The Typology of Asian Englishes

The Typology of Asian Englishes

Benjamins Current Topics Special issues of established journals tend to circulate within the orbit of the subscribers of those journals. For the Benjamins Current Topics series a number of special issues of various journals have been selected containing salient topics of research with the aim of finding new audiences for topically interesting material, bringing such material to a wider readership in book format. For an overview of all books published in this series, please see http://benjamins.com/catalog/bct

Volume 33 The Typology of Asian Englishes Edited by Lisa Lim and Nikolas Gisborne These materials were previously published in English World-Wide 30:2 (2009), under the editorship of Edgar W. Schneider.

The Typology of Asian Englishes Edited by

Lisa Lim The University of Hong Kong

Nikolas Gisborne University of Edinburgh

John Benjamins Publishing Company Amsterdamâ•›/â•›Philadelphia

8

TM

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

Library of Congress Cataloging-in-Publication Data The typology of Asian Englishes / edited by Lisa Lim, Nikolas Gisborne. p. cm. (Benjamins Current Topics, issn 1874-0081 ; v. 33) Includes bibliographical references and index. 1. English language--Asia. 2. English language--Variation--Asia. 3. English language-Social aspects--Asia. 4. English language--Study and teaching--Asia. I. Lim, Lisa. II. Gisborne, Nikolas, 1966PE3501.T97â•…â•… 2011 427’.95--dc23 2011026797 isbn 978 90 272 0252 9 (Hb ; alk. paper) isbn 978 90 272 8453 2 (Eb)

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

Table of contents Acknowledgments The typology of Asian Englishes: Setting the agenda Lisa Lim and Nikolas Gisborne The Asian typology of English: Theoretical and methodological considerations Umberto Ansaldo

vii 1

11

Aspects of the morphosyntactic typology of Hong Kong English Nikolas Gisborne

27

Typological diversity in New Englishes Devyani Sharma

49

Thai English: Rhythm and vowels Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

75

Revisiting English prosody: (Some) New Englishes as tone languages? Lisa Lim

97

Index

119

Acknowledgements The papers in this volume were originally presented at the 1st International Conference for the Linguistics of English (ISLE1), with the theme “Setting the Agenda”, in Freiburg in October 2008, in the workshop “The Typology of Asian Englishes”, organized by Lisa Lim, and we thank the conference organizers, Bernd Kortmann and Christian Mair, for their most positive support. The collection of papers was then published in 2009 as a special issue of English WorldWide 30:2. We wish especially to express our gratitude to the editor of English World-Wide, Edgar Schneider, not only for being so amenable to a thematic issue – the first special issue in EWW – but also, and more crucially, for appreciating the significance of the innovativeness of the topic and embracing our work on it; we thank him for his enthusiasm, support and trust throughout the project, and are glad to have had EWW as an ideal platform for this collection. We are very pleased that our special issue has now been selected for inclusion in the series “Benjamins Current Topics”, and we extend our thanks to Anke de Looper for her support in both this and the original publication. We also thank Umberto Ansaldo for his comments on the papers as well as his advice on other aspects of this project.

The typology of Asian Englishes Setting the agenda Lisa Lim and Nikolas Gisborne

The University of Hong Kong / University of Edinburgh

1. Why typology? Why Asian Englishes? The emergence of myriad varieties of English world-wide has resulted over the decades in scholarly approaches to categorizing the varieties into broad types, for example, the well-known and widely used Three Circles of English (Kachru 1985), or into regional groupings of Asian Englishes, African Englishes, etc. (e.g. Kortmann et al. 2004; Schneider et al. 2004; the journal Asian Englishes; the book series Asian Englishes Today; etc.). But while such categorizations are revealing when used in the ways they were originally intended, namely, for grouping and appreciating English varieties according to their diffusion, status, functions, or geography, i.e. as sociohistorical entities, they are not meant to imply that members of a group by definition share structural properties, which can only be established after close scrutiny of all aspects of grammar; nor do the groupings mean common genesis, an extremely unlikely possibility considering the diversity of contexts from which such varieties emerge. This may seem to be stating the obvious, but a substantial amount of research still examines structural features of specific English varieties on the basis of their being an “Inner Circle” or “Outer Circle” variety, as if their membership of that class defined their grammatical structure; the New Englishes have long had a list of linguistic features ascribed to them as tendencies (e.g. Platt, Weber and Ho 1984), and regional groups like “Asian” or “African Englishes” have also often been described as exhibiting common features with the implication that this is by virtue of their belonging to that group (see e.g. Lim 2009 for further discussion on this). If the object of enquiry comprises the structural features of English varieties though, then other factors — and not a reliance on such classifications — are more material. Even if the approach goes further in establishing that a certain group does have certain linguistic features in common, as in Kortmann

2

Lisa Lim and Nikolas Gisborne

and Szmrecsanyi’s (2004) survey of shared features in varieties of English,1 the question is whether these should indeed be viewed as “universals of New Englishes” or in fact be accountable for by other factors. When considering the structure of New Englishes which have evolved in — multilingual, mostly postcolonial — contexts of Asia (thus, Asian Englishes), the significant factors to be considered are: 1) the varietyâ•›/â•›ies of the English lexifier that entered the local context; 2) the nature of transmission of English to the local population; and 3) the local, i.e. substrate, languages of the community in which the New English emerges (also see Hickey 2005: 506).2 It is worth noting that, as pointed out by Schneider (2007), settlement and transmission types are clear-cut and important mostly for the early phases of settlement, but tend to become increasingly blurred with time in the increasing complexity in the development of society (2007: 25). On the other hand, the substrate languages may be seen to be the more germane factor for accounting for the grammar of the emergent English, in particular at a later — e.g. current — point in time. With this third factor being the focus of this issue, only a brief elaboration of the first two factors follows. With regard to the type of English which comes into contact with the vernaculars, in Asian Englishes, the role of the lexifier can, at least superficially, be a “control”, for the most part being 19th-century British English, in the role of missionaries, teachers, administrators, at the time of formal colonization3 in India, Singapore and Hong Kong, or a variable, e.g. in comparison with the 20th-century American English lexifier of the Philippines. We should of course be cautious of over-simplification; after all, much regional and dialectal variation would have 1.╇ Kortmann and Szmrecsanyi (2004) calculate the relative strength of representation of a feature across varieties by averaging its “feature value” for all English varieties and world regions examined; this means that, for instance, even if a feature is not found in all but in three out of a total of four Asian varieties, it receives a high “feature ratio” score (0.75) for Asia and is included as “typical” for the region. The point is that these Asian varieties have a particular feature because one or some of their substrates do, and not simply because the feature is typical for the region, i.e. it is not necessarily a “universal” of New (or here, Asian) Englishes (for detailed exploration into this, see Sharma’s paper in this volume). 2.╇ It should be noted at the outset that an evaluation of the factors relevant to the evolution of a variety must of course also involve the recognition, first, that there are other important parameters involved, such as historical and political events, sociolinguistic determinants, and identity constructions, and, next, that these may be and often are distinct across the different phases or eras in evolution, which affect the dynamics of contact and the structural features that emerge in the evolving English differently at different points in time (see e.g. Schneider 2007 for a model for Postcolonial Englishes; Lim 2007 and 2010a for the situation for Singapore English). 3.╇ Though Asia’s initial contacts with English date back to the 1600s with the East India Company (Mesthrie 2008: 24).



The typology of Asian Englishes

been present, e.g. in Irish priests and nuns in the mission schools, and non-standard varieties would also have been part of the initial input, e.g. of sailors, soldiers, tradesmen, etc.4 Attention should also be paid to the second-generation users, and thus transmitters, of the emergent English, as well as the other Asian Englishes that would have been in the equation too, e.g. the varieties spoken by Indian and Ceylonese teachers in Singapore, the Eurasian varieties spoken throughout the region, and the English of the Peranakans in the Malay peninsula; more recently, the English varieties available via the media also play a important role for all Asian Englishes. Where the nature of transmission of English is concerned, while originally via the education system and thus acquired as an L2, consideration must also be given to the subsequent spread and evolution of the variety outside of the school, for instance, in the playground between children who have acquired (a variety of) English, and amongst the local population themselves, e.g. the early adoptors such as the Peranakans speaking Peranakan English (Lim 2010b). These are clearly extremely intriguing and significant directions for investigation, but are not the focus of this collection. What the five papers take on in this volume is the third factor: they investigate the structure of Asian varieties of English by exploring the relationship between the typological profile of substrate languages in the specific linguistic ecology and the grammatical features of the emerging contact variety of English — in-depth studies of which have been noted to be in need (Kachru 2005: 119) — which are argued to be an illuminating way of exploring the similarities and differences between contact varieties of English (e.g. Ansaldo 2009; Lim 2009). Why are Asian Englishes particularly interesting as an object of enquiry in this regard? We can identify a number of reasons for why they make for an exciting read, mentioning just three here.5 4.╇ Mesthrie (2008: 27) warns that the target language was often “a varied, vexatious and moving target”. 5.╇ The reasons listed here are those most closely relevant to the issue at hand, i.e. a typological investigation of the structural features of Asian Englishes. Obviously many other factors may be identified for why Asian Englishes are interesting objects of (sociolinguistic, cultural, etc.) study; Bolton (1992), for example, highlights three obvious connections: 1) all major states in the Asian region are confronting questions related to language policy and planning, which in many cases involves the adoption of English; 2) many Asian societies share linguistic and ethnic similarities, and co-opt in language planning; and 3) in all Asian societies the English language still has strong association with higher education, internationalism, modernity, job mobility and career development; these three connections may be seen, as pointed out by Kachru (2005), as contributing to a regional profile of English in Asia and to the gradual process of acculturation of Asian Englishes. An additional factor for why Asian Englishes call for attention is their

3

4

Lisa Lim and Nikolas Gisborne

Table 1. Summary of range of lexifiers and substrates of Asian Englishes Asian Englisha

Hong Kong Singapore Malaysia India Sri Lanka Philippines

English lexifier(s)

Substrate(s)b

Austronesian Dra- Indo-Aryan Sinitic vidian BrE AmE Other Malay Filipino Tamil Hindi Sinhala Can- Hok- Mantonese kien darin ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓

a╇

The list of Asian Englishes reflected here is not exhaustive. Here they are listed in an order so as to group the Asian Englishes with similar substrates together. Of course, in some cases there may be distinguishable regional subvarieties which have different substrates and exhibit different features, e.g. north, south, and northeast IndE varieties. b╇ The list of substrates reflected here is not exhaustive but is meant to represent the main or dominant varieties in each ecology; obviously many other languages are present in each ecology. The relative dominance of each substrate is also not reflected in this table, nor is any change in dominance at different time periods shown.

1) Diverse typologies. First, Asian Englishes show diversity on two fronts: they develop in contact with a rich range of languages, which are, for the most part, genetically unrelated to English, and which have typologically different grammars, such as Indo-Aryan, Dravidian, Austronesian and Sinitic, examples of which are summarized in Table╯1. This means that the emergent Englishes have the potential for displaying features typologically distinct from those of English. It also means that the collection of substrates offers us a comprehensive range of languages in which to view the dynamics of typologies in contact. For example, if Asian Englishes A and B have substrates X and Y respectively, with features P and Q respectively, do we see P and Q in A and B respectively (but not the other way around)? Or if Asian Englishes A and B both exhibit feature Q, is it because we find Q in substrates X and Y? In all permutations afforded, such careful and specific investigation permits us to query the universality question. Also evident in the table is the fact that different Asian Englishes have certain substrates in common, which afford a comparison to demographic profile, which is overwhelming and historically unparalleled (Kachru 2005: 206): the total English-using population of Asia is now more than that of the Inner Circle, and English is the main medium in demand for bi-╛/╛multilingualism in the region (Kachru 2005: 15); in other words, comprising 60% of the world population (Kachru 2005: 206), English users in Asia are a heavyweight, and warrant serious consideration in World Englishes.



The typology of Asian Englishes

attempt to answer questions such as, for example, if Asian Englishes A and B both have substrate X with feature Q, will feature Q be selected in both A and B? Singapore English (SgE) and Hong Kong English (HKE) both having Sinitic substrates Cantonese and Mandarin are a case in point, and a question may be whether the features of Cantonese particles and Sinitic tone manifest in both SgE and HKE (Lim this volume). Whatever the answers, the possibilities afford a contemplation of the contribution of substrate typology alongside other factors in the mechanics of contact. Related to this is the fact that the contexts in which the majority of Asian Englishes emerge involve multilingual communities: such linguistic and cultural pluralism certainly holds implications not only for bilinguals’ competence and creativity in appropriating a new variety (Kachru 2005) but also for the kind of identity alignments that multilingual speakers engage in (Ansaldo 2009). 2) Dynamic ecologies. The ecologies of emerging Englishes are always dynamic, as explicitly recognized in Schneider’s (2007) Dynamic Model of the evolution of Postcolonial Englishes. What is perhaps notable for Asian (and African) Englishes is how rapidly their ecologies have changed and continue to do so, in some cases within a matter of decades. Post-independence policies in Asia have had significant and swift impact on the ecologies, and consequently on the structure, of Asian Englishes. For instance, during colonial rule, the sociopolitical situations in Singapore and Malaysia were comparable, and the long-standing lingua franca of the region, Bazaar Malay, was dominant in both ecologies, and SgE and Malaysian English were seen as similar. However, language policies in Singapore in the second half of the 20th century meant not only that English became lingua franca especially in the younger generation but also that Mandarin became dominant; this together with a later ascendance of Cantonese, due in part to immigration policies, has led to a change in Singapore’s ecology to one that is more Sinitic-dominant and much closer, for example, to that of Hong Kong (see e.g. Lim 2009, 2010a, this volume for details) — does SgE then become more similar to HKE? Such dynamism makes for interesting investigations of Asian Englishes at different points in time. 3) Different phases. The various Englishes found in Asia represent different phases of the “spread” of English and thus of evolution: many of them, e.g. those of Brunei, Hong Kong, India, Philippines, Singapore, Sri Lanka, to name a few, are institutionalized varieties of the Outer Circle, where the English lexifier has been present in the ecology since colonial times and certainly very much entrenched in the ecology. However, because of other factors — date of independence, post-independence language and education policies, resources for English education, proportion of population having access to the language — the state of the evolution of the emergent English

5

6

Lisa Lim and Nikolas Gisborne

differs, with SgE considered as having already attained endonormative stabilization in Phase 4 of the Dynamic Model, while Indian English and HKE are still only in Phase 3 of nativization (Schneider 2007).6 At the same time there is the other group of Englishes in Asia, non-institutionalized varieties of the Expanding Circle, where the presence and significance of English is increasing. Such a range of stages of nativization and stabilization of the emergent English affords reflection on whether patterns of features observed may be considered L2 learners’ acquisitional patterns or patterns not significantly different from features deriving from the substrates of nativized varieties. 2. Thrusts and themes of the collection Rather than provide a blow-by-blow account of the five papers in this volume — by Umberto Ansaldo; Nikolas Gisborne; Devyani Sharma; Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire; and Lisa Lim, in that order — here we present the collection in terms of its thrusts and foci — its typology, as it were. The first paper in this volume, however, needs special mention at the outset. Intended to set the theoretical scene, Ansaldo’s paper presents a summarized but detailed account of an evolutionary perspective needed for viewing Asian Englishes, which involves a reflection of the typological matrix in which each variety evolves, in the sense of the pool of features that defines the multilingual speech community in which language contact takes place (Mufwene 2001, 2008; Ansaldo 2009); this, the paper proposes, is best achieved by seriously considering the typology of the substrate language(s) involved in the contact situation and the competition and selection between features of different grammars in the multilingual pool. It is in this spirit that the collection of papers in this volume is best appreciated. In terms of linguistic focus, the contributions in the volume span a range of structural features: two papers delve into the sound system, both segmental and suprasegmental, examining vowel realization and rhythm (Sarmah, Gogoi and Wiltshire), and tone and intonation (Lim); the other three papers take up various aspects of morphology and syntax, such as finiteness (Gisborne), tense and aspect (past tense and progressive marking for perfective and imperfective expressions) (Sharma), the copula (Ansaldo; Gisborne; Sharma), predicative adjectives, and topic prominence (Ansaldo); and discourseâ•›/â•›pragmatics is not neglected either, with attention to discourse particles (Lim). This range of topics demonstrates how the influence of the typologies of the substrates is evident at all levels of structure in the emergent Asian English. 6.╇ In fact, it is only very recently that the existence of a “Hong Kong English”, as opposed to a variety of “English in Hong Kong” has been recognized.



The typology of Asian Englishes

A trio of Asian Englishes is converged upon by all the papers: Indian English (IndE; Sharma), Singapore English (SgE; Ansaldo; Lim; Sarmah, Gogoi and Wiltshire; Sharma) and Hong Kong English (HKE; Gisborne; Lim; Sarmah, Gogoi and Wiltshire). As mentioned earlier, all Outer Circle varieties, SgE is considered as having already attained endonormative stabilization, while IndE and HKE are still only in the phase of nativization; this allows us a look at potential differences in emerging patterns. The apparent “outlier” which is the focus of one of the papers (Sarmah, Gogoi and Wiltshire) is a variety which has not garnered much attention to date; English in Thailand is a relatively recent phenomenon and considered an L2 variety, and in this sense is a valuable inclusion as an example of an Expanding Circle variety. In short, the papers provide a view of Asian Englishes which range from extremely established and nativized varieties such as IndE and SgE to very much newer ones still considered L2 varieties, like Thai English (ThaiE). Interestingly, both similarities and differences in the patterns observed across the varieties — e.g. in vowel realizations and rhythmic patterns respectively in ThaiE and SgE — find explanation in comparable features of the substrates, lending support to the argument for looking at typologies and not at “classifications” or “universals”. Where data is concerned, the various papers conduct their investigations using a range of methods and corpora: e.g. carefully controlled data are elicited for instrumental phonetic analysis (Sarmah, Gogoi and Wiltshire for ThaiE), spoken corpora are used for intonation patterns of spontaneous speech (Lim for SgE), and the International Corpus of English (ICE) of a number of varieties is utilized for close quantitative analysis of morphosyntactic features (Gisborne for HKE; Sharma for IndE and SgE). Finally, the papers are not just descriptiveâ•›/â•›analytical, but also boldly take on what are traditional classifications or analyses. For instance, the idea of “angloversals” is challenged, with specific investigation into three of Kortmann and Szmrecsanyi’s (2004) “candidates for universals of New Englishes” (Sharma); and it is proposed that the traditional classification of English as a stressâ•›/â•›intonation language be reconsidered and, instead, that some Asian Englishes be considered tone languages (Lim). 3. To conclude, but only for now We end this introduction with a note of caution: that even while we proceed a step further in taking the typology of the substrate languages seriously in a consideration of Asian Englishes, we must not fall into yet another trap of reductionism or assume a view that is blinkered.

7

8

Lisa Lim and Nikolas Gisborne

It is important to recognize, first, that it is not just the presence (or absence) of a feature in a substrate that determines the pattern in the emergent English, but also its frequency and consistency of use in each of the substrates, as Sharma’s paper clearly demonstrates. Even if two different substrates of two Asian Englishes may have the same feature, e.g. imperfectivity marking, one needs to scratch below the surface: a different distribution of the feature in the different substrates can indeed lead to different configurations in each Asian English. Hindi has a narrow progressive form and a robust, obligatory imperfective marker, and IndE speakers consequently interpret English -ing as a global imperfectivity marker; in contrast, Chinese has highly restricted and optional imperfective markers, and SgE -ing usage thus approximates that of Standard English. The importance of frequency of a feature across substrates is also seen in instances where typological congruence can shift the balance and increase the likelihood that certain features get selected over others, as shown in Ansaldo’s paper, where the congruence between features in Hokkien and Bazaar Malay leads to those features emerging in SgE. Finally, in line with the ecology paradigm (after Mufwene 2001, 2008) explicitly ascribed to in a number of the papers (Ansaldo; Gisborne; Lim), as well as approaches such as the dynamic model (Schneider 2007) and those which contextualize Asian Englishes in their functional realities (Kachru 2005), it is also crucial to recognize that the typologies of the substrates (and of the superstrates, or adstrates) are by no means the sole or main determinant of the emergent English, constituting only a component of the internal ecology. Factors of the external ecology are just as crucial in the process of selection and subsequent replication and reinforcement of features: these include the proportion that speakers of a language comprise in a population, and their prestige; the status and penetration that the New English has in the society; whether speakers adopt an endo- or exonormative standard; the kind and extent of identity construction, and consequent stability and focussing of the emergent English; and so on; and embracing all aspects of both internal and external ecology and its dynamism (see e.g. Mufwene 2001, 2008; Kachru 2005; Schneider 2007; Ansaldo 2009) must certainly take place for a complete appreciation of Asian Englishes.

References Ansaldo, Umberto. 2009. Contact Languages: Ecology and Evolution in Asia. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. Bolton, Kingsley. 1992. “Sociolinguistics today: Asia and the West”. In Kingsley Bolton and Helen Kwok, eds. Sociolinguistics Today: International Perspectives. London: Routledge, 5–66.



The typology of Asian Englishes

Hickey, Raymond. 2005. “Englishes in Asia and Africa: Origin and structure”. In Raymond Hickey, ed. Legacies of Colonial English: Studies in Transported Dialects. (Studies in English Language.) Cambridge: Cambridge University Press, 503–35. Kachru, Braj B. 1985. “Standards, codification, and sociolinguistic realism: The English language in the outer circle”. In Randolph Quirk and Henry G. Widdowson, eds. English in the World: Teaching and Learning the Language and Literatures. Cambridge: Cambridge University Press and The British Council, 11–30. Â�Â�——— . 2005. Asian Englishes: Beyond the Canon. (Asian Englishes Today.) Hong Kong: Hong Kong University Press. Kortmann, Bernd, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider and Clive Upton, eds. 2004. A Handbook of Varieties of English. Vol. 2: Morphology and Syntax. Berlin, New York: Mouton de Gruyter. ——— and Benedikt Szmrecsanyi. 2004. “Global synopsis — morphological and syntactic variation in English”. In Kortmann et al., eds.: 1142–202. Lim, Lisa. 2007. “Mergers and acquisitions: On the ages and origins of Singapore English particles”. World Englishes 26: 446–73. ———. 2009. “Not just an ‘Outer Circle’, ‘Asian’ English: Singapore English and the significance of ecology”. In Thomas Hoffmann and Lucia Siebers, eds. World Englishes: Problems, Properties, Prospects. (Varieties of English Around the World G40.) Amsterdam, Philadelphia: Benjamins, 179–206. ———. 2010a. “Migrants and ‘mother tongues’: Extralinguistic forces in the ecology of English in Singapore”. In Lisa Lim, Anne Pakir and Lionel Wee, eds. English in Singapore: Modernity and Management. (Asian Englishes Today.) Hong Kong: Hong Kong University Press, 19–54. ———. 2010b. “Peranakan English in Singapore”. In Daniel Schreier, Peter Trudgill, Edgar W. Schneider and Jeffrey P. Williams, eds. The Lesser Known Varieties of English: An Introduction. (Studies in English Language.) Cambridge: Cambridge University Press, 327–47. Mesthrie, Rajend. 2008. “Introduction: Varieties of English in Africa and South and Southeast Asia”. In Rajend Mesthrie, ed. Varieties of English 4: Africa, South and Southeast Asia. Berlin, New York: Mouton de Gruyter, 23–31. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. ———. 2008. Language Evolution: Contact, Competition and Change. New York: Continuum. Platt, John, Heidi Weber and Ho Mian Lian. 1984. The New Englishes. London: Routledge and Kegan Paul. Schneider, Edgar W. 2007. Postcolonial English: Varieties Around the World. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. ———, Kate Burridge, Bernd Kortmann, Rajend Mesthrie and Clive Upton, eds. 2004. A Handbook of Varieties of English. Vol. 1: Phonology. Berlin, New York: Mouton de Gruyter.

9

The Asian typology of English Theoretical and methodological considerations* Umberto Ansaldo

The University of Hong Kong

This paper looks at the emergence of Asian English varieties in terms of the evolution of new grammatical features. I propose that, in order to reach a thorough understanding of how the unique combination of grammatical features that define specific Asian Englishes come about, we must approach these features from a typological and evolutionary perspective which allows us to contrast them not only with Standard English varieties but also with the Asian languages with which these come into contact. As restructured vernaculars, Asian English varieties are de facto contact languages, and, as such, evolve as a consequence of selection of features from a multilingual pool. In this pool, features of Asian varieties play a significant role in determining the output grammar and must therefore be appreciated in their own right. In order to illustrate these points, I introduce an evolutionary view of contact language formation, and I present a set of features typical of Singlish, which are all instances of replication of Asian, not English, features. Keywords: Asian Englishes, Singapore English, Singlish, Malay, Sinitic, evolution, ecology, contact languages, multilingualism, frequency, congruence

1. Introduction What are Asian English varieties (AEVs)? They are the product of the presence of English in ecologies where other non-English, non-Standard Average European (SAE) languages are also spoken. They can be cases of English L1 in contact with other languages, English L2 usage, or instances of English L2/L3 transmitted informally within a linguistically diverse ecology. In any case, looking at AEVs falls under the study of language change, in particular, change where contact between *╇ I thank Nik Gisborne, Raymond Hickey, Lisa Lim, Stephen Matthews and Devyani Sharma for comments on this and other versions of this paper.

12

Umberto Ansaldo

structurally very different languages takes place. As such, AEVs can be regarded as partially restructured vernaculars (Holm 2004) and, in order to be understood, require a theoretical approach to contact language formation. Contact language formation, simply put, is a case of language change which leads to structural differences between the input and the output grammars. We can understand these types of changes in at least three ways: 1. Change as departure from the norm, i.e. the product of something gone wrong. This typically requires “negative” explanations, in the sense of imperfect acquisition or broken transmission. 2. Change as system-internal. In this reading, speakers are passive; changes happen to the grammar because of structural imbalances, system realignments, etc. In a radical interpretation of this, context does not matter in order to understand why change happens. 3. Change as evolutionary. Here, change is a core aspect of complex adaptive systems. The reason for change is ecological variation, and the mechanisms of change are selection and replication. In light of the fact that languages normally change over time, and that there is variation within speech communities of all types, interpreting changes to a grammar as indications of abnormal acquisition or transmission, i.e. in sense (1) above, appears to go against common sense. In fact, even in cases of radical restructurings such as those observed in creole languages, it has been pointed out that explanations that rely on exceptional circumstances are usually ideologically biased rather than empirically grounded (DeGraff 2001, 2003, 2005). In relation to (2), historical linguists have time and again pointed out that it is speakers who change languages, and that explanations for language change lie ultimately in the social history of a speech community, not in its grammar (Thomason and Kaufman 1988; Janda and Joseph 2003). While system-internal dynamics may indeed influence the direction of change depending on the congruence (or lack thereof) between the systems in contact, these dynamics are ultimately heavily dependent on circumstances that are external to the grammar and originate in the history of the speakers. We can thus say that there is good reason to be critical of (1) and cautious about (2); we thus turn our attention to (3). In Section╯2, I introduce the basic assumptions of an evolutionary approach to the study of contact language formation and grammatical innovation. In Section╯3, I apply these ideas to the study of Singlish grammar, and in Section╯4, I draw the conclusions and implications for the study of AEVs.



The Asian typology of English

2. Evolution and ecology The link between evolutionary biology and language change is an old one: the parallels go back to the use of the Stammbaum model to describe language speciation (see Mufwene 2008), and are explicitly drawn out in Lass’ (1997) idea of language as an evolving system. Within grammaticalization theory, we note a strand that looks at the evolution of grammar as an instance of ritualization (Haiman 1994), a possibility already implied, for example, in Givón’s (1979a) inherently diachronic approach to grammar. But perhaps the most significant development in terms of evolutionary theory can be seen in the work of Hull (1988), who extends the basic assumptions of evolutionary theory to account for the development of all conceptual systems, in particular the development of scientific thought. In diachronic linguistics, Croft (2000) proposes a model of language change as a complex adaptive system; Mufwene (2001, 2008) approaches the evolution of new languages as products of competition and selection in differential ecologies; and Schneider (2007) draws on these evolutionary and ecological perspectives in his detailed model specifically of the evolution of Postcolonial Englishes. What these studies have in common is the recognition that languages exist in a context that is inherently variable, i.e. linguistically diverse; variability means that speakers have a certain degree of choice as to which variables, whether phonological, lexical or syntactic, to use in a given instance. While some of these choices may be conscious, for instance in terms of variables that are assigned high social value, others are unconscious, and depend on matters of cognitive salience, typological dominance as well as frequency (see Ansaldo 2009). In this sense, features of different varieties in contact can be seen as being in competition with one another; it is the cumulative result of different choices, or different outcomes of competition, that explains why languages change and why speakers vary in their usage. Let us take a closer look at the processes involved. 2.1 Selection and replication In a theory of language that acknowledges the evolutionary dimension of language, usage implies linguistic variation and thus language change. Variation can occur at three different levels (Croft 2000, 2006). First-order variants are individual variants that occur in all subdomains of grammar (phonological, semantic, syntactic, etc.) and represent the natural individual differences in speaking (the same language); first-order variation manifests itself in idiolects. When first-order variants take on a sociological significance in a community, they become secondorder variation; second-order variation can be detected in registers that are an

13

14

Umberto Ansaldo

indication of social class, gender, profession, etc. When second-order variation becomes conventionalized in a speech community and leads to divergence, we have third-order variation, i.e. dialects, or languages, depending on their political status. Just like in other complex evolving systems, change is caused by selection and replication. When communicating, speakers select among the available variables and replicate linguistic features they have been exposed to. Selection can be influenced by a combination of factors, notably the social capital of the variables, the degree of congruence between the systems in contact, and the frequency of exposure to the different variables that speakers have (Croft 2000; Mufwene 2001, 2008; Ansaldo 2009). Replication can be more or less faithful, and can result in quasi-identical or innovative replication. Whether replication is identical or innovative is largely dependent on whether the groups in contact are engaged in social convergence or divergence; the most radical instances of innovative replication, i.e. the formation of a new grammar, is the result of a process of identity formation whereby a group sets itself aside from others by projecting a completely new cultural profile, of which language is a strong indicator (Ansaldo 2009). This is also related to the linguistic diversity of the contact ecology: the more diversity there is in the features available for selection, the more potential there is for innovative replication. More often than not, identical and innovative replication are combined in contact ecologies, as contact language formation often arises out of the tension between, on the one hand, the convergence to external linguistic norms, and, on the other, the preservation of a degree of linguistic differentiation for the purpose of in-group identification (Ansaldo 2009). 2.2 Multilingual ecologies and transmission It is obvious that the selection and replication patterns of linguistic features available to speakers differ from society to society. In particular, the ecology in which transmission of features happens determines the options available (Mufwene 2001): in a monolingual ecology there will be a limited set of similar features, while in a diverse multilingual ecology there will be many features in competition with one another, i.e. many different ways of saying the same thing. We can visualize this in Figures╯1 and 2 (from Ansaldo 2009), which illustrate the difference between the options available for selection in a monolingual and a multilingual ecology, respectively, where monolingual ecologies tend to be much more homogeneous than multilingual ones. Because of the diversity of options available, multilingual ecologies are by definition a natural locus of change, and multilingual speakers are natural innovators.



