Vineeta Chand 2009 - Linguistics

Who Owns English?
Political, Social and Linguistic Dimensions of Urban Indian English Language Practices
VINEETA CHAND
B.A. (University of California, Berkeley) 1999
M.A. (University of California, Davis) 2004
DISSERTATION
Submitted in partial satisfaction of the requirements for the degree of
DOCTOR OF PHILOSOPHY
in
LINGUISTICS
in the
OFFICE OF GRADUATE STUDIES
of the
UNIVERSITY OF CALIFORNIA
DAVIS
Approved:
_____Janet S. Shibamoto Smith, Chair______
_____Shobhana Chelliah, Member_________
_____Robert J. Bayley, Member___________
_____Orhan C. Orgun, Member____________
Committee in Charge
2009
i
© Copyright by Vineeta Chand 2009
All Rights Reserved
ii
ABSTRACT
The sociolinguistic anthropology of post-colonialism fails to capture the micro-level,
realtime playing out of significant lingering but often extremely subtle effects of
colonialism on the contemporary global stage, particularly as they manifest in the finegrained interweaving of language practices and ideologies in post-colonial settings.
Drawing on sociolinguistic interviews with 35 native Hindi/Indian English (IE) bilinguals
in New Delhi, this project addresses these relationships by analyzing structural changes
in IE vis-à-vis changes in ideologies and language plans.
Quantitative variationist analysis of phonological variation in three variables—
postvocalic (r), (v) and (w)—are used in conjunction with close ethnographic and
qualitative analyses of speakers’ social identities, attitudes and linguistic ideologies.
From these I uncover significant links between language practices, competing global and
local needs, and evolving historical backdrops. Gender, age, ethno-linguistic background,
assessments of personal bilingual fluency and literacy, and domestic mobility are each
foundational elements of individual IE identity, and collectively are significant factors for
understanding systematic phonetic and phonological variation in IE.
This research tests and affirms that variationist techniques can be successfully applied to
alternative multilingual contexts. However, it challenges the assumption that oftconsidered ‘basic’ social factors, widely used in variationist studies, are adequate in
accounting for peripheral or multilingual contexts, while also problematizing the
variationist model—within which social factors are idealized as independent—given the
iii
more rigid social structures found in the Indian context and their significance in
mediating phonological variation for the three variables under investigation. Quantitative
analysis reveals that IE is progressing away from local vernacular and international
English language practices and is accompanied by the formation of an urban Indian
sociolinguistic identity: these interactions are a pivotal area of study for the Indian
postcolonial context, offering insights applicable to other post-colonial and dialect
emergence settings. Broadly, examining alternative, multilingual settings raises important
theoretical questions about the application of variationist methods in diverse contexts and
how they, with qualitative analysis, can illuminate nuanced relationships between
language change, shifts in multilinguals’ language domains, emerging local
sociolinguistic identities, language ideologies, and the long term impact of linguistic
colonialism. IE is becoming both localized and globalized, structurally, and as a
commodity within personal, local and national identity.
iv
ACKNOWLEDGMENTS
Dissertations don’t appear from a vacuum—academics are supported by their academic
mentors, grants, friends and family, all of which I gratefully acknowledge here.
Academically, I’ve had continued and amazing support from my advisor, Janet
Shibamoto Smith, who almost single-handedly opened my eyes to linguistic
anthropology and qualitative sociolinguistics. I am now a linguistic anthropologist,
thanks to you, Janet. The three additional members of my dissertation committee have
also been central to my academic development. Robert Bayley has taught me to negotiate
and successfully apply quantitative sociolinguistic methods and theories and has been
very patient with my idiosyncrasies. Orhan Orgun has supported my development as an
acoustic phoneticist, has offered invaluable help in my exploration of phonological
theory, and has also offered a welcome friendship when I’m sick of talking about ‘work.’
Finally, Shobhana Chelliah has kept me grounded in the Delhi context and has pushed me
to think more deeply about my data and my participants’ lives as integral to this project.
While not an academic, Subhash Chand has also offered enormous academic support—he
has bought books in India for me that are not available here, helped me to test my
recording setup, and has also sat through hundreds of discussions of India’s political,
economic, social and linguistic history as he has experienced them, in which he never
seemed to get annoyed with my questions and hypotheses. A big thank you to all of you.
v
I am also lucky to have been trained at the University of California, Davis, where our
Linguistics Graduate Group, with its cross-disciplinary focus, has encouraged me to
explore language from multiple perspectives.
My fieldwork and experiences in Delhi would have been much less pleasant and have
taken much longer without the support of my family there, in particular my Bhuaji, my
Dadiji and Nandini Chachi, who happily introduced me to their friends and neighbors—
thank you. My data is so rich because of the generosity of my 35 informants, who have
taught me a lot about their lives and histories as urban Delhiites.
Transcription, data coding and fieldwork were all supported by two grants, without which
this would’ve been a much more arduous process. They are a 2007 Dissertation
Fieldwork Grant by the Wenner-Gren Foundation for Anthropological Research (#7702)
and a 2008 NSF Linguistics Doctoral Dissertation Improvement Grant (#0819246). My
research assistants, Onna Nelson, Kelley Corcoran, Marcus Royce-Fulton, Alia
Piepenbrink, Jonin Cordoba and Tammy Gales, were immensely helpful and showed me
new ways of seeing the data, for which I am grateful.
Finally, academia isn’t all there is to life (at least to me!), and I would like to thank
several who have helped to maintain my sanity during this process. First and foremost,
Lunabel the Piglet has kept life ridiculous, relaxed and fascinating/distracting; this
dissertation, however inappropriate it may be, is dedicated to her. Second, my parents:
my mother has had such generosity to raise me as a multicultural and multiracial Indian-
vi
American, but also as a curious adventurer, while my father, also an adventurer in other
ways, has fundamentally taught me how to think analytically. There are a number of
other people that have been very important to me during this process who I simply can
not rank, and who I am proud to call friends and consider family: Bruce Buchanan,
Brendan Furey, Tammy Gales, Lothian Furey, John Webster, Kenan O’Brien, Lisa
Bonnici, Issac Frank, Jessie Potter, Sean McCloud, the T4T gals, and my Mill and Lark
summertime communities.
Finally, though it may be silly to thank inanimate objects or locations, I am going to. My
garden, full of life, color, chaos and growth, has been a cornerstone of my sanity during
this process and a welcome distraction when words overwhelm me.
vii
TABLE OF CONTENTS
Abstract .............................................................................................................................. iii
Acknowledgments............................................................................................................... v
Chapter 1. Introduction ................................................................................................... 1
Timeline .......................................................................................................................... 3
Pre 20th Century .......................................................................................................... 3
Early 20th Century....................................................................................................... 4
1950-1970’s ................................................................................................................ 4
Comparing Approaches .................................................................................................. 7
Intersections and Disjunctures .................................................................................... 8
The Linguistic Unit ..................................................................................................... 8
Critiques of Various Approaches................................................................................ 9
Linguistic Anthropology....................................................................................... 10
Variationist Sociolinguistics ................................................................................. 10
Are these so different? .......................................................................................... 12
The Role and Goals of Qualitative Sociolinguistic Research ............................... 12
The Total Linguistic Fact.......................................................................................... 14
This Research................................................................................................................ 16
Goals of this Research .............................................................................................. 17
Locating Research within Academic Thought on India and Sociolinguistics .............. 18
Addressing the Social Justice Agenda within Sociolinguistics ................................ 23
Chapter 2. Social, Linguistic and Historical Backdrop............................................... 26
The Diaspora of English ............................................................................................... 26
Historical Background: India........................................................................................ 28
English in India ............................................................................................................. 34
Postcolonial Sociolinguistic Theorizing ....................................................................... 39
Language and Society in the Postcolonial Context: Problems and Challenges........ 44
The Impacts of the Monolingual Paradigm .............................................................. 49
Language Ideologies ..................................................................................................... 51
Current Project .............................................................................................................. 52
Research Questions................................................................................................... 55
Chapter 3. Data Collection and Analysis Methods ...................................................... 58
Data Collection Goals ................................................................................................... 58
Target Informants...................................................................................................... 58
Researcher Influence..................................................................................................... 61
Participant Recruitment ................................................................................................ 63
Participant Population................................................................................................... 64
Negotiating Participant Background......................................................................... 65
Complexifying ‘Delhiite’...................................................................................... 66
Birthplace of the Oldest Generation ..................................................................... 68
Bi- vs. Tri-/Multilingual Participants........................................................................ 71
The Oldest Generation and the Military ............................................................... 72
Participant Breakdown.......................................................................................... 73
Data ............................................................................................................................... 73
Interview Topics ....................................................................................................... 77
viii
Interview Relationship .............................................................................................. 79
Aspects of Recording................................................................................................ 82
Target Data.................................................................................................................... 84
Structural Variables .................................................................................................. 84
(r)........................................................................................................................... 85
(v/w)...................................................................................................................... 86
Language Ideologies ..................................................................................................... 88
Analysis......................................................................................................................... 89
Transcription ............................................................................................................. 90
Transcription as Translation and Transliteration .................................................. 90
Utterance-internal Transcription Choices ............................................................. 94
Quantitative Analysis................................................................................................ 98
Qualitative Analysis................................................................................................ 100
Chapter 4. Sociolinguistic Identity, Ideologies and Alignments ............................... 104
Introduction................................................................................................................. 104
Attitudes versus Ideologies ......................................................................................... 106
Theoretical and Practical Goals .................................................................................. 108
Past Research .............................................................................................................. 110
Moving Forwards in Qualitative Research ............................................................. 113
Methodology ............................................................................................................... 115
Grounded Theory .................................................................................................... 115
Development of Qualitative Codes......................................................................... 116
Qualitative Code Families....................................................................................... 118
Coding, in Practice.................................................................................................. 118
Concepts for Analysis ................................................................................................. 119
Analysis....................................................................................................................... 120
IE............................................................................................................................. 121
Varieties of IE ......................................................................................................... 122
British Speech is ‘correct’....................................................................................... 125
All English Accents are Acceptable ....................................................................... 128
Who is a ‘good’ IE speaker?................................................................................... 135
Case Study: f27RG ................................................................................................. 140
Fake Accents........................................................................................................... 144
The Role of Hindi ................................................................................................... 147
How is Hindi devalued?...................................................................................... 149
Economic value of English ............................................................................. 149
Shuddh Hindi .................................................................................................. 154
Linking Language Competency to Cultural Knowledge ........................................ 157
‘Vernacular’ students .......................................................................................... 162
Language Maintenance Plans ............................................................................. 166
Hinglish................................................................................................................... 172
Benefits of IE .............................................................................................................. 179
Economic Benefits .................................................................................................. 179
Social Benefits ........................................................................................................ 186
Clarifying Delhi .................................................................................................. 189
IE as a Lingua Franca ............................................................................................. 190
ix
The contradictions of India ..................................................................................... 192
Conclusion .................................................................................................................. 194
Chapter 5. Postvocalic (r) in Indian English .............................................................. 196
Introduction................................................................................................................. 196
Motivations/Theoretical Contribution .................................................................... 197
Past Research .............................................................................................................. 201
(r) in IE.................................................................................................................... 201
Linguistic Constraints on (r) ................................................................................... 205
(r) in RP............................................................................................................... 205
(r) in AE .............................................................................................................. 206
Social Constraints on (r) ......................................................................................... 208
Sample Population ...................................................................................................... 208
Current Coding Practices ............................................................................................ 209
Coding Rhoticity..................................................................................................... 210
Social Constraints ................................................................................................... 211
Emergent Social Factor Groups .......................................................................... 215
Ethno-linguistic Identity ................................................................................. 216
Delhi Stay........................................................................................................ 218
Linguistic Constraints in the Current Study............................................................ 218
Token Selection and Analysis Methods.................................................................. 222
Results......................................................................................................................... 223
Overall Constraint Ranking .................................................................................... 226
Linguistic Constraints on IE Rhoticity ................................................................... 226
Social Constraints on IE Rhoticity.......................................................................... 228
The Role of Gender............................................................................................. 229
Ethno-linguistic Identity ..................................................................................... 230
Age/Occupation .................................................................................................. 232
Delhi Residence Length...................................................................................... 235
Interactions among Factor Groups...................................................................... 237
The Delhi Prestige Form......................................................................................... 237
Diachronic Analysis of IE (r)...................................................................................... 240
Evidence of Age-Grading? ..................................................................................... 241
Rhoticity across Time ............................................................................................. 243
Trills............................................................................................................................ 246
Results of Trill Analysis ......................................................................................... 246
Conclusion .................................................................................................................. 250
Chapter 6. (v) and (w) in Indian English .................................................................... 252
Introduction................................................................................................................. 252
The IE (v/w) Phenomenon .......................................................................................... 253
Hypotheses for the Historical Grounding of (v/w) ..................................................... 255
Substratum Contact Hypothesis.............................................................................. 255
England Hypothesis ................................................................................................ 261
Current Study .............................................................................................................. 265
Internal Coding Choices ......................................................................................... 265
Unmerged English Dialects .................................................................................... 265
Cross-linguistic Research on (v/w)......................................................................... 266
x
Pilot Acoustic Analysis........................................................................................... 268
Additional Internal Factors ..................................................................................... 271
Underlying Representation ................................................................................. 271
Formality............................................................................................................. 272
Syllable Strength................................................................................................. 272
Word Frequency.................................................................................................. 273
Function and Content Words .............................................................................. 276
Preceding Environment + Location .................................................................... 280
Conflate Following Environment + Location? ................................................... 282
Current Coding Practices .................................................................................... 283
Token Selection and Analysis Methods.................................................................. 287
Results......................................................................................................................... 288
(v)............................................................................................................................ 289
Social and Linguistic Constraints on (v)............................................................. 291
Following Phonetic Environment ....................................................................... 291
Syllable Count..................................................................................................... 292
Phonological variation and syllable count ...................................................... 293
Word Frequency.................................................................................................. 295
Delhi Stay............................................................................................................ 296
Formality............................................................................................................. 296
Non-Significant Factor Groups........................................................................... 297
Discussion of (v) Results .................................................................................... 298
(w) ........................................................................................................................... 300
Social and Linguistic Constraints on (w)............................................................ 301
Age/Occupation .............................................................................................. 302
Location + Preceding Environment ................................................................ 303
Syllable Count................................................................................................. 304
Word Frequency.............................................................................................. 305
Function vs. Content Words ........................................................................... 306
Following Environment .................................................................................. 306
Discussion of (w) Results ................................................................................... 307
Social Factor Groups....................................................................................... 307
Linguistic Factor Groups ................................................................................ 308
Comparison of (v) and (w).......................................................................................... 309
The reality of this ‘merger’ ..................................................................................... 309
Conclusion .................................................................................................................. 311
Chapter 7. Conclusions................................................................................................. 313
Research Question 1 ................................................................................................... 315
Research Question 2 ................................................................................................... 315
Correlations with Research Question 4................................................................... 316
Research Question 3 ................................................................................................... 317
Research Question 4 ................................................................................................... 318
(r)............................................................................................................................. 319
(v/w)........................................................................................................................ 319
The variability of variables ..................................................................................... 320
Research Question 5 ................................................................................................... 321
xi
Broader Implications................................................................................................... 322
Appendix A. Grandfather Passage.............................................................................. 323
Appendix B. IE Feature Chart .................................................................................... 324
Phonetic and Phonological Features ........................................................................... 325
Morphological Features .............................................................................................. 330
Syntactic Features ....................................................................................................... 332
Lexical and Semantic Features ................................................................................... 341
Prosodic/Suprasegmental Features ............................................................................. 345
Global Features ........................................................................................................... 347
Reading/Writing Crossover ........................................................................................ 348
Appendix C. Interview Modules.................................................................................. 352
Appendix D. Qualitative Codes ................................................................................... 360
Appendix E. Word Cloud Demonstrating IE word frequency visually................... 376
Appendix F. Word Cloud Demonstrating IE (r) word frequency visually.............. 377
Appendix G. Word Cloud Demonstrating IE (v) word frequency visually............. 378
Appendix H. Word Cloud Demonstrating IE (w) word frequency visually............ 379
References ...................................................................................................................... 380
xii
LIST OF FIGURES
2.1 Pre-Partition Indian sub-continent
2.2 Post-Partition Indian sub-continent
2.3 Languages of India
3.1 The Hindi Belt as encompassing the four states of Rajasthan, Uttar Pradesh,
Madhya Pradesh and Bihar
3.2 The Punjab region, currently divided between India and Pakistan
3.3 A visual and textual example of transcription
5.1 Young people’s alternation between the casual forms of the oldest generation
and the formal forms of their parents, the middle generation
6.1 Spectrograms of vine in IE (left) and AE (right), with [w]-like formants in IE
and [v]-like frication in AE
xiii
LIST OF TABLES
3.1 Participant demographics
4.1 Qualitative codes displayed graphically as hierarchical in Atlas.ti
5.1. Rhoticity across English dialects
5.2 Social factor coding groups
5.3 Internal factor coding groups
5.4 Overall distribution of (r)
5.5 Significant factors influencing rhotic behavior
5.6 Linguistic factors influencing r-deletion
5.7 Social factors influencing r-deletion
5.8 A comparison of non-rhotic realizations for workers with continuous and
punctuated Delhi stays
5.9 Non-rhotic realizations for workers
5.10 Overall percentage of non-rhotic tokens by age, comparing current results
with S&A
5.11 Factors which favor trill realization
6.1 Distributions between content/lexical and function words
6.2 Function words found in (v/w) tokens
6.3 Preceding environment + location conflated factor group—(w) analysis
6.4 Linguistic factor coding groups
6.5 Social factor coding groups
6.6 Overall distribution of (v)
6.7 Social and linguistic factors favoring [v] realization
xiv
6.8 Overall distribution of (w)
6.9 Social and linguistic factors favoring [w] realization
6.10 A comparison of the distribution of realizations for both variables, with
mixed realizations respectively conflated as non-dominant factors
xv
1
CHAPTER 1
INTRODUCTION
The sociolinguistic anthropology of post-colonialism fails to capture the micro-level,
realtime playing out of significant lingering but often extremely subtle effects of
colonialism on the contemporary global stage, particularly as they manifest in the finegrained interweaving of language practices and ideologies in post-colonial settings. This
project addresses these relationships through analysis of structural changes in Indian
English (IE) diachronically as they interact with changes in ideologies and language
plans, through sociolinguistic interviews with three generations of IE speakers in New
Delhi, India. It demonstrates that English ownership no longer resides only in western,
standard varieties. Increasingly, English appears in a variety of contexts, molded for local
use and legitimizing new social identities. Among the nexuses of English appropriation,
India stands out worldwide as the largest English speaking nation, with the nativized
English of India, IE, intertwined with the recent growth of India’s economy, increased
job outsourcing to India and internal valorization of western culture. However, IE loyalty
cannot be assumed: while Indians are increasingly ‘allowed’ to be native English
speakers, IE dialectal features continue to be viewed as ‘wrong,’ and ‘needing fixing.’
This research directly addresses important theoretical questions about the relationship
2
between language change and the long term impacts of linguistic colonialism in the postcolonial semiotic world.
The general approach taken here is sociolinguistics in the original, broader sense of the
term (Bucholtz & Hall 2008), tying together in Hymesian fashion the concerns of
linguistics with those of anthropology (e.g., Hymes 1974). In this sense, sociolinguistics
is a very broad-ranging field—in the course of preparing for and conducting this research,
for example, I have read about English dialects across the globe and various quantitative
and qualitative ways of circumscribing and analyzing them, and language more broadly. I
have read and considered various methodological approaches—most with their own sets
of theoretical ‘baggage’—to studying the social life and structure of language, including
multiple approaches to discourse analysis, ethnographic analysis, investigating language
attitudes and ideological analysis, social theory, power relationships, corpus based
analysis, multivariate analysis, survey-based descriptive work, historical approaches to
linguistic change, acoustic analysis, and more formalized theories of language including
Optimality Theory, and Exemplar Theory. But I have also read on, and been influenced
by, readings as widespread as the development of /r/ pronunciation in 3 year old UK
children, the reappropriation of reggae language practices by white middle class
Belgians, proposals for India’s future language policy written pre-Partition (by India’s
first Secretary of Education, and, incidentally, my great-grandfather), ethnographic
explorations of south Indian religious registers, and World Bank reports on India’s GDP
and Indian economic changes. As well, I have examined speech in unlikely scenarios—
Apu, from the Simpsons, as an icon in the west of ‘Indian speech,’ accent reduction
3
course testimonials, call center data from across the globe, IE web-based avatars,
YouTube videos describing, mocking and valorizing IE, and have had countless
discussions about people’s experiences as or with Indians. This accumulated wealth of
information is rich, but it is also so far sweeping as to appear disorienting. This, to me, is
the nature of sociolinguistics—we must learn from a wide range of disciplines because
our goal is to both understand the social life of language and to capture theoretical
insights which may help to understand far-reaching contexts in which language-in-use
has serious social and linguistic consequences. This research and some of my goals are
best contextualized by re-examining the development of sociolinguistic thought in
various fields, and to this we now turn.
Timeline
Pre 20th Century
The beginning of systematic research on social aspects of language was introduced by
Franz Boas, in the 1880’s, and focused on the relationship between language structure,
culture and worldview. Cross-linguistic and cross-cultural comparisons were conducted,
drawing on North America Native American languages and cultures (e.g., Boas 1911).
Much of this research was by necessity salvage-like, given the rapid and widespread
death of Native American languages, e.g., the infrequently published, but voluminous
research conducted by J.P. Harrington. Perhaps because of its time-sensitive nature, this
research viewed language communities as unified in their language practices. Boas’
influence on the field cannot be overstated—when he joined Columbia University full-
4
time, in 1899, his program became the first Ph.D. program in Anthropology in the U.S.
He, in turn, trained a majority of the founders of anthropology departments across the
U.S.
Early 20th Century
Linguistic anthropology continued with Boas’ students, e.g. Edward Sapir, from ~19101935, and Sapir’s students, notably Benjamin Lee Whorf, beginning in the 1930’s (e.g.,
Whorf 1941). Meanwhile, early 20th century linguistic and philological research in
Europe and North America was dominantly structural and excluded the social context
from study. This strand of research, succinctly captured in Bloomfield’s (1933)
monograph Language, focused on formal analyses of language structure and strove to
make structural analyses scientifically rigorous.
1950-1970’s
The structural descriptive approach to language lost preeminence in linguistic circles with
the advent of Generative Grammar, spearheaded by Noam Chomsky through his
dissertation (1955) and extended monograph on the same topic, Syntactic Structures
(1957)1. Generative research was—and still is—concerned with the structural nature of
language and linguistic universals, and takes as a precondition that language is an innate
system with a universal underlying structure. Methodologically, it developed from
Whorf’s analytic toolset, and is ‘merely a notational refinement’ of such (Silverstein
1979: 197). Generative research is distinct from both linguistic anthropology and
1
The popularity of generative transformational approaches to grammar coincided with the rapid creation of
independent Linguistics Departments—the staying power of this approach to analyzing language may be
linked to its seminal role within linguistics as a separate discipline in the US.
5
structural linguistics in its unit of analysis: generative research draws on linguistic
competence over linguistic performance as a data source and focuses on the syntactic
level, within which surface level variation and the social aspects of language are largely
or explicitly excluded.
Contemporary with the development of generative grammar, linguistic anthropology
continued to be influenced by Boas and Sapir, within which Dell Hymes played a
powerful role, starting in the 1950’s. He theorized qualitative aspects of sociolinguistics
and developed models and methods for analyzing the relationship of speech to human
relations and thought through genre and interactional frameworks (e.g., Hymes 1974).
Also emerging in the 1950’s and 60’s was a third vein of research on language which
investigated social dynamics of language, and, in particular, individual and societal
multilingualism, language loyalty and language maintenance. This approach, termed the
sociology of language (Fishman 1965), drew on psychology and sociology, and was
developed by the interdisciplinary social psychologist Joshua Fishman. A central concern
of research in this vein is the effect of language on society, as a tool for unification,
division and individual and group identity.
Sociolinguistic research within the field of linguistics only emerged in the early 1960’s
with William Labov’s research on language variation in New York City (1966; 1972c).
Labov challenged a critical tenet in generative circles, that linguistic variation is
unimportant, and offered a means of incorporating social concepts within quantitative
6
research on language structure and change. His approach interpreted phonological
variation across social strata in a community as systematic, mediated by linguistic factors,
and linked to diachronic sound change. These methods rely on the premise of a speech
community as sharing evaluative stance towards variation: in other words, all community
members see the same variant as more ‘prestigious,’ and all move in the same direction,
towards the prestige variant, in more formal contexts.
Labov’s analysis of variable post-vocalic /r/ pronunciation in New York City department
stores (Labov 1966) is a classic variationist study which greatly influenced the direction
of sociolinguistic research continuing today. First, r-pronunciation demonstrated a fine
grained correlation with social stratification. Second, all three socio-economic classes he
studied were more rhotic in careful speech, demonstrating a unified ideology of the
prestige variant. Third, access was introduced as influential on language practices: the
variable use of the prestige variant across the classes depended on their regular access to
the prestige variant. Fourth, linguistic environment impacted realization quality: word
final position, e.g. floor, provoked more r-ful pronunciation than coda clusters, e.g.
fourth. This research underscored that communities hold shared evaluations of prestige
variants, within which variation in use can be explained through access and linguistic
environment. Collectively, this research set sociolinguistics down a particular path,
focusing on diachronic sound change, and how it can be extrapolated through these
methods. The Department Store Study, as it came to be known, was a catalyst for
linguists to approach, model and understand language change, and dominantly, modern
sociolinguists have continued within this framework, and are focused on understanding
7
and capturing diachronic change in language through attention to internal linguistic
constraints within a sample population. With the introduction of variationist research, as
Labov’s paradigm came to be known, sociolinguistic research in anthropology, social
psychology and linguistics, while all examining variation in language practices, were
largely distinct in methodological approaches (Bucholtz & Hall 2008), and their preferred
unit of analysis.
Comparing Approaches2
Addressing methodological distinctions and foci first, Schneider (2007: 10) broadly
distinguishes two branches of sociolinguistics: macro- and micro-sociolinguistics, the
former encapsulating the sociology of language and linguistic anthropology and the latter
encapsulating variationist sociolinguistics. However, more narrow distinctions are
possible. Specifically, linguistic anthropology viewed language as a tool for social
interactions, cultural variation, different worldview perspectives and power structures,
and used qualitative and ethnographic methods to explore discourse structure and genres.
The sociology of language focused on multilingualism, language policy, nationalism,
language maintenance and shift, and hence was largely focused on ideological analysis,
discourse analysis, and on how multilingual speakers negotiate multiple codes.
Variationist research, meanwhile, focused on monolingual communities, and used
2
The three are also characterized by the range of article topics, population foci and methods presented
within their ‘prototypical’ journals: the sociology of language approach is best represented by the
International Journal of the Sociology of Language, the variationist paradigm is best represented by
Language Variation and Change, which is still edited by Labov, and linguistic anthropology is best
represented by The Journal of Linguistic Anthropology. Importantly, there are a number of journals which
offer a cross-section of research from these respective positions: e.g., Language in Society, the Journal of
Sociolinguistics, Language & Communication, Language Policy, Text & Talk, and Discourse and Society,
along with a number of journals which incorporate multiple approaches while focused on a narrower
language or region, e.g. World Englishes, English World-wide, American Speech, and the Indian Journal of
Applied Linguistics.
8
quantitative methods on a sample population to predict larger processes of diachronic
sound change as influenced by linguistic context and social stratification.
Intersections and Disjunctures
Importantly, while variationist research shares with the sociology of language an interest
in diachronic processes, the former is distinct in its focus on structural change, not on
changes in linguistic repertoires and social dynamics. Linguistic anthropology shares a
common bond with the sociology of language, in that social structures are interpreted as
multifaceted and as having unequal social authority. They are different in that linguistic
anthropology 1) continues to explore the role of language on our worldview (weak and
strong versions of the Sapir-Whorf hypothesis, now more typically termed linguistic
relativity), and 2) is centrally focused on deconstructing identity—within which language
is but one aspect—in order to expose power relationships, discrimination and
marginalization, while the sociology of language is broadly construed as any approach
which contributes to the growth of language-related knowledge. Finally, linguistic
anthropology and sociolinguistics, while attending to both social and linguistic
phenomenon, are best interpreted as distinct in their relative focus on identity formation
and change, for the former, and linguistic structure and change, for the latter.
The Linguistic Unit
There are both distinctions and overlaps in the type of data analyzed across these three
approaches. While early linguistic anthropology was directly interested in linguistic
structure, and examined differences in syntax and lexicon, later linguistic anthropology
9
research has focused on discursive interactions, on delineating aspects of register and
genre, and on collaborative and disjunctive patterns of meaning making through
language. Variationist research, in its early days, narrowly focused on phonetic and
phonological variation, necessarily so given its underlying tenet that variants have the
same denotational meaning. The sociology of language originally focused, through
Fishman’s influence, on multilingual language practices as they relate to codeswitching
and domain-specific code choice, and on the social dynamics mediating language
authorization. It draws on both large discourse units, and on smaller syntactic,
morphological, phonological and phonetic structures to interpret how social dynamics
manifest in language. However, the sociology of language can be considered a pretheoretic, exploratory frame, within which several areas were introduced, and then taken
up by sociologists, anthropologists and linguists, among others, within their own fields.
Because of its exploratory focus, and the subsequent uptake of these areas by researchers
in other fields, the sociology of language was drained of original content, and is not a
driving or important force in current sociolinguistic theorizing.
Critiques of Various Approaches
There are limitations to each of these approaches to language and society. In areas with
well spelled out methods, critiques can get very specific, while in areas which draw on
multiple methodologies, critiques are more likely to problematize the incorporation of an
array of methods as reflecting different theoretical agendas, within which no particular
theory is able to advance. As is clear below, both types of critiques are directed toward
investigations of language and society. However, because the sociology of language had
10
broad foci, multiple methodologies, a pre-theoretical frame and ideas originally
introduced within the sociology of language were then adopted and explored in related
disciplines, I found no focused critiques of this approach as a whole, and will discuss
here critiques directed towards linguistic anthropology and variationist sociolinguistics.
Linguistic Anthropology
Starting with linguistic anthropology, critiques have focused on such research as socially
focused, not linguistically focused. In some cases, this is true, but I interpret this as less
of a ‘critique’ of another field, and more of a critique of how one approach fails to
consistently contribute to another’s theoretical goals. Linguistic anthropology has also
been critiqued as lacking scientific rigor. In part this is due to the nature of ethnographic
methods—which are framed by a researcher’s own positionality in and interpretation of a
community—, and in part this is because linguistic anthropology has, to date, explored a
much wider range of communities and linguistic practices than, e.g. the variationist
paradigm (c.f. Preston 1991 for a survey of the range of languages and populations
studied through variationist methods).
Variationist Sociolinguistics
Within the variationist paradigm, the ‘linguistic’ side (if one is to overgeneralize this
research as having a linguistic and a social side) of such research has been problematized:
for example, Lavandera (1978) disputes the assumption that variants do share an
underlying meaning, while prestige, as it is traditionally interpreted and methodologically
11
approached in variationist research, does not allow a means of operationalizing—or even
accounting for—covert and overt prestige (c.f., Trudgill 1974).
On the ‘social’ side, variationists have fore fronted language internal pressures to
variation as nuanced and viable explanations for systematic diachronic linguistic change,
while both the social attributes of groups and the various styles of speech that any
speaker commands are not typically approached in such a nuanced fashion. For example,
analyses of social groupings have dominantly relied on traditional, outdated social
classifications, placing unequal emphasis on social and linguistic factors (J. Milroy
2001). In much of variationist research, speech communities are seen as socially,
linguistically (in terms of internal grammar and constraint ranking) and ideologically
homogeneous within each strata. This social and ideological homogenization of speech
communities does not allow for within group variation based on different interaction
experiences, while more nuanced approaches, drawing on locally relevant social
identities (Rickford 1987) and processes of social construction (c.f., Woolard 2008) have
been successful in nuancing social identity to uncover variation within communities (e.g.,
Llamas 2007). Linguistic analyses which incorporate linguistic ideologies have similarly
challenged past conceptions of the ideologically homogeneous speech community (e.g.,
Gal & Irvine 1995; French 2001; Wassink & Dyer 2004; Su 2008; Zhang 2008; V. Chand
to appear). Also problematic in variationist approaches is how speakers are idealized as
static in their language practices. Research on language stylization offers a valuable
counterpoint, demonstrating that speakers can and do consciously shift their speech to fit
12
the context, their interlocutors, and the personae they wish to present (e.g., Cameron
2000; Kiesling 2001; Levon 2009).
Are these so different?
The similarities and differences I have highlighted here are not universally agreed upon—
indeed, the Journal of Sociolinguistics re-approached this topic in a 2008 special issue
(Bucholtz 2003; Eckert 2008; Heller 2008; Woolard 2008)—also see Bucholtz (2003).
Opinions were offered comparing historical development, methods, theoretical foci,
preferred population samples, and researcher identity (e.g. Am I a sociolinguist or a
linguistic anthropologist?), similar to what I have offered here, but distinct in each
authors respective interpretations. The similarities and differences among these
approaches are largely a product of their distinct paths of development, and, in their
research, the range of other literature which individual researchers draw on as support. At
this point, however, there are no clear boundaries: researchers from multiple ‘fields’ are
drawing on several methodologies and theoretical approaches and offering analyses that
contribute to the goals of more than one single approach. However, within this
methodological and theoretical conflation, another important distinction is made which is
not easily explained away through a trip down academia’s memory lane: the division
between qualitative (purportedly descriptive) and quantitative sociolinguistic research.
The Role and Goals of Qualitative Sociolinguistic Research
Qualitative sociolinguistic approaches have been used by linguists, linguistic
anthropologists, and practitioners of the sociology of language to explore a range of
13
topics. Briefly, covering several dominant areas, qualitative research has explored
competing and multifaceted social alignments (e.g., Shenk 2007; Shankar 2008), the
indexical functions of language (e.g., Lippi-Green 1997; Johnstone, Andrus & Danielson
2006; Mendoza-Denton 2008; Woolard 2008), fluidity of language practices (e.g., Ito
2001; Kiesling 2001; Levon 2009), the impact of global(ized) ideologies on individuals
(e.g., Lippi-Green 1997; Woolard 1998; V. Chand in press), global and local
appropriation, reformulation and recommodification of language practices (e.g., Rampton
1995; Bucholtz 1999), linguistic bias and discrimination (e.g., Lippi-Green 1997; Purnell,
Idsardi & Baugh 1999), and has also offered contributions to applied and pedagogical
fields (e.g., Labov 1982; Blommaert 2001). In short, qualitative research has exposed
several social aspects of language use that permit us to theorize underlying social
relations.
However, the scientific rigor of qualitative research continues to be an issue. For
example, Rickford suggests that sociolinguistics has ‘a tendency to be satisfied with
observation and description, and [is] insufficiently imbued with the thirst for theoretical
explanation and prediction which drives science onward ’ (Rickford 1988). However, this
is problematic to accept at face value, even if one is narrowly committed to
understanding language variation and change. Labov’s early study of Martha’s Vineyard
highlights that some variation is not mediated by internal linguistic constraints, while it is
still socially systematic and ideologically diagnostic (Labov 1972c). While he—and the
field of variationist sociolinguistics—largely chose to continue to focus on linguistic
explanations for language change, and his own research has not reapproached the
14
exploration of ideologically motivated variation, sociolinguistics must incorporate
sociological and anthropological theories of social interaction and power as scaffolding
for theorizing language change as both a social and linguistic phenomenon. This
approach can create a middle ground between linguistic anthropology and variationist
sociolinguistics, where qualitative research can be (re)interpreted as foundational to
variationist research, and important in its own right, in that it continues to highlight the
multiplicity of emerging and existing ways people create both meaning and identity
through language.
The Total Linguistic Fact
A critical perspective on the need for such an integration, and what such analyses
require—both theoretically and methodologically—is offered by Silverstein (1979). He
highlights theoretical and methodological pitfalls inherent to past and modern linguistic
analysis, arguing that there are distortions inherent to anthropological, sociolinguistic,
structural and generative analyses which can only be remedied by addressing the ‘total
linguistic fact’. To do this, he resuscitates Whorf’s original attention to the semiotic space
of language, demonstrating that linguistic relativity is inherent to metalinguistic analysis,
rendering such analyses of language distorted.
Silverstein distinguishes two ‘functions’ of language as both a means of ‘objectifying’ or
rationalizing the world (the ‘really real’ in his words) and a negotiation of social relations
vis-à-vis social context. He makes the powerful claim that linguistic ideologies and
multifaceted functional roles of language engage with language structure diachronically,
15
and are mutually impacting. Thus, to study the ‘total linguistic fact,’ one must attend to
the metalinguistic system of structural categories and diachronic changes in
categorization, and how these are semiotically tied to indexicality and linguistic
ideologies. Although these elements are grounded in the development of sociolinguistic
theory from multiple fields, Silverstein offers a different theory of what is important
about language than any of his precursors through their integration.
Reflections on quantitative projects support this integration: Eckert admits that the use of
variationist methods in her dissertation research on Jocks and Burnouts nearly occluded
her from seeing and taking into account locally salient alignments (cited in Woolard
2008). Indeed, several of the social critiques of the variationist paradigm can be mitigated
by incorporating qualitative methodologies, e.g. ethnography—although equally
productive alternatives to ethnography are also suggested (c.f., Dubois & Horvath
2000)—to explore the local sociolinguistic reality before embarking on variationist
analysis. This permits qualitative research to be reinterpreted (and recommodified) as
pivotal scaffolding for conducting diachronic analyses of sound change. My research
resuscitates the original agenda within Labov’s Martha’s Vineyard study, following
research by Blake and Josey (2003) and Dubois and Horvath (2000), the theoretical frame
posited by Silverstein (1979; 1985), and the lessons learned from these various
theoretical and methodological approaches in isolation. I take both qualitative and
quantitative sociolinguistic research as useful heuristics, without taking on their
respective theoretical baggage, by returning to the exploration of how language
16
ideologies and locally salient social alignments interact with language variation and
change in this urban post-colonial context.
Methodologically, incorporating qualitative ethnographic and ideological data to uncover
locally relevant social distinctions and language patterns in conjunction with quantitative
variationist analysis is the cutting edge of sociolinguistic research, and is a path which
can theorize both the social and the linguistic aspects of language use (e.g. Schecter &
Bayley 2002; Wassink & Dyer 2004; Zhang 2005; Mendoza-Denton 2008). This is where
my research is focused within the development of sociolinguistic thought.
