and taken to Europe by Spanish explorers. The European word for them (in Spanish, Portuguese and French) has a simple vowel and is the source of the British pronunciation. In North America, cultivation of tomatoes developed independently and so did the pronunciation of the word, getting its long vowel probably by analogy with potato. What about the third Canadian pronunciation (2 above)? The variant tom[æ]to shares phonetic resemblances to both, with a simple vowel like the British variant but a front vowel like the American variant. It is a phonetic compromise between the other two pronunciations, called a “fudge.” Fudges occur in places where there are competing pronunciations, and people are uncertain about what they should say. The pronunciation with [æ] is distinctively Canadian compromise (or was, as it has now disappeared).

1

Linguists will notice that the pronunciations differ not only in the stressed vowel but also in the consonant that follows the vowel. It is voiceless /t/ in the first two forms but voiced /d/ in the third, because of T-Voicing, a common feature in North American varieties. The vowel difference is more salient, and the focus of my discussion here.

Saying “Tomato” in Postcolonial Canada

23

2. How did Canada get both variants? Under normal circumstances, particular dialects have one standard pronunciation for each word, as do the United States and Britain for the word tomato. The situation in Canada, the double standard, is unusual. But once we come to grips with Canada’s settlement history, it becomes readily understandable, indeed almost inevitable. I am going to summarize four major movements in Canada’s settlement history. The four movements covered a span of almost 200 years, from the earliest settlers in the 1700s to around 1910, and they are crucial to the formation of the national character. I will summarize them as succinctly as possible,2 and then I will show how these movements gave rise to the linguistic variation of tomato and other forms. FIRST NATIVE GENERATIONS. In the early 1700s, after North America was discovered by European adventurers and began receiving Englishspeaking settlers, the continent gradually accumulated sufficient permanent population along the Atlantic seaboard and on inland waterways to establish towns and small cities. The first settlers brought with them diverse accents and dialects rooted in eighteenth century Britain, but their children, the native North Americans of European descent, spoke with a more homogeneous accent than their parents. Indigenous regional markers developed in the speech of the children and the children’s children, sometimes different from the parental varieties and, it follows, different from the British varieties in Europe. These distinctive North American varieties developed wherever there was a permanent population, and in these early days, before the continent split into the political entities of Canada and the United States, it was the same in all the settled regions of the continent.

REVOLUTION AND LOYALISTS. In 1776 the Thirteen Colonies south of the St. Lawrence River rebelled against colonial rule and declared their independence, bringing about the political split that would make national entities of the United States and Canada. Under ordinary circumstances, such political splits foster and encourage cultural distinctions, including linguistic differentiation. This split did that, but with mitigating circumstances. Thousands of people in the Thirteen Colonies chose to remain British subjects, and as the revolutionary zeal of their neighbours grew, they were forced to flee the country. Many fled to England, but many more fled northward into Canada. These refugees, known in Canadian history as United Empire Loyalists, swelled the numbers in what would 2 I have discussed these movements elsewhere in greater detail (especially in Chambers 2004).

24

Chapter One

become the Atlantic provinces and, more significantly, in the inland regions on the north shore of the Great Lakes and the St. Lawrence River that would become the provinces of Quebec and Ontario. Linguistically, the result of the Loyalist influx was the further homogenization of the North American accent. The early Canadian historian William Canniff, the son of Loyalists, observed this linguistic process directly. “The loyalist settlers of Upper Canada were mainly of American birth, and those speaking English, differed in no respect in their mode of speech from those who remained in the States,” Canniff said (1869, 363), and then he added: “Even to this day there is some resemblance between native Upper Canadians and the Americans of the Midland States; though there is not, to any extent, a likeness to the Yankee of the New England States.” This last observation remains cogent almost a century-and-a-half after Canniff made it— the English variety that is most similar to Canadian English today is the variety spoken in towns and villages in western Pennsylvania, upstate New York and Ohio, the source of most of the Loyalists who settled in Ontario. COUNTER-LOYALISTS OF 1816-57. The third movement is the one that introduced the crucial outside element into the peopling of Canada. Beginning in 1816, the British governors of Canada offered generous incentives to attract greater numbers of English, Scots and Irish immigrants. Their motive was political— the Americans had raided Canadian border points in the War of 1812-14, and although the Canadians took up arms and beat back the insurgents, the governors feared that the Loyalist-descended Canadians harboured too many republican or pro-American sentiments. They set about diluting the populace with an influx of British immigrants whom they assumed would be imperialist and royalist. In this, they succeeded brilliantly. Promises of land ownership and business opportunities have attracted immigrants to Canada throughout her history, and they were especially attractive at a time when Europe was racked by cholera epidemics and potato famines. Almost 1.2 million people emigrated from the British Isles. Cowan (1961, 186) says, “This growth in the colonies had been achieved by a quarter century of transatlantic movement unparalleled in the history of modern free peoples.” The immigrants overwhelmed the sparsely populated colonies, especially inland, where the region that would become Ontario grew from about 150,000 in 1824 to at least 450,000 in 1841; Toronto, which would become the provincial capital, grew from 1,600 to 20,000 in less than 20 years.

Saying “Tomato” in Postcolonial Canada

25

URBANIZATION AND EXPANSION 1890-1910. Another massive immigration took place a few years later. This time, the motive was not so much political as economic and agrarian. Canada expanded westward into the vast regions that would become the Prairie Provinces and the Northwest Territories, and there was an urgent need for farmers and miners to exploit its abundant resources. At the same time, the eastern regions were industrializing and there was an equally urgent need for both skilled and manual labourers to keep the industries running. Thousands more immigrants arrived from the British Isles (and also from Italy, Germany, the Netherlands, Ukraine, and other European nations). The unassimilated Brits now joined the assimilated ones who had come one or two generations before them, thus thickening the layer of Anglo-Canadians superimposed on the original Loyalist-Canadian base. In summary, Canada’s first two centuries brought two distinct ethnic factions: first came settlers whose eighteenth-century accents homogenized into identifiable American varieties, including the Loyalists, and second came waves of British and Irish immigrants whose nineteenthcentury accents differed in numerous ways from the people who were already there. To return to our tomato motif, the reason for the Canadian double standard now becomes obvious. The first settlers and the Loyalists pronounce tomato so that it rhymes with potato (3 above), and the second wave pronounce tomato so that it sounds like tomAHto (1 above). In the well-entrenched Canadian ethos for tolerance, both pronunciations persisted. 3. British attitudes and Loyalist roots Ironically, the privileged and relatively well-educated British immigrants who came to Canada in the nineteenth century despised the Loyalist accent. These upper-crust immigrants constituted a minority among the thousands of artisans and labourers, but they became a vocal minority. The first use of the term “Canadian English” in any printed source occurs in an essay by the Rev. A. Constable Geikie, who immigrated to Canada with his family as a 22-year-old in 1843 and became a Presbyterian minister. In an address to the Canadian Institute in 1857, Geikie defined his subject this way: “‘Canadian English’ [is] a corrupt dialect growing up amongst our population… until it threatens to produce a language as unlike our noble mother tongue as the negro patua or the Chinese pidgeon English” (1857, 344). The North American dialect was obviously different from Geikie’s British accent, and he viewed those differences as debased or “corrupt.” In his essay, Geikie described more than a dozen differences, which he la-

26

Chapter One

beled “lawless and vulgar innovations” (discussed in detail in Chambers 1993). He was not alone. Geikie and other patricians set about trying to remake Canadian English in their own image. With hindsight, we know that they had limited success in bringing about changes that were permanent and fixed. They succeeded, for instance, in changing the name of the last letter of the alphabet from the American zee to British zed. They established some British spelling conventions, such as -our in words like colour and neighbour (for American –or) and –re in words like centre and metre (for American –er). They established tap as the name for a valve in a sink (rather than American faucet or spigot). These features are well entrenched in Canadian English, and they are the only ones that are traceable to the counter-Loyalist immigrants. The other reforms they attempted found their way into the Canadian accent as variants rather than standard forms. The British immigrants and their descendants became influential as educationists, and as a result Victorian schoolchildren were encouraged to “enunciate clearly.” The enunciation lessons generally involved replacing certain indigenous pronunciations with their British equivalents. These included pronouncing words like student, Tuesday and news with /ju/ (as NYOOZ not NOOZ, for instance), the word schedule with SH (not SK), and the word rather with ‘long AH’ (so that it rhymed with father, not with blather). For most students, these changes fell on deaf ears, but not for all. For a hundred years or so, the British variants caught on with some proportion of Canadians, thus giving rise to the fabled double standard. The pronunciation of tomato, of course, was one of the targets for change, because in some circles it was considered genteel when it was pronounced with British long AH and vulgar with Loyalist EY. The linguistic fate of the tomato variants in the twentieth century, as we shall see, represents the fate of the whole host of variants introduced by the Counter-Loyalist immigration. 4. Saying tomato in the Twentieth Century We trace the distribution of linguistic forms such as the tomato variants by eliciting their actual use in the speech of the relevant population and discovering how the uses differ (if at all) according to the age, sex, social class, ethnicity or any other relevant social attribute of the speakers. Correlating use with the age of the speakers allows historical inferences in terms of “apparent time,” that is, by noting how the variants are used by people of different ages; if younger people tend to use a different variant from older people, we infer that the form is changing because one generation preferred one variant in their formative years and the other generation preferred a different one.