The Asian typology of English

Figure╯1. Monolingual transmission (Ansaldo 2009)

Figure╯2. Multilingual transmission (Ansaldo 2009)

2.3 The abnormality of monolingual ecologies for our theories of evolution Recall the fact that most societies around the globe have been and still are multilingual (Edwards 1996). Note also that schooling and regulated language transmission is predominantly a recent feature of modern societies and, where it existed in the past, it was mostly directed at small, privileged groups. This makes monolingual, normative ecologies, such as those that have existed in Europe for the past two centuries, extremely marked from the point of view of human history. The type of language acquisition that happens in these ecologies is not “normal”, because for thousands of years this is not how languages were transmitted. If we want to understand how contact vernaculars such as AEVs develop, we must bear in mind that they evolve in multilingual ecologies in which some variety of English represents only one set of features available to speakers. In the same ecology, other grammars are present, be they Chinese, Malay, Filipino or Hindi, and

15

16

Umberto Ansaldo

grammatical features of these languages also play a role in the selection and replication processes — in line with Weinreich’s (1963) observation that in contact situations superficial multilingualism is enough for language interference. Therefore, we should not assume that English was the one and only target for non-native speakers in the evolution of AEVs; rather, the speakers that contributed to the development of AEVs were busy selecting and replicating linguistic features from a pool within which English grammar constituted but a subset of choices available. It is important to realize that normal transmission is untutored, creative and involves more than one language in most colonial settings where AEVs emerge. That all these elements need to be taken into serious consideration is clear when we look at the evolution of Singlish (basilectal Singapore English, SgE)2 in Singapore. 3. Singlish typology The variety of English spoken by a majority of Singaporeans on an everyday basis has been referred to as Singapore English (SgE), Singapore Colloquial English and, most intimately, simply Singlish. All these terms indicate that it is a spoken variety related to English that is unique to Singapore. The uniqueness is easily detected in all those aspects of its usage that distinguish different languages, or different dialects, from one another: pronunciation (or sound system), words in usage (or vocabulary), and ways in which sentences are constructed (grammar). It is considered a variety of English because much of its vocabulary is drawn from English, even if words of Malay, Hokkien and Tamil origin are also part of it. In mixing elements of different languages, Singlish takes its place among socalled contact languages, i.e. grammatical systems that evolve in multilingual ecologies (Platt 1975; Gupta 1994; Ansaldo 2004). Singapore’s ecology has been characterized by communicative systems of this type since its foundation. During colonial rule, Bazaar Malay varieties, in which Malay and Chinese features blend, flourished in the region; in Singapore, such a colloquial variety of Malay played an important role as de facto lingua franca that allowed interethnic communication until very recently (Ansaldo 2004; Lim 2007, 2010a), a function it still has in certain parts of society (Khin Khin Aye 2005). A related variety is Baba Malay, a unique blend of Hokkien and (Bazaar) Malay, typical of the Peranakan communities of Malaysia and Singapore (Ansaldo, Lim and Mufwene 2007). 2.╇ While the majority of scholarship, including the other papers in this volume, refer to this variety as Singapore English (SgE), I adopt the term Singlish, largely to underline the fact that extensive divergence from standard English is found in the basilectal variety of SgE, and it is in the features of the non-English languages that much of its grammar finds explanation.



The Asian typology of English

In addition to the above, Hokkien (southern Min group, spoken originally in Fujian province and Taiwan) and Cantonese (Yue group) have been the most important Sinitic influences in the ecology in which Singlish evolved, with Mandarin a more recent addition (Lim 2007, 2010a). With post-independence language and education policies in the 1960s and after (for details see e.g. Lim 2007, 2010a), eventually English replaced Bazaar Malay as lingua franca, and what was first an English-based lingua franca became Singlish. Today Singlish is the native language of a majority of young Singaporeans, which, together with (Singapore) Mandarin, forms the bilingual profile that defines a majority of Singaporean citizens (though additional languages are also spoken, such as Malay, Hokkien, Cantonese, Tamil, etc.), and is used in the home, on the streets, in the playground and in school. Perhaps the best way to approach Singlish grammar is to realize a fundamental difference between, on one side, English and, on the other side, Sinitic3 and Malay. While English is classified as a mildly inflectional language, meaning that it makes a certain use of morphology for grammatical purpose, Sinitic and Malay, the dominant languages of Singapore’s linguistic ecology, are typically isolating, meaning that there are hardly any morphological processes to speak of. This is true of many creole languages of the Caribbean, too, as well as of other contact languages, a fact that has led a number of scholars to argue that morphology is “lost” in the histories of these languages. However, as already pointed out by Givón (1979b), there could be another explanation: that the languages that evolve out of specific contact situation inherit the morphology of the so-called substrate languages. In the case of Atlantic Creoles these would often involve West African languages of the Kwa type, typically isolating. In the case of Singapore, as we will see, it is indeed in the nature of the non-English languages that much grammar finds explanation. In order to illustrate this, I focus on the following features: (i) zero copula, (ii) predicative adjectives (or property verbs), and (iii) Topic prominence.

3.╇ The examples that are used as illustration in the rest of this section are drawn from a number of different Sinitic languages, such as Cantonese and Mandarin, which, as mentioned earlier, are part of the Singapore ecology. While different Sinitic languages exhibit significant grammatical divergence in certain areas of grammar, with regard to the three features addressed in this paper, the various Sinitic languages behave in similar ways.

17

18

Umberto Ansaldo

3.1 Three features of Singlish grammar In the account of Singlish grammar in this paper,4 I have selected the features of zero copula, predicative adjectives (or property verbs), and Topic prominence for a number of reasons, primarily because they show innovation with respect to English varieties (though Irish English, for example, does have zero copula; see Hickey 2007), but also, importantly, because (a) they are typical of Sinitic and Malay varieties, i.e. the relevant substrates, and (b) they are typologically related, i.e. their grammatical functions are, to a degree, interrelated. As such, they are clear evidence of the fact that typologically prominent features are likely to be selected from the variables in contact; the prominence is a result of the congruence between Sinitic and Malay type, further strengthened by the systemic relation between the features in question. 3.1.1 Zero copula Consider the following sentences in examples (1) and (2). (1) This your car? ‘Is this your car?’ (2) Careful, laksa very hot. ‘Careful, the laksa is very hot’

From a normative English perspective, it might be observed that the copula to be is missing. This occurs very frequently in Singlish, and one might conclude that Singlish is incorrect vis-à-vis English (e.g. in the ‘deficit’ view of non-native varieties of English; Quirk 1990). However, from a typological point of view, what should first be recognized is that zero copula is in fact a common feature of many languages spoken in different parts of the world, as illustrated in examples (3) to (6). (3) Che1 hou2 wu4jou4 car very dirty ‘The car is very dirty’

(Sinitic, Cantonese; Goddard 2005: 103)

(4) Buku itu baru book that new ‘That book is new’ (5) Unnæhee hungak presidde kene-k he very famous person-nom ‘He is a very famous person’

(Malay)

(Sinhalese; Gair 1970: 145)

4.╇ The Singlish examples derive from data from informants, except (17) which is from the Grammar of Spoken Singapore English Corpus (GSSEC; Lim and Foley 2004).



(6) Moskva gorod Moscow city ‘Moscow is a city’

The Asian typology of English

(Russian; from Stassen 2008b)

What should be noted next is that examples (3) and (4) are particularly significant as they represent two of the languages that come into contact in the linguistic ecology of Singapore; unlike Standard English, in which zero copula encoding of predicative nominals is never allowed, Sinitic and Malay varieties hardly ever require an overt copula. It is therefore reasonable to expect that features from these languages will be available for selection to multilingual users of English in Singapore. 3.1.2 Predicative adjectives In Sinitic and Malay, zero copula relates to a second typical feature, namely the absence of a clear distinction between the word classes we know as verbs and adjectives. In Sinitic and Malay, we also find that not only are copular verbs rare and usually used for emphasis (Goddard 2005), but also that adjectives behave predicatively (Stassen 2008a), and are usually referred to as predicative adjectives (Li and Thompson 1981; Wetzer 1996), as already illustrated in examples (3) and (4). Languages which lack a distinction between a verbal and an adjectival class are typologically well-attested; out of 386 languages investigated in Stassen (2008a), more than half showed verbal behaviour of adjectives. Crucially, zero copula and predicative adjectives are common typological correlates cross-linguistically, as zero copula is a prerequisite for identifying the verbal encoding of adjectives in a language in the first place (Wetzer 1996; Stassen 1997). Also relevant to the focus of this paper is that, of the languages of the world in which zero copula and predicate adjectives correlate, there is a particularly high concentration in the (Southeast) Asian region, as illustrated in Figure╯3 (from WALS Online; Haspelmath et al. 2008) — the region in which Singlish is found. We examine this phenomenon by looking at the use of gao1 ‘(to be) tall’ in Mandarin (Sinitic). As example (7) shows, the fact that ‘tall’ behaves predicatively is indicated by the absence of the copula verb; (8) confirms the fact that ‘tall’ is indeed verbal, since it can be modified by aspect, just like other verbs as in (9) (examples from Goddard 2005: 113). (7) Ni2 hen3 gao1 you very tall ‘You are very tall’ (8) Ni3 gao1-le yi1dian3 you tall-PERF a.little ‘You’ve gotten taller’

19

20

Umberto Ansaldo

Figure╯3. Distribution of languages of the world with zero copula and predicative adjectives (from WALS Online; Haspelmath et al. 2008)

(9) Wo3 peng4-dao4-le Lin Hui I bump-arrive-PERF Lin Hui ‘I ran into Lin Hui’

In considering Singlish, similar patterns can be found, as shown in examples (10) and (11). Clearly, Singlish grammar follows Sinitic (and Malay) typology, and should be considered a language in which adjectives behave predicatively. (10) Today weather very hot. ‘Today’s weather is very hot’ (11) She good a21? ‘Is she good?’

The fact that zero copula and predicative adjectives feature in Singlish grammar must be seen as a selection of adstrate features (or substrate transfer) from the multilingual pool; in addition, the fact that both features are selected in the new grammar is justified by the typological correlation that holds between them noted above. 3.1.3 Topic prominence A third property of Singlish which can be explained as a selection of non-English material from the multilingual pool of variables available in the linguistic ecology is Topic prominence (Ansaldo 2004; Bao and Lye 2005). There is reason to believe that languages can be classified as either Topic-prominent or Subject-prominent (Goddard 2005); while Subject prominence is the prevalent option in SAE, Topic



The Asian typology of English

prominence is typical of isolating languages since they have little morphology to carry out agreement tasks, often used to identify subjecthood. This is illustrated in examples (12) to (14) from Mandarin (see Li and Thompson 1981), where zero copula and predicative adjectives may also be observed. (12) Zhang1san1 yi3jing1 jian4-guo le John already see-EXP CRS ‘Zhangsan I have already seen’ (13) Shang4 ge yue4 tian1qi4 fei1chang2 men1 last CL month weather extremely humid ‘Last month the weather was extremely humid’ (14) Xiang4 bi2zi chang2 elephant nose long ‘Elephants have long noses’

While in English subject and topic often align, in Topic-prominent languages, other grammatical elements can be topics, such as object (‘John’ in 12) and adverbial (‘last month’ in 13), in addition to subject (‘elephants’ in 14). Topic-Comment structure is also a characteristic of Malay varieties, such as Baba Malay (Lim 1988: 39–40).5 As can be seen in examples (15) to (17), speakers of Singlish also frequently structure their discourse in terms of Topic-Comment structure, following the patterns of other Topic-prominent languages in the ecology. (15) Fish you wan? ‘Do you want (some) fish (to eat)?’ (16) Expensive the durian here. ‘The durian here is expensive.’ (17) My parents old fashion a21? ‘Are my parents old-fashioned?’

(GSSEC)

3.2 Evolution and ecology as explanatory parameters In order to tie the theory discussed in Section╯2 together with the data presented so far, let us revisit what an ecological approach to Singlish entails. Firstly, as a case of contact language formation, the study of Singlish requires that we recognize that the evolution of new grammar occurs in a multilingual ecology. This means that in investigating how English is restructured in the ecology of Singapore, we 5.╇ Peranakan English also exhibits Topic-Comment structure (Lim 2010b), reflecting the TopicComment structure characteristic of Baba Malay.

21

22

Umberto Ansaldo

assume that features of English are in competition with at least the features of other dominant languages, in this case Sinitic and Malay varieties. Secondly, in order to understand what drives selection and replication of features, we need to consider the social and structural profiles of these languages. Numerically, Sinitic emerges as the strongest language throughout the history of Singapore, since the Chinese formed the largest majority from two decades into colonial rule (Lim 2007, 2010a); possibly, this also holds for its prestige, as Chinese merchants were always very influential in Singaporean society (see Lim 2007). Typologically, Sinitic and Malay dominate, because of the typological congruence that renders many features of these languages more frequent in the pool of variables that define the ecology. Numerical and typological dominance mean that Sinitic and Malay variables are more frequent and salient and thus more readily available for selection and replication. In addition, typological correlation matters in the replication of zero-copula and predictive adjectives. The competing features discussed in this section are illustrated in Table╯1. Table 1. Competing features in the ecology of Singlish Sinitic Zero copula Property verbs Topic prominence

Malay Zero copula Property verbs Topic prominence

English Copular verb Verb / adjective distinction Subject prominence

As can be seen from Table╯1, Sinitic and Malay adopt the same strategies in all three instances, while English diverges. In the linguistic ecology of Singapore, the Sinitic╛/╛Malay features win in the competition process because they have a higher type- and token-frequency in the multilingual context in which speakers of Singlish communicate. Therefore they emerge as typical features of Singlish grammar, while the English features lose out and are thus discarded (for other instances of Sinitic transfer in English, see Bao 2001, 2005; Bao and Lye 2005). Note that Singlish is by no means the only instance of contact language formation in which frequency determines to a large extent the restructuring process. Another obvious case of the dominance in particular of type-frequency in the evolution of a new grammar is found in Sri Lanka Malay, in which the congruence between Sinhala and Tamil by and large dominates the restructuring process and wins in the competition with Malay features (Ansaldo 2008, 2009; also Aboh and Ansaldo 2007). If these observations are correct, then we can say that speakers are sensitive to the dominant grammatical patterns and select congruent, type-frequent features in constructing their grammar.



The Asian typology of English

4. Final remarks I do not wish to claim that all features of Singlish are derived from its Asian substrates. This would obviously be incorrect in the light of the lexical influence, for one thing, as well as the grammatical features of English that trickle into SgE, especially visible at the mesolectal level. These include wh-movement, residual tense morphology, modal auxiliaries, etc. (see e.g. chapters in Lim 2004). The main claim that I wish to make here is that an appropriate description of Singlish grammar, and of any other Asian English variety or any other contact language for that matter, must take as a point of departure a pool of linguistically diverse features including all the dominant languages in contact. This is the approach outlined as the “Typological Matrix” (Ansaldo 2004, 2009), in which it is argued that contact language formation is the result of typological alignments in the multilingual ecology in which contact takes place (see also Mufwene 2001, 2008; Schneider 2007; Ansaldo 2008, 2009). In this approach, morphological reduction is not necessarily an instance of simplification or faulty acquisition, but rather a reflection of typological traits of isolating languages (where present) which win in the competition and selection process. Singlish has many of these, more than illustrated here, as can also be seen, for example, in the use of final particles and tonal features, which are defining features of Singlish (Lim 2007, 2009, this volume). From a morphosyntactic perspective, considering the predominantly isolating typology and the typical Asian grammatical features,6 Singlish, as an Asian English variety, is more an Asian variety with English influences than a variety of English.

References Aboh, Enoch O. and Umberto Ansaldo. 2007. “The role of typology in language creation: A descriptive take”. In Ansaldo, Matthews and Lim, eds.: 39–66. Ansaldo, Umberto. 2004. “The evolution of Singapore English: Finding the matrix”. In Lim, ed.: 127–49.

6.╇ This does not mean that we should find all features of Sinitic and/or Malay replicated in Singlish grammar. The multilingual pool of features offers far more variables than any single grammar can accommodate; this is why we need to talk about selection, because it would be logically impossible for one single grammar to incorporate all the variables of several different languages. Considering the role of frequency and typological dominance in the evolution of new grammar, we have to expect that, more often than not, features that are rare, cognitively redundant, or typologically marked within the specific linguistic ecology lose out in the competition against typologically frequent, dominant features, and will not be easily replicated in the output grammar.

23

24

Umberto Ansaldo

———. 2008. “Revisiting Sri Lanka Malay: Genesis and classification”. In K. David Harrison, David Rood and Arianne Dwyer, eds. A World of Many Voices: Lessons from Documenting Endangered Languages. (Typological Studies in Language 78.) Amsterdam, Philadelphia: Benjamins, 13–42. ———. 2009. Contact Languages: Ecology and Evolution in Asia. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. ———, Lisa Lim and Salikoko S. Mufwene. 2007. “The sociolinguistic history of the Peranakans. What it tells us about ‘creolization’”. In Ansaldo, Matthews and Lim, eds.: 203–26. ———, Stephen Matthews and Lisa Lim, eds. 2007. Deconstructing Creole. (Typological Studies in Language 73.) Amsterdam, Philadelphia: Benjamins. Bao, Zhiming. 2001. “The origins of empty categories in Singapore English”. Journal of Pidgin and Creole Languages 16(2): 275–319. ———. 2005. “The aspectual system of Singapore English and the systemic substratist explanation”. Journal of Linguistics 41: 237–67. ——— and Lye Hui Min. 2005. “Systemic transfer, topic prominence, and the bare conditional in Singapore English”. Journal of Pidgin and Creole Languages 20: 269–91. Croft, William. 2000. Explaining Language Change: An Evolutionary Approach. Edinburgh: Longman. ———. 2006. “Evolutionary models and functional-typological theories of language change”. In Ans van Kemenade and Bettelou Los, eds. The Handbook of the History of English. Oxford: Blackwell, 68–91. DeGraff, Michel. 2001. “On the origins of creoles: A Cartesian critique of Neo-Darwinian linguistics”. Linguistic Typology 5: 213–30. ———. 2003. “Against Creole exceptionalism. Discussion note”. Language 79: 391–410. ———. 2005. “Linguists’ most dangerous myth: The fallacy of Creole exceptionalism”. Language in Society 34: 533–91. Edwards, John. 1996. Multilingualism. London: Penguin. Gair, James W. 1970. Colloquial Sinhalese Clause Structures. The Hague: Mouton. Givón, Talmy. 1979a. On Understanding Grammar. New York: Academic Press. ———. 1979b. “Prolegomena to any sane creology”. In Ian F. Hancock, ed. Readings in Creole Studies. Ghent: Story Scientia, 3–35. Goddard, Cliff. 2005. The Languages of East and Southeast Asia. Oxford: Oxford University Press. Gupta, Anthea Fraser. 1994. The Step-Tongue: Children’s English in Singapore. Clevedon, Philadelphia: Multilingual Matters. Haiman, John. 1994. “Ritualization and the development of language”. In William Pagliuca, ed. Perspectives on Grammaticalization. Amsterdam, Philadelphia: Benjamins, 3–28. Haspelmath, Martin, Matthew S. Dryer, David Gil and Bernard Comrie, eds. The World Atlas of Language Structures Online. Munich: Max Planck Digital Library. Available online at http:// wals.info/ (accessed 26 Jan. 2009). Hickey, Raymond. 2007. Irish English: History and Present-Day Forms. Cambridge: Cambridge University Press. Holm, John. 2004. Languages in Contact: The Partial Restructuring of Vernaculars. Cambridge: Cambridge University Press. Hull, David. 1988. Science as a Process: An Evolutionary Account of the Social and Conceptual Development of Science. Chicago: University of Chicago Press.



The Asian typology of English

Janda, Richard and Brian Joseph. 2003. “On language, change and language change — or of history, linguistics and historical linguistics”. In Brian Joseph and Richard Janda, eds. The Handbook of Historical Linguistics. Oxford: Blackwell, 3–180. Khin Khin Aye. 2005. “A grammar of Singapore Bazaar Malay”. Ph.D. dissertation, National University of Singapore. Lass, Roger. 1997. Historical Linguistics and Language Change. Cambridge: Cambridge University Press. Li, Charles and Sandra Thompson. 1981. Mandarin Chinese: A Functional Reference Grammar. Berkeley: University of California Press. Lim, Lisa, ed. 2004. Singapore English: A Grammatical Description. (Varieties of English Around the World G33.) Amsterdam, Philadelphia: Benjamins. ———. 2007. “Mergers and acquisitions: On the ages and origins of Singapore English particles”. World Englishes 26: 446–73. ———. 2009. “Revisiting English prosody: (Some) New Englishes as tone languages?” In Lisa Lim and Nikolas Gisborne, eds. The Typology of Asian Englishes. Special Issue of English World-Wide 30(2):â•›218–39. ———. 2010a. “Migrants and ‘mother tongues’: Extralinguistic forces in the ecology of English in Singapore”. In Lisa Lim, Anne Pakir and Lionel Wee, eds. English in Singapore: Modernity and Management. (Asian Englishes Today.) Hong Kong: Hong Kong University Press, 19–54. ———. 2010b. “Peranakan English in Singapore”. In Daniel Schreier, Peter Trudgill, Edgar W. Schneider and Jeffrey P. Williams, eds. The Lesser Known Varieties of English: An Introduction. (Studies in English Language.) Cambridge: Cambridge University Press, 327–47. ——— and Joseph A. Foley. 2004. “English in Singapore and Singapore English: Background and methodology”. In Lim, ed.: 1–18. Lim, Sonny. 1988. “Baba Malay: The language of the ‘Straits-born’ Chinese”. In Hein Steinhauer, ed. Papers in Western Austronesian Linguistics No. 3. Pacific Linguistics Series A, No. 78. Department of Linguistics, Research School of Pacific Studies, The Australian National University, 1–61. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. ———. 2008. Language Evolution: Contact, Competition and Change. New York: Continuum. Platt, John. 1975. “The Singapore English speech continuum and its basilect ‘Singlish’ as a ‘Creoloid’”. Anthropological Linguistics 17: 363–74. Quirk, Randolph. 1990. “Language varieties and standard language”. English Today 21: 3–10. Schneider, Edgar W. 2007. Postcolonial English: Varieties Around the World. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. Stassen, Leon. 1997. Intransitive Predication. Oxford: Oxford University Press. ———. 2008a. “Predicative adjectives”. In Haspelmath et al., eds.: 118. Available online at http:// wals.info/feature/118 (accessed Sep. 2008). ———. 2008b. “Zero copula for predicate nominals”. In Haspelmath et al., eds.: 120. Available online at http://wals.info/feature/120 (accessed Sep. 2008). Thomason, Sarah G. and Terrence Kaufman. 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Weinreich, Uriel. 1963. Languages in Contact. The Hague: Mouton. Wetzer, Harrie. 1996. The Typology of Adjectival Predication. Berlin, New York: Walter de Gruyter.

25

Aspects of the morphosyntactic typology of Hong Kong English* Nikolas Gisborne

University of Edinburgh

English and Cantonese are the main two languages in contact in Hong Kong, together with some other minority Sinitic languages and a variety of Austronesian languages spoken by domestic helpers. Cantonese and English are typologically dissimilar in terms of word order, tense, mood and aspect marking, noun phrase structure, relative clause formation, the formation of interrogatives, and argument structure. Yet there is no work which systematically explores how these morphosyntactic typological differences are revealed in Hong Kong English (HKE). This paper explores how a typological perspective facilitates an analysis of the expression of finiteness in HKE, a significant feature because it subsumes a number of other typological facts. The analysis claims that HKE is a new English variety where the typology of the substrate is more directly responsible for the morphosyntactic features under analysis than the typology of the lexifier. Keywords: Hong Kong English, finiteness, typology, ecology, feature pool

1. Introduction Although English has been spoken in Hong Kong since the British took possession of Hong Kong Island in 1841, the variety of English known as “Hong Kong English” is really a postwar phenomenon, one of the new varieties of English Kachru (1985) describes as an “Outer Circle” variety. The sociolinguistic context giving rise to Hong Kong English (HKE) is complex. Between 1945 and 1997, Hong Kong was a British Crown Colony with administrative structures — especially those of education — run by British colonial administrators. At the same time, in the postwar period, there was significant population growth due to large scale migration *╇ I am grateful to Edgar Schneider, Lisa Lim, Umberto Ansaldo, Claire Cowie, Caroline Wiltshire and two reviewers for comments on earlier and present versions of this paper, and Anna Siewierska, Willem Hollmann and Paul Kerswill for comments on an oral presentation.

28

Nikolas Gisborne

from Mainland China. Hong Kong was never a settler colony: the British administrators, businessmen, lawyers and teachers who worked there were expected to return “home”. And so, although English was the language of administration, business, law and education in the colonial period, it was not really in contact with the languages of the indigenous populations in domestic environments. In the main, most of Hong Kong’s ethnic Chinese population acquire their English through their experience of education. For this reason, it is fair to say that English is primarily transmitted in the classroom, although, (perhaps) as there are local norms, it is not fair to state that HKE is a simple L2 variety which is acquired afresh with every generation. There is also, arguably, transmission of a local variety from one generation of HKE speakers to another, although this involves explicit instruction rather than a classic language acquisition context. These contextual facts raise a number of questions about how to study this emerging variety. Given the context, it makes sense to explore HKE as a contact variety which has emerged in a unique environment and which needs to be understood in terms of the composition of that environment. It is important to understand the typological dimensions of the languages in contact, in order to understand the linguistic environment where the new variety emerges. The two main languages which come into contact in Hong Kong are English and Cantonese — but there are additionally some other Sinitic languages spoken such as Mandarin, Hokkien and Chui Chow, and a substantial population of Filipina domestic helpers who bring different Austronesian languages to the territory. In line with other papers in this issue, I am claiming that the properties of the emerging New English can be understood in terms of the selection of a number of grammatical features from a feature pool in the ecology (Mufwene 2001; Ansaldo 2009a, b, this volume). The ecological approach I am adopting is presented in Mufwene (2001). Croft (2000) presents a similar treatment of language change in terms of population genetics, where speciation consists of the selection of features from an available body of features in the linguistic environment. In Mufwene’s (2001) approach, feature selection is discussed in terms of the external and the internal ecology. The external ecology is the sociolinguistic context of language contact; the internal ecology is the pool of linguistic features which the languages in contact contribute to, and from which the features of the restructured lexifier are drawn. The approach I am adopting here has results which show that that the grammatical features under investigation cannot best be analyzed in terms of “angloversals” (Kortmann and Szmrecsanyi 2004). So why a typological perspective? What would you understand differently or predict? My answer is in line with the other papers in this volume, and agrees with the line trenchantly argued for by both Ansaldo and Lim in their papers.



Aspects of the morphosyntactic typology of Hong Kong English

If we look at language contact in terms of the ecological approach pioneered by Mufwene (2001), we can see the innovations in the creation of the new English as being a kind of selection from a feature pool. It is necessary to see how similar the varieties are, what kinds of selections speakers make, and what their choices are if the process of new language formation is supposed to be understood. Not least, this approach affords us the opportunity of exploring New Englishes in terms of close and careful grammatical description with a keen eye on the linguistic and sociolinguistic environment, which is a useful corrective to overenthusiastic arguments in favour of “angloversals” or the so-called exceptional properties of contact varieties. Because of the history of English in Hong Kong, and because of its status, studies of HKE have primarily focused on the sociology of the variety, attitudes of the population of Hong Kong to English, codeswitching, learner errors, and the Hong Kong accent. See, for example, several of the papers in Bolton (2002). There is some work on the structure of the variety. Gisborne (2000), which looks at relative clauses, is one case study, and Hung (2000) explores the phonology of HKE as a system in toto. But there is very little work on the morphosyntax of the variety, and there are not really any useful paradigms for exploring its morphosyntax. In this paper I am specifically interested in a core area of morphosyntax. I look at the distribution of finiteness in HKE in order to establish whether it can be argued that there is a lack of finiteness in the variety which reflects the lack of finiteness in Sinitic languages. The argument has two parts: first I justify the claim that Sinitic languages do not have a finiteness contrast, and then I look at relevant data from HKE subordinate clauses to establish whether there is a suspension of the finiteness contrast in this variety. As the papers in Nikolaeva (2005b) show, finiteness is a central element in theorizing about morphosyntax. However, there are good reasons for claiming that it is not a well defined category cross-linguistically, and Hu, Pan and Xu (2001) argue that despite a number of claims to the contrary Chinese does not have finiteness. If HKE has structures like those in Chinese which are associated with the absence of finiteness, then this will argue for substrate transfer in the morphosyntax. Another point concerns the status of finiteness as a typological feature. Given its absence from the World Atlas of Language Structures (WALS; Haspelmath et al. 2008), there is a question about whether finiteness works as a feature of areal typology, in the same way as tone, for example (see Lim 2009, this volume). The rest of this paper has the following structure: Section╯2 looks in greater detail at the sociolinguistic contexts of HKE, and describes the immediate background of the variety; Section╯3 describes the relevant structures in Cantonese — in this section we argue for the position that Cantonese does not have a finiteness contrast; Section╯4 explores the relevant data from HKE; Section╯5

29

30

Nikolas Gisborne

discusses the facts in the context of recent work on the ecology of the language contact environment and presents the conclusions. 2. The sociolinguistic contexts of HKE According to Tsui and Bunton (2002: 57), approximately 96% of the population of Hong Kong is Chinese. Tsui and Bunton show that not all of the Chinese population in Hong Kong speak Cantonese, although the vast majority do. They cite the 1996 By-Census where 88.7% of the population “indicated that Cantonese is their usual spoken language, and 3.1% indicated English”. As Table╯1 shows, there has been some change since 1997.1 Table╯1. HK population aged 5 and over by usual language, 1996, 2001, and 2006 Population Aged 5 and Over by Usual Language, 1996, 2001 and 2006(1) 1996 2001 2006 Usual Language Number % of total Number % of total Number % of total Cantonese 5 196 240 88.7 5 726 972 89.2 6 030 960 90.8 Putonghua 65 892 1.1 55 410 0.9 60 859 0.9 Other Chinese Dialects 340 222 5.8 352 562 5.5 289 027 4.4 English 184 308 3.1 203 598 3.2 187 281 2.8 Others 73 879 1.3 79 197 1.2 72 217 1.1 Total 5 860 541 100.0 6 417 739 100.0 6 640 344 100.0 Note:╇ (1)The figures exclude mute persons. Source:╇ 2006 Population By-Census Office, Census and Statistics Department (Enquiry telephone no. : 2716 8025) Last revision date: 22 February, 2007

More of the population speaks Cantonese as its home language than in 1996, less of the population speaks English, and interestingly, the numbers speaking Mandarin (i.e. Putonghua in the table) have also diminished. Kwok (2004) and Lai (2005) have both found in language attitudes research that most Hong Kong residents have an emotional attachment to Cantonese and perceive English and Mandarin 1.╇ Table╯1 is Table╯140 of the Hong Kong Census and Statistics Department’s Statistical Tables and Charts, published at (25 Jan. 2009).