This Research
This research has two larger goals: 1) to illuminate some aspects of the sociolinguistic
reality of urban fluent English speakers in India—both in terms of local influences and in
terms of how globalization affects their practices and ideologies; 2) to contribute to a
central sociolinguistic and anthropological tenet—exposing underlying power structures
and inequality in sociolinguistic practices and ideologies. Importantly, dominant
theoretical and popular capitalistic writing has framed IE speakers in particular, biased
ways. I offer this research and these IE speakers voices as a humanizing counterpoint.
Global economic processes are powerful and widespread, but they fail to realistically
portray and validate other ways of conceptualizing value, authenticity, and evolving
cultural and linguistic trends (V. Chand in press). This is an equally powerful arm which
academic research can wield and which I try to do here. These speakers are not simply
giving voice to diachronic linguistic processes, but are also living their lives with their
17
multiple identities—these lives and experiences give voice to the language, and I try to
present their social setting, with all of its conflicts, changes and stability, as equally
important and valid. This research is thus framed within the development of
sociolinguistic thought and the development of modern Indian culture.
Goals of this Research
These theoretical and methodological disjunctures in past approaches to language
variation, in conjunction with Silverstein’s (1979) broader understanding of factors
involved in exploring and theorizing the total linguistic fact are my starting place for the
current project. This research is focused on upper middle class Hindi English bilinguals
(defined as speaking both languages before kindergarten and regular use of both in life
today) in south Delhi. Using early bilinguals with identical language background prevents
this study from complications related to differences in fluency or vernacular language
background, both which are issues which plague much of past literature on IE (Kandiah
1991). Socio-economic background, education and place of residence have also been
demonstrated to produce variation (e.g. Figueroa 2004; Sharma 2005) and are similarly
controlled for within this target population. Sociolinguistic interviews were designed to
capture naturalistic IE features and ideologies related to IE awareness, speaker attitudes
towards English dialects, language plans and locally perceived status of IE. Hereafter,
discussion of IE as developed from my analyses and results has a narrow meaning, and
reflects the above, native, south Delhi upper middle class population’s practices and
ideologies. My research questions, below, are explained and contextualized in the next
chapter, within a discussion of the role and current place of English in India and a review
18
of relevant IE literature, and broader range of applicable sociolinguistic, post-colonial,
and anthropological theories and methods.
•
What are speaker attitudes and ideologies towards IE, and how does IE dialect
awareness correlate with actual linguistic repertoires?
•
What social factors, alliances and categorizations do IE speakers evoke through
discussion of IE, and how are these related to IE with respect to the ideology of
standard languages with native speakers?
•
How do speaker linguistic plans, domains of language use, and the perceived
function of languages within these speakers’ multilingual linguistic repertoires
interact both with each other and with IE diachronically?
•
How do previously cited features of IE manifest in actual oral conversation? Are
they variable, what motivates this variation, and can they themselves be
quantitatively modeled?
•
How is IE structure changing diachronically? Are globalization and increased
access to RP and American English (AE) media influencing IE towards
assimilation with one of the two, or towards increased distance from these
dominant English varieties?
Locating Research within Academic Thought on India and Sociolinguistics
Here, I address how this research addresses two current strands of academic though: first,
scholarship on and in India, and second, the theoretical and methodological development
of sociolinguistics. Addressing the former, the study of English in India has taken many
19
forms, not all of which are mutually compatible, demonstrating varying degrees of
separation between English in the Indian context and Standard British Received
Pronunciation (RP) as well as varying degrees of affinity between IE and World
Englishes (e.g. Wiltshire & Moon 2003; A. Pandey 2004; Schneider 2004b). From past
literature’s multiple disjunct perspectives on the multifaceted, politicized identity and
structure of IE, within this project ‘certain distinctions will be reiterated and others
rejected,’ akin to Santa Ana’s characterization of Chicano English (1993: 4) to critically
approach and interpret past literature.
The role of the colonial background, globalization and modernity (Blommaert 2003;
Kristiansen 2003; Raj 2003; Lee 2006) on IE in a modern post-colonial context, as forces
pushing this local variety of English further from or closer to dominant standard English
varieties (J. Milroy 2001), is completely unexamined. Compounding this, what Saussure
termed the ‘external’ elements of language (e.g. its history, context, speaker identity,
Saussure 1983 [1916]: 20) and what Silverstein (1979) considers pivotal aspects of the
‘total linguistic fact’ are grossly underrepresented in examinations of the post-colonial
linguistic landscape, thus problematizing, in this instance of English in India, the majority
of prior IE research. As well, differences in inter- and intra-study speaker demographics
are possible confounding factors (Kandiah 1991) which do not allow IE literature to be
easily reconciled, within which the development of IE as a single or multiple dialects
remains unnuanced. Further, past IE research does not fully address how IE interacts both
structurally and socially within the milieu of World Englishes and the post-colonial
context.
20
Compounding these larger sociolinguistic and structural diachronic issues, there has been
little discussion of contextually determined variance in the IE speaking bilingual
community and quantitative work itself on IE is scarce. Studies of IE speaker awareness
of and attitudes towards IE (as a dialect) are both rare and narrow in focus, and the
relation between perceived IE features and actual IE speech has not been studied.
Accompanying this, work on changes in IE over time, through apparent time studies of
people of different ages (Bailey 2002) or through longitudinal studies are scarce (e.g.
Agnihotri & Sahgal 1985; Sahgal & Agnihotri 1987). The majority of earlier descriptive
work on IE is shaped by and functioning within the ideology of standardization and hence
not value-neutral, and it has, moreover, neglected a pivotal aspect of language, the
‘external’ or social character of language (Silverstein 1979; Woolard 1986; 1998; J.
Milroy 2001).
This project will advance anthropological and sociolinguistic theory in three ways, all
tied to IE’s status as an alternative marketplace. First, by focusing on the formation of
linguistic and social identity within the post-colonial context it will contribute to
understanding how alternative marketplaces react to and interact within the larger global
market, given current and historical pressures (c.f., Silverstein 1979; Bourdieu 1991;
Schneider 2007). The social value of a language is linked to the social value of the
speakers (Milroy, 2001) wherein processes of evaluation and judgment bestow value
upon both the speakers and the code (Bourdieu 1991), and dialogically then influence the
code (Silverstein 1979). This process can be understood to work dialogically (Bakhtin
1981) in both directions, with social value contested and (re)created both internally, by
21
alternative marketplace IE speakers, and externally, by dominant marketplace external
groups (French 2001; V. Chand in press). The proposed informant driven research on the
social semiotics of an IE community will shed light into how a globally subordinated
linguistic community works to (re)create, negotiate, contest and promote social identities
which then filter outwards through their language, their attitudes and their personal
investment in their speech.
Second, methodologically, this research will expand and test the efficacy of variationist
sociolinguistics in accounting for variation in alternative marketplaces. It has been noted
as problematic that ‘virtually all quantitative sociolinguistic investigations have been
carried out in standard language cultures and, moreover, mainly in monolingual
situations’ (Milroy, 2001: 546). Past literature’s conflicting contestations of IE’s status as
incorrect, a non-standard dialect, a bastardization of RP (e.g., Krishnaswamy & Burde
1998) or, contrastively, as a valid dialect (e.g., Schneider 2007; V. Chand in press), and
the near-constant comparisons with standardized English varieties like RP and AE
highlight IE’s non-standard, non-legitimized locally and globally perceived status.
Changes in domains of language use, language ownership, and perceived vernacular
competency reflect internal pressures to both support local vernacular languages and
identities as they interact with globalization, and to capitalize on English’s status within
this as a path towards power, status and financial security, and can have a very real
impact on language structure and systematic language practices (Silverstein 1979). This
research thus is moving the field forward methodologically by expanding the range of
quantitative variationist applications into the non-standard post-colonial context, while
22
also contextualizing and grounding the IE sociolinguistic setting by addressing and
incorporating IE speaker ideologies and attitudes towards and about IE. By quantitatively
modeling diachronic changes across these areas in conjunction with an accounting of the
social semiotic space, this research will highlight the interwoven and dependent nature of
such factors (Silverstein 1979) to offer a more nuanced understanding of how changes in
attitudes interact with and respectively reflect and impact language practices in
alternative linguistic markets.
And third, this research’s contribution to understanding the motivations and ideologies
particular to the Indian post-colonial context can be used to directly examine and
complexify several areas critical to anthropological sociolinguistics. Processes of
language contact, diffusion and change have each been studied in isolation (Heath 1979;
1984) to better understand the history of language, speakers and their culture. However,
less understood is which linguistic features are susceptible to diffusion and replacement,
and how these particularly malleable features (and contrarily, the features particularly
immune to diffusion) are involved in the process of sociolinguistic identity formation
(Woolard 2008). Anthropology and sociolinguistics, at this juncture, need to return to and
reexamine the potential diffusion and resiliency of linguistic features as they interact with
social and political processes through approaches like that presented in Silverstein
(1979). Synchronic and diachronic variation in IE language practices not only reflect
potential changes in language attitudes, identities and domains of language use, but are
also critically tied to understanding the relationship of language diffusion universals to
linguistic ideologies. By studying how new English dialects are formed and the direction
23
of change within a relatively short time span, negotiating both the need for a localized
linguistic identity and a code mutually intelligible with other Englishes in the global
context, I argue that the processes and features of diffusion within post-colonial contexts
are particularly important to understanding both language universals and what linguistic
features native post-colonial Indian English speakers have accessible to them for
synthesis into local sociolinguistic identities.
Addressing the Social Justice Agenda within Sociolinguistics
In the US context, AE-IE interactions (and hence awareness by AE speakers of IE, as a
dialect) are increasingly prevalent: Silicon Valley and the IT industry have brought many
IE speakers to the United States, while increased technical support and IT outsourcing
have created regular cross-continental communication between IE and both AE
specialists and laypeople alike. Accompanying this, new biases against IE (McPhate
2005) and attempts to ‘fix’ IE’s ‘unintelligibility’—akin to Cameron’s (2000)
documentation of ‘vocal styling’ in UK telephone call centers and Rahman’s (2009)
documentation of ideologies and language practices within Pakistani call centers—have
arisen (e.g. Dudley 2004; Cowie 2007). Within this context—of increased external
pejoration of IE and internal attempts to ‘fix’ IE—there are direct social implications for
English speakers in India, as ‘varieties of language do not actually have prestige in
themselves: these varieties acquire prestige when their speakers have high prestige…the
prestige attributed to the language varieties (by metonymy) is indexical’ (J. Milroy 2001:
532).
24
It is possible to confront and challenge these discriminatory stereotypes and policies
through the dissemination of both data and research results from projects such as this.
The data and analyses presented in this dissertation demonstrate the systematic and
distinct nature of IE as it is natively spoken by upper middle class south Delhiites.
Further, the public distribution of this corpus—currently, the only corpus of spoken IE to
be made publicly available—through a website will allow linguistics students and
laypeople alike to access, listen to, become more familiar with and form opinions
regarding the nature of IE through naturally spoken conversational data.
A diachronic picture of IE practices, ideologies and attitudes will thus allow us to
understand some ways in which globalization and English’s increased status worldwide
can manifest, whether current variationist techniques can be used to model linguistic
behavior within multilingual alternative marketplaces, and how language diffusion
universals are accounted for and related to process of post-colonial sociolinguistic
identity formation and contestation.
Challenging labels of illegitimacy (as they are first attributed to IE, and then transitively
linked to IE speakers) through academic research is also possible and has proved fruitful
(J. Milroy 2001). Research in this vein can thus evoke greater tolerance towards IE, IE
speakers, language variation more broadly, and the global functions and varieties of
English: the dissemination of these data and results is another goal of this research.
25
One final note is offered for readers on the organization of this dissertation. The second
chapter offers a review of past literature on IE, India and sociolinguistics more broadly,
and frames this research agenda. The third chapter details my methods, and the
motivations behind such methodological choices. The fourth chapter offers a qualitative
analysis of IE speaker attitudes, beliefs, and understandings of the local sociolinguistic
context, and will be of interest to qualitative sociolinguists. The fifth and sixth chapters
are variationist analyses of three IE structural features—(r), (v) and (w), respectively—
and may be of more interest to quantitative variationists and those interested in the
structure of IE. The final chapter, Chapter 7, sums up the findings from this research in a
broader sense, and readdresses how this research fits within and contributes to the
development of sociolinguistic thought. While I am suggesting that some sections may be
of more interest (or more easily approachable) for researchers from different
backgrounds, this is by no means an endorsement of such practices: I believe that the
areas explored across all of these chapters are inextricably intertwined, and analysis of
either would have been fundamentally weakened without the others.
26
CHAPTER 2
SOCIAL, LINGUISTIC AND HISTORICAL BACKDROP
The Diaspora of English
In the past twenty plus years, there has been an explosion of research into English(es) and
the diasporas of English, evidenced in part by the 1981 creation of the journal World
Englishes. This research is partially motivated by a growing recognition of World
Englishes (WEs), and their pervasive diversity vis-à-vis standard forms of English, such
as American English (AE) and British Received Pronunciation (RP). It is also motivated
by political and economic changes in Asian and African colonies-turned-independent
states (J. Gordon & Gupta 2004; World Bank 2006). As well, more recent attention has
focused on processes of globalization and localization, and their impact on language
practices and ideologies in WE structure and ideologies (Cameron 2000; Blommaert
2003; Heller 2003; Machin & van Leeuwen 2003; Meyerhoff & Niedzielski 2003; Bhatia
2006; Lee 2006; Cowie 2007; Pennycook 2007; Rahman 2009; V. Chand in press). As
researchers have begun to explore global(ized) and local(ized) manifestations of English,
post-colonial landscapes such as India have been brought under increasing scrutiny.
Historically, direct correlations are present between colonial rule and European
27
languages: European languages (primarily English, French, Spanish, Portuguese, Dutch
and German) were used to govern and hierarchically stratify colonists above colonized
people while the lack of an original European language native speaking populations in
these colonies further perpetuated local cross-group linguistic and power disparity
(Fabian 1986).
European colonizers used language as a means of domination, albeit not always
necessarily consciously, throughout the colonies in four specific ways (Fabian 1986).
First, the colonizing European languages were ranked hierarchically, with respect to
native languages, as superior, both implicitly and explicitly. Leading from this, the
culture and knowledge of the colonizing language were also, transitively, ranked as
superior to indigenous culture and knowledge, by their association with the European
language, European colonialism and authority (Woolard 1998). Third, the native
colonized people, at the time of colonization, did not already speak the European
languages, nor were they encouraged (or allowed, given the nature of the teaching
methods offered) to learn them fluently (Fabian 1986). Related to this, the colonized,
after learning the European tongues, were not considered native speakers of European
languages, by virtue of their ethnic and linguistic background, and therefore, lacked
authority when using the European languages (Fabian 1986). This strategic ranking and
unequal valuing of languages, combined with unequal access to the ranked languages,
and then, ethnic discrimination and devaluing of linguistic capital when it was acquired
by the native population of the colony, worked collectively to create a colonial sociopolitical disparity which continues in post-colonial contexts.
28
The colonial situation in India followed this pattern of inequity, complicating the Indian
relationship with English, the language of the British colonizers. English nativity was
unavailable as social capital within the prevailing colonial linguistic marketplace
(Bourdieu 1991; French 2001), and it is not clear that this situation has been reversed in
the post-colonial context.
Historical Background: India3
In India, English is complicated by its dual relationship with British RP, given its
disputed dialectal status as a World English, and Hindi, given its political status in India
as a legitimate governmental language. As well, the scenario is complicated by the drastic
numeric difference between India’s English speaking population (350 million) and the
population for which English is a mother tongue (178,000) (Registrar General 1991).
India gained its independence from Great Britain in 1947, an event termed Partition, at
which point the Indian sub-continent was divided into three areas comprising two nationstates: West Pakistan (now Pakistan), East Pakistan (now Bangladesh) and India, with the
two ‘halves’ of Pakistan (East and West, respectively) separated by the bulk of India (see
Fig. 2.1 and 2.2).
3
For a comprehensive account of the circumstances surround the introduction of English to India, see
Schneider (2007)
29
Figure 2.1. Pre-Partition Indian Sub-Continent (Randall 2007)
Figure 2.2. Post-Partition Indian Sub-Continent (Randall 2007)
30
This division was made via the Mountbatten Plan, and was based on religion: the
majority Hindu areas became part of India, the majority Muslim areas became Pakistan,
and the Princely States, not under the sovereignty of the British Government, were
allowed to choose which nation-state to join.
Until independence, English was the language of the British government, and was spoken
by a powerful but small minority of the Indian population. While it would have been
convenient for the newly-formed Indian government to carry on in the same language as
the British colonizers, this was not without a myriad of accompanying problems, most of
which surrounded the identity of India as a collective whole and as a newly formed
nation state (T. Chand 1944). For example, in 1946, M.K. Gandhi, a pivotal figure in
India’s fight for independence, inextricably discourages a future role for English in India
with ‘[i]t is my belief that English education has bankrupted our minds…and has left us
unprepared for courageous citizenship’ (Prabhu & Rao 2003: 368, cited in Vaish 2008:
93).
In 1950, India’s Constitution established Hindi as the national language and English as
the official language for the upcoming 15 years, until 1965. During this 15 year period,
corpus planning in Hindi was undertaken, in order to equip it to don the mantle of official
language, replacing English (Vaish 2008). However, in the 1950’s there were linguistic
riots in non-Hindi speaking regions such as Tamil Nadu (see Fig. 2.3 for a broad sketch
of the linguistic diversity of India). In light of this opposition, the Indian government
recognized that the issue of national identity and linguistic and ethnic diversity would not
31
be solved with Hindi evolving as the sole national language. This realization resulted in
the 1963 Official Languages Act and a 1968 constitutional amendment which declared
English and Hindi co-official languages—that is, languages for government activities
(Sailaja 2009). Federally, India now has no national language, however each state—many
of which were delineated based on linguistic boundaries—has their own ‘national’ and
official language combinations (Vaish 2008).
32
Figure 2.3. Languages of India (Gippert 2002)
33
Reflecting India’s linguistic pluralism, the Constitution recognizes 114 languages, and
separates them into Scheduled and Non-Scheduled Languages, while only Scheduled
Languages receive federal funding for language maintenance (Mallikarjun 2001). There
is no direct correlation between population size and Scheduled status, e.g. Non-Scheduled
Bhili is spoken by ~5.5 million, while Scheduled Konkani has ~1.7 million speakers.
However, socio-economically disadvantaged groups are disproportionally relegated to
Non-Scheduled status: these tend to be spoken by tribal communities in central and
northeastern Indian states (Vaish 2008). Scheduled status is thus a position of privilege,
and some linguistic groups have mobilized to attain Scheduled Language status. Through
such advocacy, the original fifteen Scheduled Languages have now expanded to the
following twenty-two constitutionally recognized Dravidian and Indo-European
languages: Assamese, Bengali, Bodo, Dogri, Gujarati, Hindi, Kannada, Kashmiri,
Konkani, Maithili, Malayalam, Marathi, Meitei, Nepali, Oriya, Punjabi, Sanskrit,
Santhali, Sindhi, Tamil, Telugu and Urdu, while 96 other languages are recognized as
Non-Scheduled Languages. Importantly, English has non-Scheduled Language status, but
its role is arguably more powerful, as it is also indirectly defined in the Constitution:
future language policy changes are to be determined by presidentially-appointed
Commissions who are directed to consider ‘the industrial, cultural and scientific
advancement of India, and the just claims and interests of persons belonging to the nonHindi speaking areas’ (The Constitution of India, cited in Sailaja 2009: 5).
As the colonial infrastructure has dissipated in the past 60+ years and globalization—
inextricably tied to English through outsourcing and call centers—has invigorated India’s
34
free market economy (J. Gordon & Gupta 2004; World Bank 2006), the role of English,
by now established to varying degrees within the populace, remains a point of contention.
India currently has the largest English speaking population in the world, with 350 million
speakers (Crystal 2005)— a conflation of ESL and native English speakers—which
dwarfs the most recent Census documentation of 178,000 Indians who consider English
their ‘mother tongue’ (Registrar General 1991). Meanwhile, Hindi’s co-official language
status is increasingly contentious the further south one travels, into areas dominated by
Dravidian languages. In these southern areas, English has gained popularity as a ‘middle
ground,’ not elevating one Indian cultural and ethnic identity above others. Indeed, many
of the informants in this study volunteered their views on the controversy surrounding the
co-official status of Hindi and English. As such, in examining English in the postcolonial
state of India, the role of English in government, education, the economy and personal
life, and the structure of English as potentially indigenized into a local dialect have
become politicized and contentious: this is a highly visible struggle (e.g. T. Chand 1944;
Krishnaswamy & Burde 1998; Annamalai 2004; Basu 2004; Dudley 2004; Nichols 2004;
John 2007; Kennedy 2007; Tharoor 2007).
English in India
While quantifying English use in India complicates understandings of the Indian
linguistic landscape, past structural analyses of IE have also struggled on another front.
Variability in demographics across informant populations in conjunction with data from
multiple mediums (spoken vs. written) and contexts (formal vs. casual), has made
35
broader interpolations regarding the structure of IE difficult to arrive at and assess for
validity as legitimate pan-IE structures.
Speech and writing differ systematically in structure—each has unique structural
properties (Chafe 1985; Chafe & Tannen 1987). However, these differences have not
generally been recognized within IE literature. Previous analyses and descriptions of oral
and written IE have been cited as evidence of generic IE features, without any real
reflections on differences in IE oral and written standards: research stemming from oral
data (e.g. Schneider 2000; Baldridge 2002; Sharma 2005; Sailaja 2009), formal written
data (e.g. Labru 1984; Shastri 1992; Krishnaswamy & Burde 1998), and informal
solicited written data (e.g. Sridhar 1991; Hohenthal 1998; Mehrotra 2003; Vaish 2008)
are all interspersed within past IE literature and have complicated synergistic analyses of
IE.
As well, IE fluency and nativity are assessed in different ways by different researchers
(e.g. Kachru 1982; Hohenthal 1998; Jenkins 2000), and both across study and intra-study
variation in informant population are also found in past IE literature: regional and
vernacular background (e.g. Wiltshire & Moon 2003), informants’ current country of
residence (e.g. Pickering & Wiltshire 2000; Sharma 2005), religion and caste (e.g. Kahn
1991; Coelho 1997), and education (Sharma 2005) are all potentially confounding factors
(Kandiah 1991) which do not allow past IE literature to be collectively reconciled into an
understanding of the structural nature of pan-IE features.
36
Examining the social nature of IE, sociolinguistic research and research on language
attitudes and larger societal ideologies in India has examined IE speaker identity (A.
Pandey 2004; Chelliah 2006), language practices and domains of use (Sahgal 1991;
Hohenthal 1998; Vaish 2008), language loyalty (Hohenthal 1998; Chelliah 2006), literacy
and the effects of language policy (Ramanathan 2005; Vaish 2008), similarities and
differences between local and global language ideologies (V. Chand in press), and
positive and negative associations with the speakers of different regional backgrounds in
India (Sahgal 1991; Baldridge 2002; Chelliah 2006). Importantly, akin to Preston’s work
on Americans’ conflicting views towards Southern English (under review), loyalty
towards IE cannot be taken for granted: examining Indian novels written in English,
Chelliah demonstrates that ‘[t]he very features that make Indian English an “Indian”
language are the features authors denigrate through repeated association with villains,
buffoons and losers,’ thus reinforcing IE’s ‘substandard’ status (Chelliah 2006: 8).
Baldridge (2002) also highlights a different aspect of this inconsistency between feelings
of loyalty and ownership: his informants discuss IE features, define IE as a part of ‘being
Indian,’ and understand the advantages that English offers in the global context, while
also expressing fear over the overwhelming effects English popularity will have on
vernacular languages.
Language loyalty also manifests itself through preferences for teaching models, and there
have been contrasting findings regarding the ideal model for English in India. First
examining proposals for teaching models made by linguists, Bansal (1978, cited in
Agnihotri & Sahgal 1985) recommends that a rhotic pronunciation model in schools
37
would to aid in international intelligibility, while Agnihotri & Sahgal have suggested the
opposite, that a non-rhotic pronunciation pattern, which they found in speakers from
more prestigious high schools, should be taught within poorer schools because ‘this is the
feature which marks the vernacular schoolchildren the most, and places them in a
disadvantageous position vis-à-vis the other groups’ (1985: 102). In addition to
recommendations based on linguistic behavior and international intelligibility, actual
preferences have been quantitatively studied: while Sahgal (1991) found that the majority
of informants preferred an indigenous IE model, Hohenthal found the opposite, that 70%
preferred RP, 10% preferred AE, and only 17% favored IE (1998). Loyalty towards IE is
complicated and clearly merits further research. Further, given Hohenthal’s suggestion
that inconsistencies in reported attitudes and loyalties are better understood when
examined in conjunction with domains of language use and personal language
background, research on loyalties and ideologies should be carried out in conjunction
with other data, and this nuanced combination of data is offered in this dissertation.
Individual attitudes, or personal feelings, are surrounded and supported by larger social
stereotypes, linguistic ideologies, which are themselves supported by official institutions,
popularly circulated, and officially subscribed to by government and education systems.
Language attitudes can thus be aligned with or resistant to institutionalized ideologies,
and expressions of personal attitudes and acknowledgement of larger societal ideologies
serve as naturalizing moves, wherein the beliefs come to be understood as ‘universally
true’ (Woolard & Schieffelin 1994: 58), and the relationship between individual attitudes
and societal ideologies becomes more transparent. Past research has demonstrated that
38
within the multilingual setting, linguistic ideology can influence linguistic change along
different paths, and has similarly demonstrated links to variation in patterns of language
acquisition, style-switching and change (Woolard & Schieffelin 1994). Encouraging for
the broader cross-cultural merit of the current research, ‘some of the most provocative
recent work on linguistic ideology… comes from studies of colonialism’ (Woolard &
Schieffelin 1994: 67).
However, complications arise from attempts to synthesize past IE analyses into similarly
larger sociolinguistic understandings of IE structure and practices. This is compounded
by the incomplete data on social attitudes, ownership, perceived vernacular and English
competency, patterns of language shift and maintenance, and domains of language use,
further motivating this project. A close analysis of these areas will illuminate the local
context and can reflect internal pressures both to support local vernacular languages and
identities as they interact with globalization, and to capitalize on English’s status within
this as a path towards power, status and financial security. As well, it may be that
different structural levels are differently impacted by perceptions of language ownership.
By examining generational changes across these areas, this research will highlight the
interwoven nature of such factors and offer a more nuanced understanding of how
changes in attitudes may become realized in bilinguals’ language choices, practices,
literacies, and fluency.
39
Postcolonial Sociolinguistic Theorizing
The framing for the majority of post-colonial sociolinguistic discourse is Kachru’s
Inner/Outer Circles paradigm (1985), which distinguishes Inner Circle European native
English speakers, Outer Circle speakers (ex-British colonies which have developed
nativized English varieties, recognized in modern discourse as World Englishes), and the
Expanding Circle, which primarily employs English in an English as a Foreign Language
(EFL) context. Problematically, Kachru’s tripartite paradigm of Englishes fails to
theorize language contact, language change or language socialization in the post-colonial
context. It offers no means of understanding the socio-political or linguistic processes by
which English becomes nativized, how these processes and the resulting Englishes may
vary across different post-colonial countries, or how Outer Circle Englishes differ
structurally from Inner and Expanding circle Englishes. It perpetuates the myth that
native English speakers exist only in Western, colonizing countries, while post-colonial
English-speaking countries are forever entrenched within and bound to their colonial
history. Terms used to delineate aspects of the post-creole continuum, e.g. acrolectal and
basilectal, have also been employed in academic research to post-colonial World English
varieties, as well as by non-linguists in India about Indian English, (e.g., Acrolect
Technologies 2009). While pidgins, creoles and World Englishes varieties are all
products of language contact, and research between these areas has often been dialogic
and intertwined (Schneider 2007), these terms have theoretical baggage, and are
problematic to deploy in explaining post-colonial linguistic landscapes that are not
clearly examples of creole development. Further, more recent approaches to pidgin and
creole studies have argued that creole situations are impossible to delineate as a class of
40
languages, and instead, are dialects of their lexifiers (Mufwene 2001), although this is not
widely accepted. These terms and this framework thus are not used within the current
study of native IE speakers.
Dating back to Saussure, the labels of ‘internal’ and ‘external’ elements of language
distinguish between a language’s structural system and everything else related to the
language (e.g. its history, context and speaker identity) (Saussure 1983 [1916]: 20).
Moreover, rather than just marking a distinction, the internal elements were valued over
external elements in the era of structuralism and generative grammar, and have only
become legitimate and critical areas of inquiry in linguistics in conjunction with internal
elements through early sociolinguistic and anthropological work by scholars such as Dell
Hymes (e.g., 1974), John Gumperz (e.g., 1961), Joshua Fishman (e.g., 1965) and William
Labov (e.g., 1966; 1972b; c). This has progressed to the point that it is now argued that a
complete understanding of language cannot be undertaken without addressing this social
side (Milroy, 2001). Focusing on variationist contributions to addressing society within
linguistics, Labov pushed linguists to believe that comparing variation across and within
social alignments can profitably contribute to theorizing and understanding diachronic
language change, while Hymes and Gumperz contributed in qualitative areas, by
introducing the ethnography of communication, the SPEAKING model, and their overall
emphasis on social and cultural knowledge as integral to language competency. Their
contributions are critical to the development of sociolinguistics; however, their
41
qualitative methods sometimes appear to be overshadowed in the field by the more recent
focus on quantitative variationist methods4.
While quantitative variationist studies are believed by many to be competently addressing
this intersection of social characteristics with ‘orderly heterogeneity’ (Weinreich et al.,
1968), Milroy (2001) criticizes the enterprise of quantitative variationist work for failing
to theorize social elements of language while also simplifying and relegating social
factors to a peripheral position in understandings of language standardization, language
change and practices. Quantitative sociolinguistics, entrenched within standard language
ideology, Milroy argues, will not be able to fully examine and understand sociolinguistic
prestige and standardization until it accounts for these external elements in a more
nuanced and theorized manner—in short, he is working to revitalize the qualitative
agenda put forth by Hymes and Gumperz.
However, Horvath and Sankoff’s research on the Sydney speech community offers a
important counterpoint to Milroy’s assessment of sociolinguistics’ social elements as an
oversimplification (1986). They work backwards, from patterns arising within language
to the social groupings of their informants, from which they demonstrate that basic
Parsonian social groupings (e.g. socio-economic class, ethnicity and gender) readily
channel into language practices (Horvath & Sankoff 1986). That is, by examining
structured variation in language practices without reference to social factors, divisions
between basic social grouping patterns emerged.
4
These themes are also taken up, critiqued, and elaborated on in a special issue of the Journal of
Sociolinguistics: (Bucholtz & Hall 2008; Eckert 2008; Heller 2008; Woolard 2008)
42
Variationists argue that the features which speakers demonstrate awareness of through
ideologies, by virtue of this conscious awareness, can only lead to sporadic and
haphazard changes, and will not demonstrate the large-scale, systematic changes found in
features below the level of speaker awareness (Labov 1966). However, this view has
been challenged in three ways. First, Silverstein (1979) theorizes that linguistic
ideologies and the multifaceted functional roles of language engage with language
structure diachronically, and are mutually impacting. Thus, to study the ‘total linguistic
fact,’ one must attend to the metalinguistic system of structural categories and diachronic
changes in categorization, and how these are semiotically tied to indexicality and
linguistic ideologies. Second, supporting Silverstein’s directive to integrate ideological
and social aspects of language use in context, diachronic change has been documented for
conscious, indexical linguistic features (Woolard & Schieffelin 1994; Wassink & Dyer
2004). Third, Dubois and Horvath’s (2000) analysis of the conflicting role of gender on
Cajun English development reveals a curvilinear (v-shaped) pattern of language change
in features above the level of awareness, which correlates significantly with changes in
language ideologies and sociohistory—that is, ideological difference across genders are
linked to language practices. The role of linguistic ideologies and awareness of language
variation on language change not only deserves reexamination within the context of a
community working to (re)define themselves in global and local contexts, but is a
theoretical imperative (Silverstein 1979).
Thus, operationalizing and incorporating lay-ideologies in conjunction with an analysis
of variation and change in IE features can further challenge variationist understandings of
43
ideologies as unrelated to extensive systematic change, will offer a counterpoint to
Milroy’s assessment of variationist work as perpetually marginalizing social factors and
failing to examine multilingual, non-standard, post-colonial (and hence, alternative)
marketplace contexts, and will work towards understanding Silverstein’s (1979) ‘total
linguistic fact.’
A more recent analysis theorizing stages of post-colonial English development—within
which, varieties are termed PCEs, or post-colonial Englishes—is offered by Schneider
(2007). He explores the impact of social, historical and ecological factors on PCE
development across the world, arguing that while PCEs have developed in unique
directions, they are fundamentally similar, and traverse the same developmental pathway
vis-à-vis the communicative needs of colonizers, colonists and other parties. The PCE
evolutionary cycle has five stages which are idealized as chronologically linear (although
stage boundaries are fuzzy, diverse characteristics of each stage may not occur
simultaneously, stages can overlap and/or co-occur, and the durative length of each stage
varies across contexts): 1) foundation, 2) exonormative stabilization, 3) nativization, 4)
endonormative stabilization and 5) differentiation. Within this dynamic model of PCE
development, while there are tantalizing hints that IE is approaching and/or entering
Stage 4, endonormative stabilization—e.g. endonormative attitudes are emerging, but still
contested— Schneider treats these as foreshadowing, not evidence that IE is in Stage 4.
This is because pan-Indian IE stabilization, codification and homogenization lag behind
(2007: 171-2). While useful to theorize shared developmental pathways for PCEs and
relevant to understanding the sociohistorical development of IE since its introduction in
44
India, this model is limited by its ‘big picture’ analysis of IE, for several reasons. First, it
does not address how specific IE communities may be further ahead or behind each other
in this cycle. Second, it does not address how the sociolinguistics of globalization, as an
emerging phenomenon (Blommaert 2003), may impact the PCE development process in
new and unexplored ways. Third, the utility of locating India as Stage 3 with precursors
of Stage 4 is unclear, given that regional differentiation (Stage 5) is clearly already
present in India (e.g., Agnihotri & Sahgal 1985; Wiltshire 2005)—in short, India has a
population surpassing one billion, within which any generalization has varying relevancy
across the multiplicity of Indian regional, socio-economic, and ethno-linguistic
communities. However, his synthesized incorporation of sociohistorical politics,
attitudes, ideologies and identity within dialect development lends further support to the
methodological and theoretical aims of the current project.
Language and Society in the Postcolonial Context: Problems and Challenges
Language ideology functions as a mediating link between social forms, power relations
and forms of talking, such that people enact what they believe to be ‘appropriate’ in their
speaking styles (Woolard 1986; Bauman & Briggs 1990; Woolard 1998). Particular
forms of language become associated with authority, and then work in collaboration with
authority to perpetuate a particular order, with social conditions, not language, primary,
and language deriving its illocutionary force from social conditions (Bourdieu 1991).
Within the global English context, loyalty to IE is conflicted (Baldridge 2002; Chelliah
2006) and its status contested (Schneider 2007), and IE thus functions as an alternative
marketplace for social value. A close study of this will refine and elaborate Bourdieu’s
45
(1991) linguistic marketplace to offer a more nuanced theory of the nature of alternative
marketplaces. Changes and stasis in power, agency and sociolinguistic identity can be
illuminated through sociolinguistic study, underscoring individual and group agency
within the larger linguistic marketplace (Woolard & Gahng 1990; Bucholtz 1999; French
2001; Mendoza-Denton 2002; Wassink & Dyer 2004; Blackburn 2005). Uncovering the
layered and hidden power structures manifesting in and mediating language practices
allows us to theorize post-colonial Indian English and the continuing effects of
colonialism in the contemporary world.
For instance, Kandiah (1991) highlights how English (and hence English speakers) in
India have been labeled everything within a continuum of fluency, from ‘non-native,’
‘not fluent,’ or even, ‘sub-standard’ with respect to English (e.g. Kachru 1983; 1986b;
Krishnaswamy & Burde 1998), to fluent native speakers (e.g. Hohenthal 1998; Baldridge
2002). The concepts of native and nonnative speakers are relevant here because native
speakers’ linguistic abilities are not judged in the same way nonnative are: value
judgments unidirectionally are placed on nonnative speakers’ speech in such
circumstances, such that their speech is viewed as lacking, or failing to reach the target
(Schiffrin 1994; Cameron 2001).
Indeed, ideologies about language surround us and invite us to see language and
linguistic variation in particular, stratified and commodified ways. For example, the
monoglot standard ideology (Silverstein 1996) interprets dialects as unequal, with a
single ‘standard’ and, conversely, multiple ‘non-standard’ dialects. Shuck, demonstrating
46
how verbal performance and storytelling by American college students serves to reify
Standard English beliefs, postulates an ‘IDEOLOGY OF NATIVENESS, that constructs the
category “native English speakers” as contiguous with Americans (and sometimes
British), and ‘nonnative English speakers’ as contiguous with foreigners’ (2004: 196,
emphasis in original). Within this, accents are interpreted as indicative of non-native
status and a lack of fluency, where speakers, by virtue of an ‘accent,’ are assumed to be
both non-native and non-fluent English speakers. Further, accents are understood as
incomprehensible, and it is common—and socially acceptable—in the standard language
context to complain about immigrants ‘who do not learn “the language”’ (Shuck 2004:
196). Here we begin to see a direct ideological link between accents, ‘foreigners,’ foreign
or ‘non-standard’ manners of speaking, and unintelligibility, fluency and nativity.