Saying “Tomato” in Postcolonial Canada

27

The results for tomato pronunciation come from the Dialect Topography of Canada, a large-scale sociolinguistic survey of seven Canadian regions (Chambers 2007). In Figure 1, subjects are listed by age along the abscissa or X-axis from the oldest (80 and over) on the left to teenagers (14-19) on the right with the others by decades in between (70-79, 60-69, etc.). The ordinate or Y-axis shows the percentage of subjects in each age group who say each of the variants. The variants are identified by symbols on the lines of the graph. The figure shows a dramatic decline in the use of tomAHto, the British variant in the last seven decades of the twentieth century, and a concomitantly dramatic rise in the use of tomEYto, the Loyalist variant. The thickest line on the figure combines the usage of the two ‘genteel’ variants, the Canadian fudge Æ with British AH. Combining them in this way makes the most instructive contrast because it opposes the genteel forms to the indigenous form. Looking first at the people 80 and over, the oldest group, we can see that the genteel forms were used by a majority (about 52 percent). These Canadians were born before 1913 (that is, 80 or more years before the survey date of 1993). Fortuitously, the timing of the Dialect Topography project caught the change at the moment it began reversing itself. Starting in the next decade, the indigenous pronunciation becomes the majority pronunciation, and it continues to increase its plurality in each decade after that, including a striking rise of more than ten percent among both the 50-year-olds and the 40-year-olds, that is, Canadians born in the 1940s and 1950s. The ascendancy of tom[ej]to continues steadily until it exceeds 90 percent among the 30-year-olds (born in the 1960s). After that, there is a general tailing-off, with only small increases for the youngest Canadians. Among these people, the Canadian fudge has disappeared completely. The British variant retains a small and largely inconsequential representation. The indigenous variant, the pronunciation that was planted in the usage of the first English-speaking settlers on the continent, presentday United States and Canada, including the Loyalist generations who became the first significant settlers in inland Canada, is no longer challenged by competing pronunciations. The Canadian double standard that existed for some two hundred years is wiped out, at least with respect to tomato pronunciation.

28

Chapter One

Figure 1-1- Three variants of tomato by age in all regions of the Dialect Topography of Canada

Saying tomato in the Twenty-First Century The elimination of the British variant of tomato is representative, as I have said, of a host of changes that took place in Canadian English in the second half of the twentieth century. Considering all these change together (as I do in Chambers 2004), they provide a kind of measure of the dissolution of Canada’s colonial tie to Great Britain. Culturally, the decline of Canada’s colonial ties to Britain was so gradual that it did not arouse much public awareness, and linguistically the disappearance of the British variants also took place without public awareness. The linguistic changes were not perceived in colonial terms because people are generally unaware of which variant might be British or American. The variants that disappeared were the ones that once sounded ‘genteel’ and are now perceived as pretentious; most of them happened to have British origins. Young Canadians today would be amused if one of their workmates or classmates were to say tomAHto; it is a pronunciation that would almost certainly elicit a comment, perhaps a derisive one. The British variant has disappeared in the twenty-first century, and with it the double standard has disappeared as well. Analyzing the pattern for more than a dozen linguistic changes including tomato, we find that the most significant increases in use of the indige-

Saying “Tomato” in Postcolonial Canada

29

nous variants occurred in the speech of 60-year-olds and 50-year-olds, whose formative years were the 1940s and 1950s. These were the years when Canadians were fighting alongside British troops in World War II and then shipping food and materials to Britain to allay rationing and rebuild ruins. Superficially, it was a time when Canadian sentiment for the British motherland seemed fervent. However, the subconscious trends we can discern in Canadian English reveal a deeper reality. The Canadian people were already engaged in letting go of the traces of their colonial ties. More obvious imperial symbols were legislated out of existence in the next decade. “God Save the Queen” was replaced as the national anthem by “O Canada” in 1967, and the Union Jack was replaced as the flag by the red maple leaf in 1968. By then, it was clear that Great Britain’s stature had diminished among the world powers. The British Commonwealth, the old Empire, was gone. Canada had changed too. The heterogeneous ethnic and racial mix that arrived with the millions of immigrants after 1910 transformed the old allegiances. The first four immigration waves (discussed above) had brought immigrants of Anglo-Celtic ethnicity almost exclusively. The next waves overwhelmed them. By the 1960s, AngloCeltic lineage no longer formed the majority. Today, Canada is perhaps the most multicultural and multilingual nation in the world. Postcolonial Canada came about through evolution, not revolution, and the linguistic changes in Canadian English are its harbingers. The way Canadians say tomato may seem like an innocent piece of trivia but it actually bears witness to a great transformation in the Canadian ethos.

CHAPTER TWO A CORPUS-BASED ANALYSIS OF SOME CANADIANISMS OF FRENCH ORIGIN IN CANADIAN ENGLISH MIRKO CASAGRANDA

Introduction This paper is part of a wider research project on language contact and variation in Canadian English and focuses on some lexical Canadianisms of French origin in the recent International Corpus of English (ICECanada). Being Canada an officially bilingual and multicultural country and having language contact deeply shaped its national identity over the centuries, I will show how certain French loanwords that have come into Canadian English are under-represented in some current resources. After outlining the theoretical and critical debate on language variation and language contact in multilingual settings, the first part of the paper introduces Canadian English and discusses the notion of Canadianism. In the second part, I look at Canadianisms of French origin found in Only in Canada, You Say (Barber 2008), a popular but very sound discussion of Canadianisms by the editor of the Canadian Oxford Dictionary, and then illustrate their rarity in ICE-Canada. The results necessarily open up to further investigations of different types of corpora.

Lexical variation and language contact Most recent studies in language variation and change focus extensively on grammatical structures, pragmatic features and phonological and morpho-syntactic patterns from diachronic as well as synchronic perspectives that blur the disciplinary boundaries between variationist sociolinguistics and dialectology (Bayley and Lucas 2007; Fried, Östman and Verschueren 2010; Hinskens 2006; Kügler, Féry and van de Vijver 2009; Nevalainen,

32

Chapter Two

Klemola and Laitinen 2006). The exploration of lexical variation, on the other hand, has frequently been carried out within historical linguistics and etymology studies (Cloutier, Hamilton-Brehm and Kretzschmar 2010), and structured around semantic categories that include semiasiological, onomasiological, formal and contextual lexical variation (Geeraerts, Grondelaers and Bakema 1994: 3-8). As Léglise and Chamoreau (2013) point out, variationist sociolinguists have often focused on what has been believed to be monolingual speech communities without taking into account that they are instead “heterogeneous and socially and linguistically diverse” (2). As a matter of fact, it is only by acknowledging the role played by multilingual speakers, on the one hand, and by endangered languages, on the other, that we are able to fully understand the sociolinguistic complexity of postcolonial countries where language contact is one of the main reasons behind language variation and change. If it is true that it takes synchronic variation to achieve diachronic language change within the same variety, this is also feasible for contact settings since “[c]ontact-induced changes are generally defined as dynamic and multiple, involving internal change as well as historical and sociolinguistic factors” (Chamoreau and Léglise 2012, 1). According to Matras (2012), contact-induced change is due to at least two factors, namely “gaps” in the recipient system, on the structural side, and the overall ‘social prestige’ of the donor system on the societal side” (ib.19). Furthermore, out of the four types of linguistic innovation outlined by Matras, i.e. “the insertion of lexical word-forms and lexical borrowing, replication of patterns or constructions, fusion of grammatical operators, and playful or ‘theatrical’ mixing” (ib.23), lexical borrowing is doubtless one of the most productive in contact settings: In the majority of contact situations, borrowing occurs most extensively on the part of minority language speakers from the language of wider communication into the minority language. On the other hand, one can readily identify words that have become accepted within majority language communities that derive from language shift by various immigrant groups and would thus clearly fall under the definition of ‘substratum influence’ (Sankoff 2002, 649).

This is undoubtedly the case of postcolonial and multicultural countries, especially when “languages are in contact on a long term basis” (Pakir 2009, 85) and where substrata and superstrata intertwine giving birth not only to new varieties, e.g. creoles, but also to contact-induced variation and language change.

Analysis of some Canadianisms of French Origin in Canadian English

33

As anglophone Canada has presented a “situation of linguistic-ethnic hybridity” (Matras 2012, 18) since its very foundation after the British conquest in 1759, English-speaking Canadians have quite often relied on French – along with several Aboriginal languages – to refer to what had already been named and to interact with the French colonisers. Indeed, instead of assimilating them, the British allowed the French to maintain Civil Law and Catholicism, which has resulted in the coexistence of the two cultures in the same territory and eventually led to the present-day bilingual policies. As Walter Avis first claimed in 1967, lexical borrowings have been among the most frequently employed items in the Canadian contact setting, where English occasionally behaved as a superstratum, and French as its substratum counterpart, i.e. the provider of lexical items that were unknown to English.1 As a matter of fact, since the British took over from the French many economic activities like, for instance, the fishing industry and the timber and fur trades, they borrowed quite a few French lexical items from these specialised languages. Some of them are uniquely Canadian and are thus defined as Canadianisms.