Aspects of the morphosyntactic typology of Hong Kong English

to be languages which have instrumental value, but which they are not particularly attached to. These facts are suggestive of several reasons why English does not have an “equivalent of the mesolectal or basilectal speech styles found, for example, in Singapore … since there is no equivalent range of English speech varieties in regular use by Hong Kong Chinese” (Luke and Richards 1982: 55–6, cited in Tsui and Bunton 2002: 58). After all, if local Cantonese-speaking Chinese residents of Hong Kong have no reason to use English within their community, then there are few situations of use where a local variety will develop. The broad facts about the sociology of use of English in Hong Kong also raise the question “whose norms?” — what English is used, by whom and where? One answer is given by the education system: according to Bolton (2003: 96) in 1994 “over 90% of all secondary schools were at least nominally English-medium (Johnson 1994)”. In March 1997, all but 100 schools were obliged to teach through the medium of Cantonese; Bolton (2003: 96–7) says, “[t]he figure of 100 was amended to 114 after predictable protests from schools and parents, but at present the policy remains one of providing ‘firm guidance’ for secondary schools, and of encouraging the use of Cantonese as a teaching medium”. English is also the medium of instruction in the University of Hong Kong, the oldest and most prestigious university in the territory, as it is at the University of Science and Technology, while the Chinese University of Hong Kong, the second oldest university, is officially bilingual, as is the City University of Hong Kong. This means that in order to receive a higher education, Hong Kong secondary school students need to become sufficiently proficient in English in order to follow a university course of instruction. Universities set admission standards for English by requiring an appropriate grade in the Use of English ‘A’ level. From this brief survey, it appears uncontentious to suggest that English in Hong Kong is a colonial legacy, and its presence is an artefact of the education system, and its prestige an artefact of colonial government — leaving English as the language of law and government — as well as the practices of international business. But the situations of use of English in Hong Kong are not entirely simple: Tsui and Bunton (2002) go on to cite Bacon-Shone and Bolton (1998), who identify several reasons for claiming that English is a language which, in the course of the 1980s and 1990s, is used by home speakers of Cantonese in certain socially conditioned contexts. There is a large number of Filipina domestic helpers in Hong Kong, who speak English with their employers, which makes it necessary to use English in the home; English is used in written communications such as e-mail and business memos; and “the percentage of the population who reported that they knew English quite well, well and very well, rose from 6.6% in 1983 to 33.7%

31

32

Nikolas Gisborne

in 1993 and to 38.1% in 1996” (Bacon-Shone and Bolton 1998: 76). The 2006 ByCensus shows that more of the population of Hong Kong speak English as a home language than any single Sinitic variety other than Cantonese;2 it also shows that the percentage of the population claiming to speak English as either their usual language or as an additional language rose from 38.1% in 1996, to 43% in 2001, to 44.7% in 2006.3 Bolton (2002, 2003) has claimed that Hong Kong English is a variety with its own norms and its own local creative activity. It seems true enough to say that there is a local Hong Kong accent, but it is not obvious that, in terms of the morphosyntax, Hong Kong English is anything other than an L2 variety. Of course, stable contact varieties show substrate morphosyntactic features, so showing that Hong Kong English reveals aspects of the morphosyntax of Chinese is not a knock-down argument that it is an L2 variety. But making the argument that there are features of Cantonese in the morphosyntax of Hong Kong English makes it far more possible to establish the dimensions of “autonomy” (Bolton 2002) in Hong Kong English, and to become clearer about the extent to which it is typologically similar to the native language of the local Chinese population. In the next section, we review the arguments for claiming that Cantonese, like other Chinese varieties, does not have finiteness and we look at the structure of the Cantonese NP, with a view to establishing the typological facts which will allow us to explore some of the relevant features of Hong Kong English. 3. Finiteness as a typological category, in English, and in Chinese In Section╯3.1 below, I discuss finiteness in standard native varieties of English; and in Section╯3.2 I argue (following Hu et al. 2001) that there is no finiteness in the Sinitic languages, and therefore Cantonese does not have finiteness. First, though, let us think about finiteness as a typological category. Nikolaeva (2005a) discusses finiteness as a property of clauses: in the western European tradition, it is associated with tense marking and verbal inflection, verb-subject agreement and the requirement of clauses to have a subject. Nikolaeva also points out that finiteness is associated with mood and other semantic notions, so it is not a category with clearly defined boundaries. In her survey, she observes, “there 2.╇ Table A 118 of the 2006 By-Census, “Hong Kong Resident Population by Duration of Residence in HK, Nationality and Usual Language, 2006”, available at (25 Jan. 2009). 3.╇ In connection with these facts, it is worth noting that the percentage of the population claiming to speak Putonghua (i.e. Mandarin) was 25.3% in 1996, 34.1% in 2001, and 40.2% in 2006.



Aspects of the morphosyntactic typology of Hong Kong English

seem to be some cross-linguistically valid correlations between subject requirement, subject agreement, tense, syntactic opacity, and independent clausehood” (Nikolaeva 2005a: 10). Many of the papers in Nikolaeva (2005b) are given over to the question of how finiteness might be defined theoretically so that it can have cross-linguistic validity as an analytical notion. Tense, then, is one of several properties which are possible reflexes of finiteness. The World Atlas of Language Structures does not recognize finiteness as a typological feature. It shows, for example, the distribution of languages with no pastâ•›/â•›present tense contrast; those that make a pastâ•›/â•›present tense contrast; and those that have a three-way pastâ•›/â•›presentâ•›/â•›future contrast (Dahl and Velupillai 2008). Sinitic languages are among those that show no pastâ•›/â•›present tense contrast, whereas English is listed among those languages which distinguish between past and present, but which does not have a morphologically realized future. But WALS’s discussion of these tense distinctions does not consider them in terms of a supercategory; indeed, it does not treat the more abstract category finiteness as a discrete category at all. So why say that Sinitic languages lack finiteness rather than tense alone? The specific arguments are made below, but we can consider some general observations here. Just as there are properties which converge under the rubric of finiteness, so, I hypothesize, are there properties which converge in languages that arguably do not show a finiteness contrast. For example, lack of finiteness in Sinitic languages is not only responsible for the lack of tense distinctions, but is also related to other known typological phenomena. As Ansaldo (2009b) points out, Sinitic languages are Topic-prominent rather than Subject-prominent. Sinitic languages have copula omission, which is also related to the lack of finiteness: in a language where there is no finiteness, there is no need to have a copular verb as a carrier of morphosyntactic information. And finally, it is hard to establish the lexical categories of certain elements in Sinitic languages, especially those related to predication. The categorial distinction between adjectives and verbs is especially hard to establish. Again, if a language does not have a finiteness distinction, then it is hard to make the distinction between these lexical categories, because it is the distribution under finiteness which establishes the verbal nature of a verb. Although I do not pursue these claims in a substantial way in this paper, it is likely that if HKE lacks finiteness, it will also show the Sinitic typological properties listed here. Therefore, a number of well known typological properties of Sinitic languages can be subsumed under a single parameter — which means that these same properties can be explored in HKE as a way of establishing how HKE speakers have selected from the feature pool. On the other hand, as the quotation from Nikolaeva (2005a) above shows, there are good reasons for treating tense as a reflex of finiteness, as well as the subject requirement, alongside other properties which English

33

34

Nikolas Gisborne

demonstrably has. It seems sensible therefore to assume that finiteness is the relevant morphosyntactic contrast to explore here. 3.1 Finiteness in standard varieties of English There are four main areas where English grammar shows the finiteness distinction. – – – –

Finite clauses must have overt subjects; non-finite clauses need not. Finite clauses can be independent predications; non-finite clauses cannot. Finite clauses encode a speech act function; non-finite clauses do not. Matrix predicates select for either finite or non-finite complements.

Let us take the first two points. (1)

a. She ran home. b. * ___ ran home. c. I expected [her to run home]. d. I expected [___ to run home].

Note that the finite clauses in (1a) and (1b) must have overt subjects, whereas the bracketed non-finite clauses in (1c) and (1d) do not need overt subjects. (In (1d), the “missing” subject of the infinitive clause is shared with the subject of the matrix clause.) Note too that it is not possible to use a non-finite clause outside of a subordinate context. *Her to run home is not a possible independent sentence of English. The third point is a little more complex: the syntax of the finite clauses in (2) shows that they are in turn imperative, declarative, and interrogative. (2) a. Be aware! b. He is aware. c. Are you aware?

Non-finite clauses do not encode these speech act functions syntactically although they can be embedded under heads which indicate that their non-finite clausal complements denote instructions, assertions, or questions. (3) a. He told [her to be aware]. b. He considered [her to be aware]. c. He wondered whether [___ to be aware].

In (3), tell is a verb which selects a complement that reports the giving of an instruction; consider is a verb which selects a complement that expresses a statement-like proposition; and wonder whether together select a clause that expresses a reported question. In none of the instances in (3) is the semantics of the



Aspects of the morphosyntactic typology of Hong Kong English

complement encoded in their syntax. It is true that embedded non-finite questions are slightly more difficult to show because whether does not select for non-finite clauses with overt subjects. Finally we can show that clause selection depends on the finiteâ•›/â•›non-finite contrast. This observation in turn shows that tense is a reflex of finiteness: when an English verb selects for a subordinate clause, as in (4), it selects for that subordinate clause on the basis of whether it is finite or not, not on the basis of whether it is past or present. Therefore tense is not the same as finiteness.4 (4)

a. I guessed (that) he went. b. * I guessed him to go. c. * I want that he goes. d. I want him to go.

The examples in (4) show that guess and want make different selections for subordinate clauses on the basis of a finiteness contrast: guess has to have a finite complement, whereas want has to have a non-finite complement clause. Rules about sequence of tense are orthogonal to the selection of clause type. Of these diagnostics, the most straightforward one to search for in a corpus is the last: if there is a perfect correlation between the complementation of matrix verbs in standard varieties of English and their correlates in HKE, then HKE maintains a finiteness contrast. On the other hand, if the finiteness contrast is levelled under verbs such as guess, then there is robust evidence that for (some) speakers of HKE, the morphosyntax feature system of the verb is that of Cantonese, rather than that of the lexifier. 3.2 Tense and finiteness in Chinese The argument that Chinese does not have a finiteness contrast begins with the observation that there is no tense contrast in Chinese. In this section, I follow Hu et al. (2001), whose arguments are developed for Mandarin, but which follow through for the other Sinitic languages, including Cantonese. Sinitic languages share the same relevant typological features in that they are all isolating and they all lack tense. The debate about whether Chinese has a finiteness contrast or not is largely concerned with the question of whether the mood and aspect distinctions that Chinese has are realizations of finiteness or not. In the generative literature, Huang (1984, 1987, 1989, 1998) and Li (1985, 1990) have both argued that there 4.╇ Of course, the lack of tense contrasts with imperatives such as (2a) above also shows that finiteness in English is not correlated with tense.

35

36

Nikolas Gisborne

is a finiteness contrast, taking mood and aspect as the key elements in the Chinese clause. On the other hand, Hu et al. (2001) have argued systematically against the claims made by Huang and Li, to present the position that Chinese does not have finiteness at all. In this section, I present Hu et al.’s (2001) arguments that the Chinese languages do not have a finiteness contrast. I have already observed that Chinese does not have tense. Hu et al. (2001: 1118) start by putting this claim on a systematic footing. They follow Stassen’s (1997: 350–1) arguments that there is “a typological distinction between tensed and nontensed languages”. Hu et al.’s definition of the tensedness parameter is given in (5). (5) a. If a language has a grammatical category of tense, which i. is morphologically bound on verbs, and ii. minimally involves a distinction between past and nonpast time reference, then that language is tensed. b. In all other cases, a language is nontensed.

Hu et al. (2001: 1119) also observe, “Stassen (1997) further argues that in a tensed language, the obligatory tense marking must be realized not by means of auxiliaries or particles, but by means of bound morphology on verbs, and tensed languages must meet the PAST CONDITION, which stipulates that a tensed language should have a verbal form exclusively referring to past time”. It follows from Stassen’s arguments that Chinese is non-tensed. However, as we have seen, tense alone does not determine whether a language expresses a finiteness contrast, and so showing that a language does not have tense is not the same as showing that it does not have finiteness. As Hu et al. point out (2001: 1120), “it is still argued in the literature that there is an implicit distinction between finiteness and nonfiniteness in Chinese”; the major point of their paper is to argue against the claims that there is such a contrast in Chinese. In several works, Huang (1984, 1987, 1989, 1998) argues that the occurrence of a modal or aspectual element argues for finiteness. He claims that certain verbs select finite complement clauses, whereas others select non-finite clauses. That is, Huang claims that the finiteness distinction in Chinese has at least one realization which is equivalent to the finiteness distinction in English: clausal selection by verbal heads. As a consequence of Huang’s assumptions, in his theory nonfinite clauses cannot have lexical subjects; nonfinite clauses cannot take modal predicates like hui ‘will’; and they cannot take aspectual markers like you5 either. 5.╇ Hu et al. (2001: 1122) gloss you as “ASP” because of its status as an aspectual marker; literally it means ‘have’.



Aspects of the morphosyntactic typology of Hong Kong English

That is, Huang claims that there is a conspiracy of facts that argue together for a finiteness distinction in Chinese. Hu et al. (2001) take Huang’s arguments apart. They demonstrate that both modals and aspectual auxiliaries can occur in the non-finite clauses where they are predicted not to occur, so Huang’s claims do not have empirical support. Furthermore, they demonstrate that the restrictions on the occurrence of modals that Huang observes can be accommodated in a semantic story which does not need to have any recourse to a theory of finiteness. Li (1985, 1990) argues that the finiteness distinction in Chinese “does not lie in the potential occurrence of modals in general, but in the possible occurrence of only those modals that have become tense markers” (Hu et al. 2001: 1123). Hu et al. make two arguments against this claim: they argue (i) that Li’s arguments about tense do not fit Stassen’s theory-neutral typology of tense; and (ii) that Li has failed to capture the relevant generalization because the modals she discusses, hui and yao, affect the semantics of the subordinate clause, such that they make it incompatible with the semantics of the matrix verb. In this case, then, the claim that Chinese does make a finiteness contrast can also be accommodated under a semantic story, so there is no need to exploit a morphosyntactic theory of finiteness. I refer the reader to Hu et al. for a full set of arguments. I concur with Hu et al. (2001) that the claim that there is a finiteness contrast in Chinese is unmotivated. Note that the specifics of the claim are important: in the next section I look at the selection of subordinate clauses to see whether finiteness is observable, which is the strategy that both Huang and Li adopt for Chinese. 4. Analysing the HKE data In this section, I look at some data which show that there is morphosyntactic transfer from Cantonese to HKE, including transfer at the relatively abstract level of finiteness. To get to finiteness, we can start by looking at the kinds of levelling of morphosyntactic distinctions that are common in HKE. Take the example in (6), from Gisborne (2000). (6) a. She like to go there. b. Have you try?

These examples do not reveal the form of the verb that its syntactic distribution would normally require. In (6a), the third person singular form likes is required, but instead we have like. In (6b), the perfect participle tried should be the complement of have, but instead the base form of the verb is used. There is a simple explanation of (6a) from the point of view of the language contact situation in Hong

37

38

Nikolas Gisborne

Kong: Cantonese does not have syllable-final consonant clusters (Matthews and Yip 1994: 19) and so there is often a simplification of these consonant clusters in HKE. But this explanation does not go through for (6b). The contrast between try and tried is not one that involves a consonant cluster. The examples in (6) show some differences from each other. The example in (6a) involves a failure of person marking: like is a present tense form of the verb, just not in the third person. And the contrast in (6b) is one where the base form of the verb is being used instead of a participle. As the base form can realize both present tense and the infinitive, we might assume that in an example like (6b), it is an infinitive form which is being used, with the infinitive substituted for the participle. As a result, we can only assert that the example in (6b) shows some kind of levelling of the non-finite morphosyntactic feature system of HKE. Gisborne (2000: 368) argues that this is an issue of morphosyntactic marking because of examples such as (7).

(7) I think it’s very difficult to described.

In (7), the complement of to is a participle, rather than the normally expected infinitive. From this discussion we can see that there are a few issues to factor out. I look at the realization of tense in Section╯4.1; I discuss finiteness in Section╯4.2. 4.1 Tense Let us begin by looking at tense. If tense contrasts are suspended in HKE, then it is likely that finite contrasts will be too. In main clauses, the most useful determinant of whether there is a tense contrast in HKE, as opposed to a finiteness contrast, is found where a verb is clearly used with past time reference but appears in the base form of the verb and the difference between the base form and the past form is not one that relies on a syllable-final consonant cluster for expression. Relatively good examples involve “strong” verbs, where the vowel changes, such as come and came, or those verbs where the past-tense morpheme -ed is realized as a syllable rather than a stop within a syllable cluster. So decide for decided is a good diagnostic of whether there is tense marking, whereas walk for walked is not, because walked involves a consonant cluster which may be simplified. The irregular and syllabic-past verb examples in Gisborne (2000) which are relevant to the question of tense in HKE are presented in (8)–(10).

(8) In my first year, Cats come to Hong Kong.



(9) He is born in Hong Kong and then just go to Hong Kong.



Aspects of the morphosyntactic typology of Hong Kong English

(10) China want to took … wants to take over.

In (8), the base form of the verb is used with past time reference. In (9), the present third person form of be, is, has past time reference in the first clause, and then in the second clause, the base form of go is used with past time reference. In (10), want is used as a third person present tense form, which is made clear in the correction, and the past participle is again used as the complement of to. Interestingly, in this example there is a self-correction, which suggests that there is an element of register variation within this variety of English. I looked at decide in the Hong Kong component of the International Corpus of English (ICE-HK) in order to follow up the question of whether there was levelling of tense expression. There were 63 instances of decide where it has past time reference. Past time reference is realized with decide in approximately a quarter of the total instances: 16 instances are realized as decide and 47 are realized as decided.6 There is an example in (11) which I will walk around. Recall that decide was chosen because its past tense form does not involve a syllable-final consonant cluster, so if there is no tense contrast in a HKE example, this cannot be due to interference from another domain in the grammar. (11) well we never decide which figure were out even though they out in the bill … [ICE-HK]

In this example, the past tense were indicates that decide has past time reference even though it is in the base form. This speaker has two other non-native features: plural figure (for figures), which could be because Cantonese codas do not permit fricatives (Bauer and Benedict 1997: 28), and zero copula in they out in the bill. I come back to this latter fact below. So what do examples like this indicate, and how should they be understood? It seems straightforward to claim that tense is not always expressed in HKE but that its variable expression might be a consequence of the stage of HKE as a nativizing variety in Schneider’s (2007) formulation. I return to this point in Section╯5. One point which is significant is that tense is one possible realization of an underlying morphosyntactic category. What evidence is there about the existence of that category in HKE? This is the topic of the next section.

6.╇ There are 92 tokens of decide and 109 tokens of decided which show the normal distributions of these forms (as infinitives and participles as well as tensed forms), although some non-finite instances of decide have passive semantics.

39

40

Nikolas Gisborne

4.2 Finiteness As we saw in Section╯3.1 above, the best determinant of finiteness was distribution of subordinate complement clauses. Some verbs such as guess select for finite clausal complements; others such as want select for non-finite clausal complements. Some verbs, such as expect select for both finite and non-finite subordinate clauses, as (12) shows. (12) a. We expected him to go. b. We expected that he would go.

Other verbs, such as guess, realize, and suggest, do not take non-finite subordinate clauses in standard English. The example in (13) gives some examples for realize. (13) a. We realized that he was right. b. We realized what he thought. c. * We realized him to be right.

In (13), it is clear that realize is another verb which cannot take a non-finite subordinate clause. In this case the complement clause has to be finite and headed by that, what, or its finite verb. In the ICE-HK data, none of the tokens of realize had non-finite complements, although there were instances with finite complements.7 This shows that the claim for non-finiteness in HKE is not an absolute claim. Where non-finite complements were found (with the other verbs studied here), they were never found in 100% of cases, which shows that the internal ecology of HKE has to be understood in the context of the external ecology, and that there is a clear pattern of social distribution for the feature under consideration, which is not surprising, considering that HKE is at stage 3 of Schneider’s (2007) dynamic model, rather than the stage 4 of Singapore English. The observation that non-finite complements were not found with realize, although they were found with guess and suggest, indicates that there may be lexical distribution (or diffusion) of the feature. Of course, it could also be an artefact of the corpus: there were 37 tokens of realize in the corpus, many of which were from transcripts of legal cases. As a consequence, it is reasonable to assume that there is an element of register 7.╇ The spoken corpus was searched for all the variant forms of the three verbs guess, realize, and suggest. Examples where the speaker was not a native of Hong Kong were excluded from the count, as were examples where the complement was a Noun Phrase, parenthetical examples, and (in the case of guess) instances of the noun.



Aspects of the morphosyntactic typology of Hong Kong English

variation in the facts under discussion.8 The next set of examples shows the standard pattern for suggest. (14) a. We suggested that they should go. b. We suggested that they went. c. * We suggested them to go.

The examples in (14) show that suggest requires a finite complement clause.9 Example (15), which is taken from the essay of a student at Hong Kong University, is therefore evidence of the absence of a finiteness contrast in HKE. (15) I suggest him to go.

The complement clause is non-finite, and fails to make the mood or modality contrast which is required in normal uses of suggest.10 In a search of ICE-HK, there were several tokens of suggest. Excluding examples where suggest was not complemented by a clause, as in I suggested the answer, there were 12 instances of suggest, suggests, suggesting, and suggested with non-finite complement clauses out of a total of 98 instances where it had a clausal complement. There are some examples given in (16), all of which are from ICE-HK. (16) a. It’s a beautiful space and I suggest to turn it to a cyber café but it was kind of turn down b. Can I suggest to put bill?

8.╇ There was, however, the example But uh <,> as we move along you’ll probably realize that <,> I <,> not only focusing on the empirical <,> investigation in the corpus data, which shows that although this speaker has a that-clause as the complement of realize, the complement of that is not finite. Another example where the subordinate clause had a particular morphosyntax was this: But Mr Au Yeung we must realise that in Wanchai is big area. Although the clause under that is finite in this example, as is shows, the subject is not a noun phrase, and the predicate nominal does not have a determiner. This is not an example of locative inversion. It is clear from the context that it means not ‘there is a big area in Wanchai’, but ‘Wanchai is a big area’. 9.╇ There is a complication in that for some speakers suggest may select for the vestigial subjunctive of English as in I suggest that he be back by 9 pm. The English subjunctive cannot be straightforwardly classified for finiteness. However, I think that on the whole this is not an issue in the case of HKE — Paul Kerswill tells me (p.c. Jan 2009) that the English subjunctive is on the rise in the US but that this is a recent phenomenon; during the relevant period of history, most speakers of British English will have hardly had a subjunctive. 10.╇ No doubt this works by analogy to standard patterns of non-finite complementation in English. However, that analogy does not obviate the argument that what has happened here is that the normal selection by suggest of a finite complement has been over-ridden and a non-finite complement has been chosen.

41

42

Nikolas Gisborne

c. They think that English is too hard for them so I suggest them to change to uh an Chinese uh d. I wont suggest you to to buy from the one word because… e. According to you it was Madam Ho who suggested to you to have sex with her for five hundred dollars.

These examples all work in slightly different ways. (16a) involves an arbitrary PRO subject, whose referent is not in the discourse context. The subject of to turn it to a cyber café cannot be recovered from the immediate context. The discussion is about the snooker room in the Senior Common Room at the University of Hong Kong, and the referent of PRO must be the management committee. The next example, (16b), has the same kind of structure as (16a). The example in (16c) involves a lexical subject in the non-finite clause them to change to uh an Chinese uh, so this is an unequivocal example that has the same structure as (15), as does (16d). The final example has to you after suggested, but clearly the subject of to have sex is you. Apart from the (possibly) doubtful (16b), all of these examples show the kinds of structure which indicate that the finiteness contrast is suspended for these speakers. They are all non-finite. More to the point, they are not modelled on the kind of pattern where a finite clause alternates with a non-finite clause which shares its subject with the subject of the matrix verb, as in (17). (17) a. We requested that he should go. b. * We requested him to go. c. We requested to go.

In (17), request takes a full finite clausal complement or a non-finite clausal complement with subject-sharing as in (17c). This is an alternative pattern to the one where a finite clausal complement can alternate with a non-finite clausal complement. However, the examples in (16) do not fit the pattern of (17c) — even the examples with no overt subject have a PRO subject whose referent cannot be identified with the subject of the matrix clause. To summarize the discussion, it seems that the examples in (15) and (16) suspend a finiteness contrast. There are two reasons for making this claim: the grammatical patterns we find in these examples are not modelled on patterns like that in (17), and they are entirely well-formed, except that they occur in a position where a different kind of clause is normally expected. All of the non-finite clauses in (15) and (16) are perfectly reasonable examples of non-finite clauses; it is just that we do not expect non-finite clauses to have this distribution. Now we can return to the zero copula phenomenon in (11) in Section╯4.1 which I repeat here.



Aspects of the morphosyntactic typology of Hong Kong English

(11) well we never decide which figure were out even though they out in the bill … [ICE-HK]

In the subordinate clause, out is treated as a predicator in the even though clause. Why does this matter? It matters because the example is on a parallel with the subordinate clause after decide, which has a tensed were in it. There are two parallel clauses: which figure were out and they out in the bill. The zero copula in the second clause suggests that there is no systematicity in the use of different predicators by this speaker. In one case, a tensed copula is used; in the other, there is no copula. Note too that even though is a head which normally requires a finite complement in standard varieties of English: I cannot say *even though they to be out in the bill, for example. The absence of a finite complement here suggests that the finiteness contrast is non-systematic for this speaker.11 The example in (18) below supports the general point. In this example, there is an adjectival predicate under guess.12 (18) Uh I guess his hands his hands too full and uh he had other priorities [ICE-HK]

The relevant point is that guess requires a finite complement in English, yet here not only is there a non-finite predicate, but also the predication is non-verbal, even though there is no possibility of structure sharing in an example like this. The subject-predicate relationship is between his hands and too full: we can see that the adjective phrase is a predicate here in a zero copula construction. This is important in two ways: first, it appears that this speaker does not necessarily make a lexical category distinction between adjectives and verbs; and, second, predication in this example does not require finiteness as expressed by a copula verb. To sum up, in this section I have looked at a set of data which argue that some varieties of HKE do not make a finiteness contrast, and that recognizing that lack of finiteness in HKE allows us to bring the complementation of verbs and the zero copula phenomenon under the same generalization.

11.╇ It also suggests that this speaker is capable of treating both verbs and adjectives as predicates, suggesting that, in their variety, HKE is like Cantonese — a language which does not distinguish between verbs and adjectives. 12.╇ There were only 3 out of 57 tokens of guess with this kind of structure. However, several of the clausal complements of guess were headed by if, whether and what. Excluding these examples gives a total of 3 out of 45 tokens. There were also subjectless examples such as I guess will be at least two and So I guess is just luck or if you try really hard. Finally, I have excluded examples such as to guess about how it… how it would operate from the count, treating this as an example of guess complemented by about.

43

44

Nikolas Gisborne

5. Conclusion: Language contact, typology, and the ecology of the language-contact environment in Hong Kong There are two ways of taking the examples with finiteness in the examples above. One way is to argue that these examples reveal an imperfect transfer of the grammar of Standard English to speakers of HKE. On this argument, HKE would be a reflex of imperfect L2 acquisition in the Hong Kong context. But such an argument suggests that there is little or no systematicity in HKE, and furthermore fails to make the relevant connections between the different reflexes of finiteness and how they materialize in HKE. On the other hand, if we recognize that there can be systemic transfers from the substrate13 then we can see that some speakers’ HKE has a grammar system which is typologically similar to Cantonese. Furthermore, I have argued that the right level of generalization is to look at finiteness as a phenomenon, rather than at individual reflexes of finiteness such as tense marking, copula deletion of the apparent lack of a lexical category distinction between verbs and adjectives. But, given that Singlish is another variety which shows substrate transfer from Sinitic varieties to a New English, it is necessary to think about the similarities and differences between HKE and Singlish. Ansaldo (2009b) observes that the grammar of Singlish (i.e. basilectal Singapore English) also shows systemic transfer of features of Chinese grammar. These are Topic prominence, zero copula, and blurring of the lexical category distinction between verbs and adjectives. I would claim that languages which have this bundle of typological features are also likely to be languages which lack a finiteness contrast, because the alternative bundle of features (Subject prominence, mandatory copula, clear lexical category distinctions between adjectives and verbs) involves reflexes of the finiteness distinction. In this sense, then, HKE and Singlish are similar. But there is a complication: the lack of finiteness is not systematic in HKE. For sure, I have found several examples which lack this morphosyntactic feature distinction, but it is not at all clear that the grammar has settled on one typological pattern over another. What are we to make of this? Considering the different stability of the two grammars, with HKE at stage 3 of Schneider’s dynamic model (i.e. nativization) and Singapore English at stage 4 (endonormative stabilization), we can see that English and Cantonese are still in contact in HK, and what we see is an emerging system with a considerable degree of variability. Ansaldo (2009b, this volume) makes a similar point about the two varieties. Then there are arguments to be made about how features are selected from the feature pool. I would argue that for there to be the systematic acquisition of a 13.╇ As, for example, Siegel (2003) argues.



Aspects of the morphosyntactic typology of Hong Kong English

Standard English finiteness distinction among speakers of HKE, the relevant feature would need to be strongly entrenched in the environment. As it is, the sociolinguistics of HK indicate that the most frequently found morphosyntactic features in the ecology are those of Cantonese so it is unsurprising that Cantonese morphosyntax transfers into HKE. However, the feature system of English is typological marked relative to Chinese, which might account for the relatively high frequency of finiteness contrasts among some speakers of the variety. In a feature pool approach one predicts that features that are salient in the ecology will surface in the contact grammar. Therefore, we do not only allow for the possibility of substrate transfer, but we expect superstrate features to be very salient where the superstrate is dominant. Finally, the evidence surveyed here calls into question the kind of “angloversals” approach of Kortmann and Szmreczanyi (2004) in as much as we can see that the relevant distinctions in HKE are to be found in the linguistic environment and do not emerge as a kind of default setting in the emergence of a New (contact) English. From this point of view, the typological approach, with its focus on what there is in the ecology, allows us to establish at a fine grain of grammatical description the relevant features of a variety of English such as HKE. But the most important conclusion is the importance of studying language contact of a range of diverse kinds; it is important not just to look at L2 varieties and learner errors, or straightforward examples of a vaguely defined “creolization” process. It is by exploring the full diversity of language contact situations that we can establish the processes by which language changes in contact environments and new varieties emerge.

References Ansaldo, Umberto. 2009a. Contact Languages: Ecology and Evolution in Asia. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. ———. 2009b. “The Asian typology of English: Theoretical and methodological considerations”. In Lisa Lim and Nikolas Gisborne, eds. The Typology of Asian Englishes. Special Issue of English World-Wide 30(2):â•›133–48. Bacon-Shone, John and Kingsley Bolton. 1998. “Charting multilingualism: Language censuses and language surveys in Hong Kong”. In Martha C. Pennington, ed. Language in Hong Kong at Century’s End. Hong Kong: Hong Kong University Press, 43–90. Bauer, Robert S. and Paul K. Benedict. 1997. Modern Cantonese Phonology. Berlin: Mouton de Gruyter. Bolton, Kingsley, ed. 2002. Hong Kong English: Autonomy and Creativity. (Asian Englishes Today.) Hong Kong: Hong Kong University Press. ———. 2003. Chinese Englishes. Cambridge: Cambridge University Press.

45

46

Nikolas Gisborne

Croft, William. 2000. Explaining Language Change: An Evolutionary Approach. Harlow, Essex: Longman. Dahl, Östen and Viveka Velupillai. 2008. “The Past Tense.” In Haspelmath et al., eds.: 66. Available online at http://wals.info/feature/66 (accessed 16 Feb. 2009). Gisborne, Nikolas. 2000. “Relative clauses in Hong Kong English.” World Englishes 19: 357–71. Haspelmath, Martin, Matthew S. Dryer, David Gil and Bernard Comrie, eds. 2008. The World Atlas of Language Structures Online. Munich: Max Planck Digital Library. Available online at http://wals.info/ (accessed 16 Feb. 2009). Hu, Jianhua, Haihua Pan and Xu Liejiong. 2001. “Is there a finite vs. non-finite distinction in Chinese?” Linguistics 39: 1117–48. Huang, C.-T. James. 1984. “On the distribution and reference of empty pronouns”. Linguistic Inquiry 15: 531–74. ———. 1987. “Remarks on empty categories”. Linguistic Inquiry 18: 321–37. ———. 1989. “Pro-drop in Chinese: A generalized control theory”. In Osvaldo Jaeggli and Kenneth Safir, eds. The Null Subject Parameter. Dordrecht: Kluwer, 185–214. ———. 1998. Logical Relations in Chinese and the Theory of Grammar. New York: Garland. Hung, Tony T.N. 2000. “Towards a phonology of Hong Kong English”. World Englishes 19: 337–56. Johnson, Robert K. 1994. “Language policy and planning in Hong Kong”. Annual Review of Applied Linguistics 14: 177–99. Kachru, Braj B. 1985. “Standards, codification and sociolinguistic realism: The English language in the outer circle”. In Randolph Quirk and Henry G. Widdowson, eds. English in the World: Teaching and Learning the Language and Literature. Cambridge: Cambridge University Press, 11–30. Kortmann, Bernd and Benedikt Szmrecsanyi. 2004. “Global synopsis — morphological and syntactic variation in English”. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider and Clive Upton, eds. A Handbook of Varieties of English. Vol 2: Morphology and Syntax. Berlin, New York: Mouton de Gruyter, 1142–202. Kwok, Kar-yan Bonnie. 2004. “Language attitudes in Hong Kong: The status of Putonghua and English in the 21st century”. M.A. thesis, University of Hong Kong. Lai, Mee-Ling. 2005. “Language attitudes of the first postcolonial generation in Hong Kong secondary schools”. Language in Society 34: 363–88. Li, Y.-H. Audrey. 1985. “Abstract case in Chinese”. Ph.D. dissertation, University of Southern California. ———. 1990. Order and Constituency in Mandarin Chinese. Dordrecht: Kluwer. Lim, Lisa. 2009. “Revisiting English prosody: (Some) New Englishes as tone languages?” In Lisa Lim and Nikolas Gisborne, eds. The Typology of Asian Englishes. Special Issue of English World-Wide 30(2):â•›218–39. Luke, Kang-kwong and Jack C. Richards. 1982. “English in Hong Kong: Functions and status”. English World-Wide 3: 47–64. Matthews, Stephen J. and Virginia Yip. 1994. Cantonese. London: Routledge. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. Nikolaeva, Irina. 2005a. “Introduction”. In Nikolaeva, ed.: 1–19. ———, ed. 2005b. Finiteness: Theoretical and Empirical Foundations. Oxford: Oxford University Press.