This hegemonic ideology is intertwined with social and linguistic conceptualizations of
‘us’ versus ‘others,’ understandings of (un)intelligibility, and the commodification of the
‘standard’ variety within the workforce (Silverstein 1996). Collectively, this dominant
ideology is naturalized and reified through speakers’ interactions with in-group members
versus others, as ‘semiotic movements of interpretation and construal…operate…in the
ongoing processes of production and comprehension of our own and of others’ social
behavior’ (1996: 295). These processes both create and perpetuate a particular social
reality, and can be understood as a form of erasure (Gal & Irvine 1995), where linguistic
differences are forgotten or made secondary, and presumed social differences related to
aptitude, education, and intelligence, for example, are fronted as explanations for
linguistic variation. ‘Through such naturalization, extralinguistic properties or attitudes of
47
individuals can be read in and from their participation in standardization… processes’
(Silverstein 1996: 290). Potential dialectal variation thus needs to be interpreted and
located within this social understanding of language power, economic influence and
standard language ideology, so as to not simply construe linguistic diversity as
ideologically neutral. Importantly, I have uncovered evidence that IE is ideologically
framed, and IE speakers are interpreted through this hegemonic lens by standard
language torchbearers and also, by (some) local IE speakers (V. Chand in press).
Returning to the Indian context, what constitutes a language’s native speaker is poorly
defined and means different things to different people: of relevance here, it is oftentimes
linked with the concept of mother tongue (Pattanayak 1998). Mother tongue, meanwhile,
is a term with multiple meanings which contrastively can index origin, identification,
competence and function5 (Pattanayak 1998). Defining native speaker qualifications have
thus proven tricky, and not without socio-political ramifications (Annamalai 1998). But,
by its very singularity, the term ‘mother tongue’ denies the possibility of mother tongues
and locates mother tongue as a monolingual property (Silverstein 1996).
Any quantification of English speakers in India is admittedly problematic, given the
varying means of measuring language use and fluency across consecutive censuses (U.N.
Singh 2006). Within the 1931 and 1951 Censuses, one’s mother tongue ‘was the
language first spoken “from the cradle,”’ while the 1991 Census defined mother tongue
as ‘the language spoken in childhood by the person’s mother to the person. If the mother
5
My informants understanding of mother tongue and related terms like nativity and fluency are explored
more fully in the qualitative analysis presented in Chapter 4.
48
died in infancy, the language mainly spoken in the person’s home in childhood’ (U.N.
Singh 2006)6. Beyond these cross-census differences, there are other possible scenarios
which are not even captured, e.g. the possibility of multiple caregivers using different
languages and whether the mother/caregiver is even communicating in their own mother
tongue(s) to a child. Further complicating the quantification of English speakers in India,
past Indian linguistic surveys (e.g. Grierson’s 1897 Linguistic Survey of India, the 1985
People of India Project, and several projects of smaller regional scope), have centered on
rural communities (which have fewer, if any, English speakers than large cities) and were
focused on quantifying, documenting and categorizing indigenous languages and dialects,
not ‘foreign’ languages like English (U.N. Singh 2006). Variable census questions and
definitions, along with linguistic surveys focused on rural indigenous languages, have
helped shape official and personal definitions of speakerhood and fluency in India. The
resulting census-based quantifications are hence questionable regarding their definitions
of the linguistic demographics, and more specifically, their assessment of the size of the
English speaking population in India7, and whether English can be a mother tongue in
this context.
In India, a critical view of such terms and their import in defining and delimiting groups
is being taken by the New Linguistic Survey of India (NLSI). The NLSI is comprised of a
consortium of universities and governmental agencies which will, through a pan-Indian
6
Hammering this point home, note the singular forms within these quotes “was the first language spoken
‘from the cradle,’” and “the language spoken in childhood by the person’s mother to the person. If the
mother died in infancy, the language mainly spoken in the person’s home in childhood.”
7
For example, equally problematic to take at face value are the mother tongue demographics for Delhi,
from the 1991 Census, which estimated Delhi’s population at 11 million, of which only 3622 have English
as their mother tongue (Vaish 2008).
49
linguistic survey over the next 10 years, provide a ‘a profile of the Indian linguistic space
in terms of the structure of its speech varieties, their distribution and their interactions’
which is motivated by a goal to support India’s standards of ‘responsible and responsive
governance, dignity and respect for diversity’ (Singh 2006). The NLSI advocates India as
responsive to, and, indeed, as a model for multilingual diversity. It confronts the
variability of meanings behind a mother tongue and implicitly downplays the importance
of labels like native speaker, instead embracing how ‘language has always been used
both for self-identification and as a resource for the construction of knowledge and social
capital’ (Singh 2006). This current day indigenous viewpoint is a response to and critique
of outside classifications and descriptions, which have fit less than perfectly. In this way,
the increased agency of alternative multilingual contexts like India to define significant
linguistic topics and create locally relevant classifications is an encouraging trend.
The Impacts of the Monolingual Paradigm
Definitions of the native speaker and standard language have by and large been
formulated and valued within western, monolingual settings (Milroy, 2001). These
definitions have traveled from western first world locales and now affect and categorize
English speakers in the third world. Beyond the actual definitions, the desire to delimit
who is a native speaker and what constitute standard tongues have filtered many speakers
out from consideration as ‘native’ English speakers: they are currently used to ‘explain’
multilingual contexts like India. However, the distinctions between native and non-native
speakers and standard vs. non-standard language may be less possible and have less
import in capturing something of social or measurable significance within these post-
50
colonial contexts. This is especially true given that nativization, claims to mother tongue,
and fluency are all legitimizing processes in claims to language ownership. This
process—of acquiring or establishing ownership—may run counter to reducing the
distance from other ‘native’ Englishes. How these two forces are reconciled by speakers
and how the different structural features are handled in such reconciliation are important
questions to be examined.
As well, there are heavy ideological, political and economic consequences associated
with applying such labels to an entire population, within which IE speakers’ individual
ideologies and speaker agency are relatively unexplored. The import of language
ideology vis-à-vis any given language is surely related to whether there is or is not a wellestablished standard language culture in place (Silverstein 1979; 1996; Schneider 2007).
Exploring these local attitudes and their dialogic interaction with global(ized) standard
language ideologies (c.f. V. Chand in press) will challenge standard language ideologies
of monolingual westerners as the sole repository of native English speech. As noted
above, indigenous Indian linguists are now addressing this by embracing how ‘language
has always been used both for self-identification and as a resource for the construction of
knowledge and social capital’ (U.N. Singh 2006). Within this, an alternative and related
conceptualization of language ownership, not tied to native speakerhood, and, instead
focused on how alternative marketplaces interact with other alternative marketplaces,
posits that ownership is the process by which subalterns come to ‘own’ their nonstandard
variety of English, regardless of their level of fluency (Canagarajah 2006).
51
Thus, while research internal and external to India increasingly is problematizing and
challenging the monolingual western first world speaker as the sole repository of English
ownership and nativity, there is still a large gap between theory and practice. We do not
understand how this actually plays out for IE speakers, how they understand nativity,
fluency, and the IE variety as intertwined and potentially constituting, creating or
contesting language ownership, and how these individual concepts and relationships may
be changing diachronically. My research addresses these lacunae.
Language Ideologies
Attitudes and beliefs towards languages and dialects, in conjunction with ones attitudes
towards the speakers of different language varieties, collectively comprise one’s
linguistic ideologies. As they are produced and potentially reified within sociolinguistic
interviews, these language ideologies can be both directed outwards, through
participants’ opinions and descriptions of international and intra-India sociolinguistic
groups, and internally-framed, as participants discuss their own linguistic preferences,
domain choices, fluency levels, future plans and feelings towards their linguistic
repertoire. These ideologies, because of their feedback loop with language practices,
wherein one enacts what one believes (Silverstein 1979; Woolard 1986; Bauman &
Briggs 1990; Woolard 1998), are critical to studies of language practices, especially in
understudied regions like India, where the potentially influential role of particular social
and ideological positionings is unknown. The incorporation of language ideologies and
practices within an examination of IE structural variation will highlight the
interrelationship between the external and internal elements of language integral to the
52
localized IE dialect. As well, this incorporation will offer clues as to the directions future
studies of IE should go, and what social and linguistic features are gaining in importance
within globally directed economies as both internal indexes of identity as well as external
indexes of English fluency.
Current Project
The role of the colonial background, globalization and modernity (Blommaert 2003;
Kristiansen 2003; Raj 2003; Lee 2006) on IE in a post-colonial context, as forces pushing
local varieties of English further from or closer to dominant standard English varieties (J.
Milroy 2001), is completely unexamined in India, though past work in other post-colonial
contexts have contributed more nuanced understandings of the interrelationship between
language practices, attitudes and ideologies (e.g., Woolard & Schieffelin 1994; Wassink
& Dyer 2004; Schneider 2007).
Specific to IE, there has been little to no discussion of contextually determined variance
in the native IE bilingual community, and quantitative work itself on IE is scarce. Studies
of IE speaker awareness of and attitudes towards IE (as a dialect) are both rare and
unsophisticated, and the relation between perceived IE features and actual IE speech has
not been well studied. Accompanying this, work on changes in IE over time, through
apparent time studies of people of different ages (Bailey 2002), is rare (e.g., Agnihotri &
Sahgal 1985), while no longitudinal studies have been undertaken. Earlier descriptive
work on IE is shaped by and functions within the ideology of standardization and hence
is not value-neutral. It has neglected a pivotal aspect of language, the ‘external’ or social
53
semiotics of language (Woolard 1986; 1998; J. Milroy 2001). Further, both across study
and intra-study variations in regional background, vernacular background, informants
country of residence (e.g. Baldridge, 2002; Pickering & Wiltshire, 2000; Sharma, 2001,
2005; Wiltshire, 2005; Wiltshire & Moon, 2003), religion (Coelho, 1997; Kahn, 1991;
Wilce, 1995), education, and caste (generally unmentioned, beyond Coehlo, 1997) are all
possible confounding factors (Kandiah, 1991) which do not easily allow reconciliation
across past IE literature.
Past IE research thus does not fully address how IE interacts both structurally and
socially within the milieu of World Englishes, the post-colonial context, and the current
larger sociohistorical context in which IE practices are embedded (Silverstein 1979). It
fails theoretically in generalizability with other post-colonial Englishes and descriptively
in the reliable representation of IE itself. A variationist study of IE will illuminate aspects
of IE’s internal structure and lend credence to IE’s status as native English dialect. As
well, it will allow us to directly study how post-colonial identity has and continues to
interact with colonially imposed RP, India as a linguistic area (Emeneau 1980 [1956]),
and IE’s role within an increasingly globalized Indian economy (Cowie 2007).
This research is thus centered on using sociolinguistic methods to theorize the postcolonial linguistic landscape, by focusing on IE speakers as social agents who, through
their language practices and personal attitudes, support and challenge larger socially
shared IE ideologies and understandings of IE status to create a local, nativized English
identity. Examining cross-generational, diachronic, and socially-linked variation in IE
54
forms, and related language loyalties and ideologies will all contribute to understanding
IE socially and structurally within the post-colonial Indian context as a World English
and to larger issues of post-colonial language practices, identity and alignment with
nativized Englishes.
Past IE language ideology studies (Sahgal 1991; Hohenthal 1998; Baldridge 2002) can be
faulted for their over-reliance on the written questionnaire and formulaic questions to
capture potentially conflicting and definitely complicated attitudes and beliefs, as well as
personal language choices and preferences which collectively comprise ideologies. These
mediums have been argued to be insufficient in their comprehensiveness, in particular
because ‘any kind of formal testing leads to an intensification of the effect of
observation’ (Labov 1972a: 126). However, not to throw the baby out with the bathwater,
language ideologies have clearly offered further and potentially different means of
illuminating and correlating data, if collected properly, through sociolinguistic interviews
(Woolard & Gahng 1990; Preston 2002).
Collecting attitudes through informal conversation, rather than written questionnaires,
allows participants the freedom to express conflicting viewpoints and understandings of
language choices and practices and frame the discourse in their own way, which
encourages informant agency and personally shaped identity and will offer new and
multilayered insights regarding IE speaker identity. As a first step, it is critical to link
expressed attitudes with actual language practices, a task seldom attempted, though
successfully accomplished in Wassink & Dyer’s (2004) examination of post-colonial
55
Jamaican English and Levon’s (2009) study of gay Israeli speech. This will provide a
framework for later investigation of how these attitudes play out in different settings of
peoples’ lives or across social demographics, for example, to examine how speech may
change between work and home, or the relation between age- or sex-based variations in
conjunction with IE speaker attitudes. These attitudes will thus be examined in
correlation with the results of a variationist analysis upon two structural features, as well
as speaker age, sex, and personal history. By investigating IE speaker attitudes and
structural variation, both as separate and as interrelated phenomena, this project directly
addresses important theoretical questions about the relationship between dialect
emergence, language variation and change and long term impacts of linguistic
colonialism in the post-colonial world.
Research Questions
The theoretical and methodological gaps in the past literature are my starting place for the
current project. Within this, my research focuses on upper middle class Hindi English
bilinguals (defined as speaking both languages before kindergarten and regular use of
both in life today) in south Delhi. Using early bilinguals with identical language
background will prevent this study from complications related to differences in fluency or
vernacular language background, both which are not covered in past literature (Kandiah
1991). Socio-economic background, education and place of residence have also been
demonstrated to produce variation (e.g. Figueroa 2004; Sharma 2005) and are similarly
accounted for within this target population. As well, these constraints on the sample
population, while limiting the potential generalizability of these results across diverse IE
56
populations, will allow comparison with earlier data, in order to tease apart whether
variation in a particular linguistic form is diachronically stable, and linked to age grading,
versus a change-in-progress, within which the direction of change and its relationship to
outside varieties of English will be of particular interest (Boberg 2004). Through
sociolinguistic interviews focusing on life experiences and designed to discretely capture
naturalistic IE features and attitudes related to IE awareness, English varieties, language
plans and locally perceived status of IE, I explore, analyze, and attempt to answer the
following research questions:
•
What are speaker attitudes towards and understandings of IE as a variety of
English, and how does their IE awareness correlate with their actual linguistic
repertoire?
•
What social factors, alliances and categorizations do these IE speakers bring up
when discussing IE, and how are these related to IE with respect to the ideology
of standard languages with native speakers?
•
How do speaker linguistic plans, domains of language use, and the perceived
function of languages within their multilingual linguistic repertoire interact both
with each other and with IE cross-generationally and across genders?
•
How do previously cited features of IE manifest in actual oral conversation?
Are they variable, what motivates this variation, and can they themselves be
quantitatively modeled?
•
How is IE structure changing diachronically (or, contrarily, what sort of age
grading is apparent)? Are globalization and increased access to RP and
57
American English (AE) media influencing IE towards assimilation with one of
the two, or towards increased distance from these dominant English varieties?
Changes in domains of language use, language ownership, structural choices and
perceived vernacular and IE competency reflect internal pressures to both support local
vernacular languages and identities as they interact with globalization, and to capitalize
on English’s status within this as a path towards power, status and financial security. By
examining diachronic changes across these areas in conjunction with the espoused
attitudes and ideologies of relatively homogeneous speakers from three generations, this
research will highlight the interwoven nature of such factors and offer a more nuanced
understanding of how changes in attitudes and perceived IE status may become realized
in bilinguals’ structural choices, domains of use, fluency and IE identity.
58
CHAPTER 3
DATA COLLECTION AND ANALYSIS METHODS
Data Collection Goals
The proposed research questions require examining the speech of multiple generations of
fluent IE speakers. To this end, in-depth exploratory sociolinguistic interviews with three
generations of participants8 were conducted, with two goals: 1) collection of extended
naturalistic oral IE data and 2) elicitation of informants’ linguistic attitudes, ideologies,
language related plans, awareness of IE as a dialect, and descriptions of stereotypical IE
features. Sociolinguistic interviews, supported with a background in sociolinguistic and
anthropological methods, theories and analysis tools are the only means of collecting and
correlating language practices and ideologies within the local context.
Target Informants
This study targeted a sample of linguistically, socio-educationally and regionally
homogeneous informants. With a goal of interviewing 32 upper middle class Hindi
8
Several terms (e.g. informant, participant, consultant, talker) are in popular circulation, and the differences
they may circumscribe are not entirely clear. I use the terms informant and participant interchangeably. In
some instances, when referring to the sample population as a whole, or to target demographics, informant is
more appropriate, while in other instances, e.g. when highlighting how individual opinions have shaped my
thinking about particular concepts, the term participant is more comfortable, and suggests an analysis
arrived at through mutual interaction within the interviews.
59
English early bilinguals in Delhi, India, data was actually collected from 35 individuals
from 2007-2009. Data was collected from women and men from 18-87 years of age, in
order to examine apparent time variation (Bailey 2002), to restrict for potential
confounding from an overly heterogeneous population, and to permit a real time study of
variation as potentially a change-in-progress vs. age-grading, through comparison with
past quantitative work within this population (Agnihotri & Sahgal 1985). The sample was
chronologically continuous, and did not seek to create age ‘gaps’ which may prove
significant in understandings of processes of age-grading or apparent change-in-progress.
However, four age ranges, detailed below, suggest themselves as a means of delineating
groups, given their links to historical moments of major social change, and attention will
be paid towards how these historical and social links align with other factors in the
subsequent chapters.
It is hypothesized that the following four age-delineated subsections of the sample will
capture people with different types of histories, having lived through different times and
educational experiences. The oldest generation (~70-90) was raised and educated under
the colonial system before the 1947 Partition of the Indian sub-continent into India and
Pakistan; the next youngest generation (~38-69), some recent grandparents, with children
in high school through their early 30’s, was the first generation raised and educated in
newly post-colonial India. Their children comprise the final two age-divided groups:
next, ~28-37 y.o. participants were raised and educated before the economic growth and
opening of India’s borders to western market influences and are currently working postcollege. Finally, current college students and recently employed graduates (~18-27)
60
comprise the final group, and have been raised and educated in a globalized setting
permeated with western television and print media, cell phones and the internet. While
comparing the age-delineated sections will highlight how changes in historical setting
may influence language practices and attitudes, capturing a continuous range of ages will
also allow a more nuanced understanding of the transitions between particular historically
situated settings.
Very briefly sketching out the participant demographics within the larger Indian context,
these families are characterized by their social and work networks, their international ties,
and their linguistic experiences. The three younger generations of men work in fields like
law, government service and private business, and a portion of these women similarly
work in these areas, as well as teaching, while a majority of the eldest generation of men
held positions of rank within the Indian military, and a majority of the eldest generation
of women did not work outside the home. The two younger generations are largely
geared toward working in private business and law, and both men and women are either
already working or have plans to work upon completing college. Marriage is not expected
until after men have attained some measure of financial independence, typically not
before ~28. Households are often ‘joint,’ within which multiple generations live
collectively, often with 2-3 servants for cooking, cleaning and chauffeuring. While worklife is often considered an English domain, home-life is characterized by more
English/Hindi mixing, and exclusive Hindi is used with shopkeepers and servants.
Families typically have at least one member settled abroad, in Europe, North America or
large Asian cities like Hong Kong and Singapore, although different generations’ links to
61
these emigrants are varied, as are their experiences visiting abroad. Collectively, the
target families are representative of modern, urban middle and upper class Indians in
several ways, and the results of this project, while clearly not representative of the
potentiality of IE more broadly as including rural, lower class speakers from different
language backgrounds and varying degrees of English fluency, will provide a framework
for further study of urban, globally linked IE speakers.
Researcher Influence
As a person of Indian origin striving to understand my family history, I have become
intimately familiar with India’s colonial and political history as it relates to my family’s
personal sociolinguistic trajectories; as a sociolinguistic anthropologist I have
supplemented this with comprehensive readings on the structure of IE and World
Englishes situated socio-politically in the post-colonial context, case studies of language
ideologies, and structural and ideological changes related to language practices. As I will
demonstrate, my background did not hinder the interview process, and in fact, facilitated
it in interesting ways.
Cukor-Avila and Bailey (2001) hypothesize that differences between fieldworker and
informant characteristics (e.g. race, gender) may compound upon themselves with respect
to the observer’s paradox (Labov 1966), thus undermining the validity of interview data
as representative. However, their quantitative analysis found no significant differences
which can be attributed to the race of the fieldworker. Instead, they found that the degree
of familiarity between fieldworker and informants was a stronger predictor of informal
62
speech (Cukor-Avila & Bailey 2001). Familiarity can be developed and conveyed in
many ways. Three factors substantially mitigated any potential problems related to my
American background and AE accent: (1) AE media influence and regular contact by IE
speakers in the target demographic with speakers of other English dialects, through the
diaspora Indian community and international employment based interactions, which
collectively promote familiarity with AE, (2) natural accommodation in my speech
towards IE features while in Delhi, and (3) my confirmed community membership.
Community membership within the target population is confirmed or denied (CukorAvila & Bailey 2001) through the following practices: local knowledge of shops and
marketplaces off the beaten track, comfort level with driving in Delhi, bargaining skills,
and stomach (i.e. one’s willingness to eat food cooked and sold on the streets, indicative
of one’s acclimation to Indian germs). I have developed these by annually spending time
in Delhi, both visiting relatives, and living alone. My age and gender (30 and female,
respectively) are further characteristics which could influence the interviews: the age and
gender of both participants in any interaction is likely to have some effect on the
interview process. However, the potential effects, given that these differences are uniform
across the interviews, are not strong enough to invalidate such a means of data
collection—it is doubtful that age and gender differences between the informant and
interviewer will affect the three phonological variables under investigation here, nor do
they pose an ethical or access issue.
63
Participant Recruitment
Participants were recruited by exploiting already existing social networks, using the
‘friend of a friend’ method to make initial contact within the community, and outside of
my existing Delhi social network (Milroy, 2002). The network approach to studying
variation is advantageous: it is a participant driven concept capable of dealing with
individual variation without reference to predetermined social categories (Milroy, 2002:
556). As well, it ensured a similar level of familiarity with all participants: familiarity has
been argued to be a ‘crucial factor affecting linguistic behavior in interviews’ (CukorAvila & Bailey 2001). Recruitment was done through an emphasis on interviewing
people related through first order ‘strong ties,’ those connecting friends and family, rather
than ‘weak ties,’ those connecting acquaintances (L. Milroy 2002): several of my
participants were interconnected through various social, religious and work circles and
extended family networks.
Urban living is characterized by the absence of deep interpersonal relations with every
interlocutor one encounters. This project is not a small community study, but is instead
focused on urban IE speakers interacting with and influenced by the globalized market. A
certain amount of familiarity is possible, and spending extended time with my
participants did, of course, enhance familiarity, but more normal, and a hallmark of
urbanity, is a certain level of social distance, even in most urban Delhi friendships, within
which friendships don’t necessarily extend into all the depths and reaches of life. As well,
extended interaction in different contexts, as I observed, is more possible with women of
all age groups and with men of my age group and older, while regular conversational
64
interactions with men that were younger than me proved artificial and difficult to sustain.
These gender and urban-based social constraints motivate the decision to center this
analysis in the interview setting, and on life histories, rather than through ethnographic
observations.
Participant Population
The IE data was collected between August 2007 and April 2009 within two fieldwork
trips and consists of thirty five interviews conducted with participants ranging from 18 to
87 years old. As is often the case, not all of these interviews have proved useful. Within
the ‘friend of a friend’ methodological approach to finding participants, it was
occasionally socially awkward to refuse to conduct interviews with people recommended
by past participants (typically older family members who were thought to have a ‘better’
understanding of English, or a ‘better’ or ‘more interesting’ life history), even if the
speaker did not fulfill my basic requirements (e.g., they were not from south Delhi, from
a different socio-economic background, or were not native English speakers). However, I
did include speakers with knowledge of more than simply Hindi and English: this is
discussed in the next section. I use all of my speakers’ data for qualitative analysis, while
the quantitative variationist analyses draw on data from 29 of these speakers. Each
interview lasted between 45 and 150 minutes, with the average around 75 minutes, and
each speaker signed an informed consent form prior to beginning the recording.
65
Negotiating Participant Background
I had four criteria in selecting informants for this study. First, I sought to interview
people born and living continuously in Delhi—‘Delhiites.’ Second, I sought to interview
only Hindi/English early bilinguals, because, in much of the past research on IE, there are
issues of generalizability wherein it isn’t clear whether research has captured native IE
structures, as opposed to contact induced features. This is because of India’s linguistic
(both in what languages informants are fluent in, and in their age of English acquisition),
regional, and cultural diversity. Third, I sought informants with whom it would be
possible to interview multiple generations of IE speakers within families. Fourth, I sought
to undertake data collection through the ‘friend of a friend’ method, which, although
beneficial in terms of the quality of the data, can also be problematic in that it limits the
social circles to those one is connected to via friends and family.
These criteria together proved difficult to complete in tandem, for several reasons.
Addressing the first criterion, identifying as a Delhiite is not without other social
implications. Addressing the first and second criteria in tandem, Delhi-born early English
learners born pre-Partition are rather elusive. Addressing the second and third criteria,
given that the eldest generation was typically born outside of Delhi, several participants
had additional languages in their repertoire. And, addressing the fourth criterion, this
method of meeting participants resulted in a larger population of older men with military
background than is reflected in any of the successive generations. These are all discussed
below, and require a departure from methods per-se, to briefly introduce and explore
issues of participant self-defined identity, Delhi socio-political history and local
66
understandings of Delhi and Delhiites vis-à-vis the Hindi-speaking belt which surrounds
Delhi. These will allow a better understanding of the rationale for the participants chosen
for the study, and the demographic and historical backdrop against which they are,
indeed, English-speaking Delhiites.
Complexifying ‘Delhiite’
Delhi is the capital of India, but before the 1947 Partition, Delhi was very different from
what it is now. Pre-Partition Delhi was a provincial, Hindi speaking town, not an urban
metropolis with a thriving population. My participants made clear that it is still not
known as a cultural mecca within India, and instead, is considered the seat of the
government, and a city overflowing with immigrants from the surrounding areas who
have come to Delhi to make money, or simply make it. For example, Delhi was described
as follows:
Delhi is really the center of all the neighboring states because everybody comes here for a
job eventually, and most of the immigrants are male simply because, you know. Outside
Delhi, it’s a way different universe. (M28KC 8:29-38)
And it's just that the original Delhi people are no longer here. I mean, I'm old so I also came
twenty five years ago. But, you know, like they say the original Delhi-wala, it's full of people
now from all over. They're rude. They have no sensibilities or sensitivities, and they’re, you
know, they’re -- they don't care about what’s happening in the neighborhood. They -- they
are, I don't know, very aggressive… (F52GG 4:21-25)
It’s a city of con man. @@ See this, all this culture is wasted on them, all this history is
wasted on them. They don’t look at things, they don’t see things, I can not really blame Delhi
for it, because you know, these hack populations moving in and out of Delhi for God knows
how many hundreds and thousands of years, so Delhi never really had a time to you know
make a culture of its own in that sense. Also it happens to be the political capital of the
country so all lot of hobnobbing, all that stuff, happens around there. That kind of seeps into
the place…(F39VD: 2:41-3:1)
The surrounding areas, typically referred to as the Hindi Belt (see Fig. 3.1), are associated
with poverty, backwardness and less education, and these characteristics are often
attributed to the larger Delhi population through the infusion of rural people and through
the further adoption of these characteristics by those residing in Delhi.
67
Figure 3.1. The Hindi Belt as encompassing the four states of Rajasthan, Uttar Pradesh, Madhya Pradesh and Bihar
(Tourism 2006: highlighting of states mine)
f73TM describes how she has changed, and now mirrors Delhi behavior and attitudes9:
I don’t like Delhi – what should I say? But Delhi taught me a lot of things, Vineeta. I
remember when I first came here, I was Bengali. So some of the things are inbred in me –
good behavior, not to be too aggressive, don’t use awful language or – or, or, or, particular
language which would hurt people… I was waiting there, and naturally my face and
everything got very, very different. And one lady came, a Delhiite. She gave me the first
lesson how to live in Delhi… she said this is not the way you can – you can live in Delhi.
You have to be very, very aggressive, otherwise you won’t survive here, and she laughed… I
became like a [HINDI: chuha rat], like a rat… I was hiding behind her. That – that the way I
wasn’t brought up. So this is atrocious for me. Now, you’re telling me, Mrs. M., what do you
do? I do that same kind of thing. I give a tight slap because I remember, my car was hot,
9
Hindi was occasionally used within the interviews, both by participants and myself, and my transcription
scheme—both as it relates to codeswitching and more generally—is detailed later in this chapter.
68
pushed. Three boys. I got down. Said auntie, auntie, please, please – then I said get down.
And gave that boy a tight slap… (f73TM 4:18-45)
In this respect, Delhi is home to my participants, but their attitudes towards Delhiites are
on average very negative, and it is common, within this socio-economic bracket, for some
speakers to align themselves more closely through ethno-linguistic background and
through their families’ historical origins to other regions, while others embrace a Delhiite
identity, perhaps because of the socio-political upheaval of Partition which motivated the
elder generations’ relocation to Delhi. For example, some participants don’t consider
themselves Delhiites, regardless of having been born and brought up in Delhi, while
others, after only 10-15 years in Delhi, consider themselves Delhiites. While the next
chapter explores identity as a multifaceted phenomenon, and the following two
quantitative chapters explore the relationship of various regional and ethno-linguistic
identities to structural variation, this area clearly needs more research in order to
understand how and whether the speech of Delhiites varies with respect to their selfidentification as Delhiites, as well as their out-of-town familial links.
Birthplace of the Oldest Generation
Within my population, feelings were also mixed with respect to the notion of elderly,
Delhi indigenous, early Hindi-English bilinguals. Many participants found it humorous
that I was looking for upper-middle class Hindi-English bilinguals with no other language
background in the 70+ year old age group who had lived continuously in Delhi, claiming
that such a population doesn’t exist. While they turned out to be wrong in the specifics (I
did find exactly one participant who fit this), they were correct in their overall
generalization: having a native background in English limited the oldest generation in
69
other ways: it subselects for an upper class, urban, outward looking population who did
not typically grow up in Delhi.
Delhi has changed drastically since the oldest generation was born in ~1920. PrePartition Delhi was not a large, nor an affluent city, nor were there large or prestigious
universities in Delhi pre-Partition: instead, at that point, Lahore and Bombay were
educational and cultural epicenters, where one would be more likely to find more diverse
social stratification. Importantly, it is only within such epicenters of cultural and
socioeconomic diversity that one would find a community of early acquisition English
speakers fifty years ago. Pre-Partition Delhi as a historical setting was provincial enough
that learning English at home at a young age was very rare.
As well, the Hindi-speaking belt, which better reflects the economics and social values of
pre-Partition Delhi, is not considered the most prosperous group, and is often considered
more insular or culture-bound, and hence less likely to encourage higher education or
learning English, especially amongst its females. For example, my participants discussed
the Hindi-Belt in the following manner:
So the caste system and everything is, you know, is more prevailing there in UP and smaller
towns and all. (f25GS 8:14-15)
This Hindi belt of people, there’s a particular attitude they have, you know, which is very,
very regressive. (f32NM 7:39-40)
They, they all think, that the Delhi people are completely insensitive, and you know, I think
the basic problem in Delhi is that there is very little respect for women. And that’s where -what really angers me. And I think—and I think to an extent, it's quite true. Because I don't
see that when you travel—you travel to the east—eastern part. You travel to the south. You
travel to the west. It’s not like that. I think it's this Hindi—this Belt, this Hindi Belt, you
know, of UP, Haryana, and all these places where the women has no—no status at all.
(f52GG 15:10-16)
70
Delhi-based families that were upper middle class three generations ago and speaking
English at home are thus an oddity, given the social, political and migration based
changes which have created post-Partition Delhi. Immediately before, during and after
Partition (1947) is the period when Delhi received a large influx of immigrants, for a
variety of social, religious and economic reasons. Thus, all but one of the speakers from
the oldest generation was born outside Delhi. They were dominantly from Lahore and
surrounding towns in Punjab, a region which, since Partition, has been divided and rests
partially in India and partially in Pakistan (see Fig 3.2). All of my older participants now
consider themselves to be Delhiites. Meanwhile, the Punjabi population, comprised of
both Sikhs and Hindus from Punjab, has considerable cultural and political influence in
modern Delhi: indicative of this, the current prime minister of India, Manmohan Singh, is
a Sikh Punjabi.
Figure 3.2. The Punjab Region, currently divided between India and Pakistan (Ktims 2006)
71
Bi- vs. Tri-/Multilingual Participants
A third confounding problem with my original ‘ideal informant characteristics’ was my
goal of collecting data from bilinguals, without any tri- or multilinguals, which was
problematic for very similar reasons. Given that almost all of my oldest generation
emigrated from pre-Partition Punjab (since divided into two states: Pakistani Punjab and
Indian Punjab) quite a few of them grew up with Punjabi as a third (and sometimes
mother tongue10) language. To find a population of 65-80 year old Delhiites who were
both born in Delhi and who had no experience with languages other than Hindi and
English was thus very difficult, and wouldn’t represent, if one traces the later generations
of their families, the community of Hindi/English bilinguals currently in Delhi. Thus, the
eldest generation diverges the most from my idealized target population, but was chosen
because they are the parents and grandparents of my younger participants, today’s South
Delhi upper middle class Hindi/IE population.
And, as we will see within the upcoming discussion of language domains (Chapter 4), the
oldest generation has retained their Punjabi language, and speak it with their peers and
sometimes with their children. Punjabi is losing ground with successive generations,
however, and none of the third generation (the two youngest groups interviewed) are selfdescribed as fluent Punjabi speakers. Thus, in the interests of exploring longitudinal
changes in English across generations which are representative of the actual upper middle
class south Delhi English speaking population, it was necessary to refine my original
10
Participant notions of ‘mother tongue’ are discussed more fully in the next chapter.
72
population characteristics to include Punjabi/Hindi/English elderly speakers who passed
on only fluent Hindi/English practices to their children.
The Oldest Generation and the Military
Post-Partition Delhi, as the capital of India, is the seat of both government and the
military. Many military families displaced during Partition chose to settle in Delhi after
retirement, especially in areas like Defence Colony, a neighborhood of South Delhi
where a good portion of my data was collected. The neighborhoods I collected data in
were a necessary precondition of my familial and friends contacts, and given that a large
portion of my family has been involved in government and military, my participant pool
reflects these ties. There is nothing inherently wrong or problematic about examining
families with ties to the military; however, it does mean that more of my participants had
spent two or more 3-year spans outside of Delhi, in military postings, often within India,
but occasionally abroad. Indeed, if one were to exclude families with military ties in
Delhi, it would severely reduce the potential informant population, and create other dataoriented problems. The upper ranks of the Indian military, from my participants’ points
of view, were dominated by English use at work and in the home, and children of these
upper ranked military personnel were educated almost exclusively in English. For
example,
[B]ecause most of the service officers, they speak mostly in English. Ninety percent of the
service officers and the wives and all the children always speak in English. That’s the best
way to communicate. (f64AG 19:7-9)
In this sense, these children (now adults) are ideal participants, beyond their childhood
travel, given this strong English influence.
73
Participant Breakdown
As detailed in the previous sections, the proposed speech community being examined
here clearly morphed from its planned form. These changes and choices were based on
both the larger goals of the project, and the restrictions and modifications that arose
through participant interpretations of their community and through social, economic and
historical realities: early English acquisition subselects for an eldest generation of
speakers born outside Delhi; this itself selects for multilingual speakers, not simply
Hindi/English bilinguals; this modified oldest generation is also subselected because they
are the parents and grandparents of my idealized younger generations of upper middle
class early English/Hindi bilinguals from South Delhi. These changes do not limit the
research questions proposed here, but instead, more correctly circumscribe the
community of which one can ask these questions (Paolillo 2002: 36). Table 3.1 illustrates
the participant breakdown numerically:
Data
As noted above, within the in-depth exploratory sociolinguistic interviews, there were
two dominant goals: 1) the collection of extended naturalistic oral IE data and 2) the
elicitation of input regarding linguistic attitudes, personal language related plans,
awareness of IE as a dialect, personal fluency and literacy assessments, language domain
preferences, and descriptions of stereotypical IE features. Collectively these represent
several aspects of their broader linguistic ideologies. These interviews are conceived as
conversations, and steps were taken to reduce the formality and encourage the informants
to speak freely.
74
Although sociolinguistic interviews are the classic and most common method of
sociolinguistic research (Feagin 2002; L. Milroy & Gordon 2003), and are thought to
approach, if not actually be ‘natural’ speech, this has been problematized with the
Observer’s Paradox (Labov 1972c), and the genre of sociolinguistic interviews has
subsequently been examined to better understand its effect on speech (Allbritten 2006).
Pertinent to this project’s goal of informal conversation, quantitative analysis reveals that
sociolinguistic interviews do create a setting in which the interviewer and interviewee are
both ‘insiders,’ (Allbritten 2006), and hence less formal.
Table 3.1. Participant Demographics
Social
Grouping
Number of
Speakers
(N=35)
Example
Gender
Female
Male
Age, by generation
Youth
Middle
Generation
Oldest
Generation
Age, by Sociohistorical Stage
18-27 years old
28-37 years old
38-69 years old
70-90 years old
Ethno-linguistic Background
Bengali
20
15
6
13
16
8
6
11
10
2
UP/Haryana
4
Delhi
Punjabi
Mixed
2
21
4
Other
2
Punctuated
Continuous
21
14
Delhi Stay
Originally from West Bengal or Bengali regions
of Bangladesh
Originally from northern states of UP and
Haryana
Including areas now in Pakistani Punjab
e.g., 1 parent from north India, 1 from south
India
e.g., from eastern or central states
75
In addition to the informal interviews, two other types of oral data were targeted for
collection immediately after the informal interview: a formal reading and a structured
retelling of a story. Within the formal reading, I asked participants to read the
Grandfather Passage (Appendix A), a passage designed to elicit all English phonemes in
each word position (word initially, word finally and word medially). The Grandfather
Passage can be used to capture and compare phonological variation across dialects, and
allows a point of comparison within this data set for IE phonetic analysis and inter-IE
phonetic variability related to attention to speech, which can help in uncovering prestige
variants and directions of change.
For the structured retelling, participants were shown the ‘Pear Story’ film, a short five
minute film with a basic, culturally neutral11 plotline, sound effects, but no words. This
movie was designed by Wallace Chafe for eliciting language samples from around the
world, in order to understand language variation and narrative structure (Chafe 1975;
2007). It has been successfully used to compare across and within multiple languages
(Chafe 1980, and chapters therein; Erbaugh 2001, and references therein), including IE
data collected from speakers currently residing in the US (Seale 2007), which will allow
for direct comparison and potentially, for further corroboration of past findings within the
local Delhi context.
11
Admittedly, ‘culturally neutral’ is an oxymoron, but this movie is regularly described in these terms. The
plot is about a man picking pears, within which a boy steals a basket of pears, rides off on a bicycle,
crashes, and is helped up by some other boys, who each take a pear and then walk past the pear-picker as he
realizes that one of his baskets of pears is gone.