Canadian English and Canadianisms With its double European colonization, the legacy of its Aboriginal people and the contribution of its multicultural communities, Canada is indeed a dynamic example of language – and culture – contact. Being often described as a variety in between Standard British English and Standard American English, Standard Canadian English has developed its own peculiar features and distinctive traits since the nineteenth century, when it was deemed by the Victorian visitors as “a language too remote from its British origins to preserve its sounds” (Chambers 1993: 1). Even though it is commonly considered homogeneous, Canadian English is nevertheless a linguistic mosaic as it “admits more pronunciation and vocabulary variants than other English varieties” and “appears to be a less focussed variety than it was a few years ago” (Chambers 1998: 269). Furthermore, as Tagliamonte (2006) points out, in the new millennium Canadian English has been experiencing an increase in the speed of linguistic evolution and innovation as “the young people are taking [the] 1

The same, however, can be said about the variety of English spoken in Quebec (Boberg 2010: 181-184; Poplack, Walker and Malcomson 2006) or the way English has influenced Quebec French (Blondeau 2013) and Ontarian French (Mougeon and Beniak 1991).

34

Chapter Two

changes and pushing them forward even quicker than the generation before them and changing the way Canadian English is spoken” (325-326). Along with distinctively Canadian phonological, grammatical and orthographic features, lexical Canadianisms, i.e. “Canadian terms [that], though referring to universal concepts, do not generally occur outside Canada” (Boberg 2010, 167), are among the lexical items that contribute to define Canadian English as a variety in its own right. Hamilton (1997) distinguishes among four types of lexical Canadianisms, i.e archaisms, extended meaning of existing words, creation of new words, and borrowings from other languages. Most of the latter have been imported from Aboriginal languages and French (McArthur 2002, 215-216), especially during the colonial period back in the eighteenth and nineteenth century. As Gregg (1993) affirms, the first studies on Canadianisms date back to 1912, when the Western Canadian Dictionary and Phrase Book was edited by John Sandilands. Under the title on the cover of the first edition of the thirty-two-page booklet it is made clear that the volume is meant at Explaining in Plain English, for the Special Benefit of New Comers, the Meaning of the Most Common Canadianisms, Colloquialisms and Slang, added to which is a Selection of Items of General Information Immediately Helpful to the Newcomer (Sandilands 1912, nn).

Whereas the 1912 edition comprises 32 pages, a preface and 853 entries, the 1913 edition, reprinted by the University of Alberta Press in 1977, is twenty pages longer and includes over 1,500 words and expressions with no preface. Although it reads “dictionary” in its title, the volume is more like a glossary, where names and sayings are also explained. For instance, both Herring Pond, the Puddle, and the Big Drink refer to the Atlantic Ocean. It is only in the 1950s, however, that research projects on the lexicographic study of Canadianisms developed thanks to Walter Avis, who founded the Canadian Linguistic Association in 1954 and coordinated a team whose work eventually resulted in the publication of the 1967 Dictionary of Canadianisms on Historical Principles. The 10,962 entries of the DCHP are now available online for free at www.dchp.ca and are being revised for a second edition under the supervision of Stefan Dollinger, Laurel J. Brinton and Margery Fee. The 1998 Canadian Oxford Dictionary edited by Katherine Barber (editor-in-chief) lists about 2,200 Canadianisms, including semantic shifts and loanwords. Since “Canadians kept asking [the editors of the OCD] for a separate list of words that are unique to Canada” (Barber 2008, 2), in 2008 Barber published Only in Canada, You Say. A Treasury of Canadian

Analysis of some Canadianisms of French Origin in Canadian English

35

Language, which “is not a comprehensive ‘dictionary of Canadianisms’ [even though it] includes many (but not all) of the Canadianisms in the Canadian Oxford Dictionary” (2). The Canadianisms of French origin analysed in this paper are taken from Only in Canada, You Say because it focuses on a restricted group of culturally meaningful words that speak for Canada’s cultural and linguistic uniqueness. Indeed, the book is like a semantic mosaic as it is divided into fifteen sections about the several historical, geographic and cultural tiles of Canada such as “A country with too much geography”, “Land of the silver birch, home of the beaver” and “Eat, drink, and be merry”. The definitions are taken and adapted from the Canadian Oxford Dictionary and, for this paper, have been double-checked in the Paperback Oxford Canadian Dictionary (2nd edition, 2006). Only in Canada, You Say lists 1,383 lexical Canadianisms, out of which 78 are of French origin, which means that only 5.64% of the “uniquely Canadian words” are, or derive from, French. In other words, despite the double European colonisation and the policies in the last forty years to promote bilingualism, on a lexical level French has had little influence on Canadian English and its Canadianisms, reinforcing the stereotype of the two so-called solitudes. Indeed, Canada’s official bilingualism is de facto a double official monolingualism. Most of the Canadianisms of French origin in Only in Canada, You Say are either related to cooking, e.g. poutine, fricot, rappie pie, tourtière and fèves au lard, or politics, e.g. felquiste, Bloquiste, Péquiste and Créditiste, or date back to the colonisation of the vast Canadian territory, e.g. coureur de bois, voyageur, log chute and canot du mâitre.

Corpus analysis and lexical variation Even though the small percentage of Canadianisms of French origin in Only in Canada, You Say is quite revealing about the influence of French in Canadian English, this papers aims at assessing the behaviour of such lexical items within a wider corpus. Since “a dictionary is not a neutral list of terms [but] reflects the selection and interpretation worked out by a team of experts and shapes an ideological common ground” (Palusci 2010, 60), along with word-list corpora like the above-mentioned dictionaries, also public electronic corpora have been employed here since they are highly representative of a variety and can be successfully used to carry out variationist research (Bauer 2002; Durand 2009). From a methodological point of view, thus, this case study builds on Kaunisto’s assumption that different types of corpora should be considered when analysing lexical

Chapter Two

36

variation: When one examines the lexicon from a diachronic or a synchronic point of view, I would very strongly recommend a method that includes a number of different sources of information. Alongside corpora, I will also take a close look into other sources, such as historical dictionaries, grammars, and language usage manuals (Kaunisto 2007, 4).

The International Corpus of English has been chosen because it is a conventional corpus (D’Arcy 2011: 51) of one million words, which Bauer (2002, 102) classifies as follows: Structured, electronic, both written and spoken language, comparative (in that written and spoken are both extensively covered, but also in that the various national sections of ICE can be compared), textual.

The corpus is actually made of 500 texts (300 spoken and 200 written) collected after 1989, when the ICE project first developed. The speakers are all aged 18 or above and were all born in Canada and schooled in English. However, as the online introduction to the project states, “[t]he proportions [within the corpus] are not representative of the proportions in the population as a whole” (www.ice-corpora.net/ice/design.htm), which means that there are inequalities in the way women are represented in political and legal discourse or in the way different age groups are included or excluded in the texts collected in the corpus. After being selected from Only in Canada, You Say, the 78 Canadianisms of French origin have been double-checked in the paperback edition of the Canadian Oxford Dictionary. Out of 78, eleven are not listed in the POCD and have consequently been excluded from this corpus-based analysis.2 The remaining 67 items have been subsequently searched by means of the software WordSmith and only seven of them, i.e. 10.45% of the total amount, have been found in ICE-Canada, i.e. autoroute, caribou, CEGEP, fleur-de-lys, Hôtel-Dieu, Metis, and shanty. The following table shows the definitions of the seven entries as listed in the Paperback Oxford Canadian Dictionary:

2

The eleven lexical items are: bastard canoe, batard, chicot, coordinates, coyau, Ermite, Fameuse, L’Acadie Blanc, loche, rabaska, téléroman.

Analysis of some Canadianisms of French Origin in Canadian English

37

Canadianism

Definition (POCD)

Autoroute

an expressway in Quebec, France, and other Frenchspeaking countries 3. Cdn (esp. Que.) a beverage made from red wine and whisky blanc (in Quebec) Collège d’enseignement général et professionnel, a post-secondary educational institution offering two-year programs for preparation for university and three-year training programs 1. a figure of a lily composed of three petals bound together near their bases, used as a symbol of Quebec and in the former royal arms of France. 2. the flag of the province of Quebec a name given to a hospital in French-speaking areas or to one established by a French-speaking order of nuns (esp. in Canada) a person of mixed Aboriginal and European descent 3 esp. Cnd hist. a lumberjack’s log cabin or shack. 4 esp. Cnd hist. a logging camp

Caribou CEGEP (also Cegep)

Fleur-de-lys

Hotel-Dieu

Metis Shanty

Table 2-1 Canadianisms of French origin in ICE-Canada

The orthography of the seven Canadianisms in the table follows that of the POCD and does not include the diacritical variants that are sometimes found in other dictionaries and corpora. It is indeed the case of Hotel-Dieu, which in ICE-Canada is written with the circumflex accent, i.e. HôtelDieu. Other variants like Cégép and Métis, however, are not used in ICECanada. Moreover, three out of seven entries, namely “caribou”, “fleur-de-lys” and “shanty”, are defined in the table exclusively by their uniquely Canadian meaning(s): “caribou” is known also in other languages as the reindeer living in the Arctic regions of North America, “fleur-de-lys” is commonly associated in American English to “a conventionalized iris in artistic design and heraldry” (www.merriam-webster.com), and “shanty” is commonly known in other varieties of English as “a badly built house made from sheets of wood, metal, or other thin material” (www.macmillandictionary.com). The following graph shows the occurrences of the above-mentioned Canadianisms in ICE-Canada. It should be noted that also the variants with the French diacritics have been searched in WordSmith with no occurrences but Hôtel-Dieu:

Chaptter Two

38

140 120 100 80 60 40 20 0

Fig.2-1 Occurrrences of the selected s Canadianisms of Frencch origin in ICE E-Canada.