Aspects of the morphosyntactic typology of Hong Kong English

Schneider, Edgar W. 2007. Postcolonial English: Varieties Around the World. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. Siegel, Jeff. 2003. “Substrate influence in creoles and the role of transfer in second language acquisition”. Studies in Second Language Acquisition 25: 185–209. Stassen, Leon. 1997. Intransitive Predication. Oxford: Oxford University Press. Tsui, Amy and David Bunton. 2002. “The discourse and attitudes of English language teachers in Hong Kong”. In Bolton, ed.: 57–77.

47

Typological diversity in New Englishes* Devyani Sharma

Queen Mary University of London

Recent research has aimed to integrate the investigation of vernacular universals in native English dialects with variation in postcolonial varieties of English and cross-linguistic typology (Chambers 2004; Kortmann 2004). This article assumes that any search for universals in bilingual varieties must include an assessment of the grammatical conditioning of features and a comparison with the relevant substrates. Comparing Indian English and Singapore English, I examine three proposed candidates for English universals (Kortmann and Szmrecsanyi 2004), all of which show some presence in the two varieties — past tense omission, over-extension of the progressive, and copula omission. Past tense omission is found to be genuinely similar in the two varieties and accounted for by typological parallels in the substrates, whereas progressive morphology use and copula omission are found to be divergent in the two varieties and accounted for by typological differences in the substrates. All three variable systems are explicable as substrate-superstrate interactions, tempering claims of universality in both distribution and explanation. Keywords: Indian English, Singapore English, typology, past, progressive, perfective, imperfective, copula, language transfer, universals

1. Introduction Chambers (2004: 28) describes vernacular universals as “a small number of phonological and grammatical processes [that] recur in vernaculars wherever they are spoken”. Such universals have been investigated fruitfully in monolingual varieties *╇ I would like to thank Ashwini Deo and John Rickford for earlier collaborations that informed this work. I am also grateful to Lisa Lim, E-Ching Ng, Stephen Matthews, and Umberto Ansaldo for many helpful suggestions relating to the Singaporean situation, and E-Ching Ng, Lavanya Sankaran, and Sue Fox for their assistance in finding native speakers. I am indebted to my consultants, who patiently provided judgments for this study: Huang Zhipeng, Amanda Cheung, Youping Han, Jiang Dan, Lim Chey Cheng, Huifong Chen, and Dorothy Tan.

50

Devyani Sharma

of English and recent efforts have aimed to incorporate contact varieties of English and cross-linguistic typology into this enterprise (Kortmann 2004). The search for “angloversals” in bilingual postcolonial varieties is formalized in Kortmann and Szmrecsanyi’s (2004; henceforth K&S) typological survey of shared features in varieties of English. Chambers observes of vernacular universals that [t]heir ubiquity has one of two possible explanations. Either the features were diffused there by the founders of the dialect, or they developed there independently as natural structural linguistic developments … the diffusionist explanation is implausible because of geographical spread. It is also implausible linguistically, because these features occur not only in working-class and rural vernaculars but also in child language, pidgins, creoles and interlanguage varieties. Therefore, they appear to be natural outgrowths, so to speak, of the language faculty, that is, the species-specific bioprogram. (2004: 128)

Although Chambers envisions a unified paradigm for the study of sociolinguistic, acquisitional, diachronic and typological variation (2004: 130), the two explanations above are only appropriate for monolingual situations. Bilingual varieties of English will always involve a third potential explanation — the substrate. If we find a property recurring in two unrelated bilingual varieties, it is premature to draw a universalist conclusion unless substrate transfer has been ruled out. Even if the emergent feature is known to be typologically unmarked, it may simply be prevalent in and transferred from the substrates. Similar methodological issues arise in K&S’s aggregation of English varieties to identify potential universals. In classifying the features of a given variety, they continue the practice in earlier work (Platt, Weber and Ho 1984; Williams 1987) of relying on broad presence or absence of the feature with little or no consideration of semantic or other conditioning. They acknowledge that this “bird’s-eye view approach necessarily abstracts from many details and (partly necessary) qualifications in individual varieties” (2004: 1143). However, mere presence of a feature in two grammars cannot be taken to imply genuine typological similarity unless its grammatical conditioning is comparable (Poplack 2000; Tagliamonte 2002; Rickford 2006). This article follows two basic criteria in extending the search for vernacular universals to bilingual cases: i) ascertainment of a shared typological feature through quantitative examination of its conditioning, and ii) elimination of substrate transfer explanations before making any appeals to universality. The analysis assesses three of K&S’s “candidates for universals of New Englishes” — null past tense, over-extension of the progressive, and null copula. All three features are attested to some extent in both Indian English (IndE) and Singapore English (SgE).



Typological diversity in New Englishes

First, quantitative analysis reveals systematic patterns in the degree and conditioning of each feature in the two varieties. Second, substrates are found to play a central part in determining emergent systems: the one instance of closely parallel patterning in IndE and SgE (past tense omission) is accounted for by typological parallels in the substrates in that domain; differences in patterning (progressive and copula use) are accounted for by typological differences in the substrates. Despite the challenge of higher levels of variability in multilingual speech communities (Sankoff 2002: 640), this paper shows that close quantitative and substratesensitive analysis of apparent similarities across new varieties of English can reveal them to be typologically distinct in important ways. 2. Universals in New Englishes The three features discussed in the present comparison of IndE and SgE, listed in (1), are all among K&S’s “candidates for universals of New Englishes” and are all described as shared across a number of regions, including Asian Englishes (2004: 1193): (1) a. #40: Zero past tense forms of regular verbs b. #21: Wider range of uses of the progressive c. #57: Deletion of be

K&S list zero past tense as a feature typical of Caribbean, American, Pacific, and Asian varieties (2004: 1189; see also Jenkins 2003: 26), extension of the progressive to stative verbs as typical of Asian, American and African varieties (2004: 1189; see also Platt, Weber and Ho 1984: 72–3; Williams 1987: 172–3; Jenkins 2003: 26; Melchers and Shaw 2003: 22, 158; Trudgill and Hannah 2008: 107, 130, 137), and deletion of be as typical of Caribbean, Pacific, and Asian varieties (2004: 1193). Chambers (2004: 129) lists deletion of be as one of his top four candidates for English vernacular universals. K&S calculate the relative strength of representation of a feature across varieties by averaging its “feature value” (1 = pervasive, 0.5 = infrequent, 0 = absent) for all 46 varieties and seven world regions examined. Thus, even if a feature is not found in all Asian varieties, if it is found in three out of a total four it receives a high “feature ratio” score (0.75) for Asia and is included as “typical” for the region. But what does this aggregation represent? If three of four Asian varieties have zero past marking because each of their substrates shares a narrower use for past morphology, then the high “feature ratio” score does not reflect universality. Similarly, if three of four Asian varieties exhibit substantial deletion of be, but in each case a different grammatical context conditions deletion, based on

51

52

Devyani Sharma

substrate semantics, the high “feature ratio” again does not reflect any sort of universality. Filppula, Klemola and Paulasto (2009) pay closer attention to the question of whether universals or contact determine new systems, proposing a continuum between the two and suggesting that stative progressives may fall at the boundary between these two explanations. Van Rooy (2006) similarly offers a close comparison of progressive use in two varieties of English and identifies a Bantu source for divergence in the South African variety. Examining a different feature, article use, Sand (2004) pursues a quantitative comparison of several varieties, paying attention to the substrate in each case. She finds greater substrate influence in spoken than written genres, and, by ruling out substrate explanations elsewhere, argues convincingly that a number of semantic extensions in definite article use across varieties can be accounted for by “a common tendency to expand the rules of English article use in a certain way” based on universal semantic properties (2004: 295).1 In this paper, I follow these latter approaches in pursuing a quantitative exploration of semantic conditioning and substrate effects prior to concluding universality for any feature. This approach helps to restrict broad universalist appeals to generalized L2 learning strategies as an explanation for surface similarities, as invoked by K&S: “…it is now possible to give more substance to the notion of angloversals … Mair explicitly states that some of these angloversals may be the result of learning strategies of non-native speakers, in other words properties typical of L2 varieties” (2004: 1192). Properties that may appear to be “typical of L2 varieties” may in fact be parallel cases of transfer of typologically common features in the substrates, or may be differently conditioned in each variety. This is not to deny the potential for genuinely substrate-independent universals governing the cross-linguistic behaviour of tense-aspect systems (Bybee, Perkins and Pagliuca 1994) or copula systems (Stassen 1994), but rather to distinguish among degrees of similarity in New Englishes and types of explanations offered. 3. Data IndE and SgE have markedly distinct linguistic ecologies in terms of languages in contact, their functional roles, degrees of nativization, official policies, and 1.╇ Sharma (2005) also finds a combination of substrate and universal effects in article use (K&S’s universal feature #17): some substrate transfer arises from the specificity-marking article system of Hindi but universal discourse-driven tendencies intervene where the substrate does not supply an explicit form-meaning pairing.



Typological diversity in New Englishes

language ideologies. Within his Dynamic Model of the evolution of postcolonial Englishes, Schneider (2007: 153, 161) accurately characterizes IndE as currently at an earlier stage of indigenization (Phase 3, “Nativization”) than SgE (Phase 4, “Endonormative Stabilization”). In order to compare the syntactic properties of the two varieties in this study, I try to identify relatively comparable sets of conversational data from each speech community. 3.1 Indian English The IndE data for this study come from 12 individuals, part of a larger bilingual corpus of IndE speakers. Only the subset of participants who show evidence of all three features are included here; it is important to note that many balanced bilingual speakers in the larger corpus do not show nonstandardness in copula omission and past tense use. The speakers in this study are non-English dominant and can be considered “basilectal”. None of the participants had English-medium school education, although five had a limited amount of English-medium tertiary education and all had English as a school subject. They are all dominant in their first language but have regular (often work-related), daily use of English, often with other non-native speakers. All individuals lived in India until adulthood and acquired English through formal and informal modes in India; most are small shop owners, shop employees, or are unemployed. The speech data consist of naturalistic, sociolinguistic interviews, lasting 0.5–2 hours, and ranging over topics such as work, leisure, cultural attitudes, and narratives of childhood and migration. The primary substrate language for Northern IndE is Hindi, a language of wider communication across North India, but many speakers of IndE have at least one other additional native language. All participants are speakers of Hindi; two are additionally native speakers of Gujarati and three are native speakers of Punjabi. All three languages are identical in terms of copula use and tense-aspect parameters relevant to this discussion: they all require an overt copula in all predicate contexts (Masica 1991: 336); they all inflect for imperfective and for perfective with reflexes of the original Sanskrit participles; and they all mark progressive with an auxiliary verb comparable to Hindi rahna (‘remain’) (Masica 1991: 292–302). I therefore use only Hindi as the representative substrate system in the analysis. 3.2 Singapore English The SgE data for this study are drawn from secondary sources, in particular Ho and Platt (1993; henceforth H&P), Platt (1979), and the ICE-Singapore corpus

53

54

Devyani Sharma

(with comparisons to the parallel corpus ICE-India).2 H&P use a database of conversational interviews with 100 ethnically Chinese speakers of SgE, stratified into five educational levels. All participants in H&P’s study have completed some level of English-medium education unlike the IndE participants in my database. In order to ensure comparability I focus on the basilectal end of their data. As with IndE, SgE has a number of substrate inputs. Bao (2005) takes Mandarin as the dominant input due to its promotion in education since 1970; however, as SgE began to form much earlier than independence in 1965, Hokkien, Teochew, and Cantonese are widely seen as constituting the most important formative inputs (H&P 1993: 27; Gupta 1994: 41; Ansaldo 2004; Lim 2007: 452; Yip and Matthews 2007: 236). Bazaar Malay is an additional major lingua franca, the Singaporean variety of which has been influenced by Hokkien (H&P 1993: 9; Lim 2007: 453). Kuo (1980: 41, cited in Ng 2008) reports Singaporean home language statistics in 1957 as follows: Hokkien (30%) > Teochew (17.0%) > Cantonese (15.1%) > Malay (13.2%) > Mandarin (0.1%).3 Mandarin has experienced a dramatic surge in use due to various forms of institutional promotion since independence. The Singapore Census of Population (2000) figures for reported home language in 1990 and 2000 indicate the scale of replacement of Chinese “dialects” (39.6 in 1990; 23.8 in 2000) by Mandarin (23.7 in 1990; 35 in 2000).4 This increase certainly makes Mandarin a relevant substrate language as well, but it is then important to investigate the Singaporean variety of Mandarin, which, like Malay, has been affected by contact with other Chinese

2.╇ Note that, in comparison to the ICE data, which dates from the 1990s, Ho and Platt’s speakers date from the decades earlier, and are less likely to be native speakers of SgE or of Mandarin (Ansaldo 2004; Lim 2007). Given the historical background outlined in Section╯3.2, Mandarin may therefore only be an important substrate for the newer ICE data, which is only used in Table╯6. The attention to non-Mandarin substrates for SgE elsewhere in this paper is therefore relevant, given the nature of Ho and Platt’s data (Lisa Lim p.c. 2008). 3.╇ Gupta (1994: 41) and Lim (2007: 454) warn against pure demographic inferences, however, noting that although Hokkien speakers may have dominated numerically in the mid-20th century, they may not have been the major Chinese input to SE as the Cantonese adopted English more readily and may have played a central role in the resulting variety. 4.╇ The shift to Mandarin is even more evident across age groups. According to the Singapore Census of Population, in 2000 only 4.3% ethnically Chinese 5-to-14-year-olds spoke Chinese “dialects” (18.9% in 1990). By contrast, 71.8% of above-55-year-olds still spoke Chinese “dialects” (87.7% in 1990). It is important to bear in mind that some proportion of these self-reported census data may be influenced by language ideologies rather than actual practice. Census figures remain fairly steady for Malay (14.3% → 14.1%), Tamil (2.9% → 3.2%), and English (18.8% → 23%).



Typological diversity in New Englishes

languages in Singapore. All of these substrate systems are examined in the analysis, using available published materials and native speaker consultations. 3.3 Substrate grammars for Indian English and Singapore English For the first two features examined (past and progressive morphology), we need to first consider the aspectual systems of the three language systems in question — the Hindi substrate for IndE, the Chinese substrate for SgE, and English as the superstrate. These systems are presented in Table╯1. (In all tables in this paper, parentheses in a cell indicate optionality or variation according to additional contextual factors.) Table 1. Key aspect distinctions in Hindi, Mandarin, and English past imperfective

(i) perfective (ii) neutral (i) progressive (ii) non-progressive (iii) habitual

Hindi -(y)a — rahna -ta -ta

Mandarin le — zai (-zhe) —

English — -ed -ing — —

In the domain of past tense marking, there is a clear parallel between the two aspect-prominent substrate languages, such that an overt marker is used with perfective (completive) reference. This is in contrast to the tense-prominent English system, in which the marking of past tense dominates any aspectual distinction. In the domain of imperfective marking, however, the two substrate systems diverge. The only apparent similarity across all three imperfective systems is progressive marking. By contrast, Hindi requires overt marking of imperfectivity with -ta whereas Mandarin has relatively restricted use of the marker -zhe within certain imperfective contexts. In fact, Yip and Rimmington (2004: 107) suggest that the treatment of -zhe as an aspectual marker at all is erroneous, and that it is more precisely a “manner indicator”: “zhe is suffixed to an action verb so that the resultant verbal phrase is used as a descriptive element in sentences to indicate ‘manner of existence’, ‘manner of movement’, or ‘accompanying manner’”. The form is by no means used across all imperfective contexts, as Hindi -ta is, but rather is primarily reserved for temporary result states (Sun 2006). Its use is also conditioned by prosodic factors (Yip and Rimmington 2004: 127), another significant difference from Hindi -ta, which is strictly obligatory. As noted, it can be problematic to use published descriptions of mainland Chinese Mandarin in the study of SgE. Table╯2 provides the aspectual systems of Singapore Mandarin, Cantonese, Teochew, Hokkien, and Malay (consultants;

55

56

Devyani Sharma

Prentice 1990; H&P 1993; Matthews and Yip 1994). The perfective pattern is consistent across Chinese languages and Malay. However, the restricted use of imperfective marking in Mandarin (as compared to Hindi) is, if anything, further reduced in other Chinese imperfective systems, including Singapore Mandarin; these differences are discussed further in Section╯5.3. Table 2. Comparison of Singaporean substrate tense-aspect marking Singapore Cantonese Teochew Mandarin (i) perfective le tso lio past (ii) neutral — — — (i) progressive zai gan do imperfective (ii) non-progressive (-zhe) (zyu) (do) (iii) habitual — (hoi) —

Hokkien

Malay

liau — tja — —

sudah — sedang — —

We can conclude from Tables╯1 and 2 that, although Indo-Aryan and Chinese languages can be classified together typologically as perfectivity-marking systems, they are typologically distinct in the domain of the imperfective. This makes them a useful comparative case study. If substrate languages in fact drive the emergent systems, we should expect to see the patterns in (2):

(2) a. SHARED FEATURE (-ed use): IndE and SgE will show a parallel transfer of perfective meaning to past tense morphology. b. DIVERGENT FEATURE (-ing use): IndE and SgE will not show parallel behaviour in the imperfective. i. IndE will exhibit greater use of -ing as a general marker of imperfectivity, due to obligatory overt imperfective marking in Indian substrates. ii. SgE will exhibit less use of -ing as a general marker of imperfectivity, due to the limited range and optional use of overt imperfective marking in Singaporean substrates.

I first turn to a comparison of past tense morphology in IndE and SgE in Section╯4, and then to a similar comparison of progressive morphology in IndE and SgE in Section╯5. In Section╯6, I briefly review the final example of copula absence to support the analysis. 4. Past tense omission (K&S #40) Perfective sentences denote completed or temporally bounded situations, also called events, while imperfective sentences denote unbounded, ongoing situations



Typological diversity in New Englishes

(Comrie 1976). Clausal perfectivity can be conveyed simply by lexical aspect, i.e. telic verbs, which contain an endpoint (e.g. complete, build, dismantle, drown) or by additional grammatical aspect markers, i.e. morphology that imposes a bounded viewpoint on a predicate such as perfective morphology (e.g. Mandarin le) or perfectivizing adverbials (e.g. in two minutes, all of a sudden). The emphasis in research on past tense morphology in learning situations has been on lexical aspect (see Andersen and Shirai 1996). However, Sharma and Deo (2010) argue that IndE past omission is not a universal lexical effect, but rather involves overt marking of sentential perfectivity modelled precisely on the speakers’ Indo-Aryan L1s.5 Table╯3 shows that basilectal IndE speakers clearly ascribe perfective meaning to English past morphology (modelled on perfectivity-marking in their native Indo-Aryan), and tend to avoid overt past marking in imperfective sentences. Although I treat habituals semantically as derived states in my data, they are grouped with progressives and separated from lexical statives in Table╯3, following H&P’s methodology, to ensure comparability of Tables╯3 and 4. Table 3. Past tense marking according to clausal aspect in IndE (N=702) imperfective

habitual, progressive lexical stative

PERFECTIVE

% overt past 29.5 44.2 76.6

N 132 224 346

χ2(2dof)=109.96, pâ•›≤ 0.001

Fortunately, H&P (1993: 81–3, 151) explore the semantic contexts of past omission in SgE in much the same way. A close examination of their method for classifying predicates reveals that their analysis goes beyond lexical aspect as well and incorporates most core elements of clausal aspect. They follow Bickerton (1981) in using the term punctual “in a very broad sense” to include predicates bearing the features [+completive] and [+telic]; this corresponds to the standard term perfective. Similarly, their category “stative non-punctual” includes predicates bearing the feature [+stative] and their category “non-stative non-punctuals” includes predicates bearing the features [+duration], [+activity], and [+habitual]; these correspond to standard categories of “imperfective”. Their results are therefore directly comparable to the IndE data. 5.╇ Drawing on theories of aspect in the semantics literature, Sharma and Deo (2010) criticize the excessive focus on lexical aspect alone in second language acquisition studies, particularly when the L1 in question is sensitive to clausal perfectivity. They offer various types of evidence to support the view that clausal perfectivity conditions past tense use in IndE, including tense choice in cases of misaligned lexical and clausal aspect and fine conditioning of past tense marking by perfective and imperfective types. Their analysis examines the data in Table╯3 and Table╯5 in more detail.

57

58

Devyani Sharma

Table╯4 shows a strikingly similar pattern to the IndE distribution, whereby past tense is used significantly more with perfective predicates than with imperfective predicates.6 Figure╯1 shows how similar the IndE and SgE distributions are for past tense use. Table 4.╇ Past tense marking according to clausal aspect in SgE (N=8725) imperfective

habitual, progressive lexical stative

PERFECTIVE

% overt past 14.7 36.9 56.2

As predicted in (2a), we see a direct replication in both IndE and SgE of perfectivity marking in the substrate systems. It is possible that a universal preference for marking perfectivity is emerging; however, having examined the substrates, we certainly cannot conclude this. Only further comparisons to English varieties with non-perfectivity-marking substrates can verify this. At present, the IndE and SgE may be seen as a straightforward instance of strict transfer (Lefebvre 1998; Bao 2005), as in (3), whereby the semantic component of a form-meaning pairing in the L1 is re-attached to an L2 form. (3) (i) Hindi: (ii) Chinese:

[PERF -a] [PERF le]

→ IndE: [PERF -ed] → SgE: [PERF -ed]

One difference that does emerge between IndE and SgE perfectivity marking is also directly explicable by substrate differences. SgE has grammaticalized already as an additional perfective marker (Bao 1995, 2005); this usage follows formal analogy with analytic Chinese and Malay forms le and sudah. Hindi does not have isolated forms of this kind and consequently IndE has not grammaticalized adverbs for aspectual functions. 5. Over-extension of the progressive (K&S #21) The predictions in (2b) anticipate divergence between IndE and SgE in the domain of imperfective marking. Indian speakers of English may experience substrate pressure to mark both perfectivity and imperfectivity with explicit markers (the latter is only implicitly signalled by the lack of past marking seen earlier in Table╯3). As -ing is the only marker in the imperfective domain in English, it is likely to be 6.╇ A chi-square test of significance could not be presented for the SgE data in Table╯4 as the original study only provides a composite N value.



Typological diversity in New Englishes 100

Overt past morphology (%)

90 80 70 60

Indian English (N=702) Singapore English (N=8725)

50 40 30 20 10 0

Habitual and progressive imperfective

Stative imperfective

Perfective

Predicate aspect

Figure 1.╇ Past tense use according to clausal aspect in IndE and SgE

the immediate candidate for this function. SgE speakers, by contrast, are not expected to encounter this pressure. 5.1 Use of progressive -ing in Indian English The imperfective categories examined here are progressive, stative, delimited habituals, and non-delimited habituals. I distinguish between two types of habituals because of their distinct behaviour in standard varieties of English with respect to the use of progressive -ing. In general, a sentence with a habitual predicate describes a generalization over episodes rather than reporting a particular episode (a habitual operator thus transforms any eventuality type into a state); the sub-type of delimited habituals asserts or presupposes a time-bound on the habit described, whereas the sub-type of non-delimited habituals does not. Since the form -ing in Standard English primarily imposes a dynamic, in-progress reading, the form can be used with delimited habituals due to their time-bound property, e.g. I’m eating meat these days, but not with non-delimited habituals, which are not tied to a particular timespan, e.g. *I’m eating meat. Thus, of the four imperfective categories only progressive and delimited habitual environments productively license the use of the progressive form in native Standard English varieties. A few further uses of the progressive (e.g. with temporary states I’m wanting to move back) are discussed later, but for the most part the use of the progressive is non-standard with non-delimited habituals, statives, and perfectives, e.g. *I’m eating meat, *I’m knowing the answer, *I was moving to Miami in 1998.

59

60

Devyani Sharma

Table 5. Functions of progressive forms in IndE (N=339)

imperfective PERFECTIVE

delimited habitual non-delimited habitual progressive stative

% overt past 33.1 30.8 18.4 14.8 2.9

N 112 104 63 50 10

Table╯5 presents the proportion of all progressive -ing forms across the four imperfective contexts in the IndE data.7 It is clear that a robust over-extension to non-delimited habituals and statives, i.e. to the remaining imperfective categories, has taken place in IndE.8 Examples from the data showing systematic -ing use with non-delimited habitual reference are given in (4) and with stative reference in (5). (4)

Over-extension to non-delimited habituals9 a. I have got a driver. My son driving his own car.  [IA: d103] b. Generally only dry-cleaning clothes are coming.  [PB: d35] c. There’s no Indian crowd [in Rochester] and it’s snowing.  [RS: c123] d. Every week I’m calling [my parents].  [RS: c171]

(5) Over-extension to statives a. Some people are thinking it’s a bad job.  [MM: d138] b. For sociology they were asking me for 80% … But I was only having 70%.  [DD: d108] c. Japanese patients… would not be knowing English at all.  [AM: d132] d. Then what they’ll feel is like, we are knowing each other.  [RS: c383]

7.╇ Note that this approach is the reverse of the approach taken for past tense, where all past contexts were examined for use of the past form. This is because for past tense we are interested in restrictions on the distribution of the form, i.e. omission, whereas for progressive we are interested in extensions in the distribution of the form. Furthermore, the optionality of the progressive in many standard contexts renders an examination of all potential sites of use relatively intractable. 8.╇ The few perfective verbs that occur with the progressive are come, start, and begin; although telic, these verbs are also associated with null marking of past. This may suggest that inceptives and ingressives present a semantic clash with perfective interpretations for these speakers. 9.╇ Non-delimited status was established from the discourse context by, for instance, the absence of explicit or implied temporal delimiters such as these days.



Typological diversity in New Englishes

5.2 Use of progressive -ing in Singapore English Unfortunately, H&P do not provide quantitative data on semantic contexts of -ing use, but some indication of SgE usage can be gleaned from corpus data and from qualitative comments in H&P. In the SgE corpus ICE-Singapore, very slight over-extension of -ing does occur, but in quantitative terms it is negligible as compared to the corresponding IndE uses of -ing in the parallel corpus, ICE-India. As a brief illustration, I present all be+V uses of having and knowing in ICE-Singapore and ICE-India in Table╯6.10 The difference between IndE and SgE is dramatic, but it is still noteworthy that two of the four uses of having in SgE are stative. Standard and non-standard examples of having from each corpus are given in (6) and (7) respectively. Table 6. Use of main verb having and knowing in ICE-India and ICE-Singapore

be + having be + knowing

ICE-India standard non-standard 47 141 0 24

ICE-Singapore standard non-standard 4 2 0 0

(6) Standard examples of having a. Experiential: Having my hols now and am having a great time.  [ICE-SIN: w1b-010: 545] b. Experiential: I’m having a real tough time categorizing the sex you know. [ICE-IND: S1a-048: 1131] c. Future: This afternoon we are having high tea with the Rustomjees.  [ICE-IND: W2f-016: 238] d. Habitual: So each and every party is uh having a bandh [strike] everyday.  [ICE-IND: S1a-057: 174] (7) Non-standard examples of having a. Maybe that’s why I’m having a crush on him. [ICE-SIN: w1b-010.txt: 342] b. I’m having a talk [speech] next week.  [ICE-SIN: w1b-010.txt: 342] c. We are having different beaches.  [ICE-IND: S1a-065: 670] d. He therefore uh doesn’t say that the deceased was also having O group.  [ICE-IND: S2a-067: 514] 10.╇ Eleven of the 24 uses of knowing by IndE users included modals (e.g. You must be knowing this all history, don’t you?; As you may be knowing…). These were coded as non-standard based on eight native British English consultants’ grammaticality ratings, all of which were lower than ratings given for the non-progressive equivalents. Variation did occur in the degree of ungrammaticality for these consultants, indicating increasing permissiveness with stative progressives in standard native varieties.

61

62

Devyani Sharma

SgE shows no evidence of recasting -ing as a general imperfective marker as IndE does, but (7a) and (7b) indicate slight variation in use. Further support for the view that SgE shows slight variation, but not on the scale of IndE, comes from H&P (1993: 64–6). They, too, find a number of instances of over-extension of the progressive to habitual and lexically stative contexts in their data. Some examples from their data are listed in (8).11 However, they emphasize that “in general, the state-process distinction holds for Singaporean Chinese learners of English” (1993: 189). (8)

P13.18. Like sometime dey having dinner. A18.5. You having excess (cash) in your box. P5.1. I living here. P20.37. Most of de time I speaking English down there. P10.4. I use(d) to go in and flying in and out. T3.12. Den he carrying an umbrella.

5.3 Discussion Unlike perfectivity-marking with -ed, which followed a parallel pattern in IndE and SgE, imperfectivity-marking with -ing is very different in the two varieties. In both we find some variation with respect to the semantic range of -ing, but in IndE this leads to a systemic shift to marking all imperfective categories with -ing, while in SgE such a shift does not happen. This finding immediately rules out certain types of explanations. First, an account based on general lexical-semantic levelling in contact situations is unlikely, e.g.: “South Asian English often uses the progressive … with stative verbs, so that I am knowing is possible. This can be regarded as neutralization of the stativeâ•›/â•›dynamic contrast and is perhaps parallel to the neutralization of the countâ•›/â•›noncount distinction that produces luggages” (Melchers and Shaw 2003: 141). A universalist interpretation of such levelling cannot explain the discrepancy between IndE and SgE usage. It is also unclear why IndE would neutralize a distinction that is maintained in the L1, Hindi. At the other end of explanatory models, a strict substrate reading of stativeâ•›/â•›dynamic levelling such as Bickerton’s for IndE is also ruled out: “Native speakers of Hindi frequently make mistakes such as *I am liking it… Hindi speakers apparently commit [this mistake] because in Hindi imperfective marking can be used with statives” (Bickerton 1984: 155–6). This view anticipates that SgE should 11.╇ An important difference between IndE and SgE is evident in (8), namely the absence of be in the construction; H&P (1993: 65) observe that this is based on a direct analogy with the lack of be in Chinese systems. Indo-Aryan languages require a be auxiliary with imperfective markers, and so this difference can also be ascribed to a difference in the substrates.