76
Having informants retell a story allows one to collect naturally produced data from
multiple subjects on the same topic. From these oral retellings of the film’s plot it is
possible to understand syntactic, lexical and morphological aspects of Indian English, and
how these structures may vary internal to Indian English, based on informant age, gender
and formality. After viewing the Pear Story, participants were asked to retell the plotline
in their own words. Past projects centered on Pear Story retellings have had informants
watch the movie as a group, without the interviewer present, and have then collected the
narratives through individual interviews of the following form:
Within 5-25 minutes afterward, they are interviewed individually in a different room. The
interviewer is a young woman native speaker of the same social background. Speaking the
target language, she says to each informant, 'you have just seen a film. But I have not seen it.
Can you tell me what happens in the film?' If questioned further, she says, 'just describe what
you saw. There are not any right or wrong answers.' Most speakers tell the story quite
naturally, taking around two minutes. Each description is audio or videotaped. (Erbaugh
2001)
In the current project, the form of the narrative elicitation was modified, to reflect that the
movie was shown within the context of a larger one-on-one interview, within which I
watched the film with my participants, at the end of the interview. The narrative
elicitation thus was ‘Can you tell me, in your own words, what happened in the story we
just watched? There is no right or wrong answers.’ However, given that I also watched
the film, and hence can act as a judge regarding the completeness and veracity of such
retellings, this context raises the formality of the response above that of casual
conversation. Indeed, as I will discuss in my quantitative analyses, watching the film with
participants provoked even more formal speech than the Grandfather Passage. In
conjunction with the reading passage (discussed next), this allowed for the collection of
three levels of formality from each participant.
77
The Grandfather Passage and Pear Story thus provide a means of collecting data with
varying degrees of formality, which, pre-interviews, were hypothesized as
(Interview<<Pear Story<<Grandfather Passage), but in reality manifested as
(Interview<<Grandfather Passage <<Pear Story). Labov (1966) demonstrated, in his
study of r-pronunciation in New York City department stores, that there is a positive
relationship between the degree of formality—that is, speakers attention to their own
speech—and informants’ use of prestigious linguistic forms (Labov 1972c), and this
relationship has been replicated in numerous other settings. Collecting data with multiple
levels of formality thus offers additional insight into prestige and vernacular forms, and
their continuum.
Interview Topics
Within an interview, the goals of collecting both ideological data and naturalistic speech
can at times be at odds, because asking a particular set of questions can introduce
undesired formality to the interview setting, thus increasing the effect of observation and
rendering the collected speech more deviant from ordinary conversation. There are
several ways to mitigating this while accomplishing both goals, delineated in Labov
(Labov 1972c:117), which aid in capturing excited (within which formal constraints are
overridden) and, hence, casual speech (within which formal constraints are set aside).
First, certain themes have proved fruitful in past projects, e.g. a) danger of death, b) sex,
marriage and courtship, and c) evoking moral indignation through gossip and discussions
of false accusation, collectively subsumed within Labov’s Principle of Interest and
78
Involvement (1972a: 118-122). I explored some of these within these interviews;
specifically I invited discussion of the latter two topics. The first was avoided because it
did not fit with the theme of the interview—in order to interview people, I (often) had to
explain ‘why’ I wanted to do these interviews, both to the informants and to others who
introduced me to informants. My ‘cover story’ was that I was interested in Delhi, and
hearing about individuals experiences in Delhi, and how people from different
generations might feel differently about Delhi12. Of the latter two themes, discussing
people’s recollections of their marriage proved fruitless13, while I was able to
successfully evoke discussion regarding moral indignation about certain behaviors (e.g.
fake accents, Delhi driving and the degeneration of ‘English’ in youth’s slang-filled
speech).
Second, one should learn from the responses in early interviews and incorporate and
expand on the successful topics within later interviews, subsumed within Labov’s
Feedback Principle (1972a: 122). I discovered several interesting and fruitful areas of
discussion within the early interviews, and applied them in later interviews with limited
success. I learned through this application that the success of questions varied depending
on the age of the participant, and while discussing so-called ‘fake accents’ was very
profitable with the two younger groups, discussing language plans, globalization and
‘good’ or ‘important’ aspects of life was more profitable with the older generations, who
12
Importantly, this cover story allowed for a range of tangential topics and illuminated local feelings about
Delhi, ethno-linguistic identity and the Hindi Belt, all of which proved central to understanding and
interpreting their language practices.
13
Interestingly, I think that my role as an ‘insider’ made discussion of Indian wedding ceremonies—as
something an insider would know about—less possible. This question was often dismissed, for example,
with “It was a typical traditional Indian wedding.” (m64NS 7:42-3).
79
were perhaps more eager to use the interview as a process of reflection on their own
lives.
Third, the organization and presentation of the interview is important: one should develop
an ‘interview schedule’ (in contrast with a set of formalized interview questions), within
which a set of themes are organized into modules, modules connect to each other in a
shared, overlapping network, and modules related to language ideologies are arranged
later in the interview (Labov 1972a: 123-5). My interview modules included personal
background, family, language plans, domains of language use, life in Delhi,
English/Hindi ideologies, dialects of India, language diversity in India, and Indian
English (for an expanded list, please see Appendix C). Although the personal life history
topics may appear as a marginal aspect of the interview, they often were the running
theme of the interview, with different points in individual life histories serving as
jumping off points for discussing other modules. Organizing the interview in this format
allowed me to keep the modules in my mind and navigate between them without
introducing the formality of constant reference to any papers or pre-formalized questions.
Interview Relationship
Interviews were conducted as informally as possible, and, as explained above, although I
had a preplanned range of topics I was interested in, I allowed the interview to be guided
by the direction of the conversation and the interests of the participant. Although one goal
of sociolinguistic interviews is to collect as much informant speech as possible14, I
14
Bayley offers a quick means of assessing the quality of the interview: pick a few random spots within the
interview and listen to who is talking (participant vs. interviewer). If the participant is talking each time,
80
negotiated this goal with the need to reciprocate within the interview. By having
questions be one-directional (i.e. the interviewer is the only one ‘allowed’ to ask
questions), and thus establishing a clear interviewer and interviewee relationship, the
interview can become increasing unequal, and this unequal relationship has been
problematized (Briggs 1986; Rickford 1997).
I attempted to create a situation within which the participant also felt comfortable enough
to ask me questions about my life and experiences: this created a situation more closely
resembling a natural, informal conversation and also, in some ways, mitigated my
concern that I was only ‘taking’ from my participants, and not ‘giving,’ because it
became a context of sharing. Within these conversations, I found that people were often
interested in my experiences as an Indian female in the US, and were interested in talking
about similarities and differences with their experiences within India. I also found that
younger people and women in particular were very interested in collaboratively
discussing what it is like to be young, urban and negotiating Indian culture and how it
both conflicts with and complements having a modern outlook and lifestyle. Considering
the nature of the relationship between researcher and informants within this project, I
have evidence that my participants found the conversations both interesting and thought
provoking, which encourages me to believe that the situation was not entirely one-sided
in its utility.
this is a good sign that one both elicited enough speech for quantitative analysis and captured more
colloquial speech than a formal interview (personal communication).
81
Post-interview, I have received emails from one of my participants (f50RV) continuing
the discussion started within the interview; this email conversation is developing into a
friendship. Other participants have asked me to visit when I am back in Delhi, further
evidence that these interviews and the interactions within them was not viewed hostilely,
and contrastively, as the basis for a continued friendship. Another participant, m18DK,
asked for a copy of the interview, so he could listen to our conversation at his leisure and
think about the questions and his factual responses about his family and his attitudes
towards language. After that, realizing the benefit that participants may find in these
recordings, both personally, and for posterity (many of the children and family of my
older interviewees commented that they would like to hear about their elders’ life
history), I offered copies of the digital interviews, burnt on to audio CD’s, to my
participants: several took me up on this and some families, within which I interviewed
multiple generations, were eager to listen to and compare their interviews.
These recordings can thus be considered a resource for cultural and historical
maintenance, as well as a direct example of something useful which I was able to offer
back to my participants. Whenever possible, I also helped my participants: one was very
interested in getting a digital recorder similar to mine, and I helped her to find a store in
town where she could get one; another talked about her fondness for green tea, and I have
since sent her some special toasted sesame green tea which I’m fond of, and which I had
told her about within the interview. In short, my relationship with several of these
participants is continuing, and is not one-sided, and while this work has not yet produced
82
results which may contribute to the layperson’s understanding of IE structure and related
ideologies, the interview process has proved mutually profitable and pleasant.
Aspects of Recording
Interviews were conducted either at my family’s house, in a quiet room, or at each
participant’s house, in their living or bedrooms (post-college adults preferred to be in the
living room while the college students seemed to prefer to conduct the interviews in
‘their’ space, namely, the bedroom). Interviews were recorded digitally, via a Olympus
Digital Voice Recorder WS-300M connected to two lavalier (tie clip) omni-directional
microphones (Radio Shack #33-3013), designed to be unobtrusive while collecting
quality audio material. Ironically, the biggest issue consistently faced within the
interviews was from ambient noise. While I was typically able to situate the interview far
from noisy windows, the weather proved to be almost insurmountable.
Delhi monsoon weather is notoriously unpredictable and extremely hot and humid. As
such, there were a variety of ambient sounds (ceiling and standing fans, wall-installed AC
units, and swamp coolers) which I had to contend with in my attempts to get high quality
recordings, in addition to the street sounds from living in the middle of a busy metropolis
(whistles, horns, squeaky vehicle brakes, vendors, dogs barking), now an unremitting
aspect of life in Delhi, which the microphones also picked up. Often it was impossible to
eliminate all sounds during the course of the interview, and my strategy was to reduce
ambient sounds through my placement of the clip-on lavalier microphones. Before
starting the interviews, I conducted several tests in different types of rooms with each of
83
the major sound types—by recording speakers talking while I moved the microphones
different distances closer to and away from their mouth and at different angles (e.g.
pointing upwards vs. sideways) I attempted to locate the ‘sweet spot’ for capturing the
most relevant acoustic data while minimizing outside noise pickup.
These recordings were then examined within Praat, an acoustic analysis software
program (Boersma & Weenick 2006), to both determine whether the noise could be
reduced through strategic microphone placement and which noises created more ‘noise’
within the visual presentation of the acoustic waveform, key for later acoustic analyses.
The lavalier microphones, ostensibly omni-directional, actually capture quality audio
input and less ambient noise when clipped parallel to the ground near the collar, and this
setting was used within the interviews. I did turn off all ambient sounds (fans, AC, etc.)
for a very brief portion of each interview, while participants read the Grandfather
Passage. This was done so at least a portion of the interview had high quality acoustic
data for later acoustic analyses.
After collecting this data, and given the issues surrounding ambient noise, I would
recommend that future studies in Delhi (and similar climates) take into account the time
of year they are planning to conduct interviews: November through February are likely to
be the easiest months in which to elicit speech for detailed acoustic, phonetic and
phonological analysis in Delhi.
84
Target Data
Broadly, the choice of structural variables was guided by three principles: they should be
frequent, structured, and socially meaningful (Labov 1966). In conjunction with this,
features which have been studied in other contexts are beneficial, in that they allow crossdialect comparisons (e.g., Santa Ana 1993; Santa Ana & Bayley 2004) and contribute to
understanding potential linguistic universals and constraints on variability universally.
While similar formal constraints for the choice of qualitative ideological areas do not
exist, several aspects of language background, plans and ideology have been
demonstrated to correlate with language production across different language settings
(e.g. J.R. Edwards 1982; Ryan, Giles & Sebastian 1982; Sahgal 1991; Hohenthal 1998;
Swigart 2001; Baldridge 2002), and these were likewise taken into consideration. Thus,
several structural variables and aspects of language ideology were highlighted before the
fieldwork commenced as potentially profitable for analysis, and steps were taken to
ensure that they were present within the interviews. These are outlined below.
Structural Variables
Potential structural variables of interest were chosen through two steps: first, a thorough
review of the literature on IE structural characteristics, summarized in Appendix B, was
conducted in order to understand what IE features have been noted and potentially
disputed in past literature. Second, these potential variables were then compared to
variables which have been studied cross-linguistically, or quantitatively within IE, in
order to cull variables for which cross-linguistic or diachronic comparison is not possible.
Two structural features were thus targeted for collection from the interviews and then
85
15
quantitatively modeled: (r) , (v) and (w) pronunciation. While these variables and the
general methods within quantitative variationist work will be discussed briefly within this
chapter, the full details on coding choices and methods can be found in the following
chapters, presented with the results.
(r)
Varieties of English are often termed rhotic or non-rhotic, with postvocalic [r] deletion
(e.g. [fɛdɜ] for feather) varyingly related to both prestigious and stigmatized forms, in
different contexts. While (r) deletion is stigmatized in much of the United States, it is a
feature of the prestige form in RP (Trudgill & Hannah 2002). It is variable based on both
social and linguistic features, although some studies have argued that a finer distinction,
between trill, approximant and null, is necessary (e.g. Sharma 2005). The following
social factors have had significant ties to the rate of (r) deletion in other communities:
age, sex, ethnicity, degree of formality, integration into SAE speaking communities,
regional background, socio-economic status and occupation (Labov 1972c; Myhill 1988).
Several linguistic factors have also demonstrated a relationship to [r] deletion: the
presence of another vocalic [r] in the same word (e.g. quarter can manifest as [kwɔtr̩]),
following syllable boundary (which interacts with vowel quality) or pause, syllable stress
(which interacts with vowel type), following word-boundary-plus-vowel, preceeding
vowel quality, and [r]’s status as the nuclear vowel (e.g. in bird) (Myhill 1988).
15
Parentheses are used to capture a variable with multiple realizations, e.g. (r).
86
In IE, r-pronunciation is analyzed in several contrastive ways. It is considered
nonexistent, rendering IE a non-rhotic dialect (Nihalani, Tongue & Hosali 1979: 211;
Sailaja 2009), present and environmentally conditioned, rendering IE a variably rhotic
dialect with linguistic constraints considered the primary motivation for alternation
(Bansal 1990; Gargesh 2004), or socially variable and indexing young educated females
with more years of English-only high school instruction (Sahgal & Agnihotri 1985; 1987;
Agnihotri 1994; Trudgill & Hannah 2002: 130; Sharma 2005: 208; Wiltshire 2005: 282).
This third scenario suggests that IE rhotic patterns are most strongly correlated with
social features. As well, (r) deletion has been quantitatively studied in the target
population (Agnihotri & Sahgal 1985), and, through comparison with the current data,
will permit diachronic examination of whether variation in rhoticity is a change-inprogress, is actually stable, or evidences age grading. In this case, structural variable
selection and informant population characteristics have both been influenced by the goal
of examining potential diachronic changes in IE. A full description of the rhotic coding
methods are presented in Chapter 5, along with the results of the analysis of (r).
(v) and (w)
Folk analyses of Indian English (IE) features often reference /v/ and /w/ as potentially
merged or haphazardly mixed (examples referenced within Vaid 1977; Sharma 2005;
Mahapatra 2006; Cowie 2007), while my IE speakers self-identify a v/w ‘mix’ as
characteristic of IE:
I am quite aware of it when I speak to somebody who is so-called native speaker of English,
so a Britisher, an American, a Canadian, an Australian… If I am sitting with a bunch of ten
Indian friends I’d probably be making my W’s …V’s. (m35MS, 3:43-4:5)
87
Awareness of this variable realization extends outside India, and is linked to ideologies
about ‘standard language’ (V. Chand in press). For example, India and US-based accent
reduction classes target IE v/w as a ‘problem’ for intelligibility within the American
business world (Accent Reduction Institute 2008), and AE media portrayals of IE speech,
through imitations, capitalize on this supposed ‘mixup’ to index IE speech. Acoustically,
while some research has suggested that IE has a single phoneme, the voiced labio-dental
approximant /ʋ/, which subsumes the two phonemes, /w/ and /v/ found in AE and RP,
different explanations have been offered for the actual linguistic behavior, ranging from
explaining it as a stable merger, regionally variable, environmentally conditioned
allophones or free variation.
The single methodical—albeit via aural impressions, not acoustic, analysis—IE v/w study
found the merged phoneme [ʋ] to be stably present in the speech of upper middle class
New Delhiites across two age groups (teenagers and 40+ year olds), speech styles and
language backgrounds (Sahgal & Agnihotri 1988). As well, other survey work has
identified the v/w phenomenon as a feature exclusive of English speakers from South
India and Bengali regions (Chaturvedi 1973; Cowie 2007): this phenomena is thus
potentially not found in South Indian and Bengali regions, though acoustic and/or
variationist methods have not yet confirmed this. This potential merger or ‘mixing’ is
clearly a salient IE feature both within and outside of India (V. Chand in press), and
worthy of examination within the quantitative variationist paradigm. A more detailed,
acoustically focused and quantitative analysis of the New Delhi population was hence
was undertaken, and is explored in Chapter 6.
88
Language Ideologies
The dialogic role of linguistic ideologies on language structure, while ignored in most
analyses of IE thus far, cannot be underestimated within the quest to understand the
nature of language as a social semiotic system (Silverstein 1979; Wassink & Dyer 2004;
Schneider 2007). Here, I detail several aspects of language background, plans and
attitudes which have correlated with language production across languages and settings.
Language attitudes and ideologies have both been profitably studied at individual and
societal levels, respectively, in a variety of post-colonial contexts to better understand
language shift and language revitalization efforts (Hoare 2001), changes in language
prestige and RP hegemony due to globalization (Bayard et al., 2001), Malaysian English
dialect recognition, evaluation and ownership (Newbrook 1998), shifts in language
attitudes, repertoire and use of the Hong Kong Sindhi and Sikh communities (Detaramani
& Lock 2003), changing social status of languages through loanword use (Ngom 2003),
and dialect recognition and perception (Preston 1999). Socially shared ideologies can also
change over time in their correlation with particular features: in Corby, UK and Kingston,
Jamaica, phonological variants previously stigmatized were adopted by younger
generations, having been reanalyzed to index different identities (Wassink & Dyer 2004).
Within the Indian context, sociolinguistic research has examined IE speaker identity,
language practices and domains of use, language loyalty, and positive and negative
associations with the speakers of different Indian regional backgrounds (Sahgal 1991;
Hohenthal 1998; Baldridge 2002; A. Pandey 2004; Chelliah 2006), never, however in
conjunction with a quantitative study of diachronic change in IE structure. This research
89
approach, by combining qualitative analysis with quantitative analysis of variation and
potential diachronic change in IE structure (e.g. Sharma 2005), is very rare in the field,
(e.g. Schecter & Bayley 2002; Zhang 2005; Llamas 2007), and until now, has not focused
on early bilinguals currently residing in India of any language background. The interview
techniques and specific questions from many aforementioned studies were reapplied to
the local context, to tap issues of dialect awareness, language practices, identity,
ownership and prestige for a IE as a World English, and attitudes towards the vernacular
and international varieties of English. A full list of the interview modules is available in
Appendix C, a list of the qualitative codes used during the analysis is available in
Appendix D, while a full description of methods and the resulting analysis is presented in
the next chapter.
Analysis
Transcription is the first stage of analysis, where oral speech is translated into written
text. Once completed, the interviews have been used in two ways: for their form
(quantitative analysis of the linguistic realizations) and content (qualitative analysis of the
attitudes, social identities and ideologies put forth within the interviews). Linking the
forms and content with social demographics across four age groups through standard
variationist techniques (Young & Bayley 1996; Bayley 2002) captures diachronic
changes in IE form and their relation to potential changes in ideologies.
90
Transcription
The interviews yielded a total of 49 hours of data. Transcription was then necessary, in
order to convert the data into a format more amenable to further coding and analysis.
However, ‘writing is not a direct representation of speech so much as a model of
language more generally. This model exerts a strong influence on our perceptions of what
language is or should be’ (Cameron 2001: 40). Thus, transcription itself is the first stage
of analysis, requiring many decisions on what to represent, what to exclude, and how to
represent language in a manner both useful to the analyst and faithful to the original oral
interview. These choices occur at the interactional level, including speaker interactions,
interruptions, pauses, backchannel responses and background noises, at the extended
utterance/discourse/speaker turn level, including utterance boundaries, intonation
contours, and pauses, and at the utterance-internal level, with choices in how to represent
stress, prosodic and intonation contours and emphasis, mis-pronunciations and dialectal
variations at the lexical level, lexical and non-lexical fillers, incomplete words and the
phonetic realization of language (Kerswill & Wright 1990; Du Bois, Schuetze-Coburn,
Cumming & Paolino 1993; J.A. Edwards 1993; Cameron 2001; Leech 2005).
Transcription as Translation and Transliteration
It is quite common to reflect on how narratives and discourse change in translation from
language to language and from oral to written (Chafe & Tannen 1987; Bauman & Briggs
1990; Rubel & Rosman 2003). However, the dialect-to-dialect conversion process, in
light of the unequal power relationship and uncertain ideologically understood status of
IE for its speakers and the outside world, also is worthy of consideration (Cameron 2001;
91
Fox 2004). Given that this work is centered on the enterprise of examining and promoting
the dialectal status of IE, based on oral speech, it is also important to consider how these
(potential) dialectal differences are represented within the transcripts. Are these
transcripts, then, on top of being transliterations, also translations? Do they transmute
oral, dialectal IE into a standard written English, and if so, what are the implications of
such a translation process on the ensuing analysis?
Translation is not without ideological underpinnings, and can often reflect conscious or
unconscious attempts to manipulate the text, and to encourage or diminish feelings of
distance and ‘otherness’ from the narrators (Rubel & Rosman 2003: 6). There is also
evidence that translation processes are not equal across languages: Liu argues that
languages with less socio-political clout are forced to give up more in the process of
translation, rendering the assumed cross-linguistic commensurability questionable (1999:
35). As well, she argues that noting something as different trivializes and reduces a
language’s value (Liu 1999: 21). Similarly, Cameron warns against using nonstandard
spellings to represent vernacular speech, because of the potential implicit reinforcement
of ‘stereotypes of nonstandard speakers as illiterate buffoons’ (Cameron 2001: 48), and
Fox directly wrestled with this in his depiction of Southern working-class speech:
My disciplinary commitments to linguistic anthropology, sociolinguistics, and
ethnomusicology made it imperative for me to represent the speech of my interlocutors (in
most cases) with an accurate attention to the local dialectal features of grammar,
pronunciation, and lexicon, as well as the prosodic structures of line breaks, emphasis, tone
of voice, sentence intonation, etc. This creates, however, complex problems of
representation, I am concerned here not to play into stereotypes of working-class speech as
‘lazy,’ ‘ungrammatical,’ or ‘substandard’ in any respect. (Fox 2004: xiv)
In collusion with this, there are already negative associations tied to encoding IE dialectal
features in writing: repeated from above, Chelliah has demonstrated that ‘[t]he very
92
features that make Indian English an “Indian” language are the features authors denigrate
through repeated association with villains, buffoons and losers,’ thus reinforcing IE’s
‘substandard’ status (2006: 8).
It may help to examine what sorts of transformations take place from oral IE vernacular
to the written form. Within this context, syntactic differences between IE and SAE are
brought to the forefront, while prosodic, phonetic and phonological differences are muted
(Bakhtin 1981), intentionally or otherwise, with these transcripts. However, conflicting
with this are the stated goals of this project, to document structural variation internal to
spoken IE and to understand features which IE shares and doesn’t share with other world
Englishes and internationally prestigious English dialects. How, then, can one represent
IE oral interactions so as to not perpetuate the dominant conceptualizations regarding the
‘incorrectness’ of IE and not trivialize or devalue IE structures, while not muting the
distinctive structural qualities of IE? These two, it turns out, are not mutually
incompatible, although they do require some active work on the part of the reader of such
transcripts. While such features have been used in literature in conjunction with
pejoratively viewed characters (Chelliah 2006), these transcripts will be analyzed by a
trained sociolinguist cognizant of, but not guided by, the past negative associations with
IE features.
In conjunction, one mode of analysis undertaken here—quantitative variationist
modeling—is itself a more impartial approach within which structural differences are not
interpreted as errors and are, instead understood as a separate language system. Within
93
such, by interpreting phonetic and phonological variables as conceptual and ‘structural
units’ (Chambers 2003), data—here, transcription—must be regularized to reflect the
theoretical precondition that variation is evidence for a single system. These transcripts,
thus, attempting to encapsulate a single system, hence, do not detail surface level
phonetic and phonological variation.
While the history of translation theory has reflected several concerns and understandings
of the ‘translatability’ of text, given the goals of this project, it is important understand
that transcripts will have departures from the original oral speech at several levels, and
hence, it is pivotally important to both ‘double-check’ the correspondence between the
original recordings (Nida 1964 cited in Rubel & Rosman 2003) and the transcripts and to
return to the original audio recordings throughout the analysis process.
As well, these recordings include occasional examples of Hinglish, a participant-reported
term for the Hindi-English codemixing indicative of the young, urban generation of
Indian English speakers, and examples of Hindi. While translation can render all of the
utterances into a single language, what is lost in this process? How would one determine
which, if any, or all, switches (or mixes) are meaningful, and what their meaning is?
Hindi and Hinglish structures fall outside of the scope of the present study, but will be an
interesting topic for future exploration, while changes in their roles and domains, vis-àvis IE, are explored in Chapter 4.
94
Utterance-internal Transcription Choices
Specific to this context, this project requires details at the phonetic, phonological and
morpho-syntactic levels for the quantitative analysis, and, conflicting with this, also
necessitates a transcript which is easily readable (i.e. not bogged down with cumbersome
symbols and coding), for the ideological analysis. In addition, these interviews were
collected with several goals in mind which extend beyond the confines of the current
work. I plan to draw on this data for analysis at multiple levels (phonetic, phonological,
morphological, syntactic, intonational and discourse) in the future. However, transcribing
with an eye to all of these levels is problematic: until one knows exactly how the data
will be used, overly-detailed transcription can have diminishing benefits (Cameron 2001:
46). For example, phonetic transcription takes an inordinate amount of time, requires its
own decisions to determine the narrowness of transcription, and is far less readable than a
transcript with traditional English spelling, punctuation and spaces between words. In this
vein, Leech makes a strong case that the original transcript should be free of annotations,
given multiple longitudinal uses for such corpora (2005). The current project is also
constrained by time: transcription is a slow process, and takes longer as more detail is
included. In conjunction with my future plans for analysis of these interviews, these goals
and constraints have guided the development of the current transcription policies, detailed
below.
Within the context of Indian English structure, diachronic variation, language plans and
ideologies collected within one-on-one sociolinguistic interviews, I am not particularly
interested at this point in interruptions, utterance boundaries, pause length and non-
95
lexical fillers. It has been hypothesized that IE may have intonation contours and
prosodic features not found in other varieties of English (e.g. Bansal 1990; Agnihotri
1994)16, and these prosodic and intonation contours may prove useful in characterizing
IE. However, despite the existence of programs devoted to characterizing and capturing
intonation patterns (e.g. the ToBi system, McGory 2006), personal experience has shown
that they are problematic to encode within transcripts. There exists no intra- or interdialectal unified opinion on the size or quality of a universal set of meaningful contour
shapes or how best to deal with contours, e.g., as a continuum, as fuzzy categories, or as
radial categories with clear boundaries. A pilot study would thus have to be done in
advance of coding a large set of IE data, as a dialect whose intonation contours have not
been quantitatively studied. Given this, prosodic and intonation features are not
transcribed, though they were taken into account in the quantitative analysis, as potential
preceding and following environments by returning to the audio recordings.
Non-lexical fillers have been shown to be linked functionally to turn taking, discourse
management and signals of comprehension, among other things, in American English
(Ward 2006). However, I do not plan to explore their potential utility in distinguishing
English varieties within the current project, and they have thus not been transcribed.
Utterance boundaries have been used within analyses seeking to quantify information
density in discourse (Kemper, Greiner, Marquis, Prenovost & Mitzner 2001; Baynes,
Farias, Chand & Bonnici m.s.). In conjunction, typical sentence boundaries have been
16
A complete list of past references and findings on intonation and prosody in IE is offered n the IE Feature
Chart, Appendix B.
96
used for a variety of purposes including evaluating completeness and utterance type
(Kemper, Kynette, Rash, O'brien & Sprott 1989; Kemper et al. 2001), however, these
typical sentence boundaries have proved very problematic in their application to oral
speech (Chafe & Tannen 1987; Bonnici & Chand 2007). I am not concerned, here, with
attempting to quantify information density, evaluate utterance completeness or categorize
utterance types, and have hence not used utterance final punctuation (e.g. .,!?) for any of
these specific purposes. I am interested in phonological environment, but given that
allophonic variation requires returning to the transcripts, once coding decisions are made,
in order to collect and code detailed descriptions of the phones in question, their
surrounding environments, stress patterns, intonation, etc., annotating utterance
boundaries was judged to not serve a useful purpose here.
Pause length has been quantified to help identify utterance and intonation boundaries
(Chafe 1994). As well, it has been compared cross-linguistically to understand different
culturally or linguistically motivated speaking styles, and how pauses of varying length
can be contrastively interpreted as hesitations, as turn completion, and lack of fluency.
Disputes have arisen over whether outside measures of pause length should be used,
versus speaker internal relative pause estimates, given differences in speaking styles and
rates, further complicating option of codifying pause length. Given that pause length is
not integral to the forms being studied, and is not of particular relevance to this project,
pause length is thus not encoded nor distinguished within these transcripts, although
unmeasured by apparent pauses were marked with ellipses (...), to improve readability.
97
These transcripts are formatted linearly (J.A. Edwards 1993), with each speaker on a
separate line, and include line numbering, in order to create a readable transcript within
which it is easy to return to the same spot in the transcript—see Fig. 3.3, as an example of
the transcript format. Within turn utterances are not separated, being unimportant to this
analysis, and orthographic conventions (e.g. sentence initial capitalization, contractions
like didn’t, wanna, and gonna, dashes to show a change in topic or revision and ellipses
to show a noticeable pause) have been followed to aid in readability (Du Bois et al. 1993;
J.A. Edwards 1993; Cameron 2001). Transcripts are represented word-for-word (e.g.
filling in missing words or changes to grammar have not been made to further ‘improve’
readability). Potential lexical fillers (e.g. I mean, you know) have been included from the
audio recording, as they can have various contextually driven meanings, beyond simply
as a ‘filler’ (Chun 2007). Any Hindi speech is transcribed phonetically, then translated
into English, and demarcated with brackets, e.g. [HINDI hain yes]. Unintelligible areas or
areas of questionable intelligibility are similarly labeled, with [unintelligible]. Laughter is
coded with ‘@,’ with each symbol coding approximately one syllable of laughter. Nonlexical sounds, such as environmental noises, sneezing, sighing, etc. are not transcribed
unless they are directly commented on by one of the speakers, and in these instances they
are coded. Informant identity is protected by reducing all personal participant references
to initials, e.g. ‘Mr. Singh’ as ‘Mr. S,’ while my name is used, when evoked. The
transcripts thus serve as a rough guide to the audio file, and have been used to create
more specific transcripts or coded sets of data for analysis in several veins. All excerpts
from transcripts are labeled in the following manner: individual speaker code
98
(gender.age.initials), transcript page(s): line(s), e.g. f32NM 7:14-18 is an extract of lines
14-18 on page 7 of the transcript of the interview with NM, a 32 year old female.
Figure 3.3 A visual and textual example of transcription
35
36
37
38
39
40
41
42
43
44
CK:
VC:
CK:
VC:
CK:
VC:
CK:
[HINDI hain yes] Because every – this paper comes in Hindi, whatever is – from
asperity, from this and that, everything there is Hindi. So I’ve study Hindi.
Mm-hmm. And your husband was also a Punjabi speaker?
Yes, he is Punjabi. @@
Uh-huh.
From Lahore. That’s his photo. Very handsome man.
Yes, he is. And what – was it an arranged marriage?
No. We knew each other somehow. Then my parents also, they knew – Lahore was
small place. [HINDI tho then] my father, when he was in service, he knew my fatherin-law. My father-in-law was his boss, my father’s boss. (f80CK 5:34-44)
Transcripts were initially created, double-checked against the audio files, then question
areas were triple checked, and, finally, each entire transcript was reviewed for coherence
during the initial qualitative analysis. These transcripts thus are remarkably consistent
with the audio files, given the transcription constraints laid out above.
Quantitative Analysis
While a variety of statistical tests might be used to examine linguistic variation, many are
unsuitable for the types of data collected in sociolinguistic interviews. In particular,
ANOVA tests are ‘designed to deal with the kind of balanced data that emerge from
controlled experimentation’ (Young & Bayley 1996: 254), while linguistic variables are
not balanced across conditions, rendering their analysis, through such procedures,
problematic. VARBRUL (Sankoff, Tagliamonte & Smith 2005), a specialized application
of logistic regression, has been designed to model the naturally occurring linguistic data
found in variationist studies, and has been successful in a number of contexts (Paolillo
2002; Tagliamonte 2002). Regression models test the strength of association (or
correlation) between the dependent variable and one or more independent variables
99
(factors in VARBRUL), while logistic regression allows for a test of multiple
associations with very uneven data sets. Hence, VARBRUL and ANOVA belong to
different types of statistical models. Quantitative variationist sociolinguistic techniques
(Bayley 2002; Paolillo 2002; Tagliamonte 2006) and the VARBRUL program were thus
used to analyze language variation within this project. In addition, the phonetic variation
inherent to both variables required, in conjunction with VARBRUL based coding and
analysis, acoustic analysis in order to determine the quality of the target variable.
Acoustic analysis was done within Praat (Boersma & Weenick 2006), an acoustic
analysis software program. Finally, analysis of (v/w) required exploring the role of word
frequency, for which AntConc (Anthony 2007), a concordancer software program was
used.
The coding system for alternative realizations and environments of each variable is
necessarily different across variables, given that different processes are likely to be
affecting the acoustic quality of the (r) and (v/w) variables. However, the target structural
linguistic variables were all coded for multiple factor groups: their form and function,
phonological, morphological and syllable stress context, degree of formality and speaker
demographics (Tagliamonte 2006). Within this, the Pear Story narrative and Grandfather
reading passage ensure that all interviews shared at least two identical topics and one
identical passage for direct comparison. Collectively, the variables were coded with the
goal of testing ‘the likelihood of co-occurrence of a variable form and any one of the
contextual features’ which arise from the interviews and informant demographics (Young
& Bayley 1996: 253). Care was taken to avoid creating and coding for social categories
100
with structural zeros, wherein some cells cannot be filled because of preconditions on the
code categories (Paolillo 2002:69-70). For example, if one was to code for the
relationship of the informant to the interviewer and informant gender, this would lead to
structural zeros, because there will be empty cells for some combinations, e.g. ‘male’
plus ‘mother.’ Within this population, age and occupation can create structural zeros, e.g.
there will be no 18 year olds who are retired, as can the combination of gender and
occupation, given that no men had the occupation of ‘housewife.’ This was avoided in
this project by creating a factor group which combined age and occupation, within which
occupation was defined in gender neutral terms. Some internal constraints created
structural zeroes which are best dealt with on a case by case fashion, and are discussed
more thoroughly within the following chapters. Details of the variable specific linguistic
factor groups and the social categories are found in Chapters 5 and 6.
Qualitative Analysis
While anthropologists have contributed much to theorizing ideology, in practice it has
largely been considered epiphenomenal to language production (Woolard & Schieffelin
1994: 58), and operationalizing it within variationist or anthropological work is relatively
rare (Ito 2001)—though see Silverstein (1979). Linguistic ideologies, domains of
language use, and meta-knowledge of IE have each been studied, albeit in a less nuanced
fashion, e.g. Hohenthal (1998) used an online survey with preset possible answers, while
Baldridge (2002) examined college-age IE speakers residing in the US. This project, by
both comparing across generations and genders, and using a systematic but loose
interview format, is better able to handle and account for contradictions, contrasts, and
101
gradations within espoused attitudes, and better able to understand diachronic changes
and gender-based variation. This project will thus serve as a guide for future projects in
linguistic anthropology and sociolinguistics that aim to correlate attitudes by
operationalizing language ideologies and linguistic attitudes. In the current study,
attitudes, as well as speaker age, sex, and personal history, are examined in correlation
with the results of a variationist analysis upon the two structural features.
Qualitative data was coded using Atlas.ti (Muhr 2004), software designed for organizing,
coding and analyzing large quantities of qualitative data with project-specific hierarchies
of codes. Atlas.ti works within Grounded Theory (Glaser & Strauss 1967), which allows
a project to be data driven and malleable in its conceptual and coding framework, and the
ultimate analysis: codes are created as necessary, and linked to particular quotations, with
the relationships between codes emerging during the process of coding.
In practice, this consists of the following. The choice of codes was both data driven and
theory driven—topics which have demonstrated important links with language practices
in past literature were examined within this data and the remainder of the qualitative
topics were developed directly from the interview modules and the responses. The
interviews were revisited and from emergent topics and evaluations espoused within the
interview responses, as well as the major interview modules, multiple levels of codes
have been developed. These working labels were then narrowed and hierarchically
organized (Appendix D), in order to develop a framework within which language
attitudes, language plans, domains of language use, and meta-knowledge of IE can all be
102
captured, coded, and compared across the different generations and genders to better
understand age- and gender-based variation. Given that attitudes are not binary, and can
be expressed along relative continua (e.g. positive to negative for attitudinal; constant to
zero use for language domains, etc.), each quote is both thematically and subjectively
grouped.
By using this type of searchable, hierarchical project-generated coding system, the results
of this research are, first and foremost, fully data driven, which is especially important
when examining an understudied and multifaceted linguistic setting. Further, it is
possible to link the qualitative results to the results of the quantitative analysis, in order to
understand the relationship between professed language attitudes regarding preferences
and alignment with different English varieties and actual linguistic practices. The
theoretical and methodological implications of this approach are discussed in the next
chapter, Chapter 4.
The interviews were coded for over 320 features, not all of which are discussed within
this analysis, for reasons of space. Codes were organized into 16 conceptual networks:
Demographics, Educational and Family Background, Language and Cultural
Transmission, Gender, Indian English, Key Concepts, Language Attitudes, Language
Domains, Language and Globalization, Language and India, Current Language Levels,
Diaspora Reflections, Regional Sociolinguistic Attitudes, Regional Language Variation,
Speaker Stance Variation, and Random. These supercodes and their respective codes are
discussed more fully in the following Qualitative Chapter, and a full breakdown of the
103
qualitative codes used is presented in Appendix D. We will now turn to the qualitative
and quantitative results.