Two item ms, i.e. “caribbou” and “shan nty”, are neveer employed to t refer to their Canadiian-only meanning, but to th he general onne, i.e. the weell-known North Amerrican genus raangifer and thee small dwelliing respectively. Along with “fleur-dde-lys”, thesee two Canadiaanisms of Frennch origin aree the only examples off words that have h extended d meanings coommon to oth her varieties of Engllish. The otheer four, on thee contrary, aree culture-boun nd words used in the C Canadian conttext only. In ICE-C Canada “autorroute” is geog graphically boound as it is allways associated witth the name off an expressw way in Quebecc. Out of 24 ittems, as a matter of facct, 16 are nouun phrases folllowing the strructure [propeer name + autoroute], e.g. “Bonaveenture autorou ute” and “Jeaan Lesage au utoroute”, whereas 6 fo follow the struucture [autorou ute + number]], e.g. “autoro oute Thirteen”. All thhe 24 ICE-C Canada entriess belong to sppoken monologues, in particular 222 are taken froom scripted brroadcast newss and two from m scripted broadcast taalk shows. “A Autoroute” is actually listedd also in the Merriam Webster’s oonline version as an expresssway in Francce only. This, however, may be due to the fact that it is employ yed near the Q Quebec-US bo order and has consequuently entered American voccabulary, too. The dataa suggest that,, as far as the corpus is conncerned, “auto oroute” is commonly uused as part off the proper naame of an exppressway ratheer than as a variant off the English term, which reinforces r thee assumption that such items are cuulture-bound. The same can n be said abouut “Hôtel-Dieu u”, which is associatedd with the nam me of the hosp pital in both IICE-Canada entries. e In one case, thee speaker refeers to “Sherbrooke’s Hôtel--Dieu Hospitaal”, which proves that tthe French Caanadianism is part of a propper name rath her than a variant of “hhospital”. Even thoough its etymoology is actuaally Mi’kmaq (an Eastern AlgonquiA

Analysis of some Canadianisms of French Origin in Canadian English

39

an language), the word “caribou” has been chosen for this case study because it has been introduced into western languages via Canadian French and because its second meaning is used especially in Quebec. In the 22 ICE-Canada entries, however, it never refers to the typically Canadian wine and whisky beverage and as a consequence it is not treated as a Canadianism. Cegep (122 occurrences) is the only lexical item employed consistently in the whole corpus, both in written and oral texts, both as a noun: … was that nobody who teaches in a cegep gets experience for that year …

and as a pre-modifier in noun + NPs: … like what does it mean to be a cegep teacher or why do you want to

The remaining Canadianisms, on the contrary, occur much less frequently. Whereas there is only one entry for “fleur-de-lys” in a written press report, Metis appears seven times both in press editorials, reportages and non-academic writing. It is always strongly associated with the cultural identity of the Metis community and in one case it is mentioned together with the other Aboriginal communities of Canada, i.e. the First Nations and the Inuit.

Conclusion As summarized in the following table, the low frequency of the selected items in ICE-Canada leads to the conclusion that the Canadianisms of French origin defined in Only in Canada, You Say have very little or no influence at all on Canadian English – with the possible exception of Cegep, probably due to the number of young speakers recorded for the ICE project in Quebec and Ontario. Only the uniquely Canadian meaning has been taken into account in the table and “caribou” and “shanty” have been consequently excluded:

Chapter Two

40

Canadianism

Number of occurrences

Written/spoken texts

autoroute caribou CEGEP fleur-de-lys Hôtel-Dieu Metis shanty

24 122 1 2 7 -

W/S W/S W W/S W -

Variant/ culturebound word CBW CBW CBW CBW CBW -

Table 2-2 Overview of the Canadianisms of French origin in ICE-Canada.

As shown in the last column of the table, all lexical items behave like culture-bound terms, i.e. they always fill a linguistic and cultural gap that no variant can replace. Such behaviour is probably due to the type of corpus that has been used to carry out this analysis. It would be indeed interesting – and hopefully more consistent – to extend the analysis beyond the International Corpus of English and create a written corpus made of newspaper articles where also the other Canadianisms of French origin may be searched for. So far, the data in ICE-Canada suggest that most of them are seldom used since they often refer to the colonial heritage of the country and are likely to be employed in specialised discourses only.

CHAPTER THREE HERITAGE ITALIAN IN TORONTO: ANALYSIS OF VERB VARIATION IN A SPOKEN CORPUS MARIA PARASCANDOLO

Introduction In this chapter I provide an analysis of variation and contact phenomena in the speech of three generations of Heritage Italian Speakers living in the Greater Toronto Area (GTA). Specifically, my analysis focuses on patterns of temporal/aspectual variations of verb form usage in a narrative story retelling task, namely the Frog Story (Mayer, 1969). The study is part of a larger project involving data collected from different informants using a sociolinguistic questionnaire. The paper is organized as follows. The next section deals with some problems related to the definition of heritage speakers. Implications for the use of tense/aspect markers in narrative spoken texts will be specifically presented at the interface of cognitive and semantic levels along with the methodology adopted to collect data and the criteria chosen for the corpus annotation. Finally, the criteria used for the evaluation of nonnative-like verb form usage will be discussed along with some preliminary findings.

Heritage language speakers: problems of definition Over the last few years, the term heritage language (HL) has been broadly used to refer to minority languages spoken by groups known as linguistic minorities whose members have been referred to as heritage language speakers (HLSs).

42

Chapter Three

These minorities include populations that have migrated to areas other than their home countries (e.g., Italians in Canada, Mexicans in the United States, Turks in Germany). In Canada, there is an official definition handed out by the government: if you speak a language that is neither an official language nor an indigenous language in Canada, you are a heritage language speaker as long as it is your mother tongue. In this paper we will focus on the Italian heritage speakers in the GTA. In the context of immigration, immigrant speakers, or sons and daughters of immigrant parents, may grow up acquiring and using two or more languages (Italian dialect, Italian and English) in order to meet their everyday communicative needs. According to Valdes & Figueroa (1994) they have been referred to as circumstantial bilinguals and distinguished from elective bilinguals who generally learn a second language (L2) in the classroom. At present, heritage speakers can be said to understand and use to some extent their home language even though they feel more confident in speaking the dominant language of the society in which they live. In psycholinguistics and language acquisition studies which have been carried out in the United States, a heritage speaker is a bilingual who has grown up in a family where an immigrant or minority language is spoken as the dominant language of the wider community since early childhood (Kondo-Brown, 2006; Polinsky & Kagan, 2007; Silva-Corvalán, 1994; Valdés, 2000). In this context, bilingual heritage speakers’ competence in the native heritage language (L1) is different from that of other native speakers (Montrul, 2004, 2006; Polinsky, 2006, 2008a, 2008b). According to Polinsky (2006), HLSs learn their languages in a way that exhibits a consistent pattern of simplification and loss which leads to partial acquisition of the language. Pires (2011) adds that HLSs are certainly able to attain an acceptable level of language competence, but they also exhibit greater variation than monolinguals. Therefore, HLS are defined in terms of how they speak the heritage language in the United States. Differently, in a sociolinguistic framework, we consider variation to be a central and natural characteristic of any language, so we do not make any a priori claim about the way in which HLSs speak. HLs may remain the same as, or diverge from, their source language even after being in a contact situation in the new community. In the present study, HLSs are people who belong to one of the generations defined by Walker and Hoffman (2010) as follows:

Heritage Italian in Toronto

43

1st generation refers to immigrants who arrived after the age of 18 and lived in the GTA for at least 20 years. They must have emigrated directly from the HL home country. 2nd generation includes children from the 1st generation who were born in the GTA or arrived there before the age of 5 directly from their home country. 3rd generation includes children from the second generation who were born in the GTA. Any sociolinguistic research in the Italo-Canadian community must take into account the complex nature of its linguistic repertoire that may include an Italian dialect, Italian and English. This complexity derives from the continuum along which both dialect and “standard” varieties coexist. The dialect continuum can have a local dialect at one end and a regional koiné at the other. More specifically, the Italian continuum begins where the regional dialect ends and reaches the standard variety (Alfonzetti, 1998; Berruto, 1987). This kind of research is interesting in the field of sociolinguistics as it allows the study of variation which has developed in a language contact situation. In the present study, we analyze in particular the language contact effects across three generations by keeping the familiar domain fixed as the social network. In having the familiar domain as a constraint, it is more difficult to collect data. So, in this study, we concentrated on the temporal-aspectual variation because, given the limited amount of data available, it was necessary to focus on a very common aspect of spoken language, namely the verb.

Grounding and temporality in narrative: problems and hypotheses This study considers how different generations of heritage Italian speakers deal with the verb form usage compared with native speakers in a narrative task. In our analysis of narratives we take into account the foreground/background distinction elaborated in the field of the gestalt theory of perception and applied to narrative texts by Hopper (1979) and Klein & von Stutterheim (1991). Hopper suggests that a major function of tense/aspect distinctions is to differentiate foreground events (main-line) from background (commentary) events.