Typological diversity in New Englishes

also over-generalize the English progressive, as SgE substrates also have stativity markers such as -zhe. Furthermore, as with Melchers and Shaw’s account, it does not explain why Hindi speakers level at all, instead of simply equating their own distinctive progressive form rahna with English -ing. The puzzle is therefore this: if the substrates and the superstrate all make a parallel distinction between progressive contexts (rahna in Hindi, zai and related forms in Chinese, -ing in English) and other imperfective contexts (-ta in Hindi, -zhe and related forms in Chinese, default absence of marking in English), then why is a single form adopted for both functions in one variety (IndE) at all, and why not in the other (SgE)? In this analysis, I focus on two factors: i) the extended range of uses of -ing in English, and ii) relative robustness of imperfective forms in the substrate. The first of these is shared by both varieties (giving rise to the variation we see in both SgE and IndE) and the second is distinctive in each variety (giving rise to a shift only in IndE) Together, these factors suggest an explanation based on substrate-superstrate interaction. I first consider IndE according to these two factors, and then SgE. The first step in analyzing the IndE progressive is to criticize the view, in Table╯1, that all varieties involved have comparable progressive forms. In order to do this, I move from the core set of imperfective contexts in Table╯1 to the expanded set of constructions in Table╯7. Table╯7 does not offer a semantically driven grouping of construction types; it simply serves to highlight certain differences in the distribution of imperfective morphology in the three systems in question. The first important observation in Table╯7 is the wide range of constructions involving Standard English progressive -ing as compared to the Hindi progressive rahna. Comrie (1976: 25) notes this unusually wide range and describes the Standard English progressive as “a kind of imperfective”. Certain statives occur with -ing in English, for instance, when a change in degree is involved or when the state is temporary, e.g. I’m living there but plan to move; I’m thinking we should sell it. The form -ing is also licensed with classes of predicates such as perception, posture, and location; other languages frequently focus on the result state or change-of-state properties of such classes (as in Hindi, see Table 7). The English gerund further extends the uses of -ing, and the form can even escape progressive meaning altogether in non-finite constructions, instead signalling simultaneity, e.g. Knowing that Maya was in town, I planned a party. (Comrie 1976: 36–9). By contrast, the Hindi progressive performs a much stricter function (although it does share the peripheral functions of future and of preliminary stage interpretation of achievements). In a contact situation, the English superstrate will always present learners with this complex distribution of -ing; if this distribution comes up against a narrow progressive form in the L1, this is likely to always instigate some variation caused by the L1-L2 progressive boundary mismatch and

63

64

Devyani Sharma

Table 7. Use of imperfective forms in Hindi, Standard English, IndE Type progressive preliminary stage future weather delimited habitual adverb over time adverb (simultaneity) persistent activity state habitual dress posture location temporary state nonfinite

Example He is WRITING a letter. He’s ARRIVING now. She’s LEAVING tomorrow. It’s RAINING. She’s DRIVING these days. FLYING all day, the bird got tired. He walked towards me SMILING. The wind kept BLOWING. You LOVE music. I DRIVE. She is WEARING boots. He is STANDING outside. A letter was LYING there. He is HOLDING the book. They like SWIMMING.

English -ing -ing -ing -ing -â•›/â•›-ing -ing -ing -ing — — -ing -ing -ing -ing -ing

Hindi rahna rahna rahna rahna -taâ•›/â•›rahna -ta -ta -ta -ta -ta -a (perf) -a (perf) -a (perf) -a (perf) -na (inf)

IndE -ing -ing -ing -ing -ing -ing -ing -ing -ing -ing -ing -ing -ing -ing -ing

Note: Progressive forms are marked in bold, and other imperfective forms in italics.

a resulting search for the correct semantic scope of -ing. Here we see potential for a genuine “angloversal” deriving from a peculiarity of the superstrate. The second detail to recognize in Tables╯1 and 7 is that Hindi is a strict imperfectivity-marking system, such that all finite clauses must be marked as either perfective or imperfective. The form -ta is never optional in habitual and stative contexts. This means that IndE speakers have a pervasive substrate pressure to mark imperfectivity overtly. The superstrate and substrate therefore both contribute to the following procedure. In their search for an overt imperfective marker, driven by the substrate, IndE speakers encounter -ing as a prominent candidate. Due to its extended range, the form -ing appears to equally map to rahna and -ta and IndE speakers interpret it as a global imperfectivity marker. An intriguing effect is that, although IndE results from an interaction between two grammars, the resulting system is typologically distinct from both input systems. It can be classified with other single imperfectivity-marker systems such as French, as shown in Table╯8. How does this procedure play out in SgE? Table╯7 showed that Hindi has a narrow progressive form which differs substantially from the broad scope of English -ing, and also that Hindi has a robust, obligatory imperfective marker. Table╯9 provides a similar range of constructions along with the relevant substrate morphology for SgE.



Typological diversity in New Englishes

Table 8. Imperfective marking in Hindi, Standard English, IndE, and French progressive stative habitual possession

Hindi rahna -ta -ta hona (be)

Standard English -ing (-ing) (-ing) have

IndE -ing -ing -ing -ing

French (PAST) IMPARFAIT IMPARFAIT IMPARFAIT IMPARFAIT

Table 9. Use of imperfective forms in English and Chinese languages Example He is WRITING a letter. He’s ARRIVING now. She’s LEAVING tomorrow. It’s RAINING. She’s DRIVING these days. He walked towards me SMILING. The wind kept BLOWING. You LOVE music. I DRIVE. The car is in the garage. She is WEARING boots. He is STANDING outside. A letter was LYING there. He is HOLDING the book. They like SWIMMING.

Standard English Singapore Hokkien Teochew Cantonese (and SgE) Mandarin -ing zai tja do gan -ing — — — — -ing — — — — -ing — — — (gan) -â•›/â•›-ing — — — — -ing (-zhe) — (do) (zyu) -ing — — — -ing -ing -ing -ing -ing

— (-zhe) — (-zhe) — (-zhe) (-zhe) (-zhe) —

— — — — — — — — —

— — — (do) — (do) (do) (do) —

— — — zyu zyu (gan) (gan) zyu —

Note: Progressive forms are marked in bold, and other imperfective forms in italics.

First, we can see that the mismatch between the scope of English and Hindi progressives is repeated in Table╯9; in fact, the mismatch is even more dramatic as future and preliminary stage uses of zaiâ•›/â•›tjaâ•›/â•›doâ•›/â•›gan are excluded.12 Table╯7 and Table╯9 thus share an L1-L2 mismatch that gives rise to parallel variation in the “boundary search” for -ing. H&P (1993: 189) also cite the “extended use of -ing constructions in the established varieties of English” as an influence on over-extension of -ing in SgE. 12.╇ In Standard Malay, sedang is also not used for future reference or simultaneity and is “much less frequent than English progressive” (Svalberg and Chuchu 1998: 39). Bazaar Malay is missing from Table╯9 due to lack of access to native speakers, but the column would be comparable to Hokkien as Malay does not have stativity markers. Teochew also has only one primarily progressive marker do, which shows some uses in zhe contexts, such as with posture and location (this slight difference in the dynamicâ•›/â•›stative boundary is also true for the use of Cantonese gan).

65

66

Devyani Sharma

By contrast, in the area of imperfective marking, although we find overt markers such as -zheâ•›/â•›zyuâ•›/â•›do in Chinese, their use is dramatically different from Hindi -ta. They are strictly markers of continuous states and are not used in many imperfective constructions such as habituals, simultaneity, and persistence. There is also extensive optionality in the use of these forms according to prosodic and word order factors. Some Singapore Mandarin-speaking consultants for this paper reported total optionality in the use of -zhe, most likely due to contact with Hokkien, Cantonese, and Teochew. The forms -zheâ•›/â•›zyuâ•›/â•›do are thus semantically highly restricted and frequently omitted in Singaporean usage.13 Convergence across substrates in the Singaporean case may reinforce the lack of pressure to mark imperfectivity (Ansaldo 2004). On the one hand, then, there is a resemblance to the IndE situation in the L1-L2 mismatch in the scope of the progressive form, caused by the superstrate. On the other hand, there is an absence of any strong Chinese substrate pressure to mark imperfectivity overtly, despite the presence of some imperfective markers. The first of these factors causes some variation due to a “boundary search”, but the second reduces the pressure among SgE speakers to seek an overt marker for imperfectivity, and the variety eventually approximates Standard English usage. This substrate-superstrate interaction can be schematized as in (9). The transfer of imperfective meaning to -ing — robustly in IndE and negligibly in SgE — is triggered by the general encroachment of English -ing into non-progressive territory. A substrate element remains crucial, however: the robustness of Hindi -ta as an imperfective marker gives rise to an expanded -ing system; by contrast, the optional Chinese -zheâ•›/â•›zyuâ•›/â•›do result state markers have little, if any, effect on how the SgE system stabilizes.

13.╇ Published grammars claim that -zhe is required with stative verbs such as love and with verbs of posture and activity verbs signalling states. My Singapore Mandarin consultants treat all of these as optional contexts, and so optionality is indicated for all Singapore Mandarin -zhe use in Table╯9. H&P and Bao (2005) both rely primarily or exclusively on Standard Mandarin descriptions (e.g. Li and Thompson 1981; Smith 1991), and assume a “universal Chinese grammar” (Chao 1968: 13). The present discussion shows the importance of other Chinese languages and the Singaporean variety of Mandarin in further reducing any substrate pressure to mark imperfectivity in SgE.





Typological diversity in New Englishes

(9) a. English [WIDE PROG -ing] â•›



Hindi [NARROW PROG rahna]



Hindi [ALL OTHER IMPERF -ta]



[IMPERF -ing]

IndE



[WIDE PROG -ing]

SgE

b. English [WIDE PROG -ing]

Chinese [NARROW PROG zai]



Chinese [opTIONAL STATIVE -zhe]

A lack of attention to semantic properties of morphemes and to quantitative distributions would have obscured aspects of this analysis. For instance, Bao (2005) claims that SgE is identical to standard native varieties of English in the domain of progressive -ing use due to the similarity of the English and Chinese progressive forms: Chinese has two imperfective aspect markers, pre-verbal zài and post-verbal -zhe. The former, like the English progressive, is dynamic, whereas the latter “has a static focus on states” … The imperfective aspect in Singapore English is unremarkable. It is formally identical to the English progressive, with the optional use of a copula … I was not able to collect any data from ICE-SIN which would show statives (love, believe, know) being used in the V-ing form … The stative imperfective is not found in Singapore English. (2005: 249–50)

Both the claim that the progressive in Chinese and in English are parallel and the claim that ICE-Singapore does not have instances of stative uses of -ing are complicated by the present discussion. Bao is correct in suggesting that SgE broadly conforms to Standard English behaviour, but the discussion in this section has identified empirical evidence of microvariation in SgE caused by the non-equivalence of English and Chinese progressives; this does not stabilize partly due to the influence of non-Mandarin substrate systems, excluded from Bao’s analysis. One remaining mystery is why the contact situation in (9b) does not lead to under-use of the progressive in SgE due to the highly restricted domain of zai. The sources cited suggest broadly standard rather than restricted use of -ing (although Table 6 does suggest the possibility of quantitative under-use). This can be accounted for by a version of the Subset Principle (Berwick 1985), whereby if a learner is starting from a grammar that is a subset of the target grammar, they can expand straightforwardly to the target based on positive evidence in the input. Thus, a SgE speaker may restrict English -ing to zai contexts at first, but would then encounter clear evidence for its wider use. By contrast, an IndE grammar that uses -ing as an imperfective marker already generates all standard -ing contexts and would therefore require explicit negative evidence to be restricted.

67

68

Devyani Sharma

6. Copula absence (K&S #57) In this final analysis section, I briefly summarize results from an examination of copula omission to support the present findings. Sharma and Rickford (2009) compare copula omission in African-American Vernacular English (AAVE), creoles, New Englishes, and childâ•›/â•›adult learner English. This feature has been widely assumed to be shared across creoles and learner varieties (Winford 1998: 114; McWhorter 2000: 420; Wolfram 2000: 54; Chambers 2004: 129), but Sharma and Rickford conclude that the mere fact of variable copula omission in these varieties belies underlying divergence in the grammatical conditioning of these systems. Following many studies of the deletion of be (see Rickford 1998 for a summary), I use the term “copula” here to include strictly copular be (with non-verbal predicates, e.g. nominal: She is my sister; adjectival: She is funny; locative: She is in the hall) and auxiliary be (with verbal predicates, e.g. V+ing: She is hiding; future gonna: She is gonna pay). Figure╯2 draws together rates of copula absence according to each predicate type, where possible, with a focus on Indian and Chinese substrates, and Table╯10 provides the systems of copula use in each relevant substrate language.14 Table 10. Copula used in equivalent L1 construction? Malay Cantonese / Mandarin Tamil Hindi

NOM ✗ (✓) ✗ ✓

LOC ✗ ✓ ✓ ✓

ADJ ✗ ✗ (✓) ✓

VERB ✗ ✗ ✗ ✓

It should be clear from Figure╯2 that although each variety has some degree of copula absence, the underlying system differs substantially in SgE, IndE, and South African IndE (SAIE). To a great extent, the substrate system in each case can account for relative frequencies of copula absence.

14.╇ Notes for Figure╯2 and Table╯10: The high Loc values in SgE (Malay) in Figure╯2 are based on a low total N value of 5. As individual speaker rates cannot be extracted for some of the datasets in Figure╯2, standard deviation measures could not be included. All of these data sets are discussed more extensively in Sharma and Rickford (2009). Sources of data in Figure╯2: Malay L1 SgE (Platt 1979); Chinese L1 SgE (basilectal data, H&P 1993: 48); IndE (Sharma and Rickford 2009); South African IndE (Mesthrie 1992: 158); AAVE (Labov 1972: 86); Barbadian Creole (Rickford 1992: 192). Sources of data in Table╯10: Malay (Platt 1979), Tamil (Pillai 1992: 15; Schiffman 1999: 141), Chinese (Platt 1979), Indo-Aryan (Masica 1991: 336).



Typological diversity in New Englishes

100 90 Copula absence (%)

80 70

SgE (L1 Malay, N=88)

60

SgE (L1 Chinese, N=1764)

50

IndE (L1 Indo-Aryan, N=743)

SAIE (L1 Dravidian, N=199) Creole (Barbados 1991, N=373)

40

AAVE (New York 1969, N=413)

30 20 10 0

NP

Loc

Adj

V-ing

gonna

Predicate type

Figure 2. Copula absence by predicate type (based on Sharma and Rickford 2009)

SgE (Malay L1) has the highest absolute rates out of the four bilingual varieties in Figure╯2, and Table╯10 shows that Malay does not use a copula in any of the predicate contexts. SgE (Chinese L1) has lower rates of copula absence than SgE (Malay L1) but higher than IndE, and, as Table╯10 indicates, the relevant substrates have an intermediate system of copula use as well. The two highest contexts for omission in SgE (Chinese L1) are the two contexts in which Chinese systems omit the copula.15 Basilectal SAIE (Dravidian L1) shows yet another distribution, mimicking the two highest contexts for copula absence in Tamil (see Table 10). IndE (Indo-Aryan L1) shows the lowest rates overall of copula absence, and indeed, as Table╯10 shows, the substrate languages involved have an obligatory copula with all predicate types. Copula absence is not robust in IndE: “mesolectal” or balanced English-Hindi bilinguals do not have copula omission at all, and each individual IndE speaker in the composite data in Figure╯2 was found to have a different ordering of contexts, also suggesting no strong underlying system. In contrast to these four New English varieties, the examples of an AAVE variety and a creole variety in Figure╯2 are striking in exhibiting a genuinely similar 15.╇ Ansaldo (2004, 2009) argues that copula absence across SgE substrates has a “ganging-up” effect on copula absence in the emergent variety. This is certainly true of the effect of Malay and Chinese, although in Figure╯2 we can still see traces of marginal divergences in Malay and Chinese in the corresponding sub-varieties of SgE. Ho and Platt’s data are relatively old and it is possible that SgE sub-varieties have now focussed towards a new, more unified omission pattern (Lisa Lim, p.c. 2008).

69

70

Devyani Sharma

distribution. A coarse grouping of all six systems in Figure╯2, merely based on the fact of copula absence, obscures the genuine similarity between AAVE and creoles as well as the strong evidence for specific L1 transfer in the other four systems (see Sharma and Rickford 2009 for further discussion and statistical measures). As with past omission and progressive over-extension, copula omission occurs in both IndE and SgE but quantitative analysis reveals a distinct patterning according to grammatical context in the two varieties, driven by substrate differences. 7. Conclusions This paper has argued that, in the search for vernacular universals, we cannot simply treat varieties that exhibit presence of a particular trait as identical, particularly where bilingual varieties of English are concerned. Surface similarities across New Englishes can be skin deep, diverging dramatically upon closer examination, due to substrate systems or substrate-superstrate interaction. If we genuinely aim to “identify those features which are the result of language contact, irrespective of the languages involved” (Sand 2004: 281), then it is inadequate to define a feature as broadly as “zero past tense forms of regular verbs” (K&S #40), “wider range of uses of the progressive” (K&S #21), or “deletion of be” (K&S #57). The degree and distribution of a given feature must be understood in relation to the substrate before any universal claims can be made regarding “processes [that] recur in vernaculars wherever they are spoken…” (Chambers 2004: 28). The quantitative comparison of two features — K&S #40 (past tense use) and K&S #21 (progressive use) — showed that although both features can be found in IndE and SgE, their patterning is genuinely similar in one case and substantially different in the other. The use of past tense marking to indicate perfectivity was strikingly parallel in IndE and SgE, and the substrate systems were found to be parallel as well. The use of progressive -ing to indicate imperfectivity was strikingly different in IndE and SgE, and the substrate systems were found to be different as well. Finally, K&S #57 (copula omission) was also found to occur in both IndE and SgE, as well as in other New Englishes, but in each case the grammatical conditioning is heavily influenced by the substrate language, such that some varieties do not retain copula absence as a feature of mesolectal speech (IndE) whereas others do (SgE). Certainly, repeated patterns may indeed derive from universal principles. Walker (2007) offers syntactic and processing explanations for clear quantitative parallels in variable agreement across native varieties of English, likely to be repeated regardless of the contact situation. However, for bilingual varieties of



Typological diversity in New Englishes

English it is imperative to first investigate and eliminate substrate explanations. The parallel assignment of perfective meaning to past tense morphology in IndE and SgE found in this study, for instance, is a genuinely comparable pattern yet it is not necessarily a universal. A simple substrate explanation accounts for the pattern, and only further comparison to varieties with distinct substrates can determine whether a substrate or unmarked explanation is appropriate. Similarly, even though we know that typologically progressive markers regularly grammaticalize into general imperfective markers (Bybee, Perkins and Pagliuca 1994: 140), SgE did not show the same robust over-extension of -ing as IndE did, thus ruling out a markedness explanation and supporting a substrate explanation. The curious additional complexity here is that, although it arose out of the contact between the grammars of Hindi and English, the innovative treatment of -ing as a global imperfective marker by IndE speakers results in a system that is typologically distinct from both input systems. Although this article cannot establish whether universals may additionally reinforce selected patterns found in SgE and IndE, it has shown that substratesuperstrate interactions can go a long way towards explaining emergent systems in postcolonial varieties of English, and that highly abstracted aggregations of varieties in the search for universals may obscure these sources of change.

References Andersen, Roger and Yasuhiro Shirai. 1996. “The primacy of aspect in first and second language acquisition: The pidgin-creole connection”. In William C. Ritchie and Tej Bhatia, eds. Handbook of Second Language Acquisition. London: Academic Press, 527–70. Ansaldo, Umberto. 2004. “The evolution of Singapore English: Finding the matrix”. In Lisa Lim, ed. Singapore English: A Grammatical Description. (Varieties of English Around the World G33.) Amsterdam, Philadelphia: Benjamins, 127–49. ———. 2009. “The Asian typology of English: Theoretical and methodological considerations”. English World-Wide 30(2): â•›133–48. Bao, Zhiming. 1995. “Already in Singapore English”. World Englishes 14: 181–8. ———. 2005. “The aspectual system of Singapore English and the systemic substratist explanation”. Journal of Linguistics 41: 237–67. Berwick, Robert. 1985. The Acquisition of Syntactic Knowledge. Cambridge, MA: MIT Press. Bickerton, Derek. 1981. Roots of Language. Ann Arbor: Karoma. ———. 1984. “The language bioprogram hypothesis”. Behavioral and Brain Sciences 7: 173–221. Bybee, Joan, Revere Perkins and William Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press. Chambers, Jack. 2004. “Dynamic typology and vernacular universals”. In Kortmann, ed.: 127–45. ———, Peter Trudgill and Natalie Schilling-Estes, eds. 2002. The Handbook of Language Variation and Change. Oxford: Blackwell. Chao, Yuen Ren. 1968. A Grammar of Spoken Chinese. Berkeley: University of California Press.

71

72

Devyani Sharma

Comrie, Bernard. 1976. Aspect. Cambridge: Cambridge University Press. Filppula, Markku, Juhani Klemola and Heli Paulasto. 2009. “Digging for roots: Universals and contacts in regional varieties of English”. In Markku Filppula, Juhani Klemola and Heli Paulasto, eds. Vernacular Universals and Language Contacts: Evidence from Varieties of English and Beyond. London: Routledge. Gupta, Anthea Fraser. 1994. The Step-Tongue: Children’s English in Singapore. Cleveland/UK: Multilingual Matters. Jenkins, Jennifer. 2003. World Englishes. London: Routledge. Ho, Mian Lian and John Platt. 1993. Dynamics of a Contact Continuum. Oxford: Oxford University Press. Kortmann, Bernd, ed. 2004. Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective. Berlin: Mouton de Gruyter. Kortmann, Bernd and Benedikt Szmrecsanyi. 2004. “Global synopsis — morphological and syntactic variation in English”. In Bernd Kortmann, Kate Burridge, Rajend Mesthrie, Edgar W. Schneider and Clive Upton, eds. A Handbook of Varieties of English. Vol. 2: Morphology and Syntax. Berlin, New York: Mouton de Gruyter, 1122–82. Kuo, Eddie C.Y. 1980. “The sociolinguistic situation in Singapore: Unity in diversity”. In Evangelos A. Afendras and Eddie C.Y. Kuo, eds. Language and Society in Singapore. Singapore: Singapore University Press, 39–62. Labov, William. 1972. Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press. Lefebvre, Claire. 1998. Creole Genesis and the Acquisition of Grammar: The Case of Haitian Creole. Cambridge: Cambridge University Press. Li, Charles N. and Sandra Thompson. 1981. Mandarin Chinese: A Functional Reference Grammar. Berkeley: University of California Press. Lim, Lisa. 2007. “Mergers and acquisitions: On the ages and origins of Singapore English particles.” World Englishes 26: 446–73. Masica, Colin P. 1991. The Indo-Aryan Languages. Cambridge: Cambridge University Press. Matthews, Stephen and Virginia Yip. 1994. Cantonese: A Comprehensive Grammar. London: Routledge. McWhorter, John. 2000. “Strange bedfellows: Recovering the origins of Black English”. Diachronica XVII: 389–432. Melchers, Gunnel and Philip Shaw. 2003. World Englishes: An Introduction. London: Arnold. Mesthrie, Rajend. 1992. English in Language Shift. Cambridge: Cambridge University Press. Ng, E-Ching. 2008. “English meets Malay meets Chinese: Colloquial Singaporean English”. Paper presented at the Workshop on “Interdisciplinary approaches to transfer, crosslinguistic influence and contact-induced change”, UWE Bristol. Pillai, N. Nadaraja. 1992. A Syntactic Study of Tamil Verbs. Mysore: Central Institute of Indian Languages. Platt, John. 1979. “Variation and implicational relationships: Copula realization in Singapore English”. General Linguistics 19: 1–14. ———, Heidi Weber and Ho Mian Lian. 1984. The New Englishes. London: Routledge. Poplack, Shana, ed. 2000. The English History of African American English. Oxford: Blackwell. Prentice, David John. 1990. “Malay (Indonesian and Malaysian)”. In Bernard Comrie, ed. The World’s Major Languages. Oxford: Oxford University Press, 913–35.



Typological diversity in New Englishes

Rickford, John R. 1992. “The creole residue in Barbados”. In Joan H. Hall, Nick Doane and Dick Ringler, eds. Old English and New: Studies in Honor of Frederic G. Cassidy. New York, London: Garland, 183–201. ———. 1998. “The creole origins of African-American Vernacular English: Evidence from copula absence”. In Salikoko Mufwene, John R. Rickford, Guy Bailey and John Baugh, eds. African-American English: Structure, History, Use. London: Routledge, 154–200. ———. 2006. “Down for the count? The creole origins hypothesis of AAVE at the hands of the Ottawa Circle, and their supporters”. Journal of Pidgin and Creole Languages 21: 97–155. Sand, Andrea. 2004. “Shared morpho-syntactic features in contact varieties of English: Article use”. World Englishes 23: 281–98. Sankoff, Gillian. 2002. “Linguistic outcomes of language contact”. In Chambers, Trudgill and Schilling-Estes, eds.: 638–68. Schiffman, Harold. 1999. A Reference Grammar of Spoken Tamil. Cambridge: Cambridge University Press. Schneider, Edgar W. 2007. Postcolonial English: Varieties Around the World. Cambridge: Cambridge University Press. Sharma, Devyani. 2005. “Language transfer and discourse universals in Indian English article use”. Studies in Second Language Acquisition 27: 535–66. ——— and Ashwini Deo. 2010. “A new methodology for the study of aspect in contact: Past and progressive in Indian English”. In James Walker, ed. Aspect in Grammatical Variation. Amsterdam/Philadelphia: Benjamins, 111–130. ——— and John Rickford. 2009. “AAVE / creole copula absence: A critique of the Imperfect Learning Hypothesis”. Journal of Pidgin and Creole Languages 24(1):â•›53–90. Singapore Census of Population. 2000. “Literacy and language”. Singapore Department of Statistics, Census of Population Office. <www.singstat.gov.sg> (15 Sept. 2008). Smith, Carlota S. 1991. The Parameter of Aspect. Dordrecht: Kluwer-Reidel. Stassen, Leon. 1994. “Typology versus mythology: The case of the zero-copula”. Nordic Journal of Linguistics 17: 105–26. Sun, Chaofen. 2006. Chinese: A Linguistic Introduction. Cambridge: Cambridge University Press. Svalberg, Agneta M.-L. and Hjh Fatimah Bte Hj Awg Chuchu. 1998. “Are English and Malay worlds apart? Typological distance and the learning of tense and aspect concepts”. International Journal of Applied Linguistics 8: 27–60. Tagliamonte, Sali. 2002. “Comparative sociolinguistics”. In Chambers, Trudgill and SchillingEstes, eds.: 729–63. Trudgill, Peter and Jean Hannah. 2008. International English: A Guide to Varieties of Standard English. 5th edition. Oxford: Oxford University Press. van Rooy, Bertus. 2006. “The extension of the progressive aspect in Black South African English”. World Englishes 25: 37–64. Walker, James A. 2007. “‘There’s bears back there’: Plural existentials and vernacular universals in (Quebec) English”. English World-Wide 28: 147–66. Williams, Jessica. 1987. “Non-native varieties of English: A special case of language acquisition”. English World-Wide 8: 161–99. Winford, Donald. 1998. “On the origins of African American Vernacular English: A creolist perspective. Part II: Linguistic features”. Diachronica 15: 99–155. Wolfram, Walt. 2000. “Issues in reconstructing Earlier African American English”. World Englishes 19: 39–58.

73

74

Devyani Sharma

Yip, Po-Ching and Don Rimmington. 2004. Chinese: A Comprehensive Grammar. London: Routledge. Yip, Virginia and Stephen Matthews. 2007. The Bilingual Child: Early Development and Language Contact. Cambridge: Cambridge University Press.

Thai English Rhythm and vowels* Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire University of Florida, Gainesville

We explore two aspects of English spoken by native speakers of Thai: rhythm and the vowel system, and compare each to the substrate language Thai, to target varieties of English, and to two New Englishes in Asia. Data was collected from a group of Thai speakers who participated in an interview in English, and who read a Thai paragraph, and English words, sentences and a paragraph. For rhythm, we measured the “Pairwise Variability Index” (nPVI, Grabe and Low 2002) and the proportion of time in an utterance devoted to vowels (%V, Ramus, Nespor and Mehler 1999) of Thai read speech, and English spontaneous and read speech. We find that the English of Thai speakers had stress-timed values of high nPVI, like Thai and British English (BrE), and low %V, like BrE but not Thai. Neither measure of rhythm resembled New Englishes’ more syllable-timed lower nPVI and high %V. The vowel system of Thai English revealed transfers of both quality and quantity from the substrate, resulting in a system distinct from British, American, and New Englishes. Keywords: Thai, Thai English, Singapore English, Hong Kong English, rhythm, English vowels, monophthongs

1. Introduction Although some Englishes in Asia have received extensive attention, the phonetics of Thai English remain relatively unexamined; Tsukada (2008) seems to provide *╇ We thank Jirapat Jangjamras, Donruethai Laphasradakul, and Ratree Wayland for all their help with recruiting Thai informants and providing information about English in Thailand. Thanks, too, to Lisa Lim for organizing the workshop on “The Typology of Asian Englishes” at ISLE1 in Freiburg, to members of the audience, and to Nik Gisborne, Lisa Lim, and Devyani Sharma for their comments on this work, which much improved this paper. We regret if we were unable to address all their suggestions.

76

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

the first acoustic descriptions, and that for only a subset of the vowels. We explore here two phonetic aspects of this emerging English: rhythm and the monophthong vowel system. We expect that the phonetic characteristics of this L2 English should either reflect transfer from the substrate Thai, or reflect cross-linguistic markedness constraints which are said to be shared in New Englishes; hence, we also compare the phonetics of Thai English to the characteristics of both the substrate language and to Hong Kong English (HKE) and Singapore English (SgE), two New Englishes in Asia. Thai English was chosen for its potential to distinguish transfer from markedness by its rhythmic characteristics. While the dichotomy between stress-timed and syllable-timed rhythm (Pike 1945; Abercrombie 1967) has been reanalyzed as a continuum (Roach 1982, 1998; Dauer 1983), New and L2 English varieties are often claimed to approximate the syllable-timed extreme (Schneider 2004), either due to transfer from a substrate with syllable-timing, or due to syllable-timing being unmarked. Thai L2 English may help distinguish between transfer and markedness, since Grabe and Low (2002) report that Thai’s rhythmic characteristics are mixed, based on two measures, the Pairwise Variability Index (nPVI, Grabe and Low 2002) and the proportion of time in an utterance devoted to vowels (%V, Ramus, Nespor and Mehler 1999). Using these two measures, we confirm the mixed values for the Thai substrate, and also test Thai English to compare it with British English (BrE; very stress-timed) and SgE (more syllable-timed). If transfer from the rhythmically mixed substrate Thai is dominant, we would expect Thai English to be mixed, while if unmarkedness determines L2 rhythm, we expect both measures to resemble other L2 and New Englishes (more syllable-timed). We also explore the question of transfer from the substrate vs. markednessâ•›/â•› typology by examining the vowels. In terms of quality, Thai has a more symmetric vowel system than English, lacking, for example, [ɪ] and [ʊ], and using length for contrast. We measure and plot the first two formants of the monophthongs of Thai English to compare inventories; the durations of vowels that are similar in quality are also measured to determine if length was used contrastively. The Thai English vowel system is then compared with the L1 Thai system, target native Englishes, and with New Englishes in the area with different substrates, such as HKE (Hung 2000) and SgE (Deterding 2006), to evaluate the role of direct transfer from the substrate vs. features shared throughout New Englishes (Schneider 2004). In the next section, we begin with the necessary preliminaries on English in Thailand, and the measurement of rhythm and vowels. We then present our methodology in Section╯3, and the results in Section╯4. The final section reviews the results in the context of the question of transfer vs. markedness.