104
CHAPTER 4
SOCIOLINGUISTIC IDENTITY, IDEOLOGIES AND ALIGNMENTS
Introduction
In the introduction of this dissertation I suggest that the future of sociolinguistic research
involves—necessitates—among other things, an incorporation of qualitative and
quantitative data. I view this incorporation as a methodological tool, wherein qualitative
data can uncover local alliances and social groupings that are then used within
quantitative variationist analysis, and, recursively can then be reapplied to qualitative
data, wherein social factors that are quantitatively significant in separating subcommunities of speakers can then be explored and teased apart through qualitative
reflections. This incorporation is also a theoretical tool, given the hypothesized
relationship between language ideologies and language structure (Silverstein 1979).
Further, including qualitative data permits local speakers agency in defining themselves
in their own words and allows them to challenge and negotiate how they are framed by
outsiders, both in and outside of academia. Further, incorporating such data allows one to
approach central concepts—like fluency, nativity, dialect and language—as nuanced,
potentially conflicting and importantly, politically charged terms which separate and
hierarchically stratify communities.
105
Building on all of these possibilities afforded by qualitative and quantitative research, this
chapter presents a qualitative analysis of several key sociolinguistic concepts through an
exploration of individual and shared explanations of language practices and attitudes, in
order to offer a more nuanced understanding one setting of Indian post-colonial language
practices, identity and alignment with nativized Englishes. Collectively, these can
illuminate how these IE speakers negotiate the need for both a localized linguistic
identity and a code mutually intelligible and economically profitable with other Englishes
in the global context. As I will demonstrate, variation in language domains, lay-attitudes
and perceived vernacular fluency reflect internal pressures both to support local
vernacular languages and identities as they interact with processes and endstates of
globalization (V. Chand in press), and to capitalize on English’s status as a path towards
power, status and financial security.
While much of variationist research has learned from and adheres to the findings in
Labov’s Department Store Study (1972c: Ch. 2), his study of Martha’s Vineyard (1972c:
Ch. 1) offered a crucial insight that has since been neglected by most variationist
research: speech communities do not always have shared ideologies, and ideological
differences can correlate with different language practices—indeed they can be, as Labov
found, the only means of interpreting language variation as socially meaningful and
stratified. Further, Silverstein suggests that language cannot be analytically understood
nor theorized without incorporating linguistic ideologies (1979) which, he argues,
influence language development and change. Language attitudes and ideologies thus
106
cannot be dismissed as epiphenomenal to language practices (Schieffelin, Woolard &
Kroskrity 1998). This incorporation of language ideologies and practices within a larger
examination of the sociolinguistic semiotics of IE will highlight how and whether an
interrelationship between external and internal elements of language is integral to
analyzing and understanding this localized IE dialect.
More broadly, this qualitative research sheds light on the role of language within
processes of globalization and localization in the periphery, as a globally subordinated
linguistic community creates and promotes local social identities which can then filter
outwards through their language practices and attitudes (V. Chand in press). By critically
examining what IE social and linguistic features are gaining momentum within globally
directed economies as both internal indexes of identity as well as external indexes of
English fluency and authority, this research can offer clues for directions future studies of
IE could profitably explore, and contributes to understanding what language practices are
becoming locally iconic (Woolard 2008).
Attitudes versus Ideologies
Before embarking upon an exploration of these speakers’ views, an important distinction
must be made between attitudes and ideologies. Some research has conflated individual
attitudes and larger social ideologies, e.g., by considering the government’s stance on the
role of English and vernacular languages as a de facto reflection of individual beliefs
(e.g., Krishnaswamy & Burde 1998). This approach, however, glosses over several
aspects of the sociolinguistic landscape, first because it conflates attitudes and ideologies,
107
second, because it only shallowly reflects one of several possible positions which
individuals can hold and negotiate between, and third, because it neglects how different
attitudes can influence language practices and domains of use—as aspects of
sociolinguistic identity—in different directions (Levon 2009). IE speaker agency in
endorsing and naturalizing these views vis-à-vis larger societal ideologies is thus unclear,
given past research. Examining individual attitudes in conjunction with larger social
ideologies allows conflicts and changes across generations, genders, and other social
groupings to be accounted for, by focusing on IE speakers as social agents who, through
their language practices and beliefs, both support and challenge societal IE ideologies and
status assessments to create a local, nativized English identity.
Attitudes and beliefs towards languages and dialects, in conjunction with one’s attitudes
towards the speakers of different language varieties, collectively comprise individual
linguistic ideologies. As they are produced within sociolinguistic interviews, these
language ideologies can be dialogically (Bakhtin 1981) both directed outwards,
responding to both global and local opinions and descriptions of international and intraIndia sociolinguistic groups and language practices (V. Chand in press), and internallyframed, as participants discuss their own linguistic preferences, domain choices, fluency
levels, future plans and feelings towards their linguistic repertoire vis-à-vis other local
ideologies and stances. Through qualitative analysis, then, we are able to access both
what they think of and how they respond to others, as well as how what they think about
language is mirrored (or not) in their reflections on their own language practices. These
ideologies, because of their feedback loop with language practices, wherein one can enact
108
(Woolard 1986; Bauman & Briggs 1990; Woolard 1998)—or create (Silverstein 1979)—
what one believes, are critical to studies of language practices, especially in understudied
regions like urban post-colonial India, where the potentially influential role of particular
social and ideological features is unknown.
Theoretical and Practical Goals
In this chapter, there are three overarching goals. First, I would like to flesh out this
speech community, and will do so by presenting these IE speakers’ attitudes, thoughts
and experiences. I choose to present this side of my research, and these aspects of these
speakers before I then explore their language practices because they collectively offer a
means of grounding these speakers, and contextualizing their practices as situated within
their New Delhi semiotic lives. While this order may not seem important, given that the
data and analyses are still valid in some more abstract sense regardless of their order, I
also see this as a means of fore fronting and making more primary these speakers’ lives
and experiences as an integral starting place to examining their language practices as one
aspect of their social lives.
The second goal is to problematize several concepts central to sociolinguistics (i.e.,
nativity, fluency, mother tongue, bilingualism, language ownership, dialect and language)
by examining local interpretations of IE, English, Hindi, and Hinglish language practices.
These concepts are often taken as givens: it is assumed that we all hold a shared
understanding of their meaning. However, these terms also hold power—they segment
populations in a hierarchical fashion, and social authority is not equally vested amongst
109
the divisions (Fabian 1986). My position is that these terms, because they offer social
authority, are not applied equally across the globe. Indeed, it is immediately obvious,
given the preponderance of variationist sociolinguistic research done on monolingual first
world English speaking populations (J. Milroy 2001), that some terms—and, by
extension, the populations they encapsulate—are unmarked; that is, they are ‘easier,’ in
some sense, to study, unproblematic, and perhaps even a ‘baseline’ from which more
complicated social and linguistic contexts are then compared (c.f., Preston 1991).
However, I offer two challenges to this. First, statistically, monolingual speakers are in
the minority, worldwide (R.G. Gordon, Jr. 2005). They cannot be unquestioningly treated
as more ‘representative’ of a ‘typical’ speech community (R. Singh 1998). Second, an
understanding, for example, of monolinguals or people with a single ‘mother tongue’ as
the unmarked category reflects a position of power (Silverstein 1996; Shuck 2004), but
this and related understandings are not necessarily shared across the globe in
communities where these terms are more problematic to apply dichotomously.
Experiencing these IE speakers’ voices will illuminate how these speakers can and do
dialogically (Bakhtin 1981; V. Chand in press) negotiate and challenge, both subtly and
blatantly, global definitions which often disempower and disenfranchise them. I show
that these terms are not unquestioned across the globe: peripheral and alternative
communities are also aware of the power of such terms, but do not try to ‘own’ them, and
instead, as we shall see, often scoff at their irrelevance.
110
The third goal is to highlight the future of this community as it can be extrapolated
through processes of language loss, shift and maintenance and through discourse about
and against globalized language ideologies (c.f., V. Chand in press) as a contribution to
the growing attention to a sociolinguistics of globalization (Blommaert 2003; Park &
Wee 2008). By all indications, this community will grow and have increasing influence,
both within India, and across the globe, in coming years, and it would be remiss to ignore
what indicators are present in this data about the sociolinguistic future of this population.
Collectively, these goals are ambitious within a single chapter, and apologies are offered
in advance for glossing over certain nuances and for completely ignoring several other
interesting topics that emerged within the data. After a brief discussion of past qualitative
sociolinguistic research—both within India and on other communities—I will discuss the
theoretical and methodological grounding for this analysis, and then plunge into these IE
speakers’ sociolinguistic lives.
Past Research
As briefly discussed in the last chapter, formal constraints for separating qualitative
ideological spheres do not exist, while one can draw on the aspects of language
background, plans and attitudes that have each demonstrated correlations with language
across other language settings (e.g. J.R. Edwards 1982; Ryan et al. 1982; Sahgal 1991;
Hohenthal 1998; Swigart 2001; Baldridge 2002; Coupland, Bishop, Evans & Garrett
2006). Language attitudes and ideologies have been profitably studied at both individual
and societal levels, respectively, in a variety of post-colonial contexts to better understand
111
language shift and language revitalization efforts (Hoare 2001; Wassink & Dyer 2004),
changes in language prestige and RP hegemony due to globalization (Bayard et al.,
2001), Malaysian English dialect recognition, evaluation and ownership (Newbrook
1998), shifts in language attitudes, repertoire and use by Hong Kong Sindhi and Sikh
communities (Detaramani & Lock 2003), changing social status of languages through
loanword use (Ngom 2003), codeswitching and vernacular uptake and valorization in the
Indian diaspora (Raj 2003; Shankar 2004; 2008), language maintenance, loss and shift in
the Indian diaspora (Raj 2003; Reynolds 2005; Shankar 2008), and dialect recognition
and perception (Preston 1999). Socially shared ideologies can also change over time in
their correlation with particular features. For example, in Corby, UK and Kingston,
Jamaica, phonological variants previously stigmatized have been adopted by younger
generations and reanalyzed to index different identities (Wassink & Dyer 2004)—see
also Dubois and Horvath (2000) for a similar scenario in Cajun English.
Sociolinguistic research and research on language attitudes and ideologies in India has
examined IE speaker identity (A. Pandey 2004; Chelliah 2006), language practices and
domains of use (Sahgal 1991; Hohenthal 1998; Vaish 2008), language loyalty (Hohenthal
1998; Chelliah 2006; V. Chand 2008; in press), the effects of language policy on
language practices and ideologies (Ramanathan 2005; Vaish 2008), similarities and
differences between local and global language ideologies (V. Chand in press), and
positive and negative associations towards speakers of different Indian regional
backgrounds (Sahgal 1991; Baldridge 2002; Chelliah 2006). Importantly, akin to
Preston’s work on Americans’ conflicting views towards Southern English (under
112
review), loyalty towards IE cannot be taken for granted—this is visible in Indian English
novels. Examining Indian novels written in English, Chelliah demonstrates that ‘[t]he
very features that make Indian English an “Indian” language are the features authors
denigrate through repeated association with villains, buffoons and losers,’ thus
reinforcing IE’s ‘substandard’ status (2006: 8). For example, IE features like the v/w
merger, copula dislocation, and article deletion are uncovered as satiric in a novel about
Indians living in the UK:
Vot kind of man you are? Where our invitation is? Lost in bloody post? Vhy we not
invited?.Don’t give me stupid question. Their daughter is becoming our daughter and you
give me stupid question. Vot kind of man you are? And today my friends ask me vot I’m
wearing on Saturday. Wearing to vot? How shameful this is. (Malkani 2006:241)
The IE features are readily apparent, and are selectively used to mock an Indian mother
when she faces conflict between British and Indian marriage customs: using iconic IE
features trivializes her underlying statements as out-dated and out of touch.
However, also present are Indian novels that dialogically challenge these negative
evaluations, and instead use IE features to evoke a socially and contextually authentic
Indian persona (V. Chand in press). This use of IE in novels can be understood through a
post-modern lens. John sees Arundhati Roy’s Booker prize winning novel The God of
Small Things (1997) as a ‘confrontation with English’ (John 2007: 195). While IE is used
here as a comedic and satiric device, it also offers English a local authenticity, and
through this, legitimizes local English ownership:
Indian-Englishness that is caricatured and satirized goes straight to the heart of the Indian
reader. His guilt at not being able to master the language, his ‘various troubles’ with the
language are all taken care of in this book. The guilt has been washed clean. At last, he can
have fun with English. (John 2007: 198)
113
Moving Forwards in Qualitative Research
While qualitative work has clearly made progress towards nuancing the sociolinguistic
landscape, Hohenthal suggests that research on loyalties and ideologies needs to be
conducted in conjunction with other data (1998), while other scholars like Eckert (in
Woolard 2008) and Milroy (2001) suggest that variationist research is in danger of
oversimplifying the social context: qualitative and ethnographic data are necessary
counterpoints without which neither type of data is sufficient for answering theoretical
and methodological questions. Compounding this are two additional factors. First,
Rickford suggests that sociolinguistics has ‘a tendency to be satisfied with observation
and description, and [is] insufficiently imbued with the thirst for theoretical explanation
and prediction which drives science onward ’ (1988). Second, Silverstein interprets
language structure and linguistic change as integrally tied to, and dialogic with language
ideologies, and is thus critical of any approach to language which does not include such
(1979).
Responding to these critiques of qualitative research and its goals, and quantitative
research and its limitations when undertaken without qualitative social data—and, indeed,
linguistic analysis more broadly—, the current research both builds on and expands the
reach of past work towards addressing Rickford’s urge for theoretical rigor and
Silverstein’s broader agenda, by examining how IE and vernacular language attitudes,
loyalties, vernacular fluency, and domains of use may be intertwined with synchronic
structural variation. A data driven exploration of potential structural variation and
respective links to age grading or change-in-progress, in conjunction with individual
114
reflections on a variety of language related concepts and choices is offered here. This
affords a lens to understand how globalization and English’s increased status worldwide
can manifest and affect a particular multilingual setting. This project is ultimately
focused on rigorously challenging and improving our theoretical knowledge and
predictive power within sociolinguistics, as well as spotlighting an understudied
community and region.
Further, qualitative data and analysis has already proven itself as critical to these larger
endeavors (e.g. see Chapter 3 discussion of circumscribing this community). There,
qualitative data allowed the pre-developed social, economic, linguistic and regional
delineations of the target population (e.g. what constitutes a Delhiite, acceptable variation
in language background and birthplace) to be questioned and ultimately revised to better
circumscribe a population with a shared socioeconomic, regional and linguistic identity.
In addition, as will be discussed more fully in the following two quantitatively grounded
chapters, this method of approaching data for both form and content has also allowed for
the creation and coding of additional social factor groups, which in turn has permitted a
more fine grained analysis of the strongest predictors of how structural variation is
realized in the three variables discussed in the next two chapters.
A qualitative analysis thus allows one to place more emphasis on the social semiotic
reality by focusing on social patterns (Ash 2002), social networks (L. Milroy 2002),
social identities (Mendoza-Denton 2002), and associated social ideologies (French 2001)
of and about speakers of various dialects. Thus, it steps closer to the perspective Milroy
115
and Silverstein (1979), among others, envision sociolinguistics as holding: to ‘explain the
social “life” of language and the social origins of language change’ (J. Milroy 2001: 553).
In particular, a data driven approach—discussed further in the next section of this
chapter—permits the emergence of new social elements not uncovered in previous
literature, and allows for consistently powerful social elements and concepts, e.g., ethnic
identity, language attitudes and processes of language shift and maintenance, to be
developed naturally into larger theories through patterns within the data, and not through
preexisting theories or hypotheses developed from other contexts.
Methodology
Having made a case for the role qualitative analysis can play to better portray and
ultimately understand a complicated understudied sociolinguistic setting and links
between language production and language attitudes, and as a complement to the
quantitative analyses presented in Chapters 5 and 6, I now address the methods for
qualitative coding and analysis, and the theoretical underpinning, Grounded Theory.
Grounded Theory
Qualitative codes and dominant themes were developed within Grounded Theory (Glaser
& Strauss 1967), an emergent methodology where one begins by taking notes based on
the data, and key issues, terms and associations begin to emerge. From there, by
comparing across data, a theory—or multiple theories—then emerge, and these theories
are then cross-checked against the data. From these theories, codes which encapsulate the
various aspects of the theory are created, and then applied to the dataset. Codes are
116
malleable, as one traverses the dataset, and may be narrowed or enlarged to appropriately
circumscribe the data. The next step is to sort through the codes, so that they can be
presented in order to best traverse and justify ones emergent theories.
Grounded Theory stands in stark contrast to most quantitative methodologies in that it is
not hypothesis testing, but instead, is data based, and aims to find the best theory to
account for the data set (Glaser & Strauss 1967). This emergent process also involves
how past research is used within the analysis—an overly detailed review of literature
could, it is hypothesized, constrain ones coding. Thus, I offer a broad overview of past
literature (presented above)—one that is necessarily short—and focused more on
concepts and situations that have successfully linked language practices with language
attitudes, while avoiding a thorough discussion of exactly how these practices and
attitudes are linked in past literature. While several themes are presented within this
review of literature which have proved fruitful in a variety of other context, no specific
results are fore fronted as strict hypotheses to be tested in this data set. Instead, further
literature will be drawn on within the chapter, to best explain the codes, concepts and
analysis emerging from the data.
Development of Qualitative Codes
The key tenets of Grounded Theory were used in all aspects of my data collection and
analysis. The first step in data collection was the creation of an interview scheme, within
which topics were grouped into thematic modules. As I conducted interviews, the
modules, order, and specific questions were adjusted—though not quite in a constant
117
state of flux, as the process was honed over time—to best accomplish three things: 1) to
encourage an informal environment with natural transitions between topics, 2) to explore
new concepts and experiences which had emerged from earlier interviews and 3) to shape
the interview towards the very different life experiences across the different generations.
After all of the interviews had been conducted, I listened to each of the interviews again
and jotted down notes on any key issues or themes which emerged within the
conversations—these notes were collaborative, in that both themes and particular
explanations which I felt were important, and ones that the informants had stressed or
emphasized within the interviews were noted. I next compiled these into a set of codes,
which then went through several iterations, as overlaps and gaps within the codes were
uncovered. I next explored and developed connections between codes, based on this relistening, and my notes, to better gauge how different codes could be interrelated, and
how best to capture these emergent links. From this, additional codes were created, to
capture some of the links that arose from the data. At this point, coding began, within
Atlas.ti. (Muhr 2004), software built around the tenets of Grounded Theory, and designed
for organizing, coding and analyzing large quantities of qualitative data with projectspecific hierarchies of codes. This process of coding interview transcripts often involved
referring back to the original audio files to verify interpretations and more clearly gauge
the informants’ point(s)17. As coding continued, additional codes were created, and the
17
This process of close coding required a yet another detailed examination of the transcripts vis-à-vis the
audio files and can be considered a fourth, and final, verification of the transcripts reliability, readability,
and coherence. Originally, the audio was transcribed, following the structure laid out in Chapter 3. Then
transcripts were double checked against the audio by a second person. Third, any questions or discrepancies
between the two versions were examined a third time, and a decision was made as to how best represent the
audio material.
118
coding hierarchies were adjusted, to best fit the relationships and concepts expressed
within the data. Some topics were consistently coded for multiple themes, as they tend to
work in conjunction, however no codes were completely redundant, and this narrow
approach to coding did illuminate more complicated, and even contradictory attitudes.
Qualitative Code Families
The transcripts totaled 49 hours of data, and over 325,000 words. The process of coding
resulted in 322 codes grouped within 17 code families. These codes, in conjunction with
the data, resulted in 2294 coded extracts of quotations from the corpus. For reasons of
space, each code family and code is not discussed here, while a full list of the codes used
within this project is presented in Appendix D.
Coding, in Practice
Table 4.1. Qualitative codes displayed graphically as hierarchical in Atlas.ti
119
It is possible, within Atlas.ti, to embed codes or partially or fully overlap codes, and thus
possible to code individual quotes for multiple links. The finished coding graphically
presents both the data and the codes (Table 4.1).
Concepts for Analysis
Clearly, within the confines of one chapter, this analysis cannot comprehensively cover—
or even lightly touch upon—the entire range of codes and possible viewpoints espoused
by these 35 speakers. The following analysis is thus based on a subset of the codes, but is
not limited to a subset of the speakers. Speakers’ voices are combined into natural
groupings, and how these may link to social demographics, language background
(domains of use, fluency, attitudes, etc.) are collectively examined, to better understand
how these speakers align themselves vis-à-vis a range of factors.
Examining the socio-historical and economic history of India, in conjunction with how
these speakers, among other post-colonial multilingual communities, are grouped and
understood with respect to the traditional linguistic concepts of bilingualism, definitions
of mother tongue and fluency, past literature does not make clear how this population
self-identifies or should more accurately be grouped. A close examination, taking into
account speaker agency, understandings of language ownership, loyalties, self-perceived
competency and how these link to individuals’ linguistic identity, as urban IE speaking
adults, is clearly in order. Thus far, speakers’ understandings of typical Delhiites, of the
history of Delhi as it relates to changes in the Delhi population, and the eldest group’s
birthplace as it relates to continuity across generations have already been explored, within
120
Chapter 3. Taking these as a starting point, I will next explore how their attitudes,
thoughts, and stances allow us to better understand this community, how they frame and
explain themselves as urban, upper middle class bilingual English/Hindi speakers, and
how they reconcile individual language goals and practices with changes in the Indian
economy, social structures, and India’s international success—all of which are clearly
linked to English fluency.
Analysis
Below, I purposely offer extended quotes—rather than brief sound bytes—because these
speakers’ answers are complicated, nuanced, and sometimes even conflicting. That is,
these IE speakers do not have pat, pre-formulated answers regarding how they feel about
IE, about how they do—or don’t—value IE as a valid dialect, accent or language, and
about the implications these hold for their personal evaluations of their own speech. As
these quotes show, these IE speakers are actively negotiating, reconciling, and taking
positions which take into account current and past ideologies. These responses thus must
be understood as dialogic (Bakhtin 1981) and syncretic (V. Chand in press), wherein the
responses show an accumulation of past and current views. As I will demonstrate,
individual assessments of nativity and fluency are often encapsulated within speakers’
attitudes towards IE. Therefore, we start by exploring the range of assessments towards
IE. As a final note, in all of the quotes below, some sections are underlined. Obviously,
speech doesn’t have any tangible underlining: I use underlining to emphasize critical
statements, and I refrain from repetitively noting that underlining represents my added
emphasis
121
IE
The first question which arises in examining how these speakers understand IE is whether
IE exists, to them—is it a tangible reality? Indeed, most of my informants believe that IE
does exist. Within this, IE is often characterized in laymen’s terms as a distinct accent, as
m19AS, a male university student, does:
VC: Uh-huh. So if you were talking on the phone – say somebody calls you up and they’re
talking in English, could you tell where they’re from?
AS: No.
VC: So you’d have no idea if there’s –
AS: Not unless – not unless they have a distinct accent, depending on the place. South
Indians have a distinct accent. You can tell. And you can tell if it’s Bengali.…
VC: Do you feel like the English that’s spoken in India has its own accent?
AS: Yeah, it does. (m19AS 12:6-11, 12:26-27)
This evaluation of IE as distinct from international English varieties is also described by
f18MG, an18 year old university student, as having a different ‘touch:’
There is this perception that, you know, if an Indian speaks English, and if an American
speaks English, obviously there is this huge difference, and, you know, there's an Indian
touch to it, and there is British or an American touch to it. So, yeah. (f18MG 23:17-20)
However, it was rare for speakers to use the term ‘Indian English’—f30PG, a 30 year old
working woman, was the only one to use the term without my having already introduced
it:
VC: Do you like British English?
PG: I like Indian English. (f30PG 20:9-10)
Instead, more common were descriptions of IE which did not ‘name’ IE, as f50RV, a 50
year old lawyer says:
No, I’m - I’m happy with the Indian -- the way we Indians speak. (f50RV 21:25-6)
Naming has power—it affirms that IE is a variety, not a bastardization of another variety,
like RP, however, these speakers do not name IE. While these speakers are aware that the
English in India is different from international varieties, there is no underlying consensus
122
on labels and definitions. Why? Naming it IE or Indian English does separate it from
‘English’: the lack of consensus on a name for the local—or national—variety is
consonant with the view that IE, whatever you call it, is not inherently a different beast
from the other English varieties found across the globe. Regarding definitions, they may
not agree on whose speech such a label would be capturing, e.g. their own speech, or the
‘babu next door.’ Importantly, as we shall see below, they do not refrain from naming all
codes. Hinglish—discussed below and referring to a codeswitching/mixing between
Hindi and English—is regularly ‘named’ and evoked within discussions of language
repertoire and domains of use. I will return to the implications of this dichotomy later in
this chapter.
Varieties of IE
Beyond IE simply existing, many speakers affirm what m19AS, above, said: there are
multiple types of IE which are regionally separated. m28KC, a 28 year old working man,
explains that these regional separations also correlate with the distinct linguistic and
cultural diversity of each region:
KC: Absolutely. Absolutely. See, the fact that every thirty kilometers, the language changes
in India. And there’s a different dialect, too. I mean - I mean, if you’ve been to East
India, you’ve seen the way they speak English. It’s - I mean, English is a universal
language, I would say. And in India, English is spoken in five different dialects, which
I’m sure even the English are not aware of, okay.
VC: So what are the five?
KC: I couldn’t tell them all for you. I don’t think they have any names. It’s just the way they
are.
VC: Regions or KC: You could divide them geographically, literally, from the middle. Say four parts, if not
five. I would take Central India as the fifth. Because Central India and my recent visit
there was quite shocking to see how different Central India is to east, west, … south,
north. So VC: So you’re basically saying there’s five - east, west, north, south KC: And the middle.
VC: -- and the middle.
KC: The middle is completely different. It’s - it’s incredible how different it is.
123
VC: So how is it different? Like how would you characterize-?
KC: It’s a very - it’s a very powerful culture. (m28KC 11:7-25)
A 52 year old hospital administrator, f52GG, underscores how Indians regularly try to
interpret one’s background through language practices, dress and manners: cultural and
vernacular language background is very salient in India.
VC: So when you're talking to somebody on the phone, can you pinpoint if they're talking in
English where they're from, I mean, generally? Can you say, oh, this person's -GG: Yeah, yeah. Most times.
VC: So what sort of things would someone sound like if they're from the South? Like what
would they do?
GG: It's just their accent is so -- it's so different, you know. It’s difficult to explain, but I
could get you a South Indian and a Bengali and say, you know, you can immediately -tell them to repeat the same sentence, and you can, you can pick it up. And, I don't know,
it’s difficult to explain how and why.
VC: But you're very aware of it when you're talking to people or -GG: No, I'm not aware of it, but what happens is it's very interesting because I -- it's not that
I like putting people in, in their little compartments, but I don’t do it for any other reason
but just to -- you know, it’s just, it's just so interesting. And many times, I mean, I think
nine out of ten times, people can not tell where I'm from. And when I meet people for the
first time, they all say, oh, you're, you’re, you’re Maharashtrian, are not you? And I said
no. Okay. Then you're from the South. And I said no. So I wonder what they see in me,
whether it's my, my tone, my language, my accent, my -- whatever. So asked somebody. I
said, you know, why do you say that? And she says, you know, you -- everything about
you is so South Indian because, you know, you -- you only wear -- you wear sarees, and,
you know, the jewelry you wear, and you know, whatever -- whatever she said. And
maybe I'm always wearing a South Indian silk and this -- that's what it is, but for me, I, I
like to just -- you know, just -- oh, this, this person is from, you know, from Orissa. You
can immediately -- I know -- I can even make out the difference between Orissa and
Bengal, you know? And just -- for me, it's just like an interest. You know, okay, now
18
let's see, you know, am I right? Am I not right? “ettho” [ɛt
ɛt.ɵo]
you can make out in
ɛt
India from “et-tho” [ɛtʰ
ɛtʰ.ɵo].
You know what I'm saying? This is what we -- what we do.
ɛtʰ
And here in India, you can make out people from their names, you know, which part of
the country you come from. My -- I don't know. I think it's so ingrained in us about where
you come from, you know, which part of the country you're from. (f52GG 13:26-14:7)
Clearly, language plays an important role in identifying and categorizing Indians, though,
as this example demonstrates, these categorizations are not always correct, and perhaps
Indians pay more attention to non-linguistic features in locating others as from a
particular cultural and regional demographic.
18
‘ettho’ and ‘et-tho’ are not words (that I know of), and instead, are offered as a minimal pair to highlight
the contrast between aspirated and unaspirated stops that is common to several Indian languages.
124
Beyond India, other English speaking nations are also defined as having multiple
accents—that is, India is framed as mirroring a process which also occurs across the
globe, wherein regional and cultural differences provoke a range of accents, as f64AG, a
64 year old female publisher, explains:
Now, we say the Punjabis have their particular accent. The South Indians have their
particular accent. But if we look at it in the, in the U.K., how many accents are there? I think
- I don’t know whether you’ll be able to understand the Welsh accent absolutely. Right from
Scotland to Wales and - each one has a different accent. The Cockney is different and
everything is difficult. So we have accepted it now. It doesn’t have to be the Queen’s
English. You know, it has to be - everybody’s got their own accents. (f64AG 20:21-27)
Some speakers, like f73NK, a housewife and volunteer, make a distinction between
accents and language, and suggest that across India, there are multiple ‘accents,’ but they
are all the same ‘English’:
VC: So when you’re talking to these people in, let’s say a South Indian or something, do they
sound different than North Indians to you when they speak English?
NK: Maybe accent is different. That’s all. It is all understandable.
VC: But if - say you were talking on the phone, could you pinpoint, oh, this person is from
the South versus Punjabi, versus Gujarati, or NK: Nay, nay.
VC: No? So it’s a pretty similar English that’s spoken all across?
NK: Same, [HINDI hain yes] (f73NK 14:6-13)
Across all of these explanations of IE being a different accent, or touch, and/or multiple
accents, there is a unified sense that ‘English’ is an abstract entity realized in different
forms in different locations, and that this shared ‘English’ allows for communication:
e.g., f73NK’s quote, above, ‘It is all understandable.’ In short, the views expressed above
are fairly different from global ideologies, which regularly locate the English in India as
‘wrong’ and British or American English as more ‘correct’, or inherently truer to
English’s roots (Silverstein 1996; Shuck 2004).
125
British Speech is ‘correct’
However, other speakers do not share this interpretation of all English varieties being
equal—IE’s status as a valid dialect is contested. IE and/or more abstract ‘English’
ownership correlates with whether these speakers positively evaluate IE: speakers who
don’t believe that they themselves speak ‘good’ English also don’t believe that an Indian
dialect of English can be good, and vice versa. I next explore the range of evaluations of
IE and how they correlate with individual assessments of language competency.
On one end of the continuum, f39VD, a 39 year old senior government consultant on
information technology and public policy, claims that IE doesn’t exist, and that she
prefers the British BBC accent. In this excerpt, she is relating how, when studying in
London, she visited an Indian friend, and upon returning to the school environment,
f39VD was accused by a British classmate of speaking in an ‘Indian accent,’ which she
disavows:
VD: It’s funny because even when I was studying and I was staying in one of the halls of
resident, residences and for Christmas I went over to my friends’ place and when I came
back one of my classmates said, “V. now we have a very Indian accent.” I said, “There is
nothing called an Indian accent.” He said, “What? Your accent is different now.” …
VC: Huh. If you were to, do you prefer any particular accents of English over others? Like is
there is one that soundsVD: BBC
VC: BBC?
VD: Nice clear, a lovely diction and it should be clear. I think that’s, that’s the thing. (f39VD
23:24-37)
Similarly, f25GS, a 25 year old female who works in marketing at a private hospital,
locates British English as ‘superior’ to Indian English:
I think Indian English is something like the - if I have to speak a sentence in Hindi, the same
sentence, I, you know, I know the sentence in English, it’s like that. But I know that the
English which is spoken by English and, you know, the Americans are quite different from
like what we speak here. So I think - because English is a language which has, you know,
come out from England, so I think their English is like much better from us and far superior.
(f25GS 15:37-42)
126
f39VD and f25GS stand out in their negative evaluations of IE and their positive
affirmations of British English. However, they are also separated from the rest of the
participants in another way—both fell outside the target demographic, and their speech is
not used in the quantitative analyses in the following two chapters (though, clearly, their
views are incorporated here). f39VD was excluded because she had recently relocated to
Delhi, and had grown up in a household which spoke dominantly in Hindi:
VC: When you were growing up what language did your parents speak in the house?
VD: Hindi.
VC: And did they ever speak any other languages?
VD: English.
VC: So, was it, was it mostly Hindi or like half and half?
VD: Mostly Hindi, mostly Hindi.
VC: And English would be for anything in particular or just get thrown in?
VD: Get thrown in.
VC: Um-hum. So when you were growing up then, the kids would then answer back in the
Hindi.
VD: Hindi
VC: Hindi only?
VD: Mostly Hindi.
VC: Uh-huh. And even when you were going through school?
VD: School was a mix of English and Hindi. (f39VD 14:1-16)
Meanwhile, f25GS was excluded because she is not a native English speaker, having
learned English in secondary school:
My father, he can speak Punjabi, actually, but at home, we prefer Hindi only. We speak
Hindi only, actually. (f25GS 8:35-6)
and because she is from a lower socioeconomic strata, and has lived continuously in
North Delhi, over an hours drive away from the south Delhi neighborhoods that the rest
of the speakers live:
I have like born, you can say, in a lower middle class family… Yeah, my family lives in
North Delhi. We have our own house there. (f25GS 1:7-14)
These two are separated from the rest of the speakers interviewed because they both
locate the English spoken in India as wrong, as failing to meet an outside international
target in terms of grammar or pronunciation. These speakers’ assessments are rooted in
127
monoglot standard language ideology: ‘[v]irtually everyone subscribes to the ideology of
the standard language, and one aspect of this is a firm belief in correctness…when there
are two or more variants of some word or construction, only one of them can be right’ (J.
Milroy 2001:534-5, emphasis mine). It is not unexpected that they reference the British
colonial influence as also being the source of ‘correct’ English—this coincides both with
standard language ideology (Silverstein 1996) and the ideologies encouraged (Fabian
1986) by the British colonizers (e.g. Macaulay 1920/1965).
These two speakers are further separated from my target population because they both
explain their own English competency in terms of what ‘fluent’ speakers do, wherein
their speech is located as lacking—they recognize that English is not their primary
language, while others around them do ‘own’ English in a way that they do not. Speaking
about younger generations of urban Indians, she says:
VD: They are not just more fluent in English, they’re also, it’s also the way, it’s also the way
they speak, you know. I think we, we use to think in Hindi and speak in English.
VC: Hum.
VD: You know, now they think in English, I think that’s, that’s the basic difference.
VC: Um-hum.
VD: Also their vocabulary is, is um, much greater, at least my niece is I don’t know about
other kids. And to them it seems like their primary language, you know, to us Hindi was
our primary language. This was something that we learned. You know, it was an acquired
language.
VC: Um-hum.
VD: To them it’s their primary language. (f39VD 24:11-21)
When English is one’s primary language, one has a larger vocabulary; one can
think in that language; and—demonstrated below—one doesn’t get corrected for
their pronunciation:
Yeah, but sometimes my friends, like I’ll say “balcony”. So, “Hey, what are you saying? It’s
not balcony [bæl.kə.ni
ni],
æl.kə.ni].”
Speak, in the right way. Okay. Okay. So.
ni it’s balcony [bæl
æl
(f25GS 11:29-30)
128
f25GS has internalized these corrections that her friends have made to her English—she
believes that she doesn’t ‘speak in the right way,’ as is obvious by her acceptance ‘Okay.
Okay.’ Collectively, these two speakers—who are less comfortable in their fluency,
because they did not learn English at an early age and do not speak English in intimate
circles—are the same ones who are most reluctant to valorize or even affirm the existence
of Indian English. However, as we shall see, quite a different perspective emerges when
we explore how speakers who did learn English at home and who regularly use English
in intimate domains frame IE.
All English Accents are Acceptable
Several other speakers offer a neutral assessment of different English dialects, including
IE. m52PB, a 52 year old businessman, is reluctant to hierarchically rank English
dialects:
There is no, I would not say that there is any particular way of speaking English which is
right or perfect. (m52PB 9:37-38)
As is f18AU, an 18 year old female college student:
No, not really. I think, you know, people just kind of speak English in the way they’re kind
of - you know, they’re used to basically. As in, you know, in India, people - as in, you know,
in America, people have their different accents because, you know, people around them have
the same accent. And, you know, in - in the British way of speaking, I think it’s like the
similar way. So, yeah, I don’t think there’s like a major - you know, as in, anything is sillier
or harsh or anything, I don’t think so. (f18AU 19:7-12)
f60SS, a 60 year old housewife:
See, every country has got its own accent. America doesn’t mean, “Okay, there’s a typical
American twang.” Again, it’s built up - it’s so many different - you’re from this place - eh,
eh, this place, [HINDI iska nam its name], what do you Texi- uh, Mexico, you are from
Texas, you’re from New York, they have their own twangs, you know? They have their own
ways of speaking. So there’s no typical accent. Actually, in Great Britain, again, they’re they don’t - they’re not broken up into mini-accents, but you can, you can understand. (f60SS
12:34-40)
129
And m85SPB, an 85 year old man who still runs his own business with his sons:
Actually, like -- question of liking is not there. After all, President Bush also communicates
with the London Prime Minister, and there may be slight accent -- difference in accent, but
they just communicate whatever they have to say. Little difficulty is there. If some
Switzerland Chief Minister or Prime Minister goes to Bush, he can not speak that American
English there, he will talk on English, whatever -- whatever accent they have got in their
country, they will do like that. …That is same case as here in India. (m85SPB 17:14-19)
These speakers see ‘accents’ as the product of where one lives, and draw parallels
between the various IE accents and the range of American and British accents. Language
is for communicating, and because all of these accents are still English accents,
communication between them will be successful. This choice to not rank English
varieties is further explained by m52PB, who we also heard above—for him, language
serves a purpose, and if communication is successful, then there is no reason to rank
dialects:
Personally, I feel that language was okay. It’s you know, it is for communicating so if a
language can give you a medium through which you can communicate between um, people,
it doesn’t make much of a difference, you know, even if you are going with the left of right
or and Americans do, you know, their English is what English we know of as not their
English so you know, their far away from that English which we have learned and we have
grown up with. But it doesn’t make much of a difference, again, I am saying the same thing
that the way Americans use it or you know, no harm in that. And for that matter, the way
English is used by south Indian, there is no harm in it, even if there are some of their own
words thrown in, if that person is able to communicate with another person, perfectly fine,
what’s the problem? (m52PB 9:42-10:5)
A cynic (or non-Indian) might suggest that these speakers recognize that if they were to
hierarchically rank English dialects, their own dialect would not be the highest rank given
global standard language ideologies (c.f. Silverstein 1996; J. Milroy 2001), and they are
simply choosing a means of seeing the situation which allows them to save face.