44

Chapter Three

Following the publication of Labov and Waletzky’s study (1967), a narrative has been defined as a sequence of clauses whose order mirrors the sequence of events it is related to. Such clauses contrast with both “evaluative” clauses that give the point of the narrative as an interactive speech event and subordinate clauses that do not cause the narrative to progress. In 1979 Hopper further clarifies that in a story-line, the essential events in a sequence tend to be dynamic, punctual and completive, whereas supportive background information is frequently descriptive of various states and durative situations. He supposed that if a language marks its verbs for perfective aspect (i.e. completive, non-durative, punctual), such forms tend to occur in clauses which relate to the events set in the foreground; furthermore, if it marks its verbs for imperfective aspect (i.e. durative, progressive, iterative), such forms tend to occur in the clauses set in the background. In other words, the more an event is salient to the speaker, the more it will be set in the foreground, and the less an event is salient to the speaker, the more it will be placed in the background based on verbal and aspectual markings. Besides we need to consider that grounding is a matter of perspective: foreground and background are constructed by the speaker from time to time (Berman and Slobin, 1994). Indeed, we assume that there is a correlation between cognitive level (foreground and background events) and semantic encoding in both main and secondary clauses. Specifically, we expect that foregrounded events are mostly encoded by main clauses while backgrounded events are mostly encoded by secondary clauses. In this study, a clause is the basic unit assumed for the analysis. It is identified as a unit that is headed by an inflected predicate (Berman and Slobin, 1994). Following Labov & Waletzky (1967) and Klein & von Stutterheim (1991), clauses are classified as main clause, side clause and comment clause. Even though a clause is generally identified by syntactic criteria, we adopt here semantic criteria to classify main and side clauses as well. Main clauses belong to the narrative plot-line or to the narrative skeleton, while side clauses belong to that part which is not relevant to the plot-line. Comment clauses such as personal judgments are not considered in this work. As far as verb form usage is concerned, we expect the following correlation between levels of analysis: Perfective marker > Main clause > Foreground event Imperfective marker > Side clause > Background event

Heritage Italian in Toronto

45

Data collection methodology This section presents the methodology adopted in the study. The data discussed here were collected during recording sessions performed using the Sociolinguistics interview designed by Naomi Nagy (2011) in The Heritage Language Variation and Change project (http://projects.chass. utoronto.ca/ngn/HLVC/) and a story retelling task, i.e. the Frog Story (Mayer, 1969). Following the HLVC protocol, during the sessions, the participants were put at ease by answering the questions of the sociolinguistic interview. This approach allows the informants to feel more relaxed before the elicitation of the narrative task data which is instead more formal and structured. I collected around twenty hours of speech data by sociolinguistic interview and four hours and an half by story retelling task. In total, I worked with 18 speakers grouped by families recording one speaker for each generation. I chose the family as the basic informant group, but the selection of the speakers also included the criteria mentioned before. Because of the small number of speakers, it was possible to establish a close relationship with all the participants in the study. So the effect of my presence as well as of a tape-recorder was minimized and did not affect the language used. For ethical reasons, each participant was informed about the tape-recorder before switching it on and had to sign an informed consent document, even though the attention drawn to the tape-recorder probably resulted in less spontaneous data.

Elicitation method: The Frog Story This narrative task is suitable for testing the encoding of temporalaspectual distinction because it shows a series of events that are either sequentially ordered, simultaneous, ongoing, or completed. Each speaker is asked to tell a story in Italian, based on the pictures depicted in the book and to begin with the phrase c’era una volta (‘once upon a time’) in order to set an invariable and familiar starting point for narratives. The Frog Story is a a picture book without comments representing a story with a simple structure based on a little boy who has lost his pet frog. The story involves a problem (the boy has a pet-frog which runs away), a set of actions resulting from this problem (the boy and his dog go searching for the missing frog) and a resolution (the boy finds his frog). Following Labov & Waletzky (1967), this story was segmented into episodes including the following parts:

Chapter Three

46

• • •

Subplot onset Subplot unfolding Subplot resolution

Corpus annotation The first step was the orthographic transcription of the corpus with the support of Trascrivi (http://www.trascrivi.it/), i.e. a semi-automatic tool owned by Cedat. The tagged data were later manually checked. Secondly, each text was segmented into clauses. As already mentioned, a clause is identified as a unit that is headed by an inflected predicate (Berman and Slobin, 1994). Each clause was marked semantically as main, side or comment as shown in Table 3-1. CLAUSE TYPE

CLAUSE TEXT

MAIN

un certo punto il bambino si è addormentato ‘at a certain point the boy has fallen asleep’

SIDE

e la rana stava sempre al suo posto ‘and the frog was always in its place’

COMMENT

ancora non è finita questa storia ‘this story is not finished yet’

Table 3-1: Clause type example

Finally, all inflected verb forms in the clauses were marked using the set of tags shown in Table 3-2. CODE PRE IMP PS PC PP TRA FS FC

TENSE Presente ‘present’ Imperfetto ‘Imperfect’ Perfetto semplice ‘Simple past’ Perfetto composto ‘Compound past’ Piuccheperfetto ‘Plus-perfect’ Trapassato ‘Plus-perfect II’ Futuro semplice ‘Simple future’ Futuro composto ‘Compound future’

Heritage Italian in Toronto

P pro INF pre PART pass

47

Perifrasi progressiva ‘Progressive periphrase’ Infinito presente ‘Present infinitive’ Participio passato ‘Past participle’

Table 3-2: List of tags

Non-native-like usage evaluation Although each narrative is segmented into clauses, the evaluation of non-native linguistic uses is based on the sequences of clauses. Non-native uses involving tense/aspect did not come from isolated clauses but from clausal relationships. Three rules were identified to formalize the nonnative-like verb uses. The first rule refers to the temporal sequence: 1.

In the same episode, M clause with a tense value is not sequentially ordered with the previous one.

Table 3-3 provides an example of the annotation containing a violation of the first rule.

CLAUSE TEXT CLAUSE TYPE VERB FORM TENSE DEICTIC MODE NATIVE-LIKE EPISODE

il bambino va dall’altra parte ‘the boy goes to the other side’ M va PRE Y Indicative Y EP3

e ha vista la rana ‘and he has seen the frog’

M ha vista PC Y Indicative N EP3

Table 3-3: Temporal sequence

Here is the glossed example1:

1

I followed the Leipzig glossing rules to mark the examples. However, I had to add some category labels not included in the Leipzig tagset. Specifically, I added the label PUN to mark a punctual aspectual value of a verb form and TEL to mark a telic verb.

Chapter Three

48 il the e and

bambino boy

va goes-PUN

ha he has

dall’ to the

vista seen-PFT

altra other la the

parte side rana frog

“the boy goes to the other side and he has seen the frog”

In the second clause, the speaker used a compound perfect verb form to convey a present meaning which caused a break in the referential order of the actions. Compound past in Italian is characterized by a past tense value that is relevant to the moment of speech, and by a perfective aspectual value (Bertinetto, 1986). It is used to refer to a resulting effect of a finite past action and to encode events that are still relevant at the time of speaking. Here, the past tense meaning is incorrect because it does not represent the chronological/referential order of actions, whereas the perfective aspect is correctly expressed. As in many other cases, the aspectual value seems to be more relevant than the choice of tenses in the speech of heritage speakers. Both the second and third rules are concerned with form to function mappings. 2.

Relative clause with a punctual aspect value form is used to convey imperfective aspect value (durative aspect).

Table 3-4 provides an example from the corpus.

CLAUSE TEXT CLAUSE TYPE VERB FORM TENSE DEICTIC MODE NATIVE-LIKE EPISODE

il bambino resta per terra ‘the boy stays on the ground’ M resta PRE Y Indicativo Y EP6

e il cervo che guarda il bambino ‘and the deer that looks at the boy’ S guarda PRE Y Indicativo N EP6

Table 3-4: Form to function mapping

Here is the glossed example: Il

bambino

Resta

per

terra

Heritage Italian in Toronto

49

The

Boy

stays-PUN

on the

floor

E

Il

Cervo

che

guarda

il

bambino

that-REL

looks-PUN

the

boy

And

The

Deer

“the boy stays on the ground and the deer that looks at the boy”

Third rule: 1.

Most of telic2 verbs are used in a PC form even if they convey an aoristic aspectual value.

Table 3-5 provides an example from the corpus.

CLAUSE TEXT CLAUSE TYPE VERB FORM TENSE DEICTIC MODE NATIVE-LIKE EPISODE

e poi questo cervo corre con questo bambino ‘and then this deer runs with this boy’ M Corre PRE Y Indicative Y EP6

e sono arrivati a una caduta ‘and they have arrived to a fall’ S Sono arrivati PC Y Indicative N EP6

Table 3-5: Form to function mapping

Here is the glossed example:

2

Telic means a semantic characteristic of a verb lexeme, named azionality or lexical aspect, indicating a process that is finalized to reach a goal. The distinction telic/atelic is based on the Vendlerian classification of verbs into states, activities, accomplishments, and achievements. The former two classes are traditionally grouped together as atelic predicates (e.g., love, sleep, play), because they do not involve an inherent endpoint. The latter two classes, on the other hand, are said to include a potential endpoint as part of their meaning (e.g., realize, learn), and are categorized as telic.

Chapter Three

50 e And

poi then

questo this

E And

sono they have

cervo deer

corre runs-PUN

arrivati arrived-TEL-PC

a to

con with

una a

questo this

bambino boy

caduta fall

“and then this deer runs with this boy and they have arrived to a fall”

As far as tense/aspect usage is concerned, this example represents the most frequent case of violations in the speech of heritage Italians. If a verb is telic, it tends to appear in a compound perfect form, even though it conveys an aoristic aspect. This tendency may confirm the hypothesis that the lexical aspect is more relevant than the grammatical aspect to heritage Italian speakers. This tendency was also attested in heritage Russian speakers (Polinsky, 1997, 2009). Russian native speakers express grammatical aspect via the perfective-imperfective opposition in the verb form, whereas heritage speakers of Russian tend to use the perfective morphology mostly with telic verbs. Consequently, atelic verbs tend to become invariably imperfective in a heritage grammar of Russian (Polinsky, 1997, 2009). Aspectual morphology in heritage Russian is therefore said to encode lexical (telic-atelic), rather than grammatical (perfectiveimperfective) aspect (Pereltsvaig, 2002, 2004). This seems to be a tendency for heritage Italian as well, but we do not have statistical backing to confirm this tendency yet.