Thai English

2. Background In this section, we provide some background information on English in Thailand as it would be experienced by our speakers. We also discuss measures of rhythm and vowel quality and quantity as a basis for choosing our methodology in Section╯3. 2.1 English in Thailand Unlike many countries in the South and South-East Asia area, Thailand has no history of colonization by the British. Thailand thus belongs to the Expanding Circle for English, rather than the Outer Circle, in the well-known system of Kachru (1985); that is, English is taught in the schools and used for communication primarily with foreigners, not for conversation with fellow Thais. Although English has been taught in the schools in Thailand for decades, according to Bautista and Gonzalez (2006) a lack of teachers resulted in English not being offered in schools until at least sixth grade (approximately age 11). However, “in 1996, the curriculum was revised once more, and English is now taught as a subject in Grade 1 to 12. Thus, about 99 percent of Thai students study English at school, but it appears that not many succeed in acquiring much English proficiency” (Bautista and Gonzalez 2006:â•›138). The type of English pronunciation that Thai speakers consider their target is not clear, or perhaps simply not monolithic. According to our informants, BrE has, until recently, been the predominant model in primary school and high school, with textbooks from Cambridge and Oxford University Presses and corresponding audio tapes for listening practice; however, the teachers at that level are generally native speakers of Thai, and pronunciation was rarely the focus of instruction. In college and university level English classes, there have always been more native speaker teachers and a greater variety of backgrounds, including British, American, Canadian and Australian as well. Our informants reported that American English (AmE) has become more predominant in Thailand within the past five to ten years. As the subjects recorded here had a mean age of 27, they likely received more formal instruction in BrE, but they have certainly had some exposure to AmE even in Thailand, and as they are currently studying in the US, they may have two models for English pronunciation. 2.2 Rhythm Prosodic characteristics, including rhythm, are often observed to distinguish native speakers from learners who transfer the rhythm of their L1 into an L2 (e.g. Munro 1995; Carter 2005; Tromfimovich and Baker 2006). While Inner Circle Englishes have been described as stress-timed, the rhythm of New Englishes is

77

78

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

often characterized as syllable-timed, leading to the question of whether New Englishes tend to be syllable-timed because of typologyâ•›/â•›markedness, if syllabletiming is less marked, or because of transfer, if only New Englishes with syllabletimed substrates have been examined. Thai has mixed prosodic characteristics, leading to the possibility of answering that question. The division of languages into two rhythmic categories, stress-timed or syllable-timed, has been reanalyzed as a continuum of rhythmic options, with stresstimed and syllable-timed as two extremes (Roach 1982; 1998; Dauer 1983; Ramus, Nespor and Mehler 1999; Grabe and Low 2002; Thomas and Carter 2006). In the original formulation, a syllable-timed language had approximately equal duration for each syllable, while a stress-timed language had equal duration for each foot, to keep the duration from one stressed syllable to the next approximately equal. Apart from problems such as determining the boundaries of the syllables or feet in order to test this formulation, the division did not seem to hold true; syllable durations can be unequal in languages described as syllable-timed, while the intervals between feet were not constant in stress-timed languages (Roach 1982; Dauer 1983). Dauer (1983) suggested that the perception of rhythm had more to do with contributing factors such as vowel reduction, the phonotactics of consonant clusters in syllables, and the location and acoustic characteristics of stress. In languages considered stress-timed, the variation in syllable structure and the presence of vowel reduction results in greater variability in the length of vowels. Recent work has developed acoustic measures in order to characterize the rhythmic properties of languages on a continuum. Two of these measures will be used here: %V (Ramus, Nespor and Mehler 1999) and the nPVI (Grabe and Low 2002); each of these correlates well with naïve categorizations of languages as syllable-timed, stress-timed, or as difficult to classifyâ•›/â•›indeterminate. The %V characterizes what percentage of an utterance consists of vowels: the more vocalic an utterance, the more likely it is to be perceived as syllable-timed, as it indicates a lack of vowel reduction and/or a simple CV structure disallowing lengthy consonant clusters. The nPVI compares the duration of each vowel to an adjacent vowel, to provide a measure of isochrony that does not depend on syllable or foot boundaries, but rather one that reflects that vowel reduction and variation in syllable structure contribute to greater variability in vowel duration. The higher the nPVI, the greater the perception of the variability that characterizes a stress-timed language; the lower the nPVI, the more equal the length of vowels in adjacent syllables, giving the perception of a syllable-timed rhythm. In addition to testing these measures on speech produced by native speakers, researchers have begun to investigate the rhythmic characteristics of L2 and New Englishes. While there have long been impressionistic observations that L2 and New Englishes tend towards syllable-timing, such as Wells (1982) and Kachru



Thai English

(1983), more recently acoustic phonetic measures have begun providing support. Notably, Low, Grabe and Nolan (2000) and Deterding (2001) have investigated SgE compared with BrE. Although their methodology and the details of their formula differed, both conclude that SgE is more syllable-timed than BrE. Other research includes Udofot (2003) for Nigerian English and Thomas and Carter (2006) for Jamaican English, Hispanic English, and the African-American English of exslaves (based on preserved recordings), all of which have been found to be more syllable-timed than native varieties of English. The generalization that L2 and New English varieties are more syllable-timed than BrE or AmE is often attributed to transfer or substrate influence from a syllable-timed L1 (e.g. Wells 1982:â•›572 on Carribean Creoles). The languages of Singapore (primarily Bazaar Malay and Chinese languages), Nigeria (Anaan, Ibibio, Yoruba), and Jamaican Creole tend to approximate the syllable-timed extreme, so that speakers may merely carry over a rhythmic pattern or a lack of vowel reduction in unstressed syllables. Crystal (1995:â•›176) claims that situations of use of English as a world language “have resulted in varieties of Modern English in which the syllable-timing has been transferred from the contact languages, producing a natural variety of isosyllabic English spoken as a mother tongue by large numbers of people, and viewed as a local speech standard”. However, an alternative is possible: if syllable-timing is a less marked state, it would be natural for learners to begin with that state in their L2. Crystal (1995) notes also that native speakers of stress-timed varieties use more syllable-timed rhythm in contexts such as baby-talk, perhaps indicating its status as less marked. To tease out the distinction between transfer of syllable-timing from substrate L1s and lack of markedness of syllable-timing, we need to examine the English of speakers of an L1 with mixed rhythmic characteristics. Thai English, whose rhythmic properties have not yet been measured, is an L2 variety based on a rhythmically mixed L1. In pre-acoustic phonetic research, Thai has been classified as a stress-timed language (Luangthongkum 1977, cited in Grabe and Low 2002); however, Grabe and Low (2002) report that Thai’s rhythmic characteristics are mixed. In Table╯1, we provide their findings for BrE, canonically stress-timed, and Spanish, canonically syllable-timed. SgE was measured as less stress-timed than the target BrE, a difference attributed to the syllable-timed substrate(s). With regard to Thai, their measure of nPVI for vocalic interval was high, even exceeding that of stress-timed languages such as Dutch and English, while the Thai %V was also high, resembling languages of the syllable-timed group such as Spanish and French. If transfer from the substrate is dominant, we would expect Thai English to be mixed as well. However, if unmarkedness determines L2 rhythm, we might expect both measures to show more syllable-timing than the target, similar to the values for SgE in Table╯1.

79

80

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

Table╯1.╇ nPVI and %V for BrE, SgE, Spanish and Thai, from Grabe and Low (2002); read speech, n╛=╛1 for each row nPVI (vocalic)

%V

BrE (very stress-timed)

57.2

41.1

SgE (more syllable-timed)

52.3

46.9

Spanish (syllable-timed)

29.7

50.8

Thai (mixed)

65.8

52.2

2.3 Vowels As with rhythm, the vowel systems of L2 and New Englishes are said to be influenced by the L1 or substrate vowel systems. Also as with rhythm, studies of New and L2 Englishes with distinct substrates have shown commonalities, suggesting a default system (Schneider 2004). We will compare the system of vowels used in Thai English with that of Thai and with New Englishes in the area with different substrates, to evaluate the role of direct transfer from the substrate vs. features shared throughout New Englishes. Standard Thai has nine monophthongs, each with contrastive length, as shown in Table╯2; there are also three diphthongs, /ɪa/, /ɯa/, and /ua/, which do not contrast in length (Roengpitya 2001). Table╯2.╇ Monophthongs of Standard Thai, from Roengpitya (2001:â•›1)a Heightâ•›/â•›Frontness

front

central

back

close

i, iː

ɯ, ɯː

u, uː

mid

e, eː

ɤ, ɤː

o, oː

open

æ, æː

a, aː

ɔ, ɔː

a

Tingsabadh and Abramson (1993) use the symbols [ɛ, ɛː] for the open front vowel.

According to Abramson (1962), the duration of the vowels provides the primary cue to their length. Roengpitya (2001), based on three speakers (two male and one female), provides duration measures for the short vs. long vowels, while Abramson (1962), based on two male speakers, provides formant values; both are provided below in Section╯4.2. As Thai speakers are exposed to both British and American models of English, we consider both as possible targets for vowel quality. Descriptions of the system of monophthongs that can occur in stressed vowels of the target BrE╛/╛RP and AmE generally include the vowels in Tables 3a and 3b. The vowel symbols below are placed in the chart according to their position in the IPA, although the acoustics of their pronunciation will vary depending on dialect (example measures of their pronunciation are presented in Section╯4.3).



Thai English

Table╯3a.╇ Monophthongs of BrE (Wells 1982) Height╛/╛Frontness

front

central

back

close

iː ɪ

–

ʊ uː

mid-close

e

ɜː

–

mid-open

–

–

ʌ ɔː

open

æ

–

ɒ ɑː

Table╯3b.╇ Monophthongs of AmE (Wells 1982) Height╛/╛Frontness

front

central

back

close

iɪ

–

ʊu

mid-close

–

ɜrâ•›/â•›ɚa

–

mid-open

ɛ

–

ʌɔ

open

æ

–

ɑ

a

Wells (1982) provides [ɜr], while Hillenbrand et al. (1995), whose acoustic values for AmE vowels will be used in Section╯4.1, refer to this vowel as [ɚ].

Compared to Thai, both AmE and BrE have a larger number of distinctions in quality in the open back vowel set; furthermore, BrE never has a distinction of quantity alone, as quality differences always accompany length differences. The AmE and BrE varieties also include mid-diphthongs, pronounced as [oʊ] or [əʊ] in GOAT and [eɪ] in FACE, which in many New or L2 Englishes are pronounced as long mid monophthongs [oː] and [eː], including in countries from Africa and the Caribbean (Wells 1982), as well as much of Asia including Brunei, India, Malaysia, and Singapore. Although this may seem a candidate for a markedness generalization, monophthongization may instead be due to substrate influence; for example, the substrates of SgE do not have [oʊ]â•›/â•›[əʊ]â•›/â•›[eɪ] diphthongs, and nor does SgE (Lim 2004), while the Cantonese substrate of HKE does have similar diphthongs, and so does HKE (Hung 2000). Comparison to other New Englishes may, however, reveal markedness effects in addition to substrate effects. Some potential candidates from other New Englishes are that the system is often simpler in terms of quality than BrE or AmE, with a reduced set of contrasts among the open back vowels [ʌ ɔː ɒ ɑː] and the front non-high vowels [e ɛ æ] (SgE, Deterding 2005; HKE, Hung 2000; Indian English, Wiltshire 2005). Many New or L2 Englishes also replace quality differences with quantity differences, using [iː]â•›/â•›[i] for [i]â•›/â•›[ɪ] and using [uː]â•›/â•›[u] for [u]â•›/â•›[ʊ]. In Section╯4.2, the Thai English system is described in terms of both quality and quantity, and compared with target varieties, the substrate Thai, and other New Englishes of the area.

81

82

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

3. Methodology Data were collected from a group of 12 Thai speakers; the group was fairly homogenous in terms of age and years of English education. The speakers ranged in age from 23 to 32 years (meanâ•›=â•›27); seven were from Bangkok (Central Thailand), and the others from Khon Kaen, Loei (both Northeastern), Chai Nat, Chiang Mai (both Northern), and Phuket (Southern). While there are regional differences in the Thai of the four areas, the dialects resemble each other in their systems of vowel quality and duration. Furthermore, our subjects were highly educated before their arrival in the United States (US), and thus are likely to have grown up speaking central Thai at home and/or to have adopted it while studying in Bangkok. All subjects had studied English in school and at university before their arrival in the US. They varied primarily in their contact with L1 English speakers: six have spent fewer than four months in an English-speaking country and six have spent more than 18 months in the US as of the time they were recorded. We label these two groups as “new” vs. “veteran” speakers of Thai English. For measures of Thai English, speakers were recorded reading English words, sentences, and a short paragraph. Then an interview was conducted, both to ask questions about their language background and to gather spontaneous speech. As Grabe and Low’s (2002) nPVI and %V results were on the basis of a single speaker, we wished to confirm their results for Thai with more data; thus three of the speakers were called back to read a short passage in Thai (from Tingsabadh and Abramson 1993). The recorded data of all speakers were digitized with a sampling rate of 45Hz using a CSL4400, a signal processing hardware and software package. 3.1 Rhythm There are a variety of ways to calculate rhythmic measures (e.g. Ramus, Nespor and Mehler 1999; Low, Grabe and Nolan 2000; Deterding 2001; Gibbon and Gut 2001; Grabe and Low 2002; Ong, Deterding and Low 2005; Thomas and Carter 2006), and these measures have been applied to both spontaneous and read data. As mentioned in Section╯2.2, two of these measures will be used here: %V (Ramus, Nespor and Mehler 1999) and the nPVI (Grabe and Low 2002). Grabe and Low’s (2002) investigation provides measures for, amongst others, the following varieties which are relevant to this study: one of the target varieties (BrE), the substrate (Thai) and one New English (SgE); in order to compare our results with theirs, we have adopted their formula for the most part. Note however that Grabe and Low (2002) measured only one speaker for each language, and measured the rhythm of a read passage. We have larger sample sizes, and measure both read and spontaneous/



Thai English

conversational speech for Thai English, as spontaneous speech may provide more naturalistic rhythms (Deterding 2001). Further modifications are discussed below. In order to calculate the nPVI and %V, we measured the duration of vowels and the duration of intervals between vowels. In the analysis of the spontaneous speech, three of the speakers were excluded, either because their speech was not fluent or because their utterances were rarely long enough. As noted in Thomas and Carter (2006), an insufficiently large number of syllable pairs might allow for deviations based on segmental factors such as the length of vowels or structure of syllables, so for the measures of spontaneous speech, we included only speakers for whom we had at least 200 syllables. Therefore, a total of nine speakers were analyzed for the spontaneous data; of them, four were speakers who had been in the US more than 18 months (veterans) and five fewer than 4 months (new). We then had an average of 273 syllables per speaker for the veteran speakers group (nâ•›=â•›4) and 280 syllables per speaker for the new group (nâ•›=â•›5). Following Deterding (2001), we used utterances only if they had a minimum of seven syllables without any long pause or hesitation; as these were not native speakers of English, short pauses within a breath group occurred regularly, and were included for analysis (in the intervocalic duration). The utterances were measured for vocalic duration (d), with the boundaries of the vowels in most cases determined using standard criteria such as the onset of the second formant (Peterson and Lehiste 1960). To maintain consistency in the measurements of all utterances, we used the following criterion: the transition of a vowel from the preceding consonant was considered in vocalic duration as the starting boundary, while the transition of a vowel into the following consonant/pause was not taken as part of the vocalic duration; that is, the final boundary was marked at the end of the steady state of the vowel. In order to avoid potential effects of English phrase-final lengthening, we applied the formula only to pre-final syllables. From these measures, we calculate the nPVI according to the formula in (1), as measures of variation in vocalic duration, and the %V according to (2), the proportion of time in an utterance devoted to vowels. â•…



  dk − d k+1    / (mâ•›−â•›1)   (dk + d k+1)/2 

m−1

(1) nPVIâ•›=â•›100 × ∑╇ dk 



╅╇ k╛=╛1

 ╇



m

Grabe and Low (2002:â•›520)



(2) %Vâ•›=â•›100 × ∑╇ dk / duration of utterance





kâ•›=â•›1 

Ramus et al. (1999:â•›271–2)

As shown in (1), the nPVI compares the difference in duration of the vowels in adjacent syllables; dividing this value by the mean of the two vowels’ duration

83

84

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

normalizes for rate. The absolute values calculated for each pair of syllables is divided by the total number of syllables to give a value for each speaker, or group of speakers. The formula in (2) for %V shows that it merely adds up the duration of each vocalic portion of an utterance and divides by the total duration of the utterance. Both types of measures will be compared to results reported for BrE (very stress-timed) and SgE (very syllable-timed) to situate Thai English on the rhythmic scale, and to previous measures for Thai to evaluate the extent of transfer from L1 in Thai English speakers.1 We furthermore divide our data in terms of years in an English-speaking country (new vs. veteran). We hypothesize that more experience with AmE speakers, with their high vocalic nPVI and low %V, could modify Thai English rhythm in the direction of a lower %V. 3.2 Vowels For measurements of vowel quality and quantity, the recordings of all 12 speakers were used. Using the word list of individual words articulated in isolation, the duration of the vowels was measured from the onset to offset of the second formant. Following Abramson (1962), the first two formants of the vowels were measured at a point halfway through the vowel, including the mid-close vowels which are diphthongs in BrE and AmE. Vowel formants are plotted in a table to compare inventories with Thai and varieties of English, while the durations of vowels that are similar in quality were measured to determine if length was used contrastively. 4. Results 4.1 Rhythm The results of the calculation of the rhythm measures for Thai English are presented in Table╯4, along with values from Grabe and Low (2002) for Thai in the first row. Our measure for Thai shows lower values for both nPVI and %V than that of Grabe and Low (2002). While our results are thus not as dramatically mixed as theirs, the nPVI number is still close to the stress-timed, BrE-like end of the scale, while the %V resembles the high values of a syllable-timed language such as Spanish. Thus the rhythmic characteristics of Thai do indeed appear to be mixed.

1.╇ We use BrE as the comparison here because available published results on AmE are presented as median values rather than means; however, AmE, like BrE, falls clearly on the stress-timed end of the spectrum, in e.g. Carter and Thomas (2006).



Thai English

Table╯4.╇ Results of nPVI and %V for Thai, Thai English, and BrE nPVI (vocalic)

%V

Thai (Grabe and Low 2002; nâ•›=â•›1)

65.8

52.2

Thai (read speech, nâ•›=â•›3)

54.5

49.3

Thai English (spontaneous speech, nâ•›=â•›9)

56.5

42.7

Thai English (read speech, nâ•›=â•›10)

59.6

40.0

BrE (Grabe and Low 2002; nâ•›=â•›1)

57.2

41.1

Table╯4 also includes our results for Thai English, both spoken spontaneously and from a read paragraph, the latter value being measured in a way comparable to the value measured for a native BrE speaker in Grabe and Low (2002), except that we did not include utterances shorter than seven syllables. Thai English, unlike Thai, appears more consistently stress-timed in its rhythm. The overall value of the nPVI for Thai English is quite high, remaining on the stress-timed part of the scale, which not surprisingly reflects both the targets and the substrate. The values of nPVI for Thai English, whether read or spoken, are higher than those we measured for Thai, though lower than Grabe and Low report. The values for %V for Thai English also fall at the stress-timed end of the continuum, with the values in read Thai English possibly more stress-timed than the values measured for BrE. The values for spoken Thai English fall between the values of the very stress-timed read English and the more syllable-timed value of Thai itself. In Table╯5 we divide our results by veteran vs. new speakers as well as read vs. spontaneous speech. Both groups show similar values in the read English, but they differ on both measures in their spontaneous English. By transfer from the substrate, we might expect the new speakers to show rhythmic values closer to the L1 Thai values than the veteran speakers. The fact that the new speakers have an even higher nPVI than veterans is surprising, as is the lower value for %V, as it appears that the newer speakers show values closer to the stress-timed BrE values on both measures, while the more veteran speakers perform more similarly to the speakers of SgE. None of the Thai English speakers show a %V approaching that of Thai. It is possible that the structure of Thai itself Table╯5.╇ Results of nPVI and %V for veteran vs. new Thai English speakers nPVI (vocalic)

%V

Read speech, veterans (nâ•›=â•›5)

59.6

39.7

Read speech, new (nâ•›=â•›5)

59.6

40.4

Spontaneous, veterans (nâ•›=â•›4)

54.3

43.9

Spontaneous, new (nâ•›=â•›5)

58.3

41.7

85

86

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

(length of vowels, form of syllables) contributes to the %V being so high in Thai, rather than some external prosodic characteristic. In the surprising direction of the differences between new and veteran speakers, one potential explanation is that the new speakers are more oriented towards an external norm, so that they overcompensate when interacting with native speakers. Veterans, on the other hand, may be comfortable with a more nativized variety which reflects the substrate; by relaxing towards Thai norms, they perhaps reassert their identity as Thai speakers. In this, we could extend the Norm Orientation Hypothesis of Gut (2007), who notes that the use and development of linguistic structures is related to and dependent on the normative orientation of the speakers. Her evidence comes from two New Englishes, SgE and Nigerian English; speakers of the former have an endonormative orientation and distinct patterns of consonant cluster reduction, as opposed to speakers of the latter, whose exonormative orientation results in patterns of reduction like that of L2 learners. Similarly, perhaps even speakers of an Expanding Circle English, as they develop their expertise in English, may move from an attitude of pure allegiance to the target varieties towards developing a more endonormative attitude, or at least one that does not overcompensate to avoid transfer. It is also possible to see the veteran speakers as applying a strategy used by bilinguals, trying to find a balance between the values of Thai and English rhythm, just as bilinguals maintain a balance between similar segmental speech sounds of their L1 and L2 (Flege, Schirru and MacKay 2003). 4.2 Vowels For vowels, we present measures of vowel formants (quality) and duration (quantity) for monophthongs and diphthongs below, for comparison with native and New Englishes. 4.2.1 Vowel quality In all the figures that follow, the charts of the vowel spaces are presented with F2 plotted against F1 to represent the vowel spaces in a familiar format.2 In all cases, formant frequencies in Hz are converted to the auditory Bark scale using the following formula (J. Harnsberger p.â•›c. 2008) to better reflect the relative perceptual positions of the vowels in the vowel space:

(3) Barkâ•›=â•›7 × LN(Hz / 650 + SQRT (POWER (Hz / 650.2) + 1)

2.╇ In this we follow Deterding’s (1997) arguments that this presentation is better than the alternative F1 vs. F2-F1 plot.



Thai English

First, in order to investigate the role of substrate transfer, we provide acoustic measures of the Thai L1 vowel system and target systems for the BrE and AmE L2. Figures 1a and 1b are based on the formant values provided for Thai by Abramson (1962) for the long and short monophthongs of two male speakers.

Figure╯1a.╇ Thai short monophthongs (Abramson 1962)

Figure╯1b.╇ Thai long monophthongs (Abramson 1962)

As outlined in Section╯2, there are potentially two target Englishes, BrE and AmE, that serve as norms for Thai English learners; examples of the formant values for their monophthongs are plotted in Figures 2 and 3. Note that these plots do not include the normally diphthongal mid-vowels [eɪ] and [oʊ]â•›/â•›[əʊ].

87

88

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

Figure╯2.╇ BrE monophthongs, males (Deterding 2006)

Figure╯3.╇ AmE monophthongs, males (Hillenbrand et al. 1995)

Comparing the Thai system to that of BrE, we see that amongst the front vowels, Thai provides three categories of height, while BrE requires four, possibly five if the diphthong [eɪ] is also taken into account. For the central and back, mid and low vowels, the Thai L1 provides a four-way contrast plus length; BrE has a five-way qualitative contrast among these vowels, some of which contrasts are reinforced by length as well as quality differences. The AmE system, in Figure╯3, provides slightly fewer contrasts: in the front vowel system, three heights (four with the diphthong [eɪ]), and in the centralâ•›/â•›back vowels, four monophthongs plus one diphthong.



Thai English

Table╯6.╇ Results of formant measures of Thai English monophthongs F2 (Hz)

F1 (Hz)

F2 (Bark)

F1 (Bark)

vowel

word

2463.4

375.1

14.29719

3.843589

iː

heat

2377.5

462.6

14.05727

4.635914

ɪ

hit

2288.2

532.1

13.79903

5.229974

e

hate

2412.1

536.0

14.15468

5.262823

ɛ

head

1888.2

813.9

12.51555

7.342769

æ

had

1577.4

832.4

11.33786

7.466315

ʌ

hut

1532.6

934.1

11.15187

8.115409

ɑː

asks

1380.2

536.7

10.48243

5.267856

ɜː

hurt

1188.9

608.0

â•⁄ 9.551265

5.844269

ɒ

cloth

1117.0

759.1

â•⁄ 9.171095

6.966584

ɔː

thought

â•⁄ 919.8

518.9

â•⁄ 8.027021

5.119411

oː

coat

1147.7

450.5

â•⁄ 9.335385

4.529141

uː

hoot

1038.9

483.9

â•⁄ 8.736383

4.821361

ʊ

hook

Figure╯4.╇ Thai English vowels, all speakers (n╛=╛12)

Thus the number of contrasts in Thai, as well as the locations of the vowels, differ from the targets and may provide evidence of transfer to Thai English. Our formants measures for Thai English monophthongs for each vowel, averaged over all speakers, are presented in Table╯6 and plotted in Figure╯4. Most vowels seem to be maintained in distinct locations on the averaged chart, although [e] vs. [ɛ] are close to being merged. As the BrE and AmE values plotted

89

90

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

Figure╯5a.╇ Thai English vowels, male veteran speakers (n╛=╛3)

Figure╯5b.╇ Thai English vowels, male new speakers (n╛=╛3)

above were for males, we provide plots for the male speakers alone, breaking them into groups based on length of their time in the US as in Figures 5a and 5b. From the figures of more homogeneous subgroups, we see that both groups show merger of [e] vs. [ɛ], a distinction unavailable in the substrate Thai. Similarly, the new male speakers also merge [iː] vs. [ɪ], while these two are quite distinct for the veterans. Note, too, that all speakers clearly have the low front vowel [æ], as does Thai. So in the front vowels, the new speakers tend to closely reflect the Thai vowel system, while the veterans have added a distinction for the high tenseâ•›/â•›lax vowels.



Thai English

For all speakers, the low and back vowels seem to be kept distinct, although they are not in the same places for the two groups, nor do they exactly match either of the target Englishes. For example, both the [uː] and [ʊ] of the new speakers are fairly low and central. The [uː] and [ʊ] of the veterans have moved higher and further back, but the [ʊ] is even more so than the [uː], whose location is somewhere between the BrE and AmE pronunciations. Figures 6 and 7 for SgE and HKE monophthongs provide for further comparisons with nearby New English varieties.

Figure╯6.╇ SgE monophthongs (Deterding 2006)

Figure╯7.╇ HKE monophthongs (Hung 2000)

91

92

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

The data in both tables lack equivalents for [eː] and [oː] as their authors were measuring vowels considered to be monophthongs in the target language, BrE. In the front vowel system of Thai English (Figures╯4 and 5), [æ] is noticeably distinct from [eâ•›/â•›ɛ], as in both the Thai L1 and in the target Englishes, but both SgE and HKE show merger or near merger of these front vowels. However, the substrates for SgE (Malay, Hokkien, Mandarin, and Cantonese) and for HKE (Cantonese) also lack the [æâ•›/â•›eâ•›/â•›ɛ] distinction (Lim 2004), so that both the presence and absence of [æ] in different varieties seems attributable to the influence of the substrate, rather than markednessâ•›/â•›universals. All three varieties show a tendency to merge [iː] with [ɪ] in quality. In the back vowel system, [uː] appears more back in all three non-native Englishes than in BrE. As the fronting of [uː] is relatively recent for BrE speakers, it is possible that all three have preserved an older pronunciation; for Thai, with its lack of a historical colonial dialect, this “preservation” could come from the textbookâ•›/â•›formal variety used in schools, or it may reflect the location of the back vowels in the substrate L1s. Among the non-high back and central vowels, where BrE provides for multiple contrasts, the New Englishes merge one or more, for example, [ɔː]â•›/â•›[ɒ] appear merged in HKE and close to it in SgE. Such mergers have been treated as the influence of the substrate in Hung (2000), Lim (2004), and Deterding (2006). Similarly for Thai English, both the loss of contrasts in the quality of pairs such as [iː]â•›/â•›[ɪ] and the maintenance of contrasts such as [ɛ]â•›/â•›[æ] can be attributed in Thai English speakers to the influence of the substrate Thai vowel inventory. 4.2.2 Vowel durations The durations of the vowels were measured from the read word list, and although this data was not gathered with any control for the rate of reading, the average durations as presented in Table╯7 provide evidence suggesting that the Thai English speakers maintain a distinction in duration for some vowel contrasts, in support of Tsukada (2008). The vowel symbols in Table╯7 represent the target vowels as they are normally pronounced in BrE, with key words indicating an example of a word on the reading list for the Thai English speakers. Tsukada (2008) noted that the monophthongs of Thai speakers in Australia [ɪ æ ʊ ʌ] were shorter than the corresponding monophthongs of Australian speakers, resulting in a lower ratio of short vowels to diphthongs of approximately 1:2. Tsukada suggests that this reflects a transfer to the vowel contrasts of English, as the Thai ratio of short to long is also roughly 1:2, as seen in Roengpitya’s (2001) results also provided in Table╯7. As the Thai speakers in this study reside in the US, we provide values for AmE vowel durations from Hillenbrand et al. (1995) for comparison. There are at least two cases in which the Thai speakers’ monophthongs are notably short: [ɪ] and [ʊ]. In the first case, the



Thai English

Table╯7.╇ Vowel duration for Thai English, AmE and comparable Thai vowels; mean duration in seconds English vowels

Key word

Thai English

AmE (Hillenbrand et al. 1995)

Comparable Thai vowels

Roengpitya (2001)

iː

heat

.211

.243

iː

.298

ɪ

hit

.106

.192

i

.145

e(ː)

hate

.212

.267

eː

.301

ɛ

head

.233

.189

e

.149

æ

had

.222

.278

æː

.332

ɛː

hurt

.204

.263

ɤː

.332

ʌ

hut

.118

.188

ɤ

.175

ɑː

asks

.212

.267

ɔː

.334

ɒ

cloth

.226

–

ɔ

.165

ɔː

thought

.254

.283

ɔː

.334

o(ː)

coat

.208

–

oː

.320

ʊ

hook

.106

.192

u

.150

uː

hoot

.175

.237

uː

.321

ratio of [ɪ] to [i] in Thai English approximates 1:2, as in the Thai contrast of [i] vs. [iː]. Thus it is possible that in pairs such as heatâ•›/â•›hit for [iː]â•›/â•›[ɪ] or hootâ•›/â•›hook for [uː]â•›/â•›[ʊ], the length distinction seems to be transferred from Thai into Thai English as suggested by Tsukada. However, hateâ•›/â•›headâ•›/â•›had show no such length distinction, at least in isolated monosyllables from the word list. As discussed in Section╯4.2.1, [æ] was qualitatively distinct from the other front vowels, but [eː] and [ɛ] were similar in quality; with similar quantities as well, they appear not to contrast at all in Thai English. Thus while the durational contrasts from the substrate may be transferred into the English of Thai speakers, they are not always, even when neutralization of a target contrast is the result. 5. Conclusion An examination of Thai English rhythm and vowels revealed a strong effect of the substrate Thai. These effects include the L1 rhythmic characteristic of high nPVI transferred from Thai to Thai English, the presence of the low-front vowel [æ], the absence of qualitative contrasts such as [e]â•›/â•›[ɛ] and [iː]â•›/â•›[ɪ] (for new speakers), and the location of the back high vowel [uː]. However, not every characteristic

93

94

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

from Thai survives in the Thai English of even the new speakers. The high value of %V found in Thai, which normally correlates with a syllable-timed language, is replaced with a low value resembling that of native varieties, so that Thai English is uniformly stress-timed rather than mixed in its rhythmic measures. The use of a 1:2 ratio of duration to distinguish vowels plays a more limited role in Thai English than in Thai. While a great deal of work remains to be done in describing the complete system of Thai English speakers, it is clear that as a result of differences in its substrate, the resulting L2 variety Thai English stands distinct from not only native varieties of English but also from the New Englishes provided for comparison here, SgE and HKE. The rhythm of Thai English far more closely resembles that of BrE than does SgE, due in part to transfer of Thai’s rhythmic characteristics, while the vowel systems of each reflect the L1â•›/â•›substrates of each. The presence of commonalities in the vowel systems of Thai English, SgE and/or HKE, reflect similarities in their substrates, rather than a lack of markedness or the emergence of universals. Thus, the examination of Thai English supports the notion that observed similarities among New Englishes may reflect the similarities of the substrates of the varieties examined thus far; when Englishes with divergent substrates are studied, the appearance of uniformity disappears.