However, those who were willing to rank dialects—and who did fall within my target
demographics—instead demonstrate the opposite trend, and value IE over international
varieties for being more open, balanced and liberal. For example, f30PG, a 30 year old
working woman explains:
130
PG: I like Indian English.
VC: Indian English.
PG: Because it’s a mix of both. It’s more British, but I like Indian English.
VC: Do you feel like there is any particular like characteristics things of Indian English that
the other ones don’t have?
PG: I think Indian English is far more open. It’s a little bit more open. It’s a little bit more,
it’s balanced.
VC: How so?
PG: Um, well, it’s got a mix of both, like okay for instance a word like colleague. I can even
pronounce it as “colleague” [ka.lig]. I can even pronounce it “colleague” [ka.lig̊]19. You
know what I am saying, so there is that much more flexibility whereas if I am a Britisher
or if I’m in London, I’d probably be looked down upon if I’m pronouncing a particular
word in a particular manner because the British, the Europeans are very particular about,
especially the British are very particular about how they pronounce things, so in that
sense I think it’s a little bit more liberal. It’s a little bit more open as compared to others.
American English, it’s a little difficult to understand especially when somebody’s talking
in a very - in a very fluent manner, so for a normal average person it might get, he might
take a minute a grapple with, okay what did he say, just repeat your sentence, dude, that
kind of thing.
VC: Um-hum.
PG: So that’s why I find Indian English has a right mix of both. (f30PG 20:10-30)
When one offers valorizations of IE, this seems to then require also explaining what is
‘wrong’ with other varieties. Here, AE is seen as ‘a little difficult to understand,’
especially when spoken rapidly.
IE was also located as accentless, simple, and lacking any added tones. Capturing
something common, speakers who positively evaluated IE over international varieties
were also more comfortable with English than Hindi, overall using English more and in
intimate domains, as f28SB, a 28 year old female working in the museum and art curator
field, demonstrates:
VC: And where would you rank Indian English?
SB: I think its one of the best. No doubt.
VC: Yeah.
SB: Yeah. It’s, I think many people love our Indian English, if I am not mistaken. The
American also they get very pleased with our because there is no accent, we just, we talk
the way we are I know what you think of the Indian English.
VC: Well I’m studying it right now so it’s not fair to offerSB: Right, right, American is fun no doubt, but Indian is just very simple.
VC: Um-hum.
19
These two pronunciations differ in stress (stressed syllables marked in boldface), and voicing: in the
second pronunciation, the final [g] is devoiced.
131
SB: There is no added tone or whatever.
VC: Um-hum. Would you say that you are far more comfortable in English than in Hindi?
SB: English, yeah no doubt. (f28SB 32:45-33:9)
IE is further located as ‘normal,’ ‘simple,’ and lacking a ‘twang’ or ‘accent’—all of
which aid comprehension and communication—by a 60 year old housewife:
SS: Delhi English is normal, simple Delhi English. Like maybe like where I’m speaking, I
don’t think there’s any accent in what I’m - in my vocabulary, no. It’s like that only.
They don’t have any twang or any accent or anything like that. It’s just a normal, plain,
simple English.
VC: So then SS: I think that is better.
VC: Better in what way?
SS: More simpler to understand. More simpler to speak. (f60SS 12:44-13:6)
Silverstein’s predictions about folk views about language and the functions of language
are relevant here—his model predicts that they will ‘characteristically center on the
functional capacity of words and expressions as the salient formulable interest of native
speakers in their language…denotation” (1996: 287). Indeed, these folk assessments of
IE do focus on IE’s functional capacity: it is described in terms of intelligibility, and
simplicity.
Why might they choose to valorize IE, and not an outside English variety? ‘[I]n standard
language cultures, virtually everyone subscribes to the ideology of the standard
language,’ (J. Milroy 2001:535), within which standard languages ‘acquire prestige when
their speakers have high prestige,’ (2001: 532). However, in addition to prestige, the
process of establishing some varieties as standard—standardization— and, hence,
establishing other varieties as non-standard has economic, social and political
implications (2001). There is economic, political and social profit to be made by using a
‘standard language,’ and likewise, profit to be lost if one uses non-standard language.
132
India, meanwhile, isn’t a ‘standard language culture’ for several reasons discussed in
Chapter 2, but these IE speakers also do not subscribe to the ideology of standard
language as located in international varieties like RP. Instead, they contest standard
language ideologies and the economic value which is offered to outside varieties by
locating IE as economically and socially valuable—in particular they use the same
processes of naturalization (Silverstein 1996) which standard language cultures use.
Specifically, speakers can reduce and naturalize languages, wherein naturalization can
‘anchor the process of standardization in something outside of the social organization of
language use itself, namely in psychological properties and…things to be denoted that are
independently in the world “out there”’ (1996: 288). IE speakers naturalize IE by evoking
psychological properties like ‘love’ and ‘better,’ and by abstracting IE from its speakers
to focus on IE’s denotational value—IE is ‘simpler to understand,’ establishing it as
having a higher denotational value than other English dialects. For native IE speakers, the
economic value of IE is linked to establishing it as a standard, and this is done by evoking
its value as intelligible: the more intelligible, simple, neutral and tone or accent-free a
variety is for multiple audiences, the ‘better’ it is. As we saw above, IE is defined in all of
these terms, and hence, is treated as more ‘valuable’ than outside varieties, and as a
possible new standard language.
Prestige can also be attributed and contested through particular terms. Talking about his
own speech m35MS, a 35 year old male in training to become an international ESL
teacher, separates IE as distinct from other English varieties:
I am quite aware of it when I speak to somebody who is so-called native speaker of English,
so a Britisher, an American, a Canadian, an Australian … If I am sitting with a bunch of ten
133
Indian friends after about fifteen minutes I probably be making my W’s … my V’s. (m35MS
4:1-9)
By using ‘so called,’ m35MS dialogically challenges (Bakhtin 1981) the global
attribution of native speakerhood to other communities, and not Indians: globally, Indians
are not interpreted as ‘native speakers’ of English. Clearly, he disagrees with this
ideology, suggesting that he views himself as a native speaker, however, a native speaker
of a different dialect, one within which speakers use [v], not [w].
Milroy suggests that varieties gain legitimacy, authority and value through the social
status of their speakers: ‘varieties of language do not actually have prestige in
themselves: these varieties acquire prestige when their speakers have high prestige’
(2001: 532). When challenging global attributions of social authority, however, it is
possible to reevaluate speakers through more value-laden assessments of their own
speech. f28SB, above, makes this reverse connection, suggesting that IE is ‘one of the
best,’ and ‘people love our Indian English,’ through which IE speakers are established as
socially valuable and well liked in international contexts. ‘Dominant’ languages, Milroy
suggests, are defined more by their ideologies than their internal structure: given that
these assessments of IE link internal structure to positive ideologies, it appears that the
two can work in collusion when contesting established ideologies.
Importantly, these IE speakers are referencing their speech, and make no references to
written IE as similar or distinct from outside forms. This also stands in contrast to
standard language ideology. Milroy uncovers that standard language ideology involves
abstracting language from its speakers: ‘language is not the possession of the native
134
speakers’ (2001: 537). Within this, canonical forms, often those perpetuated through
writing, are ‘right’ and native speakers acknowledge the ‘rightness of the written form
“Yes — I actually say it’s me — but I know it’s wrong!”’ (J. Milroy 2001: 537). These
IE speakers, by focusing on the spoken form, not the written form, are more subtly
disavowing standard language ideology. As well, it is possible that they do not evoke
differences or similarities between their speech and writing because they are biliterate,
surrounded by ‘bad’ English writing in the marketplace, and feel that what distinguishes
them from other Indians is their oral skills.
Finally, these positive descriptions of IE as neutral, accent-free and simple are important
for a third reason. Linguists who choose not to label language forms like ‘I seen it’ as
ungrammatical place ‘themselves outside the common culture’ (J. Milroy 2001: 536).
These IE speakers are also placing themselves ‘outside the common culture’—that is,
contesting standard language ideology—because they legitimize IE by referencing its
intelligibility, a quality with global economic value.
These framings, offered by native IE speakers, are very different from global standard
language ideologies. Comparing these speakers—who are positive or neutral about IE—
with the two non-native speakers we heard from above—f25GS and f39VD, who are
negative towards IE and view it as an incorrect deviation from their preferred target,
RP—, it is clear that global ideologies have a powerful influence on these non-native
speakers. They interpret variation in terms of standard language ideologies (Silverstein
1996; Shuck 2004), which locate IE as wrong or incorrect. The ideology asserts—and
135
they believe—that there is one correct form, which is embodied in British speech, and
therefore, forms which deviate from this, like their own speech sometimes does, are
incorrect. Because they are sometimes ‘incorrect’ they cannot claim ownership over IE—
IE doesn’t really exist to them, besides as a means of labeling and encapsulating errors.
They interpret IE in a manner very different from the native IE speakers.
Who is a ‘good’ IE speaker?
As I have demonstrated, IE speakers challenge global standard language ideologies which
locate IE as non-standard by taking up entailments and justifications which are also
evoked within standard language ideologies. Next, I will demonstrate one basis of their
beliefs: they locate IE as more proper and correct than international varieties precisely
because of the emphasis on English in Indian schools. In particular, such ‘proper’ and
‘correct’ speech is not available to all fluent IE speakers: across age groups, speakers
located this ‘correct English’ as more likely in older generations of IE speakers. For
example, below are quotes from a 25 year old working female,
I think it totally depends on the person, but I’ll say one thing. The people who are older than
us, I think their grammatical - you know, their base is really good. Because this English we
speak, I don’t think the, you know, grammar part is that much, you know, solid or strong.
But the people - I’ve seen my, you know, my uncle speaking English. The grammar thing
like is where they’re good and all, where to put which, you know. So I think they are much
better than us. (f25GS 16:21-26)
a 27 year old working female:
My mom used to complain when we were in school – I studied – most of the system in Delhi,
the education system, follows the CBSE, which is the Central Boards of Secondary
Education. And we don’t study things like Shakespeare at all. Okay? So our English is very
basic. Vocabulary is very limited. And so when we’re in school, we sit there and say things
like – after every sentence, there would be a like, you know? Like, you know, that, like, “My
mom’s gonna get really pissed off.” And she says, “What English is this?” And our stories,
even ‘til class twelve, used to be two pages long, you know? And it was no English. It was
English for the sake of knowing the language and nothing else. So basically, our English is
really very Indian English, Indianized English, pretty basic. And they, of course, were a
136
completely different level, you know, where they did literature and all. So we did not do
anything of that sort. (f27RG 12:18-32)20
a 39 year old female government consultant:
VC: Um-hum. And your, your siblings? They’re, they’re raising their kids, to speak both
Hindi and English?
VD: Yes.
VC: And do you feel like their English sounds different then say from your generation these
kids?
VD: Very.
VC: Like what sounds different in it?
VD: @ Um, when my niece, for example, recites a poem she has a very typical accent that
she uses because the teachers use that accent in class, you know, like saying tomatoes
[tə.me
me.toz]
and potatoes [pə.te
te.toz],
you know when we were kid we did not say
me
te
21
“tomatoes [tə.me
me.toz]
and potatoes [pə.te
te.toz]
,” @@ and you know or when she’s
me
te
saying, “pitter patter rain drop [pi
pi.tə
pi pæ,tɚ ɹeɪn dɹap].” I mean it is not pitter patter rain
drop [pi
pi.tə
pi pæ.tɚ ɹeɪn dɹap]22, that’s how they speak. You know, that’s how they are
taught in school. So the accent is different from day one. They learn to speak the
language differently. (f39VD 10:15-22)
a 39 year old female housewife who works part time:
VC: Do you think the younger generation sounds any better or worse?
GS: Definitely not better.
VC: Not better. Does it sound worse or -GS: May be a little bit.
VC: And this is their English?
GS: We spoke more formal English. They speak more slang. I think that is the major
difference. (f39GS 32:4-10)
a 52 year old female working as a hospital administrator:
Like I used to say that, you know, our English is better than this generation that’s gone
through school in the last twenty years because their emphasis was never on the language.
Their emphasis was on -- on -- you know, just -- I mean, I don’t know. We did senior
Cambridge so our emphasis was on grammar and poetry and literature, and you know, we
did a Shakespeare -- I mean, we did all that part. You know, English was like really major,
major. But what happens subsequently was that they did not do senior Cambridge. They did
the Indian boards, and Indian boards, the emphasis was not on English. The emphasis was
you had to take this exam, and you had to pass the exam, and which is really elementary, and
there was no emphasis on their pronunciation. I mean, they would say things like
pronunciation, and it was a really to bother me, you know? So small, small things like this.
(f52GG 15:38-16:2)
20
Although this quote does appear to simply reflect what all parents say, e.g. ‘don’t use slang,’ it is
included here because f27RG then contextualizes her speech as ‘wrong’ because it was ‘Indianized’ and
‘basic.’
21
That is, f39VD distinguishes between current youth using an American pronunciation, e.g. [tə.me
me.toz],
as
me
distinct from her pronunciation, which is [to.m
ma.toz].
22
Again, f39VD is distinguishing American from Indian (or British) pronunciation, and herself would say
[p
pɪ.tə pa.tə
pa ɹeɪn dɹap], with different vowels and (perhaps) less rhotacization.
137
and a 64 year old man who is retired from military service:
I think India by far has the best command of the language, the people who have, who are
educated. Even better than English people. I can not say what will happen ten years from
now because, in India, English language was taught with more emphasis on the grammar.
Than on the spoken English, in the initial years. That is why, you will find that the spoken
English in India is grammatically more correct than anywhere else in the world. (m64NS
20:37-41)
Importantly, as these quotes include f25GS and f39VD showing similar sentiments, it is
clear that this age-based distinction is held by native and non-native IE speakers.
However, despite this shared assessment of ‘proper’ speech as located in the older
generations, the youth are not fully reconciled about whether their speech is ‘correct’ or
not. For example, this first example highlights f18ND’s uncertainty, while the second
quote initially sounds similar, but is in fact more complicated. f18ND, an 18 year old
female university student, explains:
It’s considered that if you speak English, you are educated. You know, there’s a certain
stereotype. “Oh, you speak English.” Even in our college, both N. [another university student
and friend] and I realized because we’re English honors students, we’re considered to have
lot of attitude and we’re considered snobs because we speak supposedly good English, which
I don’t agree with, because I find myself mispronouncing so many words, and Navni makes
so much fun of me, and then she’ll do the same thing and we’ll be like,”Oh, you know, we’re
in English honors. We can say whatever we want.” (F18ND 17:43-18:3)
And m19PS, a 19 year old male university student offers:
PS: She [m19PS’s mother] came from a very Cambridge type of school and all, this
background, so she doesn’t like slangs. She always has a very proper type of English.
VC: And do you feel like your English is proper?
PS: My? No.
VC: No?
PS: @@
VC: So how would you describe it?
PS: [HINDI toota poota all broken]. I mean, broken. @@@
VC: @@@, @@ So you wouldn’t - I mean, would you feel comfortable taking a test, you
know, that asked for grammar questions or something like that?
PS: I can. I don’t have any problem.
VC: Mm-hmm. But you don’t feel like your English - I mean, you were just saying PS: No, no, no. I don’t feel like that. @@ English is okay. I can do that. @@
VC: Okay.
PS: No that bad. @@ (m19PS 15:43-16:12)
138
m19PS describes his mother’s speech as more proper, while his speech is jokingly
described as ‘broken.’ The humor he is trying to convey is further found in his echo-word
rhyming [tutəputə]. This construction can indicate various things, including intimacy
between interlocutors, and the speaker’s indifference towards the topic (Das & Subbarao
2008). Either way—whether m19PS is confiding in me, or simply doesn’t care that his
English is ‘broken’—he goes on to reaffirm some measure of comfort in English. When
questioned further, he agrees that he would be fine on the sort of grammatical test that he
just said his mother would do better at. Thus, he frames his mother’s speech as more
proper, but also believes that his own speech is fine, at some level.
However, some IE speech is located as incorrect. Speakers can obliquely demonstrate
their self-assessed fluency and competency in English by discussing and separating their
speech from those who are not ‘good’ at English. In the following excerpt, f28SB, a 28
year old female museum and art curator, separates her own speech from those whose
English is ‘not very good,’ and highlights mispronounced words—not grammatical
errors—as the reason such speech is not good. She also explains the differences in
English competency as caused by whether one grows up in a city versus a more rural
setting, and as linked to schooling and friendship circles:
SB: The kind of English that we talk its quite all right. We can make out where like city
people, we have grown up in cities but there is one girl who recently joined she is from
Orissa. She talks very softly I can not hear, understand what she is saying. Her English is
not very good.
VC: Um-hum.
SB: You know, so I can make out that she is not from a private school and everything…You
know. There are some who are very good in the English, but I have not noticed one or
two or its not, I mean, grammatically may be fine but the pronunciation some words, I
don’t know, I was just this lady person I met, I don’t want to tell her name. She is very
good and you know, doing well and suddenly she said yeah and this and then I went to
his “cabin” [kɛbɪn] cabin -- I don’t know you call it “cabin” [kɛbɪn]. I don’t know? I
139
was surprised. Oh, my god, am I hearing right and many people say that, she is not the
only one. Many words, I get very turned off.
VC: Um-hum and did that, what does that signifies just they are poor and did not have such a
good education or?
SB: That they did not talk in English as much.
VC: Um-hum so do you think?
SB: I can make out her schooling maybe, I don’t know which school she is from, but I can
make out that it may not be convent education…Because we have many schools here and
its not that school I think it’s the friends that you make.
VC: Um-hum.
SB: Practice you know.
VC: Yeah, those are the ones that you speak most with yeah.
SB: Exactly, but I met this food critic and she should be good in English because she writes
also she doesn’t know whether potato [po.ta.to] is not good. Oh gosh, so it’s very turning
off and may be [HINDI garib poor, poverty-stricken] talking, but one odd word you hear
in those few sentences. (f28SB 31:17-32:8)
f28SB clearly doesn’t like hearing English words ‘mispronounced’—it surprises her that
speakers who are grammatically ‘fine’ should do this, and it ‘turns her off’ and she
hypothesizes that mispronunciations are linked to socio-economic status. Further, this
discourse locates f28SB’s own speech as ‘correct’ in opposition to such other speech.
Youth like f18ND and m19PS directly question their own fluency and nativity, indexing
globalized ideologies about nativity and standard language—they are more influenced by
global language ideologies than speakers who are more likely to favor IE over
international varieties, touting its simplicity and lack of accent, like f28SB. Further, these
youth can, as f18ND does, correlate ‘good’ English with snobbery, locating ‘good’
English as foreign: using ‘good’ English is snubbing the local norm, IE.
These younger IE speakers, whose voices we heard above, several of whom expressed
doubt as to their English meeting ‘the target,’ are English/Hindi bilinguals. Their entire
education, through college, has been English-medium, they use English in intimate
domains and with all technology (texting, email, internet, etc.) and their self-professed
140
dominant language is English. And yet, they are uncertain of how well their English
meets external, global standards. This insecurity may be a direct result of the disparate
power relationship between ‘inner-circle’ prestige holding nations and third world postcolonial outsourcing nations like India, wherein standard language ideologies directly
affect individual and societal notions of fluency, competency and nativity. However, it
may be more complicated—it may be that their bilingual repertoire excludes them from
considering themselves as fluent or native. Importantly, monoglot and native language
ideologies narrowly locate fluent English speakers as monolinguals (Silverstein 1996;
Shuck 2004), discussed later in this chapter.
Case Study: f27RG
Within all of this self-doubt by youth, one should not give up hope that these speakers are
not simply buying into global standard, native, and monoglot language ideologies. As I
stated in the beginning of this chapter, their views are often highly conflicted, and reflect
both their interpretations of IE and themselves as seen through a global lens and as
through a local lens as one means to establish, negotiate and maintain local social
legitimacy and authority. Here, I briefly examine the discursive and dialogic explanations
and thoughts offered by f27RG, a 28 year old woman with a master’s degree in
economics who is currently working in the non-profit sector. Reconsidering her above
discussion of English schooling, f27RG highlights how her own speech is interpreted by
her mother as ‘Indianized,’ and hence ‘basic’—not a positive connotation—vis-à-vis
other international varieties of English:
My mom used to complain when we were in school – I studied – most of the system in Delhi,
the education system, follows the CBSE, which is the Central Boards of Secondary
Education. And we don’t study things like Shakespeare at all. Okay? So our English is very
141
basic. Vocabulary is very limited. And so when we’re in school, we sit there and say things
like – after every sentence, there would be a like, you know? Like, you know, that, like, “My
mom’s gonna get really pissed off.” And she says, “What English is this?” And our stories,
even ‘til class twelve, used to be two pages long, you know? And it was no English. It was
English for the sake of knowing the language and nothing else. So basically, our English is
really very Indian English, Indianized English, pretty basic. (f27RG 12:18-30)
Importantly, f27RG values India, and has no strong desire to leave India, evidenced in:
But yeah, if I had continued to - like let’s say if I thought of a PhD, I probably would have
gone out. But for me, it was never something cool and, you know, fantastic, that I should get
out of India and study, you know…. (f27RG 7:29-31)
and:
I mean, I’ve never found any other place as exciting. I can not see myself living anywhere
else. So my struggle is, you know, here, that, you know, I wish all this away, and - but no,
not really thought about moving out and staying somewhere else, no. (f27RG 10:28-30)
Through the next few quotes, the above emphasis placed on staying in India is tied to,
and naturalizes her language background and language skills as common and locally
authentic.
She first contextualizes her English, and her mother’s reaction to ‘inadequate’
coursework which resulted in her ‘basic’ English skills by describing her language
background and current comfort with Hindi and English:
RG: Actually, right from the beginning in my family, we’ve mostly spoken in English, so my
Hindi’s pretty bad. And a lot of Hindi that I’ve picked up has been when I started
working, ‘cause a lot of my work was in Hindi. So I’m more fluent in English than I am
in Hindi. So even at home, we mostly speak to each other in English.
VC: And do you do any reading in Hindi, any RG: No, I can not. I - unless absolutely necessary. Again, work-related. Otherwise, if I just
pick up a book, it’ll take me years to finish it, ‘cause it’s…
VC: @@@@
RG: - I’m just not comfortable. (f27RG 7:37-45)
She has acknowledged that her English fluency (and competency) are challenged by
some (e.g. her mother), reflecting standard language ideologies. However, f27RG also
interprets her language practices as one aspect of her local cultural upbringing—within
which IE is her native language, and the language she is most comfortable with—and
142
suggests that her English home interactions must be interpreted as part of a larger, urban,
modern, and ultimately Western ethos:
You know, for example, my family, now, just the four of us, we’ve had a very Western, very
urban upbringing, and that’s why a lot of English and no Hindi, and that’s why things like,
you know, no restrictions on clothes or who you’re meeting, who - you know, who’s your
friend and where are you going. (f27RG 9:18-22)
She views her language practices as emerging from the local context, as part of a set of
modern urban practices, and hence, valid, in the local setting.
Her commitment to her local identity and local roots is further evidenced in her
explanation of the social unacceptability of rapidly adopting globally prestigious accents
for localized discourse. She explicitly disavos and mocks such behavior (known as ‘fake
accents,’ and explored further in the next section of this chapter), challenging standard
language ideologies which locate social authority in other English varieties:
The other is even funnier. This woman, she went to the airport and back to drop someone or
receive someone, and she came back with an [American] accent. Someone who went to
America for a three-week-long vacation came back with an accent….@@ (f27RG 12:44-6)
Meanwhile, she is aware that her comfort with English—and how English is (or was) an
important aspect of her identity and development—does not mirror the range of fluencies
in India:
English. Everyone wants to speak in English, you know? And you actually, people - I mean,
I was probably in school at one point of time, to be honest, I was a little, you know, “I need
to speak in English,” and, you know, “Hindi is not a cool language to speak in.” And I see a
lot of people - and now, I have to make an effort with Hindi, because I never really learned
by Hindi in school. But everyone wants to speak in English, so whether you go to a
restaurant - and as because when you go to a restaurant, if I call a waiter, I know that the
person is more comfortable in Hindi, so I’m gonna order in Hindi. But I have to kick a few
friends under the table to say, “You know, speak in Hindi. That person is not comfortable.”
We’ve - English has become a first language for most of us, and we forget that we’re talking
to people who may not be comfortable in English. (f27RG 11:42-12:5)
143
Clearly, IE and India are not homogeneous contexts for f27RG, and, instead she sees
regionally and culturally distinct pragmatics for language practices across the subcontinent—Stage 5 of Schneider’s (2007) dynamic model of PCE development:
I think we’re all, we're [India] all very, very different in terms of everything, whether it’s
language, resources, and all. When I go to the South, I do feel like I have a handicap because
I can not interact with people. You know, I’m - I do know - you know, even if I, even if I
speak in English, can I speak the English I speak usually, or do I need to, you know, bring in
some Hindi words, or say “da,” like people in Bangalore, “What da”? You know? Do I need
to do that? Do I need to say - like in Delhi, if I need to catch an auto, I say “bhaiya”. In
Bangalore, I’d probably say “boss”, because “bhaiya”, they would beat me up. You know,
Bhaiya means goon. So, I mean, you’re aware of these cultural specifics. So but you do feel
that a lot. You’re like, “Okay, how am I gonna talk to anybody? How am I gonna ask for
directions?” (f27RG 5:33-42)
F27RG acknowledges that a seemingly simple task, asking for directions, is in fact
complicated and important for assessing how successfully she can communicate in
English across India. She is aware of cultural, pragmatic, and lexical differences which
separate India into regions in which she is more or less comfortable using English for
communication.
Language, in this context, and for f27RG, is a tool for communication, as the above
example demonstrates; a tool to ease social distance, as her preference for Hindi when
ordering at a restaurant demonstrates; a tool for demonstrating and asserting local social
authority and legitimacy, and conversely, for attempts to tap into international regimes of
authority, as the fake accent example demonstrates; a product of where and in what social
context one is raised, as her discussion of language background as tied to an urban,
modern family demonstrates; a tool to be acquired, or forgotten, as its use in the
workplace varies, as her discussion of her Hindi literacy shows; and something which can
evoke and adopt a local flavor or touch, while also be devalued as basic through such
transformations, as her musings on her English schooling reflect. As I have demonstrated
144
above, f27RG is simultaneously digesting standard language ideologies as imposed
through school, family, and other popular discourse, while she also reflects on 1) her
identity, as an Indian with no emigration plans, 2) socially inappropriate attempts by
others to use internationally prestigious English varieties in the local context and 3) the
range of English skills she encounters within India. Her relationship to English is clearly
nuanced, conflicted, and dialogic with both local and global ideologies. The next section
explores more fully one theme introduced here: how IE speakers’ interpretations of local
authority and legitimacy are tied to language ownership, language access, and linguistic
practices.
Fake Accents
These youth dialogically acknowledge that international normalized language practices
and their associated social authority are not available for local uptake—they are
inauthentic in the local setting. They critically respond to local speakers who quickly
‘acquire’ international language practices, akin to the ‘cultural cringe’ towards American
language practices found in New Zealand (Meyerhoff & Niedzielski 2003: 536). In India,
discourse centers on ‘fake accents’ which typically emerge in high school and college age
IE speakers who unnaturally ‘acquire’ an American accent through limited contact with
AE speakers or travel to the US. These youth each had personal favorite stories
highlighting the ridiculousness of ‘fake accents’ and their response to such ‘wannabes,’
some of which I share here. First, f27RG, discussed above, offers two examples of ‘fake
accents’:
145
VC: When you were in school or in college, did you ever see anybody having a fake accent?
Like RG: Yeah. @@@@ There are all kinds of people. Okay. I had this - I knew someone in my
college, when I was doing my bachelor’s in Hindu College…This woman, she went to
the airport and back to drop someone or receive someone, and she came back with an
accent. Someone who went to America for a three-week-long vacation came back with an
accent. So there were all sorts. @ And I’ve had friends who’ve been there for so long
who don’t have an accent. So I think it’s, again, got to do with, you know, how cool you
think you are and emulating the West, like we’ve always done.
VC: And do people ever have like a British accent, or is it always American?
RG: American, surprisingly, yeah. (f27RG 12:33-13:5)
These speakers are not simply aware of such practices. Instead, they also reject and
denigrate such practices, as f18AU, a female university student, explains:
AU: Obviously you can not really, you know, kind of just go up to a person and say, ‘You
have a fake accent,’ but, you know, behind their backs, they obviously, you know, kind
of go like, you know, ‘This girl, she has a fake accent, and, you know, she’s such a
wannabe,’ this and that. That’s - basically, they just talk. Nothing really, as in, they don’tVC: So they wanna be- what do they wanna be? I mean, do they wanna just be American or AU: I don’t know. That’s what I don’t get, as in, you know, if you’re like an Indian, you
should be, you know, kind of proud to be an Indian. (f18AU 19:31-38)
They even actively call others on their ‘speedy’ acquisition of an American accent,
as another college student, 18 year old f18ND relates:
ND: And behind their backs, you can make fun of them, but some people are so direct,
they’ll just be like, ‘You did not have an accent before. What happened to you? You just
went to the U.S. for two months on holiday.’
VC: @@@ So –
ND: And it’s like, ‘Oh, yeah, I have got cousins there, actually.’ And you’re like, ‘Oh, that’s
not a very good excuse’… you can be a social embarrassment if you do that. We will
embarrass you.
VC: So do people – do they try it and then they knock it off, or do they keep it up?
ND: Some people, it’s very genuine. I mean, some people who’ve lived abroad, you know,
they can not help it. But other people here, we make fun of them and they stop after some
time. They deny it, but then they stop. (f18ND 16:2-18)
These stories and reactions tell us several things: first, they demonstrate an awareness
that language practices are not something one can change with the season—they reflect
one’s history and background, e.g., ‘some people who’ve lived abroad…they can not help
it [their AE accent].’ This in itself challenges globalized language ideologies by
explaining language variation through the socio-historical milieu in which local language
practices emerge and are learnt. Second, the fact that all the ‘fake accent’ stories related
146
to me involved the (improper) acquisition of an American accent speaks to the power and
global social authority of the US, and American accents: local IE speakers are aware of
such global commodification. Finally, ‘fake accent’ discourse illuminates IE’s local
social value. Youth who use ‘fake accents’ are interpreted as being embarrassed of their
heritage and nation and IE is taken up as a symbol of national pride and linguistic
confidence. For example, young Indians are expected by their elders, like 85 year old
m85SPT, here, to sound ‘Indian’:
VC: Is it all right with you if some young people sound British and some sound American,
or do you feel like they should sound Indian?
SPT: No, they should sound Indian, if they are so good they should not lose their nationality
pattern and all that. (m85SPT 12:27-30)
This ideological construction of IE as locally authentic, as indexing patriotism and pride
in India, and as a valuable local social commodity challenges globally imposed language
ideologies and assessments of social authority.
Further, it is interesting to note the difference between the international variety regularly
chosen by fake accent practitioners, AE, and the variety which is valued by non-native IE
speakers outside of my demographic, RP. These non-native speakers are operating under
a different set of values—they favor RP as more correct, while fake accent practitioners
favor American English as more prestigious.
Given that these IE speakers are sophisticated enough in their understanding of accent
acquisition to denigrate fake accents, it is not unexpected that they not only accept, but
favor IE, for its local authenticity. However, there are two additional reasons these
younger IE speakers may offer complicated, contradictory statements about their English.
147
First, their insecurity about their English competency may be linked to the role English
and Hindi play in their lives: older speakers believe that while the youth are gaining in
English competency, they are also losing their ‘Indianness,’ in part through their lack of
sophisticated and shuddh Hindi—that is, pure Hindi. Second, their uncertainty may be
linked to the emergence and rising popularity of Hinglish, a mixed hybrid form which
incorporates Hindi and English. These are explored in turn.
The Role of Hindi
The IE youth I spoke to are all fluent in spoken colloquial Hindi, but express varying
degrees of uncertainty and disfluency regarding their literacy skills. This is recognized
across the generations, by f18AU, an 18 year old female university student:
AU: We do have a Hindi class which we just have to pass. So yeah, we do have a Hindi
class, which no one really pays any attention to.
VC: Why not?
AU: I don’t know, because we actually have like so many other subjects and other things to
kind of focus upon that something, you know - something like Hindi, people don’t really
kind of pay attention to it and, you know, don’t give it much of importance. (f18AU
11:45-12:4)
By f27RG, the 27 year old whose conflicts towards English we examined above, who,
while initially suggesting that her Hindi is ‘bad,’ later qualifies this, and explains that she
isn’t comfortable with Hindi literacy, but uses ‘pure’ Hindi at work:
RG: Mostly in English. Actually, right from the beginning in my family, we’ve mostly
spoken in English, so my Hindi’s pretty bad. And a lot of Hindi that I’ve picked up has
been when I started working, ‘cause a lot of my work was in Hindi. So I’m more fluent in
English than I am in Hindi. So even at home, we mostly speak to each other in English.
VC: And do you do any reading in Hindi, any RG: No, I can not. I - unless absolutely necessary. Again, work-related. Otherwise, if I just
pick up a book, it’ll take me years to finish it, ‘cause it’s
VC: @@@@
RG: I’m just not comfortable… But I know my Hin - I speak in Hindi only when I’m
training, and then I’m speaking the clichéd Hindi, you know, the perfect - the [HINDI
shuddh pure] Hindi. (f27RG 7:37-45)
148
And by f59MS, a 59 year female who is currently completing a master’s degree in
Education focused on alternative and culturally-relevant pedagogical models:
VC: So you think there’s a growing population of kids nowadays, like in Delhi, that don’t
speak -MS: Hindi.
VC: -- Hindi.
MS: Yeah.
VC: But they all have to get around like at the shops and stuff?
MS: So then that’s just the basic requirement for that. I mean, Hindi is a rich, rich language,
and so the vocabulary is very limited, you know. You ask them to express feelings in
Hindi or, you know, phrases or adjectives, they’re at a loss.
VC: Uh-huh.
MS: And they don’t understand the words a lot. Like Hindi news, if somebody were listen to,
most of the kids today would not understand what the news reader was saying.
VC: I don’t understand the news.
MS: Yeah.
VC: But I can understand when people are talking.
MS: Yeah. Yeah. So that’s why I’m saying that, you know. So it’s not just idiomatic Hindi,
but even the -VC: Just the terms that are used.
MS: -- normal -- yeah. Yeah. The terms and phrases that you use.
VC: Yeah.
MS: They don’t understand at all. (f59MS 23:9-29)
This generational difference—where youth (myself included!) are uncomfortable with the
Hindi terms used on the news—is confirmed by f18MG, an 18 year old female university
student:
MG: For example, if it’s the news, I'll stick to English because the, you know, the words in
Hindi, I just don’t understand very complex words. I don’t understand them in Hindi.
And yeah, I prefer watching the news in English, definitely. I never go to the Hindi news
channels, go to the English one. And like if I have to watch a movie, then I’ll go to the
English channels, or if there some Hindi movie coming. It’s both, actually. If there's a
Hindi movie coming that I really want to watch, I'll watch that. So, there's no bias there.
VC: Will you have trouble understanding it and all?
MG: No. The Hindi movie, Hindi, that’s perfectly alright. The Hindi news, you know, there
are some words which are very, very high level. So, yeah. (f18MG 19:37-46)
These quotes illustrate that IE youth are not personally comfortable nor viewed as
comfortable with their Hindi literacy and shuddh Hindi skills, and, instead prefer English
for literary domains and news. Literacy has an interesting, non-linear relationship to
language domains. For example, it would be an overgeneralization to claim that Hindi is
relatively absent from intimate domains, like the home, given that Hindi speaking
149
servants inside the home are a common aspect of these speakers’ lives. However,
speaking with servants and other such interchanges requires oral, informal Hindi, not
written or formal Hindi. Cross-generationally, these IE youth are interpreted as having
less developed literary skills and less competence with shuddh Hindi than older
generations. As I will next explore, there is a stark difference between both which
domains and the proportion of domains in which older generations and youth use Hindi
versus IE, within which there is a growing trend for youth to see Hindi as irrelevant. We
will now explore the circumstances underlying why this is.
How is Hindi devalued?
The devaluing of Hindi can be understood in terms of the increasing value of English;
both economically, with job opportunities and financial success linked to English, and
socially, with English found in ever expanding Indian domains and English fluency
internationally valuable as a global social commodity. As well, this can be understood in
terms of the distinction speakers make between colloquial spoken Hindi and more formal
shuddh Hindi.
Economic value of English
Exploring how English skills increase one’s economic opportunities, f18AU, a female
university student explains:
VC: Like what’s wrong with the Hindi medium schools?
AU: I don’t know, because I feel these days, you know, the English is basically I feel like,
you know, the international type of language, as in, you know, it’s kind of understood all
over the world. And Hindi is basically kind of confined to India, as in, you know, not
even like - as in, I don’t know, if you’re don’t - if you’re not like, you know, well-versed
in English, then I don’t think you can kind of, you know, go places that way.
VC: So - so people kind of view English as like a ticket to success?
AU: Not a ticket to success, but definitely, you know, kind of a step towards success. @@
150
VC: @ So a necessary ingredient?
AU: Yeah, definitely. (f18AU 14:28-37)
This is endorsed by f28SB, a female working as a museum and art curator:
SB: Because we are encourage English all the more because so many opportunities for jobs
are thereVC: Um-hum.
SB: For call centers and we have the English Institutes that are coming in the last few years.