Non-native like usage: other examples In our corpus there are many non-native uses that we do not mark as non-native because our interest is mainly focused on the tense/aspect usage violations. Here are shown two frequent non-native uses not considered in this study: a) dal from the

lack of verb-subject agreement: nido nest

è is

uscito escaped-M.SG

“from the nest is escaped the bees”

le the

vespe bees-F.PL

Heritage Italian in Toronto

51

b) auxiliary selection with intransitive verbs: per by

caso chance

il the

cane dog

ha has- AUX

caduto fallen down

“by chance the dog has fallen down”

Results In this section we present initial findings resulting from the analysis of the data. Figure 3-1 shows the distribution of clauses in the corpus. The graph shows that first generation speakers feel more confident with Italian (as we expected), because they speak and comment in their L1 more frequently than the other generations do. The graph also shows that main clauses are more frequent than side and comment clauses in all the generations.

Figure 3-1: Distribution of clauses

Figure 3-2 shows the distribution of verb forms in the corpus. As you can see, the present tense is the most used tense in the corpus, whereas the simple past is completely absent, even if speakers came mostly from a central-southern area of Italy where the use of this tense is particularly widespread. The distribution of this verb form confirms the expansion of the compound past at the expense of the simple past in heritage Italian, even in the case of narrative texts.

52

Chapter Three

The Italian present tense can cover a wide range of uses. It can refer to situations that coincide with the speech time such as an event which begins in the past and continues in the future and, typically in narrative texts, it can occur to describe past events as historical present tense. Therefore, this finding is not inconsistent with what we would expect from Italian native speakers. Moreover, the imperfect tense appears not to be frequently used by the informants.

Figure 3-2: Distribution of forms (all generations)

Figure 3-3 shows the distribution of forms across generations. When we look at the distribution of verb forms across the three generations, we observe a significant decrease in the use of the present tense in both the second and third generations. We can possibly explain this effect as being due to the higher level of competence of the first generation speakers. They seem to be aware of the fact that in Italian it is acceptable to tell a story using the historical present, but both the second and third generations appear to be more conservative in the use of the past, and are not familiar with using the historical present.

Heritage Italian in Toronto

53

Figure 3-3: Distribution of verb forms among generations

Figure 3-4 shows the different levels of competence across the generations. First, we observe that both the second and third generations used a higher number of non-native verb forms with respect to the first one. Moreover, the graph shows that non-native verb form usage is more consistent when comparing the third generation with the second one. It appears that in the second generation, the highest variability in the competence of the speakers occurs. In this generation, in fact, very competent speakers can be found together with speakers whose level of competence is similar to the one exhibited by third generation speakers.

Figure 3-4: Non-native like usage variations across generations

54

Chapter Three

Conclusion and future work Overall, two main findings can be reported from these data. The first one deals with the increase in the number of non-native-like uses of verb forms across generations, while the second one is the significant decrease in the use of the present tense in the narrative task starting from the second generation. Furthermore, we observed a tendency to stabilization in the number of non-native-like uses between the second and the third generation. Future work will primarily focus on a comparative study of those data with data from native Italian speakers who will perform the same narrative task in Italy. A further goal of future research will be to compare the results of this study with the narrative sections contained in semispontaneous data collected using the sociolinguistic interview.

SECTION II VARIATION AND CHANGE IN AFRICA

TOWARDS A STANDARD WEST AFRICAN ENGLISH GERHARD LEITNER

In this section we will move on to Africa with a closer look to West African countries where English was first brought by missionaries and traders and later by colonisers. In these countries it is only after the postcolonial period that English has acquired the status of dominant language of politics, communication, education, and of social advancement in general. Before this last development, English had played an insignificant role among non-elitist social groups and a prominent as well as widespread use among leading social classes in some regions. It is worth noting that English developed across African regions in a quite different way as first, second and pidgin variety and under different circumstances that have determined its true nature. In particular, English came very early to West African countries through trade routes and the slave trade; in fact it became stronger with the repatriation of former slaves in Sierra Leone and Liberia from where they started mission work along West Africa’s coastline (Leitner 2012). It developed into West African Pidgin English as means of communication to facilitate interactions with local peoples. It was only in the nineteenth and in the twentieth century that English acquired an important status in Sierra Leone, Nigeria, Ghana along with the English Pidgin becoming a symbol of prestige and power (Schneider 2011, 135f.)1. Today in most of Britain’s former colonies English is spoken as either a second or official language with different degrees of penetration. To mention one example, in Nigeria, which is the most anglicized country, the local Pidgin is perceived as an informal variety of English (ib.139) and in Ghana English is the official language along with its stigmatised Pidgin. 1

To understand the complex language context in Africa, it is worth recalling the classification Samarin (1962, 56) proposed for classifying a lingua franca as natural, pidginized, creole or artificial. By natural language is meant a language acquired by normal processes of enculturation. When a natural language is acquired as a second language it may lose some of its linguistic features and become a simplified / pidginized variety.

58

Section II

In particular, the English language in Nigeria is the major means of communication in every aspect of society, including education, commerce and governance, but most importantly, it is the lingua franca among the over 200 ethnic nationalities that make up the Nigerian state (Ogoanah 2011, 201). The forms and functions of this variety have been well documented in the literature, mostly focusing on phonology, syntax, and lexicosemantic analyses. As a result, it is generally agreed that a distinct variety called Standard Nigerian English now exists and is used by educated groups of speakers also at university level of education. In Africa, it is possible to identify a general trend in the development of English which is rapidly growing, spreading further and getting indigenised [in which] two trends are most noteworthy: a process of grassroots diffusion and nativization, and young people’s love of mixing it with local languages (ib.140).

In Nigeria, for instance, the distinction is getting blurred with linguistic structures continuously borrowed from English to pidgin and vice-versa. There is no one African English. English is best seen as the aggregate of different traditions and a highly heterogeneous entity. It has become a literary language that reflects important aspects of African identities […] West African experts suggest that West African features are quite regular and could become features of Standard West African English. Endonormativity seems to be on the horizon (Leitner 2012, 221ff).

From this perspective along with the theoretical framework of the book presented in the general introduction, the findings reported in the following chapters can be interpreted from a cross-linguistic perspective as well as generalisations in the domain of linguistic typology. As a matter of fact, the two papers in this section give a significant contribution to the understanding of the process of standardization Nigerian English and Ghanaian English are currently undergoing towards the identification of Standard West African English. Brato presents recent auditory findings suggesting that specific affricated realisations in Ghanaian speakers of English are on the rise in various positions and that these may become a supra-ethnic Ghanaian marker. Also Mazzoli presents the results of a research on the tonal analysis of the item DE in Nigerian Pidgin aiming to isolate specific verbal property items in order to attest their fluctuating semantic content (stative/nonstative) and labile syntactic behaviour (intransitive/transitive).

Towards a Standard West African English

59

From a different standpoint, Mazzaferro addresses the issue of the status and functions of postcolonial Englishes beyond the boundaries of their speech communities dealing with the spread of English-speaking communities from former British and American colonies in countries (e.g. Italy) where English is historically a foreign language (EFL).

CHAPTER FOUR A PILOT STUDY OF ACOUSTIC FEATURES OF WORD-FINAL AFFRICATED /T/ AND /TS/ IN EDUCATED GHANAIAN ENGLISH THORSTEN BRATO

Introduction Ghana is a West African country located at the Gulf of Guinea and borders Côte d'Ivoire, Burkina Faso and Togo. In 2012 it had a population of about 25 million. Although more than 80 languages are currently spoken in Ghana, English is the only official language while nine languages have the status of so-called government-sponsored languages (Lewis et al. 2013). This means amongst other things that these languages are used as means of instruction in the early years of schooling. Contact with English speakers goes back as far as the late 16th century (Huber 1999: 86–95). Huber suggests (personal communication) that in terms of Schneider’s (2003) model of postcolonial Englishes Phase 1 (Foundation) lasted from about 1632 when the first English trading posts were established to the 1840s when some of the local kings gave judicial power to the British. Phase 2 (Exonormative stabilisation) lasted until Ghana gained its independence from Britain in 1957. Currently the country is located in Phase 3 (Nativisation), but there are first indications that the variety is moving into Phase 4 (Endonormative stabilisation) as there is both a growing body of literature written in English by writers born and educated in Ghana and complaints (most notably Nimako 2004) about the demise of the English language as Ghanaianisms (cf. e.g. Dako 2001) are gaining ground. On the phonological level the most striking feature of the nativisation process is the affrication of /t/ to [ts]. This results in pronunciations such as [tsaim] for time and [b‫ܭ‬tsa] for better. In word-final position it may lead to a (near-)merger as in the pair kit [kits] vs. kits [kits]. /t/-affrication is nei-