References Abercrombie, David. 1967. Elements of General Phonetics. Edinburgh: Edinburgh University Press. Abramson, Arthur S. 1962. The Vowels and Tones of Standard Thai: Acoustical Measurements and Experiments. Bloomington, Indiana University Research Center in Anthropology, Folklore and Linguistics. Bautista, Maria Lourdes S. and Andrew B. Gonzalez. 2006. “Southeast Asian Englishes”. In Braj B. Kachru, Yamuna Kachru and Cecil L. Nelson, eds. The Handbook of World Englishes. Malden, MA: Blackwell, 130–44. Carter, Phillip M. 2005. “Quantifying rhythmic differences between Spanish, English, and Hispanic English”. In Randall S. Gess and Edward J. Rubin, eds. Theoretical and Experimental Approaches to Romance Linguistics. Amsterdam, Philadelphia: Benjamins, 63–75. Crystal, David. 1995. “Documenting rhythmical change”. In J. Windsor Lewis, ed. Studies in General English and Phonetics: Essays in Honour of Professor J.D. O’Connor. London, New York: Routledge, 174–9. Dauer, Rebecca M. 1983. “Stress-timing and syllable-timing reanalyzed”. Journal of Phonetics 11: 51–62. Deterding, David. 1997. “The formants of monophthong vowels in Standard Southern British English pronunciation”. Journal of the Interational Phonetic Association 27: 47–53. ———. 2001. “Letter to the editor. The measurement of rhythm: A comparison of Singapore and British English”. Journal of Phonetics 29: 217–30.



Thai English

———. 2005. “Emerging patterns in the vowels of Singapore English”. English World-Wide 26: 179–97. ———. 2006. “The north wind versus a wolf: Short texts for the description and measurement of English pronunciation”. Journal of the International Phonetic Association 36: 187–96. Flege, James E., Carlo Schirru and Ian R.A. MacKay. 2003. “Interaction between the native and second language phonetic subsystems”. Speech Communication 40: 467–91. Gibbon, Dafydd and Ulrike Gut. 2001. “Measuring speech rhythm”. Eurospeech 2001 — Scandinavia, 95–9. Grabe, Esther and Low Ee Ling. 2002. “Durational variability in speech and the rhythm class hypothesis”. Papers in Laboratory Phonology 7: 515–46. Gut, Ulrike. 2007. “First language influence and final consonant clusters in the new Englishes of Singapore and Nigeria”. World Englishes 26: 346–59. Hillenbrand, James M., Laura A. Getty, Michael J. Clark and Kimberlee Wheeler. 1995. “Acoustic characteristics of American English vowels”. Journal of the Acoustical Society of America 97: 3099–111. Hung, Tony. 2000. “Towards a phonology of Hong Kong English”. World Englishes 19: 337–56. Kachru, Braj B. 1983. The Indianization of English. Delhi: Oxford University Press. ———. 1985. “Standards, codification, and sociolinguistic realism: The English language in the Outer Circle”. In Randolph Quirk and Henry G. Widdowson, eds. English in the World: Teaching and Learning of Language and Literature. Cambridge: Cambridge University Press, 11–36. Lim, Lisa. 2004. “Sounding Singaporean”. In Lisa Lim, ed. Singapore English: A Grammatical Description. (Varieties of English Around the World G33.) Amsterdam, Philadelphia: Benjamins, 19–56. Low, Ee Ling, Esther Grabe and Francis Nolan. 2000. “Quantitative characterizations of speech rhythm: Syllable-timing in Singapore English”. Language and Speech 43: 377–401. Munro, Murray. 1995. “Nonsegmental factors in foreign accent”. Studies in Second Language Acquisition 17: 17–34. Ong, Fiona Po Keng, David Deterding and Low Ee Ling. 2005. “Rhythm in Singapore and British English: A comparative study of indexes”. In David Deterding, Adam Brown and Low Ee Ling, eds. English in Singapore: Phonetic Research on a Corpus. Singapore: McGraw-Hill Education (Asia), 74–85. Peterson, Gordon and Ilse Lehiste. 1960. “Duration of syllable nuclei in English”. Journal of the Acoustical Society of America 32: 693–703. Pike, Kenneth. 1945. The Intonation of American English. Ann Arbor: University of Michigan Press. Ramus, Franck, Marina Nespor and Jacques Mehler. 1999. “Correlates of linguistic rhythm in the speech signal”. Cognition 73: 265–92. Roach, Peter. 1982. “On the distinction between stress-timed and syllable-timed languages”. In David Crystal, ed. Linguistic Controversies. London: Edward Arnold, 73–9. ———. 1998. “Some languages are spoken more quickly than others”. In Laurie Bauer and Peter Trudgill, eds. Language Myths. London: Penguin, 150–8. Roengpitya, Rungpat. 2001. “A study of vowels, diphthongs, and tones in Thai”. Ph.D. dissertation, University of California, Berkeley. Schneider, Edgar W. 2004. “Global synopsis: Phonetic and phonological variation in English world-wide”. In Edgar W. Schneider, Kate Burridge, Bernd Kortmann, Rajend Mesthrie

95

96

Priyankoo Sarmah, Divya Verma Gogoi and Caroline Wiltshire

and Clive Upton, eds. Handbook of Varieties of English. Vol. 1: Phonology. Berlin, New York: Mouton de Gruyter, 1111–37. Tingsabadh, M.R. Kalaya and Arthur S. Abramson. 1993. “Illustrations of the IPA: Thai”. Journal of the International Phonetic Association 23: 24–8. Thomas, Erik R. and Philip M. Carter. 2006. “Prosodic rhythm and African American English”. English World-Wide 27: 331–55. Trofimovich, Pavel and Wendy Baker. 2006. “Learning second language supersegmentals”. Studies in Second Language Acquisition 28: 1–30. Tsukada, Kimiko. 2008. “An acoustic comparison of English monophthongs and diphthongs produced by Australian and Thai speakers”. English World-Wide 29: 184–211. Udofot, Inyang. 2003. “Stress and rhythm in the Nigerian accent of English: A preliminary investigation”. English World-Wide 24: 201–20. Wells, John C. 1982. Accents of English. Cambridge: Cambridge University Press. Wiltshire, Caroline. 2005. “The ‘Indian English’ of Tibeto-Burman language speakers”. English World-Wide 26: 275–300.

Revisiting English prosody (Some) New Englishes as tone languages?* Lisa Lim

The University of Hong Kong

Many New Englishes are spoken in what can often be considered multilingual contexts in which typologically diverse languages come into contact. In several Asian contexts, one typological feature that is prominent in the multilingual contact situation (the “ecology”) is tone. Given that tone is recognized as an areal feature and is acquired easily by languages in contact, the question that arises is how this is manifested in the prosody of these New Englishes. Recent work has shown that contact languages, including English varieties, evolving in an ecology where tone languages are present do indeed combine aspects of tone languages. This paper attempts to go a step further, in suggesting not only that such varieties should not be viewed as aberrant in comparison to “standard” English but recognized as having their own prosodic system partly due to substrate typology, but also that in the consideration of New Englishes — here, Asian (but also African) Englishes — the traditional view of English as a stressâ•›/â•›intonation language need to be revisited and revised, to consider some New Englishes as tone languages. Singapore English (SgE) is presented as a case in point, with the presence of tone demonstrated in the set of SgE particles acquired from Cantonese, at the level of the word, as well as in the intonation contour which moves in a series of level steps. A comparison is then made with Hong Kong English, another New English in a tone-language-dominant ecology, with a consideration of typological comparability as well as difference due to the dynamic nature of SgE’s ecology. Keywords: Singapore English, Hong Kong English, Asian Englishes, New Englishes, typology, ecology, intonation, tone, tone language *╇ I thank Edgar Schneider, Bao Zhiming, and two anonymous reviewers for comments on an earlier, slightly different version of this paper, and Umberto Ansaldo, Nik Gisborne, Edgar Schneider and Caroline Wiltshire for comments on this paper. I am grateful to the following for sharing their work: Zhu Shenfa and Stephanie Wagenaar for some of the instrumental work on SgE intonation and particles respectively, Ng E-Ching and Wee Lian Hee for their thoughts on tone in the SgE word, and K.K. Luke for his material on HKE intonation.

98

Lisa Lim

1. Why revisit English prosody? Why revisit English prosody indeed? This might well be the question that immediately arises upon encountering this paper. It is not particularly novel, after all, to recognize that the prosodic patterns of varieties of English differ: recent years have seen extensive attention to intonational variation in English (IViE) within the British Isles (see e.g. Grabe 2004), and different intonation patterns in varieties of English around the world, including New Englishes, have long been noted (e.g. Wells 1982; Schneider et al. 2004). Closer to home — the focus of this issue — in the past two decades, too, a significant amount of research has made headway in providing much-needed and important descriptions of various prosodic aspects of Asian Englishes (see e.g. the summary overviews in Mesthrie 2008b: for Indian English, Gargesh 2008; Malaysian English, Baskaran 2008; Philippine English, Tayao 2008). For Singapore English (SgE), research has encompassed areas of intonation (e.g. Deterding 1994; Low 1994, 1998; Lim 1997, 2000, 2001, 2004a; Goh 1998; Zhu 2003), stress (e.g. Low 2000; Lim and Tan 2001; Tan 2003; Lim 2004b), and rhythm (e.g. Low, Grabe and Nolan 2000; Deterding 2001). So why is a revisiting of English prosody necessary? The motivation for this is three-fold. First, as has often been the pattern in some scholarship, such descriptions of structural features of New Englishes have tended to be, at worst, in relation to what they lack in comparison to native varieties of English, and, at best, how they differ from native varieties while still adopting notions and terms which may not be at all relevant for New Englishes, such as tonic stress. While such observations of prosodic patterns have certainly been valuable in the growing body of material for the description of a variety, I believe that viewing patterns in New Englishes from the vantage point of patterns documented for native Englishes can do a disservice to the New English varieties in that such an approach tends to miss representing patterns which reflect structural features from the Asian substrates. Instances where a serious attempt to describe and explain such distinctive patterns in these New Englishes in their own right are few and far between, though in very recent years, a number of younger scholars have been taking this up (more of this later). Second, descriptions of structural features of various regional groups of New Englishes — African, South Asian, Southeast Asian, East Asian, etc. — have tended to manifest in a list of features which the members in each group — or across a number of groups — share (e.g. Kachru and Nelson 2006 on Asian Englishes; Mesthrie 2008a on African and Asian Englishes), as if by virtue of membership of that particular group (also see Lim 2009). While in (summaries of) such widely encompassing, ambitious volumes, a certain amount of reductionism is certainly necessary, even desireable, for providing broad brush strokes of groups



Revisiting English prosody

of varieties, and while elsewhere in the volumes (e.g. in the separate chapters on individual African and Asian Englishes in Mesthrie 2008b) there may be an indication of the factors for such patterns, some mention of why these systematic phonological similarities or phonetic differences amongst or between the groups still needs to be made, on the basis of the typologies of the substrate (and superstrate) languages.1 Finally, a clear recognition of the typologies of the substrate languages of some New Englishes where one or more of the substrates are tone languages must surely lead to a serious consideration of what this means for the New English. While English is classified a stress or intonation language, New Englishes which manifest some presence of tone in their prosodic system must warrant at least some consideration for whether they may be classified instead as tone languages, as is suggested for SgE (Lim 2007a, b, 2008a, b).2 This paper is thus an attempt to address these three concerns. To best appreciate the patterns observed in the prosody of some Asian Englishes, I find it useful to subscribe to an ecological approach (after e.g. Mufwene 2001, 2008; for a detailed account of applying this approach to the understanding of Asian Englishes, see Ansaldo 2009a, b, 2010; for a model which assumes this perspective for Postcolonial Englishes, see Schneider 2007), which adopts ecology as a metaphor from population genetics and biology. In this approach, the emergence of contact-induced varieties such as SgE can be regarded as speakers making selections from a pool of linguistic variants available to them in a contact setting. This feature pool consists of the sum total of the individual forms and variants that each of the speakers involved, with different language backgrounds and varying linguistic experiences, brings into the contact situation. Which variants from this feature pool are chosen as stable elements of the newly emerging variety depends on the complete ecology of the contact situation, which comprises both external and internal factors. External ecology involves components such as the numerical (demographic) relationships between speech communities, the social relationships, involving issues of power or prestige distributions, between the participants, 1.╇ I wish to add that, even as I make this point, I have the greatest amount of respect for these invaluable collections of descriptions of varieties of English and the authors and editors who put them together. 2.╇ To my knowledge, the idea that SgE, as a New English, be considered a tone language was for the first time seriously postulated in Lim (2007a:â•›468–9), and then more explicitly proposed in Lim (2007b, 2008a, b). (Note though that Killingey (1968) suggests that SgE word stress should not be discussed on the grounds that Malayan [Singapore + Malaysia] English is “a tone language” but later (Killingey 1972) withdraws the statement, cited in Bloom 1986:â•›428; note also though that “Malayan English” of four decades ago is a different animal from SgE today.) Later, other similar statements have been independently proposed or assumed (Ng 2008, 2009; Siraj 2008; Wee 2008a, b).

99

100

Lisa Lim

as well as attitudes towards the participants and/or their languages; and the quality and quantity of communicative events. The two main external factors impinging on the ecology in which SgE has emerged, namely migration patterns and language policies, which have both had effects on aspects such as population proportions and dominance of languages over different eras, are explored in other work (see Lim 2010a). Internal ecology involves the nature of the linguistic input elements, surface similarities and typological degrees of relatedness between the languages involved, and it is this aspect of the ecological approach that is relevant to the investigation undertaken in this paper. This paper is structured as follows. In Section╯2, I first provide a brief illustration of various aspects of the prosody of one Asian English: SgE. Rather than attempt an analysis which tries to find equivalences to native Englishes — e.g. describing the frequencies of various intonation contour types or tones, or noting the placement of tonic syllable or stresses, always compared to patterns in British English (BrE) — I aim to examine those patterns which are particular to SgE. In Section╯3, I consider an explanation for why such patterns manifest in SgE by appealing to the internal ecology in which SgE has evolved, specifically examining the typology of the substrate languages, addressing in particular the idea of the presence of (Sinitic) tone. In Section╯4, I then go on to consider the implications for understanding the prosodic patterns of other Asian Englishes — but not simply because these varieties belong to the same regional group, but rather because it so happens that there are elements of their ecologies, viz. the typologies of their substrate languages, which share common traits which contribute to such prosodic patterns. I also consider the implications that such findings hold for work on English intonation, as well as for theoretical classifications of intonation and tone languages. In this regard, what I argue for is a revisiting of English prosody, in a more radical reconsideration of the prosody of some Asian Englishes, in terms of viewing such varieties not as stress or intonation languages but as tone languages. 2. Like Chinese mε55?3 SgE has often been anecdotally described as if it “sounds like Chinese” (Bloom 1986:â•›430, citing Killingey 1968). This intriguing observation has not, however, until recently, garnered serious investigation based on an examination of substrate typology as to why this may be the case. This section presents a summary overview 3.╇ The SgE particle mε55 indicates (often mock-) surprise or incredulity for the proposition it is attached to; also see example (5a). Here the translation of the section heading would be ‘Does it really sound like Chinese?’, in response to the description in the first sentence of this section.



Revisiting English prosody

of the evidence derived from a number of fronts: discourse particles, words, and utterances. 2.1 Discourse particles Perhaps the most obvious presence of tone in SgE is that found in the discourse particles. Again, although the particles have long been acknowledged in most scholarship as coming from the (southern) Chinese languages (e.g. Platt 1987; Gupta 1992; and see Lim 2007a for a comprehensive overview), and the question of whether the particles themselves carry (lexical) tone was posed early in scholarship on the particles (Platt 1987), it is only in very recent work that these two issues have been seriously addressed (Lim 2007a, b, 2008b, 2010a). It has been argued that the larger set of SgE particles, namely hor, leh, lor, ma, meh, have their origins in Cantonese, and were acquired into SgE in a later era compared to the earlier particles lah, ah and what, and have carried into SgE (their original) Sinitic tone (Lim 2007a). Here we focus on this Cantonese set, examples of which are provided below in (1a) to (1e); note that the particles are transcribed together with their tone, represented as pitch level numbers, a practice proposed in Lim (2007a).4 These are accompanied in each case by an example of the corresponding particle in Cantonese from which it derives (from Matthews and Yip 1994:â•›347, 348, 352).5 A comparison of the SgE and Cantonese particles in the (a) and (b) pairs reveals striking parallels in segmental form, tone and meaning. In (1a), for example, the SgE particle hɔ24, which always occurs with a rising tone, and which asserts a proposition, making clear a positive response from the addressee is expected (Wee 2004:â•›124; Lim 2007a), is matched in Cantonese in its hó particle, with the same rising tone and indicating an expectation of the addressee’s confirmation (Matthews and Yip 1994:â•›347). The SgE particle lɔ33 in (3a), which occurs with midlevel tone, and which indicates obviousness, and in negative contexts inevitability or resignation (Wee 2004:â•›123; Lim 2007a), is similarly matched by a Cantonese 4.╇ SgE data for particles and utterances derive from the naturally occurring data of the Grammar of Spoken Singapore English Corpus (GSSEC; Lim and Foley 2004), except (4a), from Wong (1994). The tones of the particles are represented as pitch level numbers 1 to 5 where, in the Asianist tradition, the larger the number the higher the pitch; thus 55 represents a high level tone, 24 represents a rising tone, and so on. 5.╇ The transcription of examples (1b) to (5b) is provided as in the source (Matthews and Yip 1994), which uses the Yale system. Rising and falling tones are shown by rising and falling accents; high level tone is indicated by a level accent; no tonal indication is given for the mid level tone, and is inserted after the vowel to indicate all low-register tones (low rising, low level and low falling).

101

102

Lisa Lim

particle lo in (3b) with mid-level tone and meaning of resignation (Matthews and Yip 1994:â•›352).

(1) a. A: But it’s beautiful in that… how… I mean, Finn got a chance to realize himself, right? B: He’s quite innocent la21 hɔ24? Innocent. ‘He’s quite innocent, don’t you agree?’ [asserting proposition, expecting agreement] b. A: Géi leng a hó? quite nice PRT PRT ‘Pretty nice, huh?’ [expecting confirmation] B: Haih a. is PRT ‘Yes, it is.’ (2) a. A: My parents will disown me a22 if I marry someone Caucasian or Indian. My parents very what. ‘My parents will disown me if I marry someone Caucasian or Indian. My parents are really impossible.’ B: *** very old fashion a21. A: My parents very old fashion a21? Then your parents le55? ‘Are you saying that my parents are old fashioned? Then what about your parents?’ [indicating comparison, ‘what about?’] b. A: Dī gāsī maaih saai béi yàhn la. CL furniture sell all to people PRT ‘The furniture has all been sold.’ B: Ga chē lē? CL car PRT ‘What about the car?’ [meaning ‘what about’?]



(3) a. A: But um I might stop working for a while if I need to, if I need to la21, especially for looking after kids. B: But for me, I won’t stop working lɔ33. The most I won’t give birth to kids lɔ33. For the most I don’t marry lɔ33. ‘In my case, (even if I have children to look after) I won’t stop working. In the worst of cases, I won’t have children. In the worst of cases, I won’t get married.’ [indicating obviousness, resignation]



b.

Revisiting English prosody

Ngóh mjī dím syun lo I not-know how act PRT ‘I really don’t know what to do.’ [indicating resignation]

(4) a. A: How come you call me? ‘Why did you call me?’ B: You page for me ma22. ‘You paged for me, after all (as you know) (so naturally I’m returning your call).’ [indicating obviousness] b. A: Bīngo lèih ga? who PRT ‘Who’s that?’ B: Ngóhdeih sān lóuhbáan āma. our new boss PRT ‘Our new boss, of course.’ [indicating obviousness] (5) a. A: No la21! He’s using Pirelli, you don’t know mε55? ‘No, he has Pirelli tyres; didn’t you know that?’ [indicating surprise, scepticism] B: Really? Don’t bluff. b. sīnsāang wah mh dāk ge mē? teacher say not okay PRT PRT ‘What, did the teacher say it wasn’t okay?’ [expressing surprise]

As evidence for the claim that the particles occur in SgE with their original (Cantonese) tone, Figure╯1 provides an illustration of the pitch contour of one SgE particle, from the utterance maybe it like what you say lɔ33, with the particle lɔ33 visualized clearly as being realized with level tone; instrumental analysis confirms that the particle lies in the middle of speakers’ pitch range, and hence can be categorized as mid-level. By virtue of the tones originally from Cantonese — comprising mid-level and high-level tones as well as rising and falling tones — the particles introduce very particular pitch levels or contours into the SgE prosodic system. As the particles occur in phrase-final position, this means that there are thus very particular boundary tones in SgE.

103

Lisa Lim 500

Pitch (Hz)

104

0 0.901855

1.71591

Time (s)

Figure╯1.╇ Mid-level tone of SgE particle lɔ33

2.2 Word level At the level of the word, some very recent work, all of which assuming at the outset that SgE “has tones” (Wee 2008b), has suggested that SgE has tone in addition to stress, with tone being predictable from stress (Ng 2008), with a high level tone being assigned to the final syllable (Ng 2008; Wee 2008a), as can be seen in the words in example (6) (from Ng 2008; Wee 2008a, b).6 (6)

cat, see `manage, `teacher in`tend, a`round `Singapore, `managing `origin, bi`lingual o`riginal, se`curity o`riginally

55 / H 33–55 / MH 11–55 / LH 33–33–55 / MMH 11–33–55 / LMH 11–33–33–55 / LMMH 11–33–33–33–55 / LMMMH

H M

M

H

M

H

Figure╯2.╇ MH tones in SgE word `normal, in sentence-initial, -medial and -final position (from Ng 2009) 6.╇ The tones on each syllable in example (6) are represented in pitch level numbers as well as in the phonological tradition where L = Low tone, M = Mid tone, and H = High tone.



Revisiting English prosody

This word-level tonal pattern is also shown to be independent of sentence position (Ng 2009), as illustrated in Figure╯2. 2.3 Phrase / Utterance Finally, let us consider SgE prosody at phrase level, where, echoing the observation in Lim (2004a:â•›42–8), a characteristic pattern in the intonation contour may be analysed as comprising sequences of sustained level steps or level tones which step up or down to each other, rather than glide more gradually from one pitch level to another. An illustration of such a pattern is provided in Figure╯3, which constitutes the intonation on the utterance I think happier, where it is evident that the pitch steps up abruptly to a high level pitch for think, and then steps down again for



|╅╛╛ai╅╛╛|â•…â•…â•…â•…â•… tiŋâ•…â•…â•…â•…â•… | â•…â•…â•…â•…â•…â•…â•…â•… hæpiəâ•…â•…â•…â•…â•…â•…â•…â•… |



|â•…â•…  juâ•…â•…   |â•…â•…â•…â•…â•…â•…â•…â•… tol ╅╅╅╅╅╅╅╅╇ |╅╅╅╇ miâ•…â•…â•…â•… |

Figure╯3.╇ Sustained level step pattern in SgE utterance I think happier

Figure╯4.╇ Sustained level step pattern in SgE utterance You told me

105

106

Lisa Lim

happier. Similarly in Figure╯4, the utterance You told me moves in a series of sustained level tones, each of which is at a slightly higher pitch than the previous one.7 3. But is it tone? Having now viewed the evidence for a tone-like presence in SgE at the levels of particles, words and utterances, a number of issues need to be addressed. For the more sceptical amongst us, the first question is what the likelihood is that this is indeed tone that is observed in SgE. A number of reasons come together to confirm that this is in fact a more than credible prospect. First, suprasegmental features, including tone, are documented as being susceptible to being acquired in contact situations (Curnow 2001). Tone is often acquired in a non-tonal language by borrowing or imitation due to the presence of tone in the broader linguistic environment (Gussenhoven 2004:â•›42–3), such as in Middle Korean due to the prestigious status of Chinese in society then (Ramsay 2001) — here it should be noted how this is related to dominance in the external ecology. As a result of this, tone has been noted to be an areal feature, occurring in genetically unrelated languages spoken by geographically contiguous speech communities, as in Africa and South East Asia (Nettle 1998; Svantesson 2001). The next point to recognize is that the linguistic feature of tone is certainly present in the ecology of Singapore. If we consider just the main players in the ecology — by this I mean the languages which are dominant — we have Bazaar Malay, Hokkien, Mandarin, and Cantonese (Lim 2007a, 2010a; Ansaldo 2009a, b, 2010), where, with the latter three being Sinitic varieties, tone languages are clearly in the majority. Tone is thus a salient typological aspect of the feature pool. Other work has shown that dominant traits do influence the output (Thomason and Kaufman 1988), as seen, for example, in the case of Sri Lanka Malay (SLM) where, while pidgin-derived Malay is SVO, Sinhala and Tamil are both SOV, and the resulting Sri Lanka Malay is SOV; similarly, agglutinative morphology emerges in SLM because it is salient in two of the three adstrates, Sinhala and Tamil (Ansaldo 2008, 2009b). Additionally, if we consider external ecology, in Singapore it is the Chinese who form the largest proportion of 78% of the population, and have been such a majority since the early 20th century (Lim 2007a, 2010a). On both counts then, namely, of the proportion of tone languages, and the proportion of speakers of these languages, tone dominates in the ecology. Moreover, tone is high 7.╇ SgE has been described as displaying a frequent use of (BrE) level tones (Goh 1998), but in such a description the pattern seems to be construed in analytical terminology for native Englishes, as mentioned in Section╯1.



Revisiting English prosody

in markedness, in terms of the feature having a heavy functional load, or, put in terms of Matras’ (2000) model of categorial fusion, it is pragmatically dominant, which also makes it a more likely target for being acquired (Matras 2000:â•›577). In all, it is most possible for tone to be acquired in SgE, given the feature’s dominant presence in the ecology, both internal and external. Finally, observations similar to the SgE case study have been made for other contact varieties whose substrates involve tone languages, in particular languages arising from contact situations involving European accent languages and African tone languages. One such example which has been argued strongly for in this regard is that of Saramaccan, an Atlantic maroon creole spoken mostly in Surinam, generally classified as an English-based creole, though its lexicon shows substantial Portuguese influence, with Gbe and Kikongo as substrates. There is evidence for a split lexicon in Saramaccan where the majority of its words are marked for pitch accent, with an important minority marked for true tone (Good 2004a, b, 2006). Similarly, Portuguese-lexifier Papiamentu shows use of both contrastive stress and contrastive tonal features which operate independently from stress (Kouwenberg 2004; Rivera-Castillo and Pickering 2004; Remijsen and van Heuven 2005), and the Austronesian language Ma’ya is also documented as a hybrid system involving both contrastive stress and tone, a result of contact with tonal Papuan languages (Remijsen 2001:â•›43). Lest one may argue that “creoles” are categorically different from “varieties of English” or of any other “language”, there is also the example of Roermond Dutch, which has a Germanic-style stress system, but also a lexical tonal contrast, in that words may have no tones or a single H tone (Yip 2002:â•›257), as well as Nigerian English, whose prosody is also suggested to be a mixed system that stands “between” an intonationâ•›/â•›stress language and a tone language (Gut 2005): its pitch inventory is reduced compared to BrE, and the domain of pitch appears to be the word, with high pitch triggered by stress, thus resembling a pitch accent language. In short, what is being suggested for SgE is nothing inconceivable, but a phenomenon that has been documented in other contact languages with substrate typology that is comparable in terms of involving tone languages. 4. Asian Englishes on the agenda The more significant question, I believe, that follows from all the above is this: What does this herald for the linguistics of English and the setting of the agenda for current and future work?8 In what follows I consider two sets of issues arising 8.╇ This question is posed in the spirit of the 1st conference for the International Society for the Linguistics of English (ISLE1), with its theme “Setting the Agenda”; the workshop “The

107

108

Lisa Lim

from our observations of tone in SgE, which provide a response to the three concerns raised in Section╯1, with the first being more theoretical in considering the implications for descriptive models of English intonation and for categorizations of intonationâ•›/â•›stress and tone languages, and the second focusing on implications for the study of Asian Englishes. 4.1 English intonation models and tone language classification First, in the consideration of New Englishes — here, Asian (but also e.g. African) Englishes — I believe that we can attempt to stop using BrE- or “native-English”based models. Certainly there are models such as ToBI and INTSINT which have been used for diverse languages; however when it comes to New Englishes, even if scholars have noted the problems in applying, e.g. the British model of intonation to the transcription of New Englishes (e.g. for SgE: Deterding 1994; Low and Brown 2005), and have suggested that such models are not appropriate for describing other Englishes, there is still a tendency to describe the system with regard to a native English standard. Further, and perhaps more interestingly, the traditional view of English as a stress/intonation language needs revising. As is increasingly recognized, distinguishing between so-called stress languages, accent(ual) languages and tone languages is in fact not clear-cut. While traditionalists may still cling to the classic divide, many cutting-edge scholars have become more amenable to regarding these categories as being more loosely or broadly defined. For instance, most now agree that the category of accent languages does not group languages of a typologically coherent class (Hyman 2001a; Gussenhoven 2004), and take the position that the so-called accentual languages are just a subclass of tone languages (Yip 2002:â•›4). More significant for the purpose of this paper, tone languages are most recently defined much more broadly than before: following Hyman (2001b:â•›1368), “a language with tone is one in which an indication of pitch enters into the lexical realisation of at least some morphemes”, regardless of the density of lexically contrastive tones on words; lexical tonal marking, after all, has been noted to be of gradient nature (e.g. van der Hulst and Smith 1988). This opens up possibilities for more fluid considerations of tone languages, and of “combinations” of characteristics of what traditionally are considered stress languages and tone languages. It has, after all, been noted that tone and stress are “two separate phonological dimensions — tone being basically a property of segments, and stress not — which may well occur combined in the same language in quite a variety of ways” (Arends, Typology of Asian Englishes” in which this paper was presented was organized (by this author) for the conference.



Revisiting English prosody

Muysken and Smith 1995:â•›329); and that drawing a dividing line between languages with contrastive tone on (almost) all syllables and languages with tone contrasts in more restricted locations in the word is difficult (Gussenhoven 2004:â•›47). In the light of the possibilities outlined in Section╯3 and released from the constraints of more traditional categorization noted above, how then may we view the prosodic system of SgE? For the set of particles, we first return to the question first explicitly raised by Platt (1987:â•›394) two decades ago: Do the particles have independent (lexical) tone? Based on the linguistic and sociohistorical evidence presented in this paper and in Lim (2007a, 2010a), it is quite clear that the later set of SgE particles have been acquired into the SgE system in their entirety, including the tone they have originally in Cantonese; further, they must be used with that form, and not with any other pitch pattern, for the meaning required, regardless of the intonation pattern of the utterance in which they are found. This, I argue, is reason enough for accepting that they indeed comprise a subset of items which have tone (though not lexically contrastive tone — note, as outlined in the preceding paragraph, that this is not necessary for considering this to be tone).9 These tonal items are then situated within what is possibly a different prosodic system — one that is more of a stressâ•›/â•›intonation language, in which pitch functions in a system of intonation relatively comparable to the forms and functions identified in other “standard” varieties of English such as BrE (Zhu and Lim 2002; Zhu 2003; Lim 2004a:â•›39–42). Such a phenomenon is noted by Gussenhoven (2004:â•›46) as one of three typologically special cases where tone languages are concerned, namely when there is lexically specified tone in intonation-only languages. An example of this situation is when there are tonal specifications in the “segmental” lexicon for particles which invariably appear with a particular intonation contour, such as Dutch sentencefinal [hɛ], which expresses an appeal for agreement, which always appears with H after the pitch accent H*L on a preceding word (Kirsner and van Heuven 1996); similarly Bengali has focus-governing particles which come with their own pitch accent (Lahiri and Fitzpatrick-Cole 1999), i.e. they must be lexically specified for tones, which crucially constitute morphemes in their own right and do not form part of the representation of the segmentally represented morphemes, unlike lexical tone (Gussenhoven 2004:â•›46). These both constitute situations not unlike our case in point. The observation of tone at the SgE word level matches the second of the typologically special cases identified by Gussenhoven (2004:â•›45–6) in which languages 9.╇ Some scholars may still be cautious about whether pitch is being used linguistically here to warrant calling the phenomenon tone (e.g. Bao Zhiming p.c. Dec 2006), but I hold that the evidence and arguments outlined in sections 3 and 4.1 justify referring to it as tone.