We keep seeing the adds that you are not getting the job you want because of English so
learn English from us and I heard somebody of doing that you know, so in fact, one of
my old colleagues in my first job his English was very poor, the way he was writing and I
would keep correcting him because he comes from Haryana you know. He is brought up
in Hindi environment so I keep telling him please go for the training and you get a good
job, you know, because you don’t want to stick to IT all the time, you want to rise up in
the managerial level.
VC: Um-hum.
SB: So, English is very important. (f28SB 33:19-31)
English proficiency means one has better job opportunities, and this correlation is
encouraged in English training course advertisements. Vernacular students, who I will
explore further below and who come from Hindi dominant rural areas, states like
Haryana, are especially stigmatized as having ‘poor’ English (Ramanathan 2005).
Further, the connection between language skills, job opportunities and prestige is made
explicit with f39VD’s explanation:
It is just that any language that’s your ticket to job gains prominence, you know. I mean, we
have French being spoken in England for 600 years because it was the court language or in
India before independence, Urdu and Persian was spoken and taught because those were the
court languages. Today English is a court language so that’s a language you learn; it is as
simple as that. (f39VD 21:38-42)
while opportunities for higher education are also correlated with English by 73 year old
m73KBS:
VC: And they went - so they went all the way through with English medium schools?
KBS: My children? Yes.
VC: Yeah. And did you like that? Would you rather they did some years of Hindi or KBS: I have no problem.
VC: No?
KBS: They - in fact, for all practical purposes, English medium was necessary for proper
study and proper understanding. (m73KBS 6:15-21)
151
And 62 year old retired pilot m62SCS:
SCS: No, these days, even in schools the emphasis is to learn English. That is the language
that is spoken all over the world if they want to go for higher education anywhere unless
they’re good in English, they can not compete over there. So most of the homes, even our
homes, our grandchildren, they are encouraged to speak English because Hindi
automatically they learn. Why? Because there are servants in the house they speak in
Hindi itself so that language they can learn easily so the emphasis is to learn English
these days.
VC: Do you think learning Hindi is important?
SCS: Basically, our country’s language is Hindi, so you have…know that language. You
represent a particular country, you have to know the culture of that country so Hindi you
have to know, but English is more important if you want to become something in the
world. (m62SCS 7:31-42)
While m62SCS suggests that English is necessary for higher education, he also suggests
that Hindi is also necessary, in order to know the culture of India, and to appropriately
represent India.
Meanwhile the opposite view is also offered, that English fluency is necessary to
represent India well. Here, m39GS, a housewife who works part time explains that
English fluency is expected for the Prime Minister, because through it, he represents
India in a positive light:
VC: And what about the Prime Minister?
GS: Our Prime Minister is very well educated. He speaks well.
VC: Does he normally give speeches in English or in Hindi?
GS: English.
VC: English. Do you think that’s appropriate?
GS: In between, we had a President who was a Sardarji, Gyani Zail Singh, who couldn’t
speak English at all.
VC: No?
GS: Which I felt was really bad because he goes abroad, he represents India, he should be
able to speak the language.
VC: So would he take a translator or -GS: No, he could -- he would speak broken English, not correct English. That, I felt, was not
good. (f39GS 28:10-22)
In representing India, a past President’s ‘broken English’ put him at a disadvantage: he
failed to present India in a positive light because of his ‘broken’ and ‘not correct’ English
skills.
152
Beyond not being a necessary ingredient to success, Hindi is just not needed, evidenced
by 27 year old working f27RG:
RG: And I have friends who’ve really forgotten their Hindi, can not speak a single sentence
in Hindi.
VC: And do they mind or are they okay with the way they’ve forgotten their Hindi?
RG: I think they’re okay with it because they don’t need it right now. I mean, they don’t
need it at all. (f27RG 14:35-9)
And 30 year old working f30PG:
PG: No. I can not read Hindi very comfortably. Writing is far-fetched.
VC: Um-hm. And do you feel like that’s, like do you regret that or do you feel like that’s
okay because you did not need it?
PG: I won’t say I regret it. I would say I did not need it. (f30PG 5:3-5)
As well, both the importance of getting good marks in school and youths reinterpretation
of Hindi as only useful as a spoken language for commercial transactions with lower
classes have played a role in youth choosing not to continue studying Hindi. For example,
18 year old university student m18DK explains:
VC: And you stopped taking Hindi when it stopped being compulsory.
DK: Yeah as soon as I could.
VC: How come?
DK: You just need to know how to speak Hindi in India
VC: So what did they teach you in class?
DK: Literature, grammar it’s actually pretty boring. English was much, much more
interesting always Hindi was not good.
VC: Oh, so the class wasn’t fun.
DK: Class wasn’t fun, I mean the teachers were usually not nice Hindi teachers have this
reputation of being…
VC: Being what?
DK: Weird, and strict and make you work hard. Hindi is actually you only need to know
how to speak Hindi here. Because, I mean, every thing in today like all documents and all
the government thing and every thing is English basically. So.
VC: Um-hum. Do you think people, if everybody keeps doing this of stopping in ninth with
taking Hindi do you think they all gonna forget?
DK: No. Nobody will forget and it won’t make any difference. People even had Hindi until
tenth and there tenth board percentage, their board exam percentage dropped because of
Hindi. I had French I got ninety in French that was my top, that was my best subject,
French. So I really enjoyed it.
VC: So the Hindi class is hard.
DK: Hindi is hard and you can not score that much in Hindi. Not that score. (m18DK 2:273:2)
153
And 18 year old university student f18AU:
VC: So why - you stopped in class eight?
AU: Yeah.
VC: So why’d you stop?
AU: Because we had an option of taking Hindi or a second language, and I chose French.
That’s why.
VC: And then some people in tenth would then switch to French?
AU: No, actually, it was - I think, yeah, class nine was like the time where you had to
choose between Hindi or your second language. So second language was basically
French, German, and Sanskrit. So people could choose between either of the two and
they had to continue in class ten.
VC: And did most people pick Hindi or French or what?
AU: I think people picked the second languages, as in, most of the people.
VC: Why do you think they did that?
AU: I don’t know, actually. Because, you know, we’d all heard that, you know, Hindi kind
of gets really tough, you know, as you go into like higher classes. And, you know, it’s
basically that the tenth boards also matter a lot and, you know, we’d heard that people
don’t really score well in Hindi. So that would - that also basically made a difference, so
that’s why we opted for the second language.
VC: So what kind of things do they do? Is it mostly writing and reading in these Hindi
classes or AU: In Hindi classes, basically, we have this, uh, book and we have like grammar and we
have writing. So basically, everything combined. It’s not really much, actually, the
syllabus, but people don’t really wanna do it. @ I mean, neither do I. So yeah. Because
we have like so many other stuff to do other then attending Hindi classes. So, ‘cause even
today, I had - my classes finished at four and I had Hindi from like four to five. And just
because of the attendance, I was, you know, kind of there in the Hindi class. So that’s
why. Because we have like so many other things to do. We have like, you know, regular
tests every day for like our, you know, proper math classes that we don’t really have the
time for Hindi. That’s why. (f18AU 2:14-42)
Here, Hindi is considered a ‘hard’ subject, one that can impede on economic success,
because university admissions depend on high school exam scores, and university
admission is highly competitive.
As well, Hindi classes are avoided (in some senses, this is a classic diglossic situation)
because written Hindi doesn’t have relevance to these youth—it is only seen as useful as
an oral form to the same speaker, f18AU:
AU: We do have a Hindi class which we just have to pass. So yeah, we do have a Hindi
class, which no one really pays any attention to.
VC: Why not?
AU: I don’t know, because we actually have like so many other subjects and other things to
kind of focus upon that something, you know - something like Hindi, people don’t really
kind of pay attention to it and, you know, don’t give it much of importance. (f18AU
11:45-12:4)
154
This is not a random phenomenon—none of the younger generation of college-age
students that I interviewed had continued in Hindi after it wasn’t compulsory in school (at
the eighth standard, for all of them). English, meanwhile, is considered ‘interesting,’
showing that it has symbolic local value, and is identified as being important for
deciphering government documents, and more broadly, as one skill necessary for
economic success—this is explored in detail after we explore the association between
types of Hindi and the relationship between language loss and cultural loss.
Shuddh Hindi
Shuddh—pure, or uncontaminated—Hindi is a common theme for these speakers, who
separate colloquial spoken Hindi, which they are to various degrees competent in, from
shuddh Hindi, which would require schooling to acquire, and which several of them lack,
but want to learn, at some level, as 18 year old university student f18MG muses:
It depends on what I want to watch. If it's something that - for example, if it’s the news, I'll
stick to English because the, you know, the words in Hindi, I just don’t understand very
complex words. I don’t understand them in Hindi. (f18MG 19:37-9)
And 18 year old university student f18ND also explains:
ND: Actually, I wanna relearn Hindi myself again. So yeah, I think it’s kind of important to
- because a language has like a lot of cultural meaning to it. Like something like a
[HINDI: jhoota dirty, polluted (slang)] in Hindi, how you are gonna translate that into
English? You can not. There’s no concept for it. It comes in the caste system. So, I mean,
we - you can not say spoiled, because it doesn’t signify everything. So it’s kind of
interesting that way.
VC: Mm-hmm. Do you feel like you’ll be able to? Like are you confident in your level of
Hindi now or ND: I’m not very confident in my level of Hindi, but I’m thinking maybe I can take a two,
three month course, because I just need to go back to familiarizing myself with the
language. Like I can read in Hindi, but it takes me at least ten minutes to read two pages.
And [HINDI shuddh pure] Hindi, that’s like the pure, good Hindi, it’s harder to read. The
way we speak is very different from good literature Hindi.
VC: So what you wanna study is the reading and writing?
ND: Yeah. (f18ND 20:1-14)
155
But, what is shuddh Hindi—is it more ‘pure’ because it doesn’t have English words
mixed in? Below, f30PG’s fiancée, who is from a more rural Hindi speaking region, but
currently lives in the US as a highly successful businessman, is also located as not
speaking shuddh Hindi. Therefore, shuddh Hindi does not separate urban and rural
speakers. F30PG goes on to explain that while he is very comfortable with Hindi, neither
of them speaks shuddh Hindi:
PG: But he is very, very comfortable with Hindi. He is more comfortable than I am.
VC: Um-hum. And when you two talk together, is it normally in English or Hindi, or say
mix?
PG: It’s mix. When we chat it’s English, when we talk on the phone, it’s Hinglish. It’s a mix.
VC: Is it ever totally Hindi or is itPG: Not really. Not completely Hindi. I mean every sentence would at least have couple of
words thrown in from both the languages.
VC: Um-hum.
PG: So, no it’s not like a shuddh Hindi sentence. No, no. It’s pretty much mixed. (f30PG
5:30-37)
Shuddh Hindi thus isn’t simply avoiding English codeswitches. Instead, returning to
f18ND’s explanation, above, shuddh Hindi is interpreted in a different way, and
explained in the following quote. f32NM, who used to work for Aaj Tuk, an Indian TV
station on an English channel, locates shuddh Hindi as having ‘big words’ which she did
not know:
Yeah, all in English. All in English, except for Aaj Tuk, which was in Hindi, and we used to
struggle like rats to do it. And because, you know, we would write scripts in English and then
go to one Hindi guy and get it translated. You know, the shuddh Hindi and, you know, big
words, et cetera. (f32NM 16:1-4)
As well, m64NS, a 64 year old retired military man, correlates it with larger words, but
also suggests that these words won’t necessarily help you to be understood across the
country:
Hindi, again, is a language, which is spoken differently in different parts of the country. You
can have a very high flown Hindi which probably nobody will understand what you are
talking about, but usual spoken Hindi that’s simple, mix of Hindi/Punjabi so it is a very
simplified day to day. (m64NS 15:36-39)
156
Shuddh Hindi thus isn’t seen as a part of normal Hindi speech, nor is it even intelligible
to the bulk of India’s Hindi population. Instead, it is linked to more formal registers, like
Hindi news reports through references like ‘high flown’.
Indeed, it is complicated to define shuddh Hindi—as one Indian NRI put it, shuddh Hindi
is
Hindi without any bastardization from other languages—using less Urdu words, and more
Sanskrit words. It’s harder…most Hindi speakers don’t speak shuddh Hindi.’ .
Shuddh Hindi is linked to India’s political history—during British colonization this same
speaker explains, ‘Hindi did not move forward…terms for new things were in English.’
Indeed, shuddh Hindi originated immediately after Partition and India’s independence in
1947, during the 15-year planned move from English to Hindi as India’s national
language. At this point, Hindi’s lack of vocabulary for, among other things, new
technologies and ideas which had been introduced during the colonial era became
important. Corpus-based language planning steps (Vaish 2008) were taken to remove
English and Urdu terms—because Urdu was now linked to a different country and
religion, Muslim Pakistan—and to replace them with newly created Hindi terms, and
with preexisting Hindi terms which had fallen out of use by affluent, more westernized
populations. Shuddh Hindi terms were disseminated through media, for example, via
India’s first TV channel, Doordarshan, and continue to be used on Hindi news channels,
but were not—and still are not—features of natural Hindi as spoken by the masses.
While shuddh Hindi exists, it is not commonly spoken—it is a ‘standard’ without a
community of speakers. It is thus not surprising that these IE youth don’t speak shuddh
157
Hindi—it requires formal training, and is contrived. Their evocation of shuddh Hindi is
framed within standard language ideologies, wherein colloquial Hindi is non-standard
and hence not ‘good,’ and shuddh Hindi is standard, and correct. Lacking shuddh Hindi
competence, they will never be able to see themselves as good, or correct Hindi speakers.
Thus, it isn’t clear, from their descriptions, which index shuddh Hindi and standard
language ideologies, how fluent in Hindi they actually are, because these assessments are
ideologically constructed.
Linking Language Competency to Cultural Knowledge
The youngest generations’ discomfort and disengagement with shuddh Hindi and Hindi
literacy has powerful ramifications for how these youth are interpreted. The older
generations in particular see and disapprove of the youth’s lack of advanced Hindi
vocabulary and literacy. They interpret urban fluent IE youth as losing their culture, of
failing to understand, appreciate, and embody India’s rich historical, cultural, and
linguistic background because they are (ostensibly) less fluent in Hindi, as 39 year old
f39VD explains:
Increasingly, as I see more and more people moving away from their own language to
English I see them losing a part of their culture, a big part of their culture because language
carries nuances of culture you know, and I see that with my, my own niece, who is fourteen,
you know. And sometimes when I use certain words which for me are like common place
and she turns around and says, “Huh?” And I’m like, “But you live in Hindi heartland, how
come you don’t know this?” and she is like, “But we never speak Hindi like that.” You
know, because there is so much emphasis on speak English correctly speak English, speak
English, speak English, but you lose part of your culture and you tend to, we tend to learn
English at the cost of our own language. I think that’s, that’s the other thing that’s happening
so I hope we will be able to do something about keeping some parts of our language this and
that. (f39VD 20:29-39)
These youth are not learning about their heritage by experiencing it through Hindi use,
and elder generations are saddened by this, as 73 year old housewife f73TM explains:
158
VC: And do you - do you feel like they speak as much or more English than you spoke in
your TM: They speak English all the time. They speak English in such a way, even my servants
can answer them back.
VC: Do you like that?
TM: They miss a lot. I sometimes tell them you don’t even know your origin, your language,
but if - a time will come, perhaps you will miss, but it will be too late. I would rather
suggest give it a try from time to time instead of regretting it. (f73TM 13:23-30)
These youth are also evaluated as ‘non-Indian’ by more rural, Hindi speaking
populations, as 28 year old KC, who works for an international company, demonstrates:
KC: They don’t consider you an Indian, in fact. In fact, they don’t - I am frequently mistaken
for anything but Indian.
VC: Like what?
KC: People assume I’ve lived abroad all my life. People assume I wasn’t born here. People
assume - I’m not very, very Hindi either. So, you know, when I’m traveling in India, I
have a problem communicating with people beyond the point. I don’t speak the slang.
And there’s different slang in every state, of course. You have to be quite fluent at Hindi
to manage that slang. Because everybody will speak pure Hindi, which for me, coming
from my mouth really, that doesn’t sound like pure Hindi so they assume there is - and I
also belong to a social strata which is slightly affluent, and that doesn’t work very well
when you’re in the rural areas in India. That goes against you. So you’d rather be humble.
(m28KC 9:45-10:9)
Younger, less fluent Hindi speaking youth are assessed as failing to be ‘humble’ through
their inability to use Hindi slang, which marks them as having been raised abroad (which
is not true), as being affluent (which is true). The Indian diaspora in particular is cited as
a situation where English fluency is at a premium, and Hindi fluency is not. Diaspora
youth are assumed to be non-fluent Hindi speakers, who do not need Hindi, because
‘everyone knows English,’ as a 73 year old housewife demonstrates:
VC: Now, when your son in the - or your nephew in the U.S. gets married, do you think it’s
important that his children speak Hindi?
NK: I don’t think so. Everybody knows English. But maybe they will pick up themselves they would want to pick it up, you know? Otherwise, everybody knows English. (f73NK
18:11-14)
f73NK does suggest that some diaspora youth would want to pick Hindi up, but wanting
to pick it up locates them as not having acquired it so far. This ‘desire’ is echoed in the
following quote, by 80 year old housewife and volunteer f80CK, who suggests that some
diaspora youth are actively interested in Hindi, because of its cultural links:
159
VC: And what if they did not learn Hindi also? So pretend they only learned English. Would
that be bad or is that okay also?
CK: [HINDI ye kya he what this is] they’d learn Hindi, and they, children, the girls, that she
wears Indian clothes. She loves the Nani - my daughter comes here. She take all the
[HINDI unintelligible] with a lot of work and all that. And the other girl, they cry to
seeing her with-- children like this, all this…[HINDI ki that], where for you get the
clothes Now whenever she takes, she takes some for her cousin also. Maybe she wants
them. This time she came, [HINDI aaj today] she wanted to buy [HINDI pile anklets with
bells]. She bought [HINDI pile anklets with bells]. Silver jewelry, bangle, and they took
some for her friends also. And that’s good. Sweet children. (f80CK 10:16-25)
Returning to the situation in India, the eldest generation places responsibility for their
grandchildren’s’ lack of Hindi fluency on their own children, the middle generation,
suggesting that their children did not encourage Hindi in the home enough while their
grandchildren were growing up. For example, m81SPT, an 81 year old retired engineer
who currently volunteers as a science and math teacher to underprivileged ‘slum kids’
explains:
VC: So their parents did not talk to them in Hindi when they were growing up?
SPT: No, they don’t. They, even now they don’t, I tell them, why do you speak them in
English? In house they speak English. They all just speak in English the same way when
we used to talk in our own language and like to talk in Hindi. We love our children, now
they used to talk with them in English and they, themselves speak some in Hindi or
English. There is the difference. How we will be losing our Hindi and ourVC: So when you ask them, why do you speak in English to your children, what do they say?
SPT: Nothing. They say that my daughter doesn’t know Hing- Hindi, that’s why he’s saying
that. @@@
VC: And how do you feel about people losing their Hindi?
SPT: Very bad.
VC: Do you feel like they are losing more? Like their culture? I mean, are they still Indian?
SPT: They’re still Indian, but when they don’t speak Hindi, don’t, they don’t learn the
culture off our ancestors our then naturally you don’t know what, who was Rama, who
was Buddha, who was all that. Harishchandra, they don’t know. They don’t know this
history. Whatever is writ- (m85SPT 13:5-20)
In particular, people who are not literate in Hindi do not read and experience the
Ramayana and Hindu mythology more broadly, as these quotes (directly above, and
below) highlight. Some parents personally blame themselves for this, citing their own
lack of Hindi fluency, and even framing it in terms of their own discomfort with Hindi, as
this 52 year old hospital administrator and mother does:
160
VC: Do you ever read a Hindi…?
GG: I can not, no.
VC: No. Do you have -- do you do any reading in Hindi, like books or -GG: No.
VC: No.
GG: I can not because, you know, I -- I have to concentrate on each word. And by the time I
reach the end of the sentence, I can read it, but it takes, just takes forever, you know, and
it will never get done. So I don’t have the patience to...
VC: And when you're out shopping -- I mean, if there's signs that are in Hindi -GG: Yeah, I can read.
VC: That's -GG: Yeah, yeah, yeah. I can read Hindi. I can speak Hindi. I can, you know, communicate
in Hindi. But it's not my first choice. And I don't think I've ever bought a Hindi book or a
Hindi magazine or a Hindi comic or -- no, now that I think about it, I don't.
VC: And your daughters, did you ever see them with those? Did you ever buy them Hindi?
GG: I bought them Hindi dictionaries because they've always -- always, you know, run into
trouble with vocabulary and this and that and -- no.
VC: So they, they did not read comics or little stories or anything in Hindi?
GG: No. Even, you know, all the Hindi mythology -- the Hindu mythology, all the stories
was all in English. This Amar Chitra Katha -- everything's in English now, and it's so
simple and no, not really. Oh, god, I sound like a horrible snob. @@@@@ (f52GG
10:41-11:16)
Hindi literature is not introduced, and Hindi fluency is so low that Hindi dictionaries are a
necessary component for the youngest generation, to help them with their vocabulary.
Further, f52GG’s final statement ‘Oh, god, I sound like a horrible snob’ suggests that she
sees her failure to have her children read Hindu mythology in Hindi as disrespecting
Indian culture: children can not get the same cultural knowledge through English
translations.
Why did not these parents push Hindi at home? These parents do not themselves read
Hindi literature, in part because ‘everything’s in English now.’ Has this been a problem
for these parents? No—f59MS, below, who is working in later life on a master’s degree
in Education, has ‘never felt a lack’ of her Hindi literacy:
MS: I can not read Hindi.
VC: So they did not teach you that in school or you just did not keep it up?
MS: Just enough. Just -- I mean, English because I think we only had Hindi up to the eighth
standard, not in the ninth, tenth, eleventh, and twelfth.
VC: Uh-huh.
MS: So not in higher secondary, only in the middle school. And that also is very basic.
161
VC: So you’re fine with street signs and stuff like that, but a whole novel would probably beMS: Even street signs, I don’t think I can read them.
VC: But there’re so many signs in Delhi that are -- no?
MS: I’ve never felt a lack of it, of not knowing Hindi, but -VC: That’s interesting.
MS: Simple Hindi, maybe I’ll be able to read, but -VC: Sound it out?
MS: You know, too many of the matras and things and I’m a little confused. And I’m slow.
Like if I’m going behind a bus and I have to read the destination, it takes me time to.
VC: Yeah.
MS: Yeah. So my Hindi is not very good. (f59MS 25:6-24)
These upper middle class, urban parents have not placed a premium on Hindi literacy in
their own lives, and have likewise skated over the possible benefits of Hindi literacy for
their children. Given this situation, it is not surprising at all that their children are not
literate in Hindi, are less familiar with Hindu mythology and culture, and may question
their own Hindi fluency vis-à-vis shuddh Hindi.
As well, circumstances can also work in collusion with individual Hindi literacy and
fluency. For example f64AG’s husband was in the Air Force, and they moved around the
country on various postings while the children were growing up. Meanwhile, she is
uncomfortable with Hindi—she ‘did not do any reading in Hindi.’ This has a more
powerful effect on children’s’ acquisition of Hindi when they are posted in non-Hindi
speaking regions of the country:
AG: No, no. Not at all. No, I did not do anything - I did not do any reading in Hindi. It’s
mostly English, mostly English. And again, you know, like when we move around, the
boys, they studied in English medium schools and I have been - I’ve been the one looking
after their studies mostly. So as we’re tutoring them all the time, it’s all English. We
don’t get a chance. Only when they had to do the Hindi in the Sanskrit, I could teach
them a little bit, but mostly, it was English.
VC: Mm-hmm.
AG: So, you know, we never bothered to do Hindi. We did not learn much Hindi. Or, rather,
read in Hindi, not learn. Read much in Hindi. (f64AG 13:38-46)
The urban youth who I interviewed are perceived as speaking English all the time, and
not fluent in Hindi which, to their elders, is not good because it involves a loss of culture
162
and shared history. Importantly, the elders don’t specifically cite illiteracy as causing
cultural loss23—to them, spoken Hindi fluency and knowledge of advanced shuddh Hindi
vocabulary are critical. Meanwhile, the youth claim to be able to get along in spoken
Hindi—they have enough Hindi to suit their needs—but are disengaged from any active
interest in formal Hindi literacy and are only abstractly interested in learning shuddh
Hindi. That is, none of them is taking—or has taken—a class on or in Hindi, or done
some similar activity to improve their knowledge of spoken shuddh Hindi or Hindi
literacy. Given that none of the three generations that I interviewed in this community
emphasize Hindi literacy, a wider disconnect is likely to emerge, separating urban upper
class IE speakers from rural, lower class Hindi-medium educated speakers (e.g.,
Ramanathan 2005; Vaish 2008). These urban affluent Indian communities are likely to
further shift into English dominance, and have, for the most part, already ‘lost’ their
formal spoken shuddh Hindi and Hindi literacy.
‘Vernacular’ students
In fact, there is already a contrast made between two communities vis-à-vis Hindi fluency
and literacy. Youth who are not English dominant are lumped together and called
‘vernacular’ students, as a 73 year old housewife describes, reflecting on her grandson
and his generation:
23
Most of the oldest generation were themselves taught Hindu mythology orally, by their grandparents.
However, this tradition stopped with that generation—for example one 62 year old IE speaker remembers
‘When I was growing up, my [HINDI dadi paternal grandmother] told us all the stories of the Ramayana
at home. It was a regular thing. But my mother didn’t, and then I didn’t [with my own children].’ The
current oldest generation thus grew up learning their culture orally, but their parents (that is, now-dead
great-grandparents of the current youth) didn’t continue this tradition, and nor have they, with their own
children. The current youth’s lack of cultural knowledge, then, reflects changes bigger than their own
illiteracy—this is a product of technology. People with TV and radio and the internet don’t sit around all
winter retelling stories, and these oral stories are being lost.
163
They don’t care if it is Gujarati speaking. They call them - what - what do they call them, the
students? Vernacular - vernacular student. They named them. (f73TM 13:32-4)
‘Vernacular’ students are much more likely to have attended a non-English medium
school and university, and thus be more fluent in a vernacular Indian language and less
fluent in English. Labeling these students—who are also typically from more rural areas
and lower socioeconomic groups—as vernacular students (above) or Hindi medium
students (below) sets them apart from the urban youth, who are English dominant.
However, the sociolinguistic gap separating this fluent—and affluent—IE community
from the vast majority of the Indian population also has its downside. This IE community
is not ‘typically Indian’ in a larger sense: they do not represent the majority of India in
their language skills, in their cultural competence, which was challenged above, and
because of their socioeconomic background—in short, the population I examine here is in
the minority, across India. Vernacular students, who comprise a much larger population
within India, are both physically and socially separated from these IE youth, a process
which continues in the university setting. m18DK, an 18 year old studying political
science, explains that his courses are taught twice, once for Hindi medium students, once
for English students like him:
DK: We have a few Hindi medium students that means they studied all their subjects in
Hindi, so what happens is the class is one hour long so first half an hour will be in
English and the next half an hour will be in Hindi. So it’s actually nice because in last
half an hour for us is like free, we can talk, we can work.
VC: So you don’t have to listen, they teach the same thing twice.
DK: Yeah.
VC: First in English and then in Hindi.
DK: Yeah, they have to, because there are Hindi medium students.
VC: So these students are gonna graduate having taken all their classes in Hindi. But isn’t it
a English medium college or is it both?
DK: Both.
VC: I see.
DK: Exams are also in the both the mediums.
VC: I see.
DK: We have a choice of answering in Hindi or in English and the questions are also in
English and Hindi, so.
164
VC: And those people who are in Hindi medium, why do you think that they haven’t
switched to English?
DK: Because they were in schools that did not have English as a medium, probably.
(m18DK 15:20-38)
m18DK doesn’t hang out with these Hindi medium students, and attributes their
enrollment in the Hindi medium courses as a result of their earlier education being in
Hindi. However, this is not simply a choice the vernacular students have made—to
continue their studies through Hindi medium courses. Lower class and rural students are
systematically restricted from English courses throughout their schooling (Ramanathan
2005), and thus are restricted from the social, educational and economic benefits which
are associated with English competency. For example, English seats are limited in
university courses, and students with less English background, like 25 year old f25GS—
who as we learned earlier, is from a lower-socioeconomic background and lives in a
Hindi speaking household—often find themselves shunted back into vernacular medium
courses without any recourse:
GS: Then again, in my graduation, I had done it in through Hindi medium.
VC: Oh.
GS: It’s just that my two years, I have spent in English medium, then my graduation, I spent
- like again, I went to the Hindi medium. So then there were some problems I couldn’t
join - I couldn’t do it because the seats were full, so I couldn’t do it in the, you know,
English medium. So I had no choice. I had to do it in Hindi medium. (f25GS 12:5-10)
Even though f25GS wanted to study in English, she wasn’t able to—clearly, there is an
‘English Vernacular divide,’ as Ramanathan (2005) so aptly titles her book, one which
increasingly also separates the groups through their Hindi literacy. Fundamentally, this
continued sociolinguistic stratification is understood as linked to affluence, urban/rural
distinctions, and ‘better’ cultures, as f60SS, a 60 year old housewife explains:
VC: Mm-hmm. And are there areas that are known as being better at English speaking and
worse at English speaking? Like are there certain areas you’d go to and feel really
confident everybody’s gonna be speaking like good English?
165
SS: That’s, that's, again, depending upon the education, and it depends upon the social
cultural strata which you come from. Because the higher - the better the strata, better the
culture - the person, how cultured he is, English is bound to be better, bound to be
grammatically right, every way correct. And if you - for example, I asked this girl [her
servant] to speak. She doesn’t know two words, but she can understand. If I tell her,
“Come here,” she can understand. But she cannot tell me “come”, “I’m coming.” She can
not say that.
VC Mm-hmm. And that’s true across India?
SS: And again, college and school, that plays a very vital role. The better your college better means to say, I would put colleges in different, this is a average college and this is
a good college and this is an extremely good college. Obviously, like my daughter, I told
you - or granddaughter, they’re not allowed to speak any Hindi. Their schools they go to
are very expensive schools, private schools. They have to speak in English. There is no
two ways about Hindi. They won’t let you. So as a result, the child learns English. So
again, it depends upon - you can say it depends upon--money. The better the college, the
better the school, the more expensive the fees, the better the teach--quality of teachers.
See, it’s the kind of steps, you know? The quality of teacher. The principal will also come
from a good background. So the better the background, the better the teachers are able to
teach your children. Then those kids learn better English. Then those - there are some
schools which are free schools, like the government pays for them. They’re free. The
tuition is free. Everything is free. Are the teachers going to be dedicated? Not at all. If the
child learns English or Hindi, how are they bothered? They are not bothered. Why are
they not bothered? Because the government pays their fees. They are not answerable to
the parents. If you’re answerable to the parents, like in private schools, you’re
answerable. You know if the child learns a bad word, then the parents come charging to
the school and demanding to know that, where has that child learned that word? But in a
private school - in a school which is free, like government school, no, the teacher will
probably tell the parent, “Take the child out. We are not interested in teaching.” So it
depends upon the social strata, education, money. So many conditions involved with that.
(f60SS 11:28-12:16)
More expensive private schools—the ‘better’ schools—prohibit their students from
speaking Hindi, and through this, and the diligence of their teachers, the children learn
‘better’ English. Meanwhile government run public schools, which are tuition free, and
hence more accessible to lower socioeconomic strata, are shackled with less dedicated
teachers who have no personal drive to make sure their students learn English, as
opposed to the de facto language which these children enter school speaking, Hindi.
The youth of the IE community I studied both place themselves and are also interpreted
by their seniors as separated from less urban, vernacular speaking Indians, because of
their English fluency, and the quality of the education they have received (which depends
largely on their financial prosperity). And yet, while f60SS correlates better teachers with
166
learning better English, there is also respect for the person who is literate in both English
and Hindi (Shobhana Chelliah, p.c.). While these IE youth are seen as having diminishing
Hindi fluency, ideologically Hindi is a valuable component in one’s repertoire, even if it
is not a much-used component. I next explore two ways Hindi is valued.
Language Maintenance Plans
Language ideologies and loyalty often manifest through future plans for language
maintenance. Here, some speakers, like this 39 year old, are rather matter of fact that
their (hypothetical) children will acquire the language(s) spoken by their parents:
VC: So, what if somebody had married like a Marathi speaker?
VD: The kid would have learned both the languages, so what’s the big deal?
VC: I am just asking if you think they would have learned both.
VD: They would have learned.
VC: They would have learned.
VD: They would have learned! (f39VD 16:206)24
Meanwhile, f18AU, an 18 year old university student, places an importance on Hindi,
regardless of where one is living, but also sees speaking the local language as important
for communication:
VC:
AU:
VC:
AU:
VC:
AU:
VC:
24
Huh. So, so when you have kids, let’s say - let’s talk far, far future.
Mm-hmm.
You plan to teach them Hindi, it sounds like.
Yeah.
Yeah. And what if you moved, say, down to Bombay or something?
Mm-hmm.
Do you think that they should learn Marathi25?
While my repetition of f39VD’s comment may appear, decontextualized as it is here, to be directly
questioning her, in a manner which could be interpreted as rude, in fact she had just finished relating how it
was OK that a diaspora family member’s children were growing up as English monolinguals, explaining
that it was ‘the parents choice.’ I then posed a hypothetical question about her siblings, to which I then got
a very different response, that ‘[t]hey would have learned [both]’. While I was momentarily confused and
repeated her, in retrospect, f39VD is saying that what is OK in other families—be they diasporic or local—
is not the same standards which are held to in her family, who would encourage and/or expect children to
speak both of their parents’ languages.
25
Marathi, an Indo-Aryan language in a different sub-family from Hindi, is just one example of languages I
suggested, because it is a dominant language in other Indian states. I also used, for example, Oriya, an
Indo-Aryan language from yet another sub-family which spoken in Orissa, an eastern state, and Malyalam,
a Dravidian language spoken in Kerala
167
AU: If basically - see, actually, if it’s kind of a medium of instruction there - as in, you
know, I mean, if people kind of - like, you know, like Hindi is like very prominent here in
Delhi, as in, you know, everyone speaks Hindi most of the time. So if people speak
Marathi over there, then I guess they should, because they to make themselves
understood and they have to understand when people speak. So I think yeah. (f18AU
16:14-25)
Interspersed with speakers’ positive associations with Hindi is an awareness of the
communicative function of language: if one moves outside the northern India Hindi
Belt region, Hindi may maintain its personal cultural significance, but other
languages become more important for daily communication.
Meanwhile, other speakers, like f50RV, a lawyer with NRI (non-resident Indian) diaspora
family and a grown daughter working in Bombay, are motivated to pass Hindi on to the
next generation because it marks a child as special, as Indian, with an accumulation of
other cultural artifacts which link one to India:
VC: So if your - say your daughter shifted to the U.S. and got married, do you think it’s
important for her to raise her children speaking Hindi?
RV: Well, I think it’s nice to know your own language somehow. Because then that -- I
mean, I think it’s important to have a link with your own roots, your own country. It’s
important to know the language and to speak it also.
VC: Uh-huh.
RV: Like it’s important to eat Indian food also. Okay. I'll, I enjoy American food also, but
it’s somehow like -- it’s also nice to enjoy Indian food if you’re an Indian.
VC: Uh-huh. Do you think that -RV: Why lose touch with the -- you know, because I can not speak my own language in the
sense what my mother speaks, Derewal, as well as I would have because they never
spoke with us in that language. And maybe I did not pick it up as well, but it’s -- I mean,
this language will be lost, no, soon, if I don’t speak this language.
VC: Uh-huh.
RV: And, similarly, I think Indian food is - its, it’s an art by itself so why should we lose it,
you know? And if - it’s nice to have your own individuality also. (f50RV 23:18-33)
f50RV’s parents both spoke Derewal (or Multani—the name varies depending on which
parent I asked), a Punjabi/Multani dialectal mixture which is dying out. Her parents
actively chose not to teach f50RV and her siblings Derewal/Multani, as her 71 year old
mother explained,
168
No, I did not teach them, because when we came here in Delhi, Hindi was prevalent.
Everybody used to speak Hindi. So we used to speak in Hindi and when they went to school,
they started speaking in English also. (f71ST 6:7-9)
f50RV feels the loss of her parents’ language—it speaks to her heritage, in the hills
between Punjab and Pakistan, and it is, at this point, inaccessible to her. Given this
historical framework, which she evokes above, it is not surprising that she believes that
children should be taught their heritage languages. True to this belief, f50RV did teach
her children Hindi.
However, several, while firm in their desire to promote Hindi in the next generation, are
also aware of the reality surrounding them, where youth are not maintaining their Hindi,
as 39 year old f39VD explains:
VC: So if you were to have kids then you would speak to them inVD: Hindi and English.
VC: Both?
VD: Um-hum.
VC: And what if, say, for example hypothetically, what if your husband said, “no, no I only
want them to learn English?” Would you argue with that or?
VD: I would argue with that, yes.
VC: So you feel strongly?
VD: I do feel strongly about this.
VC: Um-hum. Have you ever discussed this with your siblings that have kids?
VD: When they speak to them in Hindi as well as English so I am okay, I mean, I keep
pulling my sister’s leg because my niece reads only English novels and I am like why
don’t you read Hindi but she is not, apart from I mean she does get excellent marks in
Hindi exams but she doesn’t really read Hindi per se, you know, but I mean that’s her
child, the way she wants to bring it up. (f39VD 21:43-22:11)
Meanwhile, the IE youth I interviewed are nostalgic about Hindi in the grand scheme of
things and several have vague plans to teach their children Hindi, as part of their ‘roots’
as 18 year old f18MG explains:
VC: When - if you have kids, do you think - do you want them to learn Hindi also?
MG: Yeah, I would definitely. Yeah.
VC: And what if you ended up shifting outside of the country? Do you think it would be
important then?
MG: See, of my kids, I would want them to definitely know Hindi, because that’s our native
language. But English is very necessary because, you know, they're going for a job, they
169
going out of - they're stepping out of their house, they need to know English. And like
you said, if I'm shifting to some other country, then definitely English is something that's,
I think, really, really necessary. Yeah. And Hindi is something that is ours, you know?
It’s our language, so would want them to know Hindi as well. Yeah. (f18MG 19:12-21)
These plans evolve to be more pragmatic as they approach childbearing age, and begin to
realize that they would have difficulties passing Hindi on to the next generation, because
of their lack of fluency in it. For example, f30PG is about to be married to a US based
NRI, and, reflecting on her Hindi, she realizes that it will be difficult to teach her children
Hindi without the surrounding Hindi culture of Delhi:
PG: No. I can not read Hindi very comfortably. Writing is far-fetched.