62

Chapter Four

ther mentioned in the early studies by Strevens (1954; 1965) and Sey (1973) nor does there seem to be any anecdotal evidence1 prior to Adjaye (2005) whose data was collected in the 1980s. Two impressionistic studies (Huber & Brato 2008; Failer 2010) have recently addressed this matter quantitatively and have shown that [ts] is spreading rapidly through the lexicon. Furthermore, the adoption of the variant seems to be socially stratified in that female speakers show significantly higher usage rates of the innovative variant than male speakers (see section 0 for a more thorough discussion). If this is the case – and more recent data collected in 2013 (Huber personal communication) seems to confirm the previous findings – this would suggest that what we currently see in Ghana could be a case in which a variety bypasses features typical of the transition from phase 3 to phase 4 in Schneider’s (2003) model and shows traces of phase 5 (differentiation into ethnic or other social varieties) ‘too early’. On the other hand it is well-known that females often are the driving force of linguistic change with males lagging behind and therefore we may just witness a change in progress into phase 4 which presently happens to show gender-based variation patterns. However, the present study is not so much concerned with this question. Rather, it aims to provide an analysis of some acoustic-phonetic features of /t/-affrication in educated Ghanaian English (henceforth GhE). The focus of this paper lies on the acoustics of [ts] in word-final position in comparison to realisations of the phonological /ts/ cluster in the same position. It relies on methods used in the description of /t/-frication and /t/affrication in other varieties of English, most notably Liverpool (Sangster 2001), Australia (Tollfree 2001) and RP (Buizza 2011; Buizza & Plug 2012), but will also apply more general concepts of the acoustics of affricates and fricatives to the Ghanaian context. A focus will be on the question whether differences found between males and females are socially or biologically determined. The more phonetic approach is also mirrored in the choice of data collection methodology. The data stems from only seven speakers who were recorded in a sound-attenuated booth reading out a 251-item wordlist particularly designed for this study.

/t/-affrication in varieties of English Plosive affrication refers to a phonetic process in which the release of the closure is not made rapidly but gradual. This usually results in a homorganic fricative sound in the release phase. As Cruttenden (2001: 160) 1

I am grateful to Magnus Huber for pointing this out to me.

A Pilot Study of Acoustic Features of Word-final Affricated /t/ and /ts/

63

points out “[i]n some varieties of English the alveolars /t,d/ may frequently be heard in affricated form [ts,dz]” in all positions in the word. In wordfinal position, affricated forms can be distinguished from the sequence /ts/ which occurs in plural or third-person forms such as kits [kܼts] and hurts [h‫ޝܮ‬ts] mainly by the “brevity of the friction associated with the affricated plosives” (160). Overall /t/-affrication has received comparatively little attention by scholars working on varieties of English. Studies which discuss acoustic parameters of /t/-affrication in English include Sangster (2001) and Watson (2007) on Liverpool, Tollfree (2001) on Australian English and Buizza (2011) and Buizza & Plug (2012) on RP. Sangster was particularly interested in how affricated forms of /t/ and /d/ could be distinguished from the fricatives /s/ and /z/ as well as the stop-fricative clusters /ts/ and /dz/. In her first experiment she compares the durations of the friction in the stop release to those of the fricatives and finds significant differences with fricatives being consistently longer in duration than lenited (affricated) stops. The second experiment is concerned with affricated variants of /t/ and /d/ vs. /ts/ and /dz/. Here, she compares the proportional durations of the hold and release phases. Compared to /ts/, the duration of the frication is proportionally significantly longer for [ts] variants. This pattern does not hold for the voiced pair (Sangster 2001: 407–409). Buizza’s (2011) study focuses on frication and affrication of /t/ in unscripted RP speech. The results of her auditory classification show a great variety of variants. Fricated forms are found most frequently on par with aspirated forms. Other variants (glottal stops, elisions, affricated stops, etc.) occur much less often. Besides an auditory classification in which fricated forms were the most frequent just before aspirated variants, Buizza (2011: 21–23) also carries out a range of acoustic analyses. She compares aspirated [t‫]ހ‬, affricated [ts] and fricated [t࡜ ] variants with regard to their mean durations in the hold and release phases, their mean amplitudes in the release phase and their mean centres of gravity (COG, see above). There are significant length differences between all three variants; affricated tokens are the longest, followed by aspirated and fricated variants. Affricated tokens have a significantly higher amplitude than aspirated and fricated variants, which do not differ significantly. The results for COG do not show significant variation, but values for fricated and affricated /t/s are higher than aspirated ones.

64

Chapter Four

/t/-affrication in Ghanaian languages and Ghanaian English All Ghanaian languages belong to the Niger-Congo language family and can be further divided into the Kwa and Gur sub-groups. The former are usually found in the south of the country, the latter are predominant in the north. The data for the current study comes from seven speakers from four different L1 (Akan, Ewe, Ga and Dagbani) backgrounds. The first three belong to the Kwa group; the latter is a Gur language. While in all four languages /t/ has phonemic status, the concrete realisation varies greatly. In Akan, the dominant L1 spoken by about two thirds of the population as their mother tongue or as a second language, the place of articulation is alveolar (Adjaye 2005: 162), but a dental articulation is suggested in older work by Christaller (1875 [1964]: 7). When /t/ occurs before /i/, /ܼ/ and /‫ܭ‬/, speakers of the Fante dialect have an affricated allophone [ts] (Adjaye 1989: 30).In speakers of Ewe, mainly found in Eastern Ghana and bordering Togo. /t/ is generally unaspirated and dental but aspirated variants may surface idiolectally (Westermann 1954 [1930]; cited in Adjaye 2005: 23). In addition, in Ewe /ts/ has phonemic status, however because of the language’s syllable structure does not occur wordfinally. Before /i‫ޝ‬/, the second element is retracted yielding a /t‫ݕ‬/ realisation (Duthie 1996: 15). For Ga (Adjaye 2005: 44) and Dagbani no affricated variants of /t/ or a phonemic /ts/ cluster have been reported. Previous work on /t/ in GhE by Adjaye (2005: 161) suggests ethnic variation in that Akan speakers prefer an alveolar plosive [t], but speakers of Ewe and Ga generally use a dental variant [tࡧ ]. In syllable-final position /t/ may be fully or partially glottalised or elided (Huber 2004: 858; Adjaye 2005: 167). The exact distribution of affricated variants of /t/ is not very clear. Adjaye (2005: 161) argues that [ts] “occurs in some individuals’ speech, particularly Akan and Gã” without being more precise about any ratios, but “in all positions except when precededby word-initial /s/ or followed by tautosyllabic /r w j/.” Huber(2004: 858) on the other hand suggests that only speakers of the Fante dialect of Akan “sometimes transfer this allophony [of affricated variants before high vowels] to English”. The first quantitative auditory analysis of /t/-affrication in GhE was carried out by Huber & Brato (2008) who analysed about 1,600 tokens of wordlist and reading passage data from 24 speakers (6 each with a Ga, Fante, Twi and Ewe background). They found clear style and gender patterns. In the wordlists [ts] was found much more frequently, the same is true for female speakers, who were clearly leading in the adoption of the affricated variant. Unlike previous findings, they found that a following

A Pilot Study of Acoustic Features of Word-final Affricated /t/ and /ts/

65

high vowel did not increase the likelihood of [ts]. More surprising, however, was the fact that there was no significant effect of L1. A very comprehensive analysis of the variation patterns of /t/ in GhE is provided by Failer (2010: 37-42; 46-51), who analysed data from twelve speakers of GhE from three ethnic backgrounds (Akan, Ewe, Ga) and a range of professions and age groups. She analysed three speech styles: wordlist, reading passage and sociolinguistic interview. There is a stylistic cline in the use of affricated forms. In the wordlist data more than two thirds of all tokens are [ts], in the reading passages there are 52% and in conversational style the figure stands at just under 37%. What is more interesting, though, is that she finds clear gender patterns with females consistently using affricated forms much more frequently than males. The differences between the three ethnic groups (Akan, Ewe, Ga) with regard to /t/-affrication were also quite striking, particularly in comparison to Adjaye (2005). Failer attests large-scale affrication in all three in all styles and by all three ethnic groups. In Adjaye’s data, these variants were largely restricted to Akan and Ga speakers. In fact, in her wordlist data, Ewe speakers (78%) have the largest amount of [ts], followed by Gas (68%) and Akans (56%). A preliminary auditory analysis of five speakers from the current sample (Brato & Schmidt 2012) confirms Huber & Brato’s (2008) and Failer’s (2010) findings with regard to the high frequency of affricated tokens in wordlist style; about 66% of all variants in all positions were identified as [ts].

Methodology Seven speakers of educated GhE participated in the study. They were between 21 and 26 years (mean: 22.7 years, sd: 2.1 years) and at the time of recording were students at the University of Giessen, Germany. Five of the speakers were exchange students who only spent a semester in Germany, two others (AKF1 and AKF2) were enrolled in a Master’sdegree. All speakers were born and raised in Ghana and had spent almost all their lives in the country. The speakers (see Table 1) come from four different ethnic (Akan, Dagaare, Ewe and Ga) and linguistic (Twi, Dagbani, Ewe and Ga) backgrounds whose relevant linguistic features were outlined above. This sample is of course by no means representative of GhE and additional data from more speakers would be highly desirable.