109

110

Lisa Lim

have non-distinctive word-based tone; an example of this is Noon, a language of Senegal, which predictably has an H-tone on the penultimate syllable of every word (Soukka 2000). Again, the SgE word which specifies an H-tone on the final syllable of each word parallels this pattern, and thus can be considered a tone language in this regard, even if typologically peculiar. 4.2 The typology of Asian Englishes The preceding discussion also leads us to a number of thoughts about the study of Asian Englishes, which can take us in at least three different, though closely related, directions. In what follows, I consider 1) typological comparability across different Asian Englishes, 2) the dynamic nature of ecologies, and 3) the differences that can arise between Asian Englishes as a result of the changing of ecologies, and therefore typologies, and therefore feature selection. Each of these can — and should — be discussed at far greater length, but within the constraints of this paper can only be very briefly addressed here. 4.2.1 Typological comparability If we accept the ecological approach, then we can recognize that, in theory, if emerging English varieties have ecologies which are similar — for example, if the feature pools for the two ecologies include languages with similar typologies — then it is neither unlikely nor surprising if the two emergent Englishes exhibit linguistic features which are similar.10 One does not have to look far to find a reallife example: the ecology of Hong Kong, in which Hong Kong English (HKE) is emerging, includes Cantonese, English and Mandarin, where Cantonese is clearly dominant in both external and internal ecologies; linguistic features such as Sinitic tone (from both Cantonese and Mandarin) as well as discourse particles (in particular from Cantonese) can thus be expected to be dominant in the feature pool. This may be seen to be quite similar to Singapore’s ecology in the recent era where Mandarin and Cantonese can be seen to be more dominant. Unsurprisingly then, common features in the emergent varieties may be identified. Cantonese particles have been observed in HKE (James 2001), which, as in SgE, are used with their Cantonese tone. Further, already more than two 10.╇ We need of course to exercise caution in too easily assuming similarity between different ecologies: this is underlined in Ansaldo and Lim (2008), who, after noting possible similarities in the ecologies of SgE and HKE, also point out how they differ, both in terms of external ecology — Singapore being English-dominant but Hong Kong being extremely Cantonese-dominant; SgE having attained Phase 4 of endonormative stabilization, and HKE being only in Phase 3 of nativization (Schneider 2007) — and internal ecology.



Revisiting English prosody

decades ago, it was suggested that “the English intonation system is reinterpreted on the basis of the Cantonese tone system” (Luke and Richards 1982:â•›60), with more recent work demonstrating that stress and intonation in HKE is more than a simple process of approximation, but rather one of transforming the system into one based on tones, by the assigning of (lexical) tones to syllables for word stress (Luke 2000), to give a basic template of LHL!, and the applying of a computation for compound/linking words, phrases and sentences, which results in what is described as a “choppy” as opposed to a smooth intonation contour for HKE (Luke 2008). The patterns parallel those of SgE’s level stepping tones, described in Section╯2.3. 4.2.2 Dynamism The comparability sketched above between SgE and HKE should not, however, be assumed to have always been the case. Ecologies are dynamic — as modelled for Postcolonial Englishes in Schneider (2007), and noted for SgE in Lim (2008a) — and changes in factors in the external ecology can result in changes in the typological mix of the feature pool and consequently the ensuing competition and likelihood for feature selection (see Ansaldo 2009a). While Sinitic varieties are certainly dominant in Singapore’s ecology in recent decades, in an earlier era during colonial rule and in the years prior to and just after independence, the dominant language would have been Bazaar Malay, as lingua franca for interethnic communication (as well as Hokkien for intraethnic communication in the Chinese community) (see Lim 2007a, 2010a for details). In this, and other aspects of sociopolitical history, Singapore’s ecology was relatively comparable up until around mid-20th century with Malaysia’s, and certainly not with Hong Kong’s; correspondingly, SgE and Malaysian English were also considered extremely similar, evidenced in descriptions of that period (e.g. Platt and Weber 1980; Platt, Weber and Ho 1983), at least until around the same time, but perhaps no longer. 4.2.3 Difference The preceding two points of typological comparability between varieties and dynamism in ecologies together help account for instances of lack of fit. Even as we recognize that SgE and HKE exhibit similarities in terms of the presence of (Cantonese) particles and tonal prosody, which are explained by their current comparable ecologies in which Cantonese amongst other (Sinitic) varieties is dominant, we discover that upon closer examination the patterns start to diverge. In HKE, H tones are located on stressed syllables, and L tones on unstressed ones (Luke 2000, 2008; Chen and Au 2004; Wee 2008a), illustrated in example (7); this contrasts with the pattern for word-level tone in SgE, described in example (6), where H tones are located on the final syllable.

111

112

Lisa Lim

(7) in`tend `origin, `photograph o`riginal

(8) I saw the manager this morning

11–55 / LH 55–11–11 / HLL 11–55–11–11 / LHLL LHHHHHHHL!

Similarly, at phrase level, while HKE would have a pattern involving a sequence of tones as in (8), based on the basic LHL! template and subsequent computation (Luke 2008),11 SgE tends to prefer prominence on the phrase-final syllable such that the pitch is perceived as relatively high: no significant decrease in fundamental frequency is measured compared to the initial syllable of the phrase-final word (Low 2000); such a maintenance of pitch or movement to high(er) tone phrasefinally is also observable in Figure╯4.12 How can we understand these diverging patterns in SgE and HKE? I suggest that the word- and phrase-final prominence noted in SgE is due to the influence from Bazaar or Baba Malay in an earlier era when, as mentioned in the preceding section, it was dominant in the ecology (also see Lim 2008a). While no comprehensive study of the prosody of (Bazaar) Malay in Singapore is available,13 there has been much research on other Malayâ•›/â•›Indonesian varieties (see e.g. various chapters in Gensler and Gil to appear).14 While findings for word stress are diverse, a number of studies do point to prominence on the penultimate and/or final syllable; and at phrase level there is general consensus that prominence is located phrase-finally (with acceptability increasing closer to the right edge of phrase-final word) (e.g. Goedemans and van Zanten to appear); a similar pattern is also noted in Singapore’s Baba Malay (Wee 2000). While it may indeed seem curious that this earlier Malay influence appears to be maintained in spite of more recent Sinitic influence, in the ecology paradigm, the Founder principle suggests that the founder population in an ecology exerts a strong influence in features which persist in the emergent variety. A possible hypothesis is thus that it is the Babaâ•›/â•›Bazaar Malayspeaking Peranakans, as the early English speakers in Singapore, whose influence 11.╇ Different boundary tones of H% or L% would then apply depending on the context (Luke 2008). 12.╇ Experiments investigating emphatic and contrastive stress in SgE also demonstrate that speakers do not place prominence on the contrastive element as in “standard” Englishes but systematically locate pitch prominence utterance-finally (Lim and Tan 1999; Lim 2004b). 13.╇ Though Ng (2009) is now doing instrumental work on Singapore Malay and Bazaar Malay word prosodic patterns. 14.╇ What is represented in this paragraph is necessarily an extremely summarized account of Malayâ•›/â•›Indonesian prosodic patterns documented in the literature.



Revisiting English prosody

is seen in SgE in such instances (Lim 2011; see Ansaldo, Lim and Mufwene 2007 and Lim 2010b for details on the Peranakan community and their languages Baba Malay and Peranakan English);15 similar word-â•›/â•›phrase-final prominence is in fact noted in Peranakan English (Lim 2010b).16 5. Closing remarks In conclusion, what has a revisiting of English prosody — in the light of some Asian Englishes — revealed? We can certainly recognize that there is tone in some Asian Englishes, if the typological pool of the languages in contact allows for it. Tone is clearly observed in SgE particles and in the SgE word, and SgE intonation shows contours constituting sustained level tone movements; similar observations apply to HKE. While tone as a structural feature must surely derive from the Sinitic languages in the ecology of SgE, the clearest example being the SgE particles, differences in actual tonal patterns in SgE vs. HKE at both word and phrase level suggest that the story is more complex and warrants further, deeper investigation in future work: for instance, that in SgE it is not just the selection of Sinitic tone (Lim 2008c) from the feature pool, but an interaction between that and features of pitch prominence patterns in other languages dominant in the ecology, such as Babaâ•›/â•›Bazaar Malay. In any case, a revised view of English prosody, even in a weak version, must still recognize that there can be presence of tone in some New Englishes (here Asian, but also African), even if English is by traditional definition a stress language. In a strong version, we can venture further in rethinking this traditional classification of English prosody and accept that some Asian Englishes can indeed be tone languages. In short, once we recognize the significance of ecology, then it is not difficult to recognize that anything is possible for the typology of the emergent contact variety of English — not a cavalier or anarchic free-for-all, but an infinite potential within the possibilities that the typologies of the substrates afford — including tone.

15.╇ I thank Umberto Ansaldo and Salikoko Mufwene for highlighting the likelihood of the Peranakans as the founder population for SgE. 16.╇ This line of investigation outlined in this section is certainly intriguing and far more complex than can be presented given the scope of this present paper, and is taken up in other work (Lim 2011).

113

114

Lisa Lim

References Ansaldo, Umberto. 2008. “Revisiting Sri Lanka Malay: Genesis and classification”. In K. David Harrison, David Rood and Arianne Dwyer, eds. A World of Many Voices: Lessons from Documenting Endangered Languages. (Typological Studies in Language 78.) Amsterdam, Philadelphia: Benjamins, 13–42. ———. 2009a. “The Asian typology of English: Theoretical and methodological considerations”. In Lisa Lim and Nikolas Gisborne, eds. The Typology of Asian Englishes. Special Issue of English World-Wide 30(2):â•›133–48. ———. 2009b. Contact Languages: Ecology and Evolution in Asia. (Cambridge Approaches to Language Contact.) Cambridge: Cambridge University Press. ———. 2010. “Contact and Asian Varieties of English”. In Raymond Hickey, ed. The Handbook of Language Contact. Oxford: Blackwell, 498–517. ———, Lisa Lim and Salikoko S. Mufwene. 2007. “The sociolinguistic history of the Peranakans: What it tells us about ‘creolization’”. In Umberto Ansaldo, Stephen Matthews and Lisa Lim, eds. Deconstructing Creole. (Typological Studies in Language 73.) Amsterdam, Philadelphia: Benjamins, 203–26. ——— and Lisa Lim. 2008. “Converging Asian Englishes? Singapore English, Hong Kong English and the languages in their ecologies”. Paper presented at the 14th Conference of the International Association for World Englishes (IAWE). Hong Kong. Arends, Jacques, Pieter Muysken and Norval Smith. 1995. “Conclusions”. In Jacques Arends, Pieter Muysken and Norval Smith, eds. Pidgins and Creoles: An Introduction. Amsterdam, Philadelphia: Benjamins, 319–30. Baskaran, Loga. 2008. “Malaysian English: Phonology”. In Mesthrie, ed.: 278–91. Bloom, David. 1986. “The English language and Singapore: A critical survey”. In Basant A. Kapur, ed. Singapore Studies: Critical Surveys of the Humanities and Social Sciences. Singapore: Singapore University Press, 337–452. Brown, Adam, David Deterding and Low Ee Ling, eds. 2000. The English Language in Singapore: Research on Pronunciation. Singapore: Singapore Association of Applied Linguistics. Chen, Charles C. Jr. and Ching-Pong Au. 2004. “Tone assignment in second language prosodic learning”. Paper presented at Speech Prosody 2004. Nara, Japan. (accessed 5 Feb. 2009). Curnow, Timothy Jowan. 2001. “What language features can be ‘borrowed’?” In Alexandra Y. Aikhenvald and Robert M.W. Dixon, eds. Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. Oxford: Oxford University Press, 412–36. Deterding, David. 1994. “The intonation of Singapore English”. Journal of the International Phonetic Association 24: 61–72. ———. 2001. “The measurement of rhythm: A comparison of Singapore and British English”. Journal of Phonetics 29: 217–30. Gargesh, Ravinder. 2008. “Indian English: Phonology”. In Mesthrie, ed.: 231–43. Gensler, Orin and David Gil, eds. to appear. Malayâ•›/â•›Indonesian Linguistics. London: Curzon Press. Goedemans, Rob and Ellen van Zanten. to appear. In Gensler and Gil, eds.: 35–62. Goh, Christine C.M. 1998. “The level tone in Singapore English”. English Today 14(1): 50–3. Good, Jeff. 2004a. “Split prosody and creole simplicity: The case of Saramaccan”. Journal of Portuguese Linguistics 3: 11–30.



Revisiting English prosody

———. 2004b. “Tone and accent in Saramaccan: Charting a deep split in the phonology of a language”. Lingua 114: 575–619. ———. 2006. “A twice-mixed creole? Tracing the history of a prosodic split in the Saramaccan lexicon”. Ms, Max-Planck-Institute for Evolutionary Anthropology, Leipzig. Grabe, Esther. 2004. “Intonational variation in urban dialects of English spoken in the British Isles”. In Peter Gilles and Jörg Peters, eds. Regional Variation in Intonation. Tübingen: Niemeyer, 9–31. Gupta, Anthea Fraser. 1992. “The pragmatic particles of Singapore Colloquial English”. Journal of Pragmatics 18: 31–57. Gussenhoven, Carlos. 2004. The Phonology of Tone and Intonation. Cambridge: Cambridge University Press. Gut, Ulrike. 2005. “Nigerian English prosody”. English World-Wide 26: 153–77. van der Hulst, Harry and Norval Smith, eds. 1988. Autosegmental Studies on Pitch Accent. Dordrecht: Floris. Hyman, Larry M. 2001a. “Privative tone in Bantu”. In Kaji, ed.: 237–57. ———. 2001b. “Tone systems”. In Martin Haspelmath, Ekkehard König, Wulf Oesterreicher and Wolfgang Raible, eds. Language Typology and Language Universals: An International Handbook. Berlin, New York: Walter de Gruyter, 1367–80. James, Gregory. 2001. “Cantonese particles in Hong Kong students’ English emails”. English Today 17: 9–16. Kachru, Yamuna and Cecil L. Nelson. 2006. World Englishes in Asian Contexts. (Asian Englishes Today.) Hong Kong: Hong Kong University Press. Kaji, Shigeki, ed. 2001. Proceedings from the Symposium Cross-Linguistic Studies of Tonal Phenomena: Historical Development, Phonetics of Tone, and Descriptive Studies. Tokyo: Tokyo University of Foreign Studies, Institute for Language and Cultures of Asia and Africa. Kirsner, Robert S. and Vincent J. van Heuven. 1996. “Boundary tones and the semantics of the Dutch final particles hè, hoor, zeg, and joh”. In Crit Cremers and Marcel den Dikken, eds. Linguistics in the Netherlands 13: 133–45. Kouwenberg, Silvia. 2004. “The grammatical function of Papiamentu tone”. Journal of Portuguese Linguistics 3: 55–69. Lahiri, Aditi and Jennifer Fitzpatrick-Cole. 1999. “Emphatic clitics and focus intonation in Bengali”. In Wim Zonneveld and René Kager, eds. Phrasal Phonology. Nijmegen: Nijmegen University Press, 119–44. Lim, Lisa. 1997. “Prosodic patterns characterising Chinese, Malay and Indian varieties of Singapore English”. Ph.D. dissertation, University of Reading. ———. 2000. “Ethnic groups aligned? Intonation patterns of Chinese, Indian and Malay Singaporean English”. In Brown, Deterding and Low, eds.: 10–21. ———. 2001. “Ethnic group varieties of Singapore English: Melody or harmony?” In Vincent B.Y. Ooi, ed. Evolving Identities: The English Language in Singapore and Malaysia. Singapore: Times Academic Press, 53–68. ———. 2004a. “Sounding Singaporean”. In Lim, ed.: 19–56. ———. 2004b. “Everything you wanted to know about how stressed Singaporean Englishes are”. In Somsonge Burusphat, ed. Papers from the Eleventh Annual Meeting of the Southeast Asian Linguistics Society. Arizona: Program for Southeast Asian Studies, Arizona State University, 429–44.

115

116

Lisa Lim

———, ed. 2004c. Singapore English: A Grammatical Description. (Varieties of English Around the World G33.) Amsterdam, Philadelphia: Benjamins. ———. 2007a. “Mergers and acquisitions: On the ages and origins of Singapore English particles”. World Englishes 26: 446–73. ———. 2007b. “Singapore English is a tone language meh55? SE particles and the hybrid prosody of a contact variety of English”. Paper presented at the English Language Research Seminar Series, Department of Linguistics and English Language, University of Edinburgh. ———. 2008a. “Dynamic linguistic ecologies of Asian Englishes”. Asian Englishes 11(1): 52–5. ———. 2008b. “English can be tone language meh55? Singapore English what21! SE particles and the hybrid prosody of a contact variety of English”. Ms, University of Amsterdam. ———. 2008c. “Revisiting English prosody: Sinitic tone in Singapore English”. Paper presented at the Typology of Asian Englishes Workshop. 1st International Conference on the Linguistics of English (ISLE1). Freiburg, Germany. ———. 2009. “Not just an ‘Outer Circle’, ‘Asian’ English: Singapore English and the significance of ecology”. In Thomas Hoffmann and Lucia Siebers, eds. World Englishes: Problems, Properties, Prospects (Varieties of English Around the World G40). Amsterdam, Philadelphia: Benjamins. ———. 2010a. “Migrants and ‘mother tongues’: Extralinguistic forces in the ecology of English in Singapore”. In Lisa Lim, Anne Pakir and Lionel Wee, eds. English in Singapore: Modernity and Management (Asian Englishes Today). Hong Kong: Hong Kong University Press, 19–54. ———. 2010b. “Peranakan English in Singapore”. In Daniel Schreier, Peter Trudgill, Edgar W. Schneider and Jeffrey P. Williams, eds. The Lesser Known Varieties of English: An Introduction. (Studies in English Language.) Cambridge: Cambridge University Press, 327–47. ———. 2011. “Tone in Singlish: Substrate features from Sinitic and Malay”. In Claire Lefebvre, ed. Creoles, Their Substrates and Language Typology. (Typological Studies in Language.) Amsterdam/Philadelphia: Benjamins, 271–87. ——— and Joseph A. Foley. 2004. “English in Singapore and Singapore English: Background and methodology”. In Lim, ed.: 1–18. ——— and Tan Ying Ying. 2001. “How are we stressed?! Phonetic correlates and stress placement in Singaporean English”. In John A. Maidment and Eva Estebas i Vilaplana, eds. Proceedings of PTLC2001: Phonetics Teaching and Learning Conference 2001. London: University College London, 27–30. Low, Ee Ling. 1994. “Intonation patterns in Singapore English”. M.Phil. dissertation, University of Cambridge. ———. 1998. “Prosodic prominence in Singapore English”. Ph.D. dissertation, University of Cambridge. ———. 2000. “Is lexical stress placement different in Singapore English and British English?” In Brown, Deterding and Low, eds.: 22–34. ——— and Adam Brown. 2005. English in Singapore: An Introduction. Singapore: McGraw-Hill Education (Asia). ——— and Esther Grabe. 1999. “A contrastive study of prosody and lexical stress placement in Singapore English and British English”. Language and Speech 42: 39–56. ———, Esther Grabe and Francis Nolan. 2000. “Quantitative characterisations of speech rhythm: Syllable-timing in Singapore English”. Language and Speech 43: 377–401.



Revisiting English prosody

Luke, Kang-kwong. 2000. “Phonological re-interpretation: The assignment of Cantonese tones to English words”. Paper presented at the 9th International Conference on Chinese Linguistics, National University of Singapore. ———. 2008. “Stress and intonation in Hong Kong English”. Paper presented at the 14th Conference of the International Association for World Englishes (IAWE), Hong Kong. ——— and Jack Richards. 1982. “English in Hong Kong: Functions and status”. English WorldWide 3: 47–63. Matras, Yaron. 2000. “How predictable is contact-induced change in grammar?” In Colin Renfrew, April McMahon and Larry Trask, eds. Time Depth in Historical Linguistics. Vol. 2. (Papers in the Prehistory of Languages.) Oxford: The MacDonald Institute for Archaeological Research, 563–83. Matthews, Stephen J. and Virginia Yip. 1994. Cantonese: A Comprehensive Grammar. London: Routledge. Mesthrie, Rajend. 2008a. “Synopsis: The phonology of English in Africa and South and Southeast Asia”. In Mesthrie, ed.: 307–19. ———, ed. 2008b. Varieties of English 4: Africa, South and Southeast Asia. Berlin, New York: Mouton de Gruyter. Mufwene, Salikoko S. 2001. The Ecology of Language Evolution (Cambridge Approaches to Language Contact). Cambridge: Cambridge University Press. ———. 2008. Language Evolution: Contact, Competition and Change. New York: Continuum. Nettle, Daniel. 1998. Linguistic Diversity. Oxford: Oxford University Press. Ng, E-Ching. 2008. “Malay meets Chinese meets English: Where does colloquial Singaporean English word-level tone come from?” Paper presented at the Language Contact workshop, Bristol. ———. 2009. “Borrowing tones: Word-level tone in Colloquial Singaporean English”. Paper presented at the Society for Pidgin and Creole Languages (SPCL) conference, San Francisco. Platt, John. 1987. “Communicative functions of particles in Singapore English”. In Ross Steele and Terry Threadgold, eds. Language Topics: Essays in Honour of Michael Halliday. Vol. 1. Amsterdam, Philadelphia: Benjamins, 391–401. ——— and Heidi Weber. 1980. English in Singapore and Malaysia: Status, Features, Functions. Kuala Lumpur: Oxford University Press. ———, Heidi Weber and Ho Mian Lian. 1983. Singapore and Malaysia. (Varieties of English Around the World T4.) Amsterdam, Philadelphia: Benjamins. Ramsay, Robert. 2001. “Tonogenesis in Korean”. In Kaji, ed.: 3–17. Remijsen, Bert. 2001. Word Prosodic Systems of Raja Ampat Languages. Utrecht: LOT. ——— and Vincent J. van Heuven. 2005. “Stress, tone and discourse prominence in Curaçao Papiamentu”. Ms, Leiden University. Rivera-Castillo, Yolanda and Lucy Pickering. 2004. “Phonetic correlates of stress and tone in a mixed system”. Journal of Pidgin and Creole Languages 19: 261–84. Schneider, Edgar W. 2007. Postcolonial English: Varieties Around the World. (Cambridge Approaches to Language Contact). Cambridge: Cambridge University Press. ———, Kate Burridge, Bernd Kortmann, Rajend Mesthrie and Clive Upton, eds. 2004. A Handbook of Varieties of English. Vol. 1: Phonology. Berlin: Mouton de Gruyter. Siraj, Pasha. 2008. “Intonation in Singaporean English is stress dependent”. Paper presented at the 3rd conference on Tone and Intonation (TIE3), Lisbon.

117

118

Lisa Lim

Soukka, Maria. 2000. A Descriptive Grammar of Noon: A Cangin Language of Senegal. Lincom Europa. Svantessen, Jan-Olof. 2001. “Tonogenesis in South East Asia: Mon-Khmer and beyond”. In Kaji, ed.: 45–58. Tan, Ying Ying. 2003. “Acoustic and perceptual properties of stress in the ethnic subvarieties of Singapore English”. Ph.D. dissertation, National University of Singapore. Tayao, Ma. Lourdes G. 2008. “Philippine English: Phonology”. In Mesthrie, ed.: 292–306. Thomason, Sarah G. and Terrence Kaufman. 1988. Language Contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Wee, Kim Soon Gabriel. 2000. “Intonation of the Babas: An auditory and instrumental approach”. B.A. Honours thesis, National University of Singapore. Wee, Lian Hee. 2008a. “Phonological patterns in the Englishes of Singapore and Hong Kong”. World Englishes 27(3/4): 480–501. ———. 2008b. “Notes on SgE tone”. Ms, Hong Kong Baptist University. Wee, Lionel. 2004. “Reduplication and discourse particles”. In Lim, ed.: 105–26. Wells, John C. 1982. Accents of English. Vol. 3: Beyond the British Isles. Cambridge: Cambridge University Press. Wong, Jock Onn. 1994. “A Wierzbickan approach to Singlish particles”. M.A. dissertation, National University of Singapore. Yip, Moira. 2002. Tone. Cambridge: Cambridge University Press. Zhu, Shenfa. 2003. “Intonation in Singapore English: An auditory and acoustic analysis of four sentence types”. Ph.D. dissertation, National University of Singapore. ——— and Lisa Lim. 2002. “Intonation in Singapore English declaratives: An auditory and acoustic analysis”. Paper presented at the 13th World Congress of Applied Linguistics (AILA2002), Singapore.

Index A Angloversal╇ 2, 28, 29, 45, 50, 52, 64 Asian Englishes╇ 1–8, 11, 51, 75, 97–100, 107, 108, 110, 113, 116 Austronesian╇ 4, 27, 28, 107 auxiliary be╇ 68 B Bazaar Malay╇ 5, 8, 16, 17, 54, 65, 79, 106, 111–113 C Cantonese╇ 4, 5, 17, 18, 27–32, 35, 37, 38, 39, 41, 44, 45, 54–56, 65, 66, 68, 81, 92, 97, 101, 103, 106, 109, 110, 111 Chinese╇ 8, 15, 16, 22, 28–32, 35–37, 42, 44, 45, 54–56, 58, 62, 63, 65–70, 79, 100, 101, 106, 111 Chui Chow╇ 28 competition╇ 6, 13, 14, 22, 23, 111 congruence╇ 8, 11, 12, 14, 18, 22 contact languages╇ 11, 16, 17, 79, 97, 107 contrast╇ 29, 33–39, 41–45, 62, 76, 80, 81, 84, 88, 89, 92, 93, 107–109, 112 copula╇ 6, 17–22, 33, 39, 42–44, 49–53, 56, 67–70 copula be╇ 68 creole╇ 12, 17, 50, 68–70, 79, 107 creolization╇ 45

D dynamic model╇ 5, 6, 8, 40, 44, 53 E ecology╇ 3–5, 8, 11, 13–17, 19–23, 27, 28, 30, 40, 44, 45, 97, 99, 100, 106, 107, 110–113 ecological╇ 12, 13, 21, 28, 29, 99, 100, 110 English vowels╇ 75, 89, 90, 93 evolution/evolutionary╇ 2, 3, 5, 6, 11, 12, 13, 15, 16, 21, 22, 23, 53 F feature pool╇ 27, 28, 29, 33, 44, 45, 99, 106, 110, 111, 113 finiteness╇ 6, 27, 29, 32–38, 40–45 founder principle╇ 112 frequency╇ 8, 11, 13, 14, 22, 23, 45, 112 H Hindi╇ 4, 8, 15, 52, 53, 55, 56, 58, 62–69, 71 Hokkien╇ 4, 8, 16, 17, 28, 54, 55, 56, 65, 66, 92, 106, 111 Hong Kong English (HKE)╇ 5, 6, 7, 27–30, 32, 33, 35, 38–41, 43–45, 75, 76, 97, 110–113

inner Circle╇ 1, 4, 77 intonation╇ 6, 7, 97, 98, 99, 100, 105, 107, 108, 109, 111, 113 isolating languages╇ 21, 23 L language transfer╇ 49, 73 length╇ 76, 78, 80, 81, 83, 84, 86, 88, 90, 93 M Malay╇ 3, 4, 5, 8, 11, 15–23, 54, 55, 56, 58, 65, 68, 69, 70, 79, 81, 92, 98, 99, 106, 111, 112, 113 Mandarin╇ 4, 5, 17, 19, 21, 28, 30, 32, 35, 54–57, 65–68, 92, 106, 110 monophthongs╇ 75, 76, 80, 81, 86–89, 91, 92 multilingualism╇ 4, 11, 16 N New Englishes╇ 1, 2, 7, 29, 49, 50, 51, 52, 68, 69, 70, 75–78, 80, 81, 86, 92, 94, 97–99, 108, 113 Non-native English╇ 92 O Outer Circle╇ 1, 5, 7, 27, 77

P I particles╇ 5, 6, 23, 36, 97, 101, imperfective (aspect)╇ 6, 8, 103, 106, 109, 110, 111, 113 49, 53, 55–60, 62–67, 71 Indian English╇ 6, 7, 49, 50, 53, past tense╇ 6, 38, 39, 49, 50, 51, 53, 55–60 70, 71 55, 59, 81, 98

120

Index

Peranakan╇ 3, 16, 21, 112, 113 predicative adjectives╇ 6, 17–21 present tense╇ 33, 38, 39 progressive (aspect)╇ 6, 8, 49–53, 55–67, 69–71 prosody, prosodic╇ 55, 66, 77, 78, 86, 97–100, 103, 105, 107, 109, 111–113 Putonghua╇ 30, 32 R rhythm╇ 6, 7, 75–80, 82–86, 93, 94, 98 S selection╇ 6, 8, 11–14, 16, 19, 20, 22, 23, 28, 29, 35–37, 99, 110, 111, 113 simplification╇ 23, 38 Singapore English╇ 2, 5, 7, 11, 16, 18, 23, 40, 44, 49, 50, 53, 55, 59, 61, 67, 75, 76, 97, 98, 101╇

Singlish╇ 11, 12, 16, 17–23, 44 Sinitic╇ 4, 5, 11, 17–20, 22, 23, 27, 28, 29, 32, 33, 35, 44, 100, 101, 106, 110–113 stress╇ 7, 75–80, 84, 85, 94, 97–100, 104, 107–109, 111–113 stress language╇ 107, 108, 113 stress-timing╇ 75–80, 84, 85, 94 substrate╇ 2–8, 17, 18, 20, 23, 27, 29, 32, 44, 45, 49, 50–56, 58, 62–64, 66–71, 75, 76, 78–82, 85–87, 90, 92–94, 97–100, 107, 113 superstrate╇ 8, 45, 49, 55, 63, 64, 66, 69, 71, 99 syllable-timing╇ 75,76, 78–80, 84, 85, 94 T tense╇ 6, 23, 27, 32, 33, 35–39, 43, 44, 49, 50–53, 55–60, 70, 71, 90

Thai╇ 7, 75–82, 84–90, 92–94 Thai English╇ 7, 75,-77, 79–85, 87, 89, 90, 92–94 tone╇ 4–7, 29, 97, 99–113 tone language╇ 7, 97, 99, 100, 106–110, 113 topic prominence╇ 6, 17, 18, 20, 22, 44 topic-comment╇ 21 typological dominance╇ 13, 22, 23 typological variation╇ 50 typology╇ 1, 5, 6, 8, 11, 16, 20, 23, 27, 29, 37, 44, 49, 50, 75, 76, 78, 97, 100, 107, 108, 110, 113 U universal╇ 2, 4, 7, 49–52, 57, 58, 62, 66, 69–71, 92, 94, Z zero copula╇ 17–22, 39, 42–44