Chapter Four

66

Speaker AKF1 AKF2 EWF1 GAF1 DAM1 EWM1 EWM2

Gender female female female female male male male

Age 26 25 21 21 22 23 21

L1 Akan (Twi) Akan (Twi) Ewe Ga Dagbani2 Ewe Ewe

Table 1: Speaker sample

The speakers read out Deterding’s (2006) ‘Wolf passage’ followed by a 251-item wordlist. In the current paper only results for the wordlist are reported. The words were randomised and presented to the participants on a tablet computer mounted onto a music stand automatically in two-second intervals to avoid that speakers already see the next word(s) and the length of the list, and make sure that the speakers’ speech rate was held constant. The data was collected in a recording studio at the University of Giessen, Germany3. Speech samples were recorded in a sound-proof booth with a Neumann TLM 103 microphone and digitised directly to an Apple Macintosh computer running Pro Tools at a sampling rate of 44,100 Hz. The annotation and both the auditory and acoustic analyses were carried out in Praat (Boersma & Weenink 2012). The data was segmented manually on the basis of visual information from the oscillogram and spectrogram as well by auditory analysis. Relevant segments were identified and their components (e.g. hold phase and release phase) and their boundaries marked4. Two experiments were carried out. In the first experiment the variants of word-final /t/ were identified in an auditory analysis. These were 2

This speaker gives his ethnicity as Dagaare, but lists Dagbani as his first language and Dagaare as his second language. For both parents the order is reversed. In Dagaare /t/ is also alveolar; affricated variants are not attested (Somé 2003: 44). 3 I am grateful to Sebastian Schmidt and Timoty Hörl for their support in the recording process. 4 In 60 tokens preaspirated variants [‫ހ‬t] were attested. 39 occurred in /t/, another 21 in /ts/. Preaspiration is characterised by “a period of voicelessness at the end of the vowel, nasal or liquid preceding the onset of the stop closure” (Ladefoged & Maddieson 1996: 70) and can be identified in the oscillogram by its aperiodic pattern and in the spectrogram by its concentration of energy in the higher frequency regions. Preaspiration will not be further considered in the current paper although it occurs frequently in lenited stops in mainly female speakers of other varieties of English cf. e.g. Jones & Llamas (2003) on Middlesbrough English.

A Pilot Study of Acoustic Features of Word-final Affricated /t/ and /ts/

67

[t‫]ހ‬: an aspirated alveolar/dental voiceless plosive [ts]: an affricated alveolar/dental voiceless plosive [t]: an unaspirated alveolar/dental voiceless plosive [t࡜ ]: a fricated alveolar/dental voiceless plosive [tࡘ ]: an unreleased alveolar/dental voiceless plosive In experiment 2 a range of acoustic-phonetic parameters of affricate variants of /t/ are compared to those of the /ts/ cluster. In a first step the mean durations of the hold and release phases and the overall duration are discussed. For /ts/ the hold phase is referred to as the section from beginning of the oral closure to the release burst (i.e. [t]; the release phase is covered by the [s] articulation. In addition, zero-crossing rate (ZCR) and centre of gravity (COG) are being considered. (cf. e.g. Sangster 2001; Jones & Llamas 2008; Jones & McDougall 2009; Buizza 2011 for similar approaches on the affrication and frication of /t/ in varieties of English and ĩygis et al. 2012 on a range of Czech and Polish affricates). Prior to the acoustic analyses the speech recordings were downsampled to 22050 Hz. Frequencies below 2000 Hz were filtered out with a pass Hann band filter and 20 Hz smoothing in Praat. ZCR was calculated over the whole duration of the release phase, i.e. the fricated part in [ts] and the fricative part in /ts/. COG was measured in a temporal window of 50% of the duration centred at the midpoint of the release. ZCR is a measurement which calculates how often the speech signal crosses the 0 dB threshold in a given time and is usually measured in zero crossings per second. It can be used to identify sounds on a sonority scale. Voiceless fricatives have a higher sonority than voiceless affricates and voiceless plosives. In other words, and in the current context, this means that the more [s]-like a sound is, the higher will be its ZCR (cf. e.g. Harrington 2010: 61–63). Comparing ZCR in the release phases of [ts] and /ts/ this means that we would expect it to be higher in the stop-fricative cluster than the affricated stop. ZCR is measured here only for the raisers, i.e. those zero crossings in which the signal crosses the threshold rising from a negative to a positive value and over the complete duration of the release phase. COG is also referred to as the first spectral moment. Spectral moments analysis was first proposed by Forrest et al. (1988) as a means of classifying obstruent spectra. It assumes that a complex spectrum can be described by four statistical parameters, the so-called moments. They are the mean, standard deviation, skew and kurtosis. The spectrum is modelled as a probability density function – or histogram – consisting of as many individual bars as are set as the sampling rate (22,050 in this case). Both Har-

68

Chapter Four

rington (2010: 297–304) and Thomas (2010: 109f.) provide good nontechnical descriptions of spectral moments analysis. Spectral moment analysis has been used in a number of subsequent studies discussing acoustic properties of plosives, fricatives and affricates (amongst others by Jongman et al. 2000; Jones & McDougall 2009; Kirkham 2011; Buizza 2011). In the current study, only the first moment (COG) is considered because it has the greatest explanatory value and also because a more detailed analysis would be beyond the scope of the current paper. COG describes the mean frequency of the spectrum and “is computed by multiplying each frequency sample by its spectral energy, summing those products and dividing that number by the sum of the energies in the sample.” (Thomas 2010: 109). Similar to ZCR measurements, sounds with a greater sibilance are characterised by a higher value. This means that we would expect /ts/ to have a greater COG than [ts]. The statistical analysis was carried out in R (R Development Core Team 2013) using two-way analysis of variance (ANOVA). The acoustic measurements were selected as dependent variables. The release variants ([ts], /ts/) and gender (female, male) as well as their interactions form the independent variables. In order to identify which factor levels differed significantly Tukey-HSD post-hoc tests were applied to the results of the ANOVA.

Findings This section presents the results of the analyses described in the previous section. The first part will outline and discuss the findings of experiment 1, the auditory analysis of /t/. This will be followed by the results of the acoustic experiment carried out.

Experiment 1: auditory analysis of word-final /t/ Fig.4-1 shows the results of the auditory analysis of variants of wordfinal /t/ separated by speaker. Affricated variants are by far the most common, accounting for over 83%. Aspirated variants are found in about 11.6% while the other three variants identified ([t], [t࡜ ] and [tࡘ ], subsumed in the figure as other) only play a subordinate role, jointly accounting for just 5.4%. Despite the small sample size, the general pattern is quite clear. For all speakers except EWM1, [ts] accounts for at least four out of five tokens and two speakers (AKF2 and EWM2) use it near-categorically. What is striking is that this also includes the Dagaare speaker from the north. For

A Pilot Study of Acoustic Features of Word-final Affricated /t/ and /ts/

69

100 N= 316

[aspirated t] [affricated t] other

80

% Variant

60

40

20

0 AKF1

AKF2

DAM1

EWF1

EWM1

EWM2

GAF1

Fig.4-1. Distribution of variants of word-final /t/ by speaker (in %) this ethnic group /t/-affrication has not yet been mentioned. Whether this is idiosyncratic or part of a pan-ethnic trend towards [ts] in GhE remains to be confirmed in subsequent studies. A careful comparison to Failer (2010: 37) shows a strong increase in affricated variants of about 15 percentage points. There are a number of possible explanations. One very likely cause is the difference in the means of collecting the data. Failer’s data were collected in the field following a sociolinguistic interview with a native Ghanaian speaker. The data for the current study were collected in a recording studio by native Germans. This would suggest that the affrication of /t/ is introduced to the system from ‘above’ in Labov’s (2006) terms, i.e. by a (relatively conscious) decision so that the more formal data collection method triggers this variant more frequently. Another possible explanation for these findings is the greater social homogeneity of the current sample. All speakers are students from middle-class backgrounds and are on average about eight years younger than those in Failer’s study. Incidental observations suggest that younger speakers use [ts] more frequently than the older generations.

Experiment 2: acoustic analysis of word-final [ts] and /ts/ In the current section the focus is on a range of acoustic measurements which allow distinguishing affricated variants of word-final /t/ from the stop-fricative cluster /ts/ in the same position. We will first turn to the

Chapter Four

70

500 450

affricated t

/ts

affricated t

/ts/

257.8

384.2

238.1

426.6

131.3

266

116.5

328.4

120.1

118.2

119.2

97.2

400

Duration (ms)

350 300 250 200 150 100 50 0 female

male

Fig.4-2. Mean duration of the hold (grey) and release phases (white) and overall duration of word-final [ts] and /ts/ (in ms) separated by gender

durational measurements, shown here separated by feature and gender in Fig.4-2. The mean duration of the hold phase is in grey, the mean duration of the release phase in white. The values at the top indicate the overall duration. All measurements are in ms. There is only a small mean difference (1.9 ms) in the duration of the hold phase in the female group. It is about 120.1 ms for [ts] and 118.2 ms for the cluster. In both variants, the standard deviation is high at about 38.3 ms in the former and 39.4 ms in the latter, indicating there is a large variation in the actual realisation. Further analysis looking at more – particularly internal – factors such as phonological context – will hopefully yield an explanation for this pattern. Overall, this suggests that for female speakers hold duration is not a distinctive feature. In the male speakers on the other hand, the difference is much larger. Variants of affricated /t/ have a mean duration of 119.2 ms compared to only 97.2 ms for the /ts/ cluster. The standard deviations are equally large than those of the females. The differences found constitute a very significant interaction effect of variant and gender, F (1,460)=7.518, p