VC: Um-hm. And do you feel like that’s, like do you regret that or do you feel like that’s
okay because you did not need it?
PG: I won’t say I regret it. I would say I did not need it. So, I mean, yeah, my only concern
is if tomorrow if I have to teach Hindi to my children, I will be a little handicapped
because I am not good at it. So, I will probably take my mother’s help or mother-in-law’s
help, but I don’t have any regrets because all said and done, it’s very patriotic to say that
you have to, you should be good in Hindi and you should respect Hindi and all that, that’s
fine. I am not saying I don’t respect it, but let us face reality today. You don’t know
English, you can not get anywhere in life, so I mean I don’t want to be like a
pseudopatriot and say no, no, no. I really regret it. I don’t because I am not that bad. I can
read it. I mean I might take an hour to write a paragraph, but I can write it. So, there is no
need, right. There is nothing, which I am doing today, which is in Hindi, so it’s okay as
long as it doesn’t harm my career and doesn’t harm, and as far as speaking is concerned, I
am fairly comfortable, very, very comfortable, so no problem. (f30PG 5:2-16)
f30PG locates those who say Indians should all respect Hindi and be good in Hindi as
‘patriots,’ however she suggests that the situation is more complex. While one can
respect Hindi, English is necessary to get ahead in life, and Hindi is likely to be lost, in
this situation.
Meanwhile, respect is also interpreted in other ways. About to get married, m28AG,
talking about his future children and his siblings’ future children, says:
AG: It should be their choice. I mean if he comes here he will, I mean, they will
automatically learn Hindi. It’s their choice you know it really doesn’t matter if, I mean,
my nephew turns out to be English speaking, I mean, you know how does it matter.
VC: Um-hum. And is that how you and your soon to be wife feel also?
AG: Yeah, see, I mean, honestly you expect I mean, you expect your nephew to be
respectful to you that’s it and the person who, kids these days are very mature. They’re
170
very, I mean, you don’t want brats. I mean, you don’t mind kids as brats, but you don’t
mind like them to brats, brats types you know. Like notoriously brat bratish about things.
As long as they you know are respectful it should be fine. And if they speak English,
Hindi doesn’t matter. I mean, I normally, speak English the entire day now. (m28AG
8:31-40)
m28AG hopes that his children and his nieces and nephews will be respectful towards
their elders, but sees no reason to push Hindi on them—he himself speaks English ‘the
entire day now.’
One 32 year old mother of two, who worked before having children, is now dealing with
her daughter’s demand that the family speak more Hindi at home:
And I’m, you know, struggling to kind of, you know, tell her that it’s okay to speak Hindi.
Of course we should speak Hindi, but, you know, Hindi, you will always learn in school.
You know? It’s the standard, you know, [HINDI matri bhasha mother language] of the
country and what have you. But, you know, if - since you speak English well and you know
you can, and speaking good English is an advantage, you shouldn’t let go of that, you know?
But how does one try and convince a four and a half year old? It’s a bit difficult, you know?
(f32NM 14:22-27)
Hindi is portrayed as something one ‘should’ learn and f32NM is dialogically (Bakhtin
1981) responding to local—and perhaps national Indian—ideologies—which she doesn’t
share—that forge this link. However, as is clear from her next statement, she herself
believes that English is crucial for advancement, and needs to be nurtured in the home,
while Hindi ‘you will always learn in school.’ However, this isn’t true: as we saw above,
several prestigious schools don’t allow one to speak Hindi at school, and the youngest
generation—having just finished high school—are not very comfortable in their Hindi,
beyond basic communication.
As well, some speakers, like 27 year old f27RG, value bilingualism, but do not feel that
Hindi in particular would be beneficial or necessary for their (future) children:
VC: Mm-hmm. Do you think it’s important, like if you have kids, for them to learn Hindi?
171
RG: I think it is, yes.
VC: Mm-hmm.
RG: I think, yeah.
VC: Do you - what if you did not live - like say you got married and ended up living outside
of India. Do you still think it’s important or, or less so?
RG: I think it would be nice, but I’m not gonna grill them over it. I think it would be nice,
you know, to know another language, I think, just for the fact of knowing another
language.
VC: Mm-hmm. So if they happened to - like say they lived in a French - say you lived in,
you know, France, then would learning French be just as good as learning Hindi?
RG: Absolutely. Absolutely, yeah. (f27RG 7:46-8:10)
Hindi, especially Hindi literacy, is portrayed as unimportant to future success. College
age students value Hindi, and abstractly believe that it would be ‘good’ for their children
to know it, but do not have plans which operationalize this, given disconnections between
their own Hindi fluency (and literacy) and these goals. Meanwhile, all speakers in their
30’s recognize that English will be central to their children’s’ success. Some respect
Hindi, and within this plan to recruit their parents and in-laws’ help in teaching Hindi to
their children, while others see no need to push Hindi on the next generation, either
because the children will learn it outside the home, or because it simply doesn’t matter.
These anecdotes suggest that in India and abroad, Hindi maintenance will decrease in the
next generation.
Collectively, these youth are viewed by the older generations as having lost part of their
cultural heritage given their continued involvement with English, and, while bilingualism
is idealized as good, speakers are loosing such bilingual fluency themselves. Their plans
to pass Hindi on to the next generation are conflicted, and sometimes unrealistic, while
they unanimously value English as a path towards success. However, as we saw much
earlier, these youth view their parents’ and grandparents’ English as better or more
172
formal. What, then, is left for them? As I shall next discuss, youth are embracing a hybrid
form—Hinglish—as theirs.
Hinglish
Schneider dates the explosion of Hinglish, a local term to describe a mix of Hindi and
English, to the mid-1990’s, in conjunction with the introduction of cable TV and youth
channels (2007:170-1). Hinglish is evoked by all generations, while it is highlighted by
30 year old f30PG as the local norm for youth—their most common mode of
communication:
VC: Um-hum. And then, do you and your brother speak in English or in Hindi?
PG: Hinglish is more like it.
VC: @@
PG: Yeah, Hinglish. When we generally fight, we fight in English where otherwise it’s more
like Hinglish. I mean starting my generation and everybody is very, very fluent in
English. In fact, we speak more English than we speak Hindi but it’s a mix even with my
cousins and my other brothers, sisters everybody, it’s more or less Hinglish. (f30PG 4:406)
And an 18 year old female university student:
We use a lot of slang and we also use a lot of Hindi, Hinglish slang. (f18ND 18:36-7))
And a male 19 year old university student:
VC: So with your friends, you speak English, Hindi or AS: Hinglish.
VC: Huh?
AS: Hinglish.
VC: Hinglish. Do you feel like - like does anybody ever get annoyed and say just speak one
language or just speak the other language?
AS: No, nobody does because everybody talks the same way…
VC: So would you say you speak all the time in English, half the time in English?
AS: Half the time.
VC: And the rest of the time is in Hindi?
AS: Hindi. Hinglish. (m19AS 13:43-14:3, 13:12-16)
Youth thus not only acknowledge codeswitching, but have a term for it—Hinglish,
which, in a sense, serves to institutionalize and normalize the process of codeswitching.
Youth often brought it up in these interviews without being directly questioned, as their
173
normal communicative choice. However, standard and monoglot language ideologies
also come into play with Hinglish: these ideologies are acknowledged, and also
dialogically challenged, as an 18 year old university student reflects:
And sometimes, the way we speak, my parents will look at me like, “What did you just say?”
I mean, you know, like it doesn’t make any sense if you actually think about it. We’re like,
“Oh,” you know. And we’re not so bad. It’s acceptable. (f18ND 18:37-9)
f18ND evokes standard monoglot language ideologies and evokes her parents’ reflections
on Hinglish by suggesting that Hinglish mixes ‘don’t make any sense if you actually
think about it.’ However, she dismisses the criticism she just evoked, by referencing the
local and reframing her Hinglish evaluation with ‘we’re not so bad,’ concluding that
Hinglish is ‘acceptable.’ In short, for these youth, there is not a strong negative local
value associated with Hinglish for her generation. Yet, this attitude towards Hinglish is
not shared across the generations:
GS: Here -- in fact, I normally find, even in Devangi and Chandrika’s school, they’re
speaking more of sort of a Hinglish -- it’s mixed Hindi and English. They do not speak
proper English.
VC: So they just switch back and forth constantly?
GS: Yeah. And maybe even in the same sentence, a few words may be English, a few of
them, Hindi.
VC: And that’s with all the children their ages speaking?
GS: Yes….
GS: I feel you should speak the language correctly, but whether you’re speaking Hindi or
English, but speak it properly. I don’t like this mixture.
VC: I see.
GS: It’s sort of -- if you don’t know either of the two languages properly, then -VC: So do you tell your daughters to pick one?
GS: Yes, I do.
VC: What do you say to them?
GS: I just keep sort of -- if they’re speaking in Hindi and putting some English words, I sort
of repeat the same sentence in full Hindi or in full English, sort of.
VC: Uh-huh. And your husband, does he do the same or -GS: No.
VC: So he doesn’t care as much?
GS: He grew up here. (f39GS 21:19-26, 21:39-22:6)
From these quotes, it is evident that Hinglish is valued differently across age groups.
Older generations condemn Hinglish as ‘messy’ or a sign of non-fluency in either
174
language. As well, as f39GS suggests above, and f30PG, below, confirms, Hinglish is not
a pan-Indian phenomenon:
Like for instance my office itself has offices in Bombay and Bangalore and Chennai. We just
opened a Chennai Office, so I’m speaking to my colleague in Bangalore. I can make out
because he wouldn’t understand my Hinglish, mostly they don’t. (f30PG 16:44-6)
This is not unsurprising, given that Hindi is not a required nor native language for most
people in the south. However, in other regions, mixes of different local languages with
English are found, e.g. English and Tamil.
While Hinglish is highlighted as used by youth, it is also common in older generations,
and the business world, as the preceding quote, by f30PG, and the following quote, by 39
year old f39GS, demonstrate:
GS: At this NGO I am working for, I find it’s mixed over there.
VC: Is that harder to understand than -GS: No.
VC: No.
GS: What happens is anytime there’s a slightly difficult word in one language, you switch to
the other language. So, you know, understanding is not the problem. The problem is that
you’re restricting your own vocabulary because you don’t bother to think -- you know, if
you think for a second, I’m sure you can find the correct English word or the correct
Hindi word for it.
VC: Uh-huh.
GS: But you’re just not bothering to take that much - to make that effort.
VC: Uh-huh. So it’s a bit of laziness?
GS: I think so….
VC: And when you’re talking to your friends, like these people you go out to lunch with
once a month, would that be -- what would that language -GS: Mostly English.
VC: Maybe a bit of Hinglish?
GS: Most -- yes. Actually, quite a lot of Hinglish now that I think about it . It would be, I
think, mostly Hinglish.
VC: So when you’re in this situation, all the women around you are speaking this Hinglish,
and you’re responding either in English or in Hindi? Or do you then switch into Hinglish
also?
GS: No.
VC: You don’t?
GS: I try to avoid.
VC: So you don’t like that?
GS: Maybe I do at times, but I try and avoid it. (f39GS 22:29-41, 23:30-43)
175
f39GS views Hinglish as lazy, because, as she puts it, ‘anytime there’s a slightly difficult
word in one language, you switch to the other language’. While she doesn’t like Hinglish,
f39GS eventually admitted, at the end of this quote, that she does occasionally use it.
Importantly, f39GS, the most vocal nay-sayer about Hinglish, was excluded from
quantitative analysis because she grew up in Bombay—thus, her interpretation can be
viewed as standing in contrast to how most Delhiites feel about Hinglish, and as adhering
more strictly to outside standard and monoglot language ideologies.
Meanwhile, Hinglish is specifically construed by the oldest generation as a deficiency, as
one 60 year old grandmother explains:
Then my daughter, her daughter - she has one daughter of about four years. That little girl
also goes to an English medium school. So the girl does know more English than Hindi, but
at home, her grandparents speak in Hindi. So the child is a bit confused, you know? What is
English and, I mean, who do I speak to in English and Hindi? So she has now got a mix of
speaking Hindi at home and English in school. And when she comes here - and I try and
insist that she speaks in English, because I feel that is a language one should know. That’s a
universal language. So she comes, she does speak. But again, not too fluent, but she’s able to
make herself understood. And so it’s kind of a mixture. English-Hindi mixture. (f60SS 3:1624)
The oldest generation often evokes the future value of English, encouraging English and
discouraging Hindi. These negative views of Hinglish, dominantly held by the oldest
generation—though don’t forget f39GS, above, who grew up in Bombay—, are
particularly striking because their interviews were peppered with Hinglish, while the
younger generations, like 28 year old m28AG, below, seem to have a stronger control
over their (non)use of Hinglish in interviews:
VC: So you feel more comfortable in English now orAG: It’s not comfortable, but it’s the way I am talking to you right now.
VC: Um-hum.
AG: I mean, honestly, I am mentally saying that you know, do not shift to Hindi, but in
normal condition I will keep talking in English then suddenly shift to Hindi and then get
back to English. Doesn’t really matter. (m28AG 8:45-9:4)
176
For the youth, in these interviews Hinglish was often used as a functional tool restricted
to terms which don’t translate well. It is not clear how to reconcile the fact that they both
valorize Hinglish and did not speak in Hinglish in the interview, while the elder
generations are less positive towards Hinglish, but used it more in these interviews. It
may be that youth are better at ‘monitoring’ their own speech, or it may be that the older
generations were more comfortable in the interviews, and hence more likely to switch
into Hinglish. Supporting this, the interviews with older generations were more
collaborative and lasted longer than the interviews with younger generations.
This cannot be separated from the overwhelming hours that younger generations work in
the private sector: their hectic, workaholic lifestyles may play a role. In the following
excerpt, f32NM is venting about her husband, who works extended hours, and who really
doesn’t have time to be a husband to her and a father to his two children:
VC: Does your husband work long hours?
NM: Yeah.
VC: Because P. [a co-worker of f32NM’s husband] works NM: yeah.
VC: - a lot. And she seems to work when she’s at home and NM: Yeah. Yeah, yeah, all the time. 2:00 in the morning, 5:00 in the morning, all of that. He
comes home at 11:00 and he’s still on the phone, and I’m like, “Can you shut the damn
thing off?” You know, I was in labor, he was on the phone, you know? The baby was
born, he had the laptop. I said, “I’m gonna throw that damn thing out.”... But, you know,
now, when you’ve got like wailing kid and potty on your hands and, you know, diaper
not there, and you just don’t have enough hands, and your husband’s on the phone or
doing a presentation, you’re like, “What the fuck?” You know? You’re just - you know
what I mean?
VC: Yeah.
NM: So I don’t care what the people in his office - I don’t know how much it costs all of
them, you know, because - I mean, and - I mean, you know, and all these so-called single
people, it’s still okay, because, you know, they get home and they just have to eat and
sleep and that’s it.
VC: Right. They’re not NM: But he has hundred one million things to do, you know? And people don’t just - they
just don’t - in India, they don’t respect that if a person’s gone home, just lay off. You
know? You can speak to the guy in the morning. You know? It’s - you know, world is not
gonna come crashing down, you know? You’re not George Bush. Even he goes off on
holiday.
VC: Every weekend.
177
NM: Yeah, every weekend. You know? So you know what I mean?
VC: Yeah.
NM: So they’ll call him at 10:30 at night, at 11:00. I mean, he’s burping the baby or he’s
whatever or - you know, I mean, there’s 1 million things to do. I’m talking to him about a
check he has to write or, you know - we don’t - the romance has gone out of our lives like
that. I mean, there’s absolutely no time. I mean, weekends are spent taking one kid to the
doctor, then the other kid to the doctor, then the groceries, then the this, then entertaining,
or going out somewhere and it’s all a process. You know, get yourself ready, get the kids
ready. Have the dogs been fed? Has the next, you know, meal been cooked? All of that.
And you get into the car, and suddenly there’s a call, and that’s it. You know, just when
I’m thinking, “Okay, in the car, I have maybe half an hour to actually talk to my husband
and figure out who is he now,” you know? And there’s a damn call. It’s very
irritating….But having said that, one has to have the ability to just take out that much
time for yourself and your family. You know? (f32NM 25:27-26:33)
This is common: several of my interviews with this working generation took place at very
odd hours—after 11pm, for example. While f32NM suggests that this workaholic
lifestyle is because her husband’s co-workers are single, it appeared to me that several of
the married 30-somethings I interviewed faced similarly brutal workplace schedules.
Now, imagine being in the shoes of f32NM’s husband, with two young children, and a
hectic work schedule, which clearly affects his limited family time, and suddenly being
asked to sit and talk for an hour or two in the context of these interview—it is very
conceivable that these younger, working IE speakers view this interview process as an
extension of their workday, as another obstacle to spending time with their families.
Indeed, I never interviewed her husband—he was too busy. Coming off work, as most of
them were just before the interviews, they are also more likely to be in ‘English mode.’
This area clearly needs further investigation, but tracking down this generation is very
difficult, and I have no clear suggestions on how to accomplish this.
178
Meanwhile, even more complicated language combinations and labels are also evoked—
and contested—as local norms by some of eldest generation (who, remember, also often
have Punjabi background, as this 85 year old does):
FS: English is mixed with Hindi and Punjabi so you can say Hinglish Punjabi.
SPB: Hinglish Punjabi
FS: Punjabi-Hinglish
SPB: In Punjabi English, we’ll say.
FS: Hinglish.
SPB: This English is different from the other English you speak, but whoever is Punjabi
English. @@@
VC: So when you take walks in the morning, is that what you speak, this Punjabi English?
SPB: As I told you, we are all mostly Punjabis, then why should we talk in English? We
enjoy in -FS: It's not English, it's Punjabi-Hinglish.
VC: Ah-ha.
SPB: Punjabi-Hinglish.
FS: It’s all combined. The words of Punjabi -SPB: One word of English and other Punjabi.
FS: It's Hinglish, not English.
SPB: Not English. This is what we talk?
VC: Uh-huh.
SPB: And they also enjoy. You can not talk to them all this and that in English, no, I think.
We talk in Punjabi with one or two words which we can not translate in Punjabi, English
word we will use. @@
VC: Uh-huh. @@@@@ (m85SPB 14:22-43)
Punjabi English or Hinglish Punjabi, either way, multiple languages are being mixed in
regular conversation. There is a functional utility to this—‘We talk in Punjabi with one or
two words which we can not translate in Punjabi’ then expressed in English.
Collectively, while the youth are dominantly positive about Hinglish, older generations
are conflicted, with purists, like f39GS negative towards Hinglish because it
demonstrates ‘laziness,’ while others, like f60FS and m85SPB more neutral or positive
about Hinglish. Over time, this non-fluent use of Hindi, within Hinglish, as it is now
practiced by urban youth may lead to an increased dependency on English, and language
shift towards English in the next generation of urban, affluent Indians—I am curious
what direction(s) of shift and development will emerge in the coming years.
179
In tandem with this, there are benefits associated with IE, several of which were evoked
above—English is a path to success, and will not be acquired without focused English
use at home. We now explore more fully these perceived benefits of IE fluency.
Benefits of IE
IE offers several benefits to its speakers, and their valorization of IE can be interpreted in
terms of these. In particular, three themes were regularly evoked within the interviews: IE
holds economic value in the global market—as opposed to Hindi’s perceived global
value—, IE is associated with a newly formed urban Indian identity, and, finally, IE holds
a lingua franca role within India, which allows speakers across India to communicate.
Economic Benefits
The first major—and most constantly evoked, as we have already seen—benefit is IE’s
value as a global economic commodity: it is a variety of English, and English skills are
highly marketable. In India, almost all of the higher paying jobs require strong English
skills, while globally, English skills are a necessary precursor to working with
international companies and customers. For example, Lakshmi N. Mittal, CEO of India
based ArcelorMittal Steel Co., the world’s largest steel company, told employees ‘Our
business long ago evolved from being local to being global. We need, therefore, a
common language to help drive the business forward. Fluent command of English is
indeed a priority, and it is a valued asset that may expand your career opportunities’
(Smerd 2007). The economic value of English in this setting cannot be understated, and I
180
offer several quotes to justify this statement. First, by 64 year old m64NS, who is retired
from military service:
My personal opinion is that, yes, English being an international language, it is always an
asset. That if you understand English, if you can communicate well in English, it is always
an asset. (m64NS 20:22-4)
An 18 year old university student:
AU: I feel these days, you know, the English is basically I feel like, you know, the
international type of language, as in, you know, it’s kind of understood all over the world.
And Hindi is basically kind of confined to India, as in, you know, not even like - as in, I
don’t know, if you’re don’t - if you’re not like, you know, well-versed in English, then I
don’t think you can kind of, you know, go places that way.
VC: So - so people kind of view English as like a ticket to success?
AU: Not a ticket to success, but definitely, you know, kind of a step towards success. @@
VC: @ So a necessary ingredient?
AU: Yeah, definitely. (f18AU 14:29-37)
Another 18 year old student, who views English as economically important both inside
and outside of India:
But English is very necessary because, you know, they're going for a job, they going out of they're stepping out of their house, they need to know English. And like you said, if I'm
shifting to some other country, then definitely English is something that's, I think, really,
really necessary. (f18MG 19:17-20)
A 28 year old woman working in the museum and art curator field, who highlights its
importance—and access, through English training institutes—for the masses:
SB: Because we are encourage English all the more because so many opportunities for jobs
are thereVC: Um-hum.
SB: For call centers and we have the English Institutes that are coming in the last few years.
We keep seeing the adds that you are not getting the job you want because of English so
learn English from us and I heard somebody of doing that you know, so in fact, one of
my old colleagues in my first job his English was very poor, the way he was writing and I
would keep correcting him because he comes from Haryana you know. He is brought up
in Hindi environment so I keep telling him please go for the training and you get a good
job, you know, because you don’t want to stick to IT all the time, you want to rise up in
the managerial level.
VC: Um-hum.
SB: So, English is very important. (f28SB 33:19-31)
A 39 year old working part time in the NGO sector:
VC: I see. Do you feel like if somebody doesn’t learn English, that there’s certain jobs
they’re routed into or certain jobs -GS: Yes.
VC: So what sorts of jobs if you don’t know English would you end up in?
181
GS: Would you not get? Oh, I think most of these corporate offices and all do want good
English, especially the multinational companies. They do want a good English-speaking
background. And most offices, these days, do ask for good English. (f39GS 29:36-42)
A retired military and commercial pilot:
SCS: No, these days, even in schools the emphasis is to learn English. That is the language
that is spoken all over the world if they want to go for higher education anywhere unless
they’re good in English, they can not compete over there. So most of the homes, even our
homes, our grandchildren, they are encouraged to speak English because Hindi
automatically they learn. Why? Because there are servants in the house they speak in
Hindi itself so that language they can learn easily so the emphasis is to learn English
these days.
VC: Do you think learning Hindi is important?
SCS: Basically, our country’s language is Hindi, so you have know that language. You
represent a particular country, you have to know the culture of that country so Hindi you
have to know, but English is more important if you want to become something in the
world. (m62SCS 7:31-42)
In short, English is understood as necessary ‘if you want to become something in the
world.’
Several participants from the oldest generation demonstrated past success which they
attribute to this belief, relating how English fluency has positively influenced their own
success, and has motivated the educational choices they’ve made for their own children.
For example, one 87 year old describes how English impacted his life as a British soldier
during World War II (before India gained its independence and had its own military):
RKB: Actually, the language plays a very important role in everybody’s life. And I found it
so right from my beginning, when I went to my college. And I found that even in the
school, if I had not been made to read, write, and speak English in the school, I couldn’t
have gone into the army, because in the army, in those days, all officers were British.
And there are about fifteen officers in each battalion. So when I became an officer, I was
one amongst the fifteen. One Indian and fourteen British officers. And if it were not for
good knowledge of English and my school and college, I couldn’t have managed to be an
officer in the army. It helped me a good bit, because those officers, British officers, were
not educated.
VC: They were not?
RKB: No. They could only speak English because they were from Britain. But I was
amongst those who had been to college and, uh, I could speak as well as them or
probably better than them. So it helped me a good bit during my profession, as long as it
was British Army. Well, once we reverted back to Indian Army, things changed then, it
was [unintelligible]. But it helped me a lot because when I was in the Middle East, we
had a mixed lot of, um, British soldiers, American soldiers, African soldiers, Australian,
New Zealand, all - but all speaking English. Those who wanted to have any job done, it
182
was mainly through English language. So it was a good, nice thing for me. (m87RKB
19:32-20:2)
While another 85 year old man explains how English has impacted his life, his
children’s’ lives, and his grandchildren’s’ lives:
VC: And was your BA degree, the courses, were they in English or in Hindi?
SPB: English.
VC: English.
SPB: No question of Hindi. I did everything in English. English subject, then political
science, et cetera. Other subject I took in English.
VC: Uh-huh.
SPB: And that is why, because I was brought up in a village, remote village. My children,
they got very good education. My elder son H. is a master of MA. He did in history. And
my second son, who is in USA, he's [unintelligible] engineering. He did here in Delhi.
And third son is also graduate. He's in Toronto. And my daughter, she also studied with I don’t know where. She also got this graduation. Because I myself could not do even
college in village, I was very keen, get all my children should get proper education.
Although it was difficult for me to give so much education because I'm getting a not so
much good salary at that time.
VC: Uh-huh.
SPB: But I was very keen that they should be graduates, everybody. And her husband is
master in history.
SPB: They all got proper education.
VC: Uh-huh.
SPB: That’s why they're very successful, and secondly, I gave them the education in
missionary schools and colleges.
VC: Are those the best?
SPB: So that they can speak their English very well in America, and they say you did very
good thing, Papa. You gave us good education from a small village you came and gave us
when we are very successful in this country. In America, we are -- we can talk to
anybody. There's no problem, and we are very well fixed up and getting all their houses,
property, everything, and money-wise, they are very, very happy.
VC: So do you feel like English is a very important part of -SPB: Yes, certainly -- English is foremost. Everybody should learn. Not only go in this
private school, but they should get -- I should say get education proper, very good
colleges also, famous colleges. My grandson, I am proud to say, he went from
Massachusetts to economical -- college of economics in London also. (m85SPB 4:2-5:11)
The oldest generation believes that a ‘proper’ education in English is necessary—it
allows one to live in other countries, to communicate with other English speakers, to be
financially prosperous, and, ultimately, to be happy. Interestingly, m87KB, above, points
out that his English, in the 1930’s and 40’s, was ‘better’ than British military personnel,
simply because he was educated. There is a strong link forged between education and
183
English here—no speaker suggested that one could be considered ‘educated’ without
English skills.
Meanwhile, while f59MS agrees that IE is viewed as an economic asset, she sees
downfalls to this belief:
MS: And I believe that a lot of parents say that, no, no, give them the English books, even
the little ones.
VC: And that’s because it’s -- you’ll get better jobs or -MS: Economic, it’s English.
VC: But getting a better job typically also means like leaving a village and going to work
elsewhere -MS: Uh-huh.
VC: -- and people don’t mind that?
MS: No, they don’t. They -- I mean, they seem to think that this is where the prosperity is.
(m59MS 24:1-9)
In past quotes we have seen that f59MS does believes that English offers access to
financial success. However, here she acknowledges that this is not a given, and is
problematically applied by other parents: this is clear through her pointed choice of
pronouns, switching from ‘I’ to ‘they’ and her explanation that she ‘believes’ that
parents encourage English books ‘ because they ‘seem to think’ that such leads to
‘prosperity.’
In fact, some Indian communities are thought of as being left behind with the move
towards English: m81SPT separates urban ‘good’ English speakers who use
English in intimate domains from rural Hindi speakers, because scientific
developments are all in English. He speaks from a position of first hand knowledge,
given that he is a retired government engineer who currently volunteers as a science
and math teacher for underprivileged ‘slum kids:’
Those who are in urban areas, they, those children or those brought up in urban area, they
speak English well, and in a different, between themselves. And, but majority of India is not
184
urban people. There are villages also where they don’t know English at all. Even they hardly
understand Hindi. And therefore, no doubt, for educated people or for the scientific
development, for other development now the English has become the media. Now, at this
stage when the science has built up so much and we are also built up so much space, atomic
energy, all those things we have developed ourselves and agriculture. And all feel that the
English as become very important to us now. (m81SPT 11:44-12:6)
m81SPT does, however, reiterate that English competency is necessary to secure a good
job in the urban Delhi context.
Contradicting m81SPT, IE is even seen as accessible to non-fluent lower classes by a 73
year old woman who is a lifelong volunteer. She is very involved in the religious and
social teachings of Sai Baba, within which she is devoted to helping ‘slum kids’ and
underprivileged communities:
VC: Right. So you’re also saying that across India, people are using English as a common
language. Do you think it should become a national language?
NK: It is already a national language, leave aside politics.
VC: Uh-huh.
NK: It is a national - all the books, everything is in English, you know. And I mean,
politicians can say anything, what they like. Wherever you go, English is being used, and
it will be used.
VC: Yeah.
NK: It doesn’t - it doesn’t belong to now, you know, only U.K. and all that. It is a common
language.
VC: And do you feel like it’s being used by more people than when you were younger?
NK: Uh-huh. People are trying to learn it.
VC: And do you think that’s good?
NK: Yes. Why not? It opens more vistas for them. Even if slum area child wants to learn
English, even if servant wants to learn English. You should see how my servants writes
his account, in English. So, it’s good. Why not? Because they also feel they can get a job
anywhere if they know a little English better, the- their prospects are better. They could
get a place in embassy or somewhere, so. And they also now say thank you. When we say
thank you to them, they also thank us. I think it’s good.
VC: Uh-huh.
NK: It is only narrow-mindedness to think that English only belongs to England -- it belongs
to all, and there’s so much literature. You do not know so much Sanskrit or Hindi as you
know about English. (f73NK 14:31-15:7)
f73NK believes that English belongs to all of India, in that it can open vistas for all
Indians. Her explanation must be interpreted vis-à-vis her continued commitment to how
globalization and modernization affect and can help rural and underprivileged
185
communities. English ownership is linked to the economic advantages it can bring, and
this declaration of ownership does not link English to a particular socioeconomic,
regional, urban vs. rural group. f73NK also supports this stance by discussing the
language practices in her volunteer and devotional work in southern India as involving
English and other southern Indian languages, within which she frames English as a
‘common language’ for Indians:
VC: And what about the Sai Baba community? Is that - is most of the discussions in English
or - I don’t even know what his native tongue is?
NK: [HINDI nehi no] it’s not a Sai Baba’s community. It was, the local languages also, and
English also. Whatever the audience is.
VC: I see.
NK: Some days, the audience will not understand English. Then we have to talk in local
language, which is not only Delhi but South India, you know, Karnataka, Tamil Nadu, so
it can be their language, but most of Indians know English. And English is a common
language. I many not know Tamilian, Tamil and I may not know Telugu, but I know
English. You see, that way English is a common language now. I think we’ve gotten so
used to English now. I don’t think even there’s any difference [HINDI ki English main
bat keri thi ki Hindi main if I was speaking in either English or Hindi]. Sometimes we
consciously - if some people don’t understand, then we talk in Hindi, you know.
VC: Right.
NK: And mostly you find even the workers also understand English better. Didn’t you notice
this thing?
VC: I think it depends on where I am that I’ll see it more or less. It seems more in South
India.
NK: They understand English. (f73NK 13:34-14:5)
f73NK, when comparing speaking in Hindi versus English, is referencing their
contentious history, wherein Hindi was supposed to take over for English, in 1965, as an
official national language. She explains that English has become such a lingua franca in
India: Indians have ‘gotten so used to English now.’
These IE speakers recognize the immediate economic benefits derived from their English
fluency, and they also recognize the long term necessity, given the increased
globalization of world markets, and are planning for their own and their children’s’ lives
186
and language skills accordingly. I now explore the social benefits related to English
fluency.
Social Benefits
The second reason IE is valued by IE speakers is that it indexes an urban, authentically
acquired Indian persona, a commodity which is used in two ways. In international
contexts, IE speakers feel that their unique and recently established identity, as elite,
educated, English speaking Indians is transmitted in part through their language practices.
Using IE shows that they are proud of this newly emergent identity, and are valid agents
on the global economic stage. This resonates strongly with research on other alternative,
multilingual settings—for example, Guatemalan vendors actively chose to use Spanish
with customers, over a vernacular, to establish their credibility and authority in the
marketplace (French 2001). Further supporting this, we have seen that ‘fake accents’ are
strongly discouraged in the local context—IE, then, has become a badge of pride,
showing that these speakers are proud to be Indians. They are not, by using IE, affirming
the myriad of global ideologies which frame India and Indians as poverty stricken, poor,
yet culturally rich, a region with filled with religious and cultural conflict. Instead, using
IE indexes a different, emerging persona in the local context: that of a worldly individual
who is not constrained by their cultural, ethnic, caste-based and religious background in
determining their language patterns, lifestyle and outlook.
Interpreting IE as locally valuable, as a means of asserting an urban upwardly mobile and
globally aware—yet still distinctly Indian—identity can be seen as part of a larger social
187
shift. There is evidence from several corners that modern Indian urban culture(s)
has/have developed: speakers now identify as ‘Indian’ over narrower ethnolinguistic,
religious and regional identities (Raj 2003). For example, these urban Indians are
increasingly using English in more intimate domains, like family and friendship, as
demonstrated in several of the above quotes. With these types of social changes also
come internal valorization of IE and specific IE features indexing a modern urban Indian
identity. For example, increasingly common in popular Indian literature are rejections of
an outside standard, and support for IE as locally relevant, as one Indian author and
former UN diplomat writes:
After our chhota-pegs we sign chit-books; the next day we don our dhotis and Ghandi-topis
and do pranam when felicitating the PM at his daily darshan…As far as I’m concerned,
Indianenglish Zindabad! (Tharoor 2007)
Zindabad expresses accolade, enthusiasm and approval, in this case, for IE. Ironically,
zindabad is not shuddh Hindi, given that it is originally an Urdu term. This lends further
support to the above quotes and analysis, wherein modern identity(ies) are drawing on
multiple linguistic resources. This enthusiasm in and approval for IE, and how it relates
to local authenticity, is demonstrated in my data. For example, there is evidence that IE
speakers are proud to have a local authentic English dialect, one which distinguishes
them from international English speaking communities. In this vein, having an IE
‘accent’ opens an avenue for discourse: people like this 64 year old woman working in
publishing can then talk about what it means to be Indian, how and why they are proud of
India:
VC: Do you like the Indian accent better than, say, British and American?
AG: Yeah. Yeah, definitely. I mean, American accent, I don’t think, even if I stayed there I’ve been there for - I had nine - ten trips I’ve been there, stayed for a long, long time, but
I could not change my accent. I would prefer my accent. Why should I change - I mean,
sometimes, you have to change a little bit to make them understand because they don’t
understand. But the thing is, you know, you don’t have to, because I feel that they’re also
188
fascinated by you. If you speak in your own accent, they want to know, and then you can
explain to them, “This is that.” And that fascinates them. It’s not that you have to keep up
with them, make them understand, change your accent, that’s not there. (f64AG 19:3442)
In keeping with this shared English speaking urban affluent identity, English offers a way
to avoid being pigeonholed or categorized as from a particular Indian demographic, and
the stereotypes which can come with that, as f52GG, a hospital administrator relates:
GG: And when my -- when my daughter, older daughter was I think in class one or
something. She came home and she said, Ma, where are we from? And I'd say, R.
[addressing her daughter], why? She'd say most of my -- in class they were asking me
where I'm from. I said what do you mean? And she said, you know, somebody's from
Gujarat, somebody's from -- there were a lot of Bengalis in this school. And someone's
from Bengal, and someone's from the South. Where are we from? I said just tell them
you're from India. So she was okay. She went back, then she came back after a couple of
weeks -- no, no, no, but now the teacher wants to know where I'm from because I have an
unusual surname, and no one could figure out where this was from. So they're trying to
put her in, in some kind of a slot, and they were unable to do that. So I said just go back
and tell them. And then when I went for a parent-teacher meeting, the teacher actually
asked me. Mrs. G., can you please tell where you're from? I said, oh, yeah, you know, R.
[f52GG’s daughter] was asking me but -- she said, no, we we’re a little confused. We
kind of discussed it in the -- in the -- in the class -- what, what do they call them, the, the
teachers' room -- and that where is this? Has anyone heard of this surname? Where do
you think they are from? So we had this big discussion. So then we told your daughter
‘go find out from your parents where you're from.’ So -VC: Did you tell her?
GG: I don’t know if we told her, but I think eventually she found out. But I don't think that’s
important, yeah.
VC: Do you think there's a lot of stereotype towards particular regions that are -GG: Yeah, yeah. (f52GG 14:8-29)
Here, f52GG asserts an identity which is not linked to ethnic, regional or cultural
background, and is simply ‘Indian.’ Similar assertions are common in this affluent urban
bilingual community. This identity is tied to and possible because of their English
nativity—these speakers see themselves as united in their English: they socialize with
people from diverse regional and linguistic backgrounds who they encounter in their
community. This identity is not simply social—it also can improve work relationships:
If you have a mix of sitting officer from some officer from Andhra, Tamil, or some other
northeast then they co- again, you come back to the common language and that’s English
(m64NS 8:12-14)
189
Clarifying Delhi
However, many argue that Delhi English practices differ from much of the country, and
link the local rise of English to Delhi’s notoriety as a ‘class conscious city,’ within which
English was initially taken up because it sounded ‘cooler’ than vernacular languages. For
example, reintroducing a quote, f27RG explains Delhi’s relationship with English:
RG: So it’s become a very class-conscious city, I think.
VC: And how does it play out in language? You mentioned that.
RG: English. Everyone wants to speak in English, you know? And you actually, people - I
mean, I was probably in school at one point of time, to be honest, I was a little, you
know, “I need to speak in English,” and, you know, “Hindi is not a cool language to
speak in.” (f27RG 11:40-4)
Echoing similar themes, a 28 year old man explains the popularity of English in Delhi in
terms of the federal government:
So - so, I mean, English you can imagine when it’s spoken in other places like that where
there’s no government domination. Like in Delhi, of course, everybody’s here, right? It’s not
so in these areas, and there tends to be a mixing between the local language and the local
form of what is accepted in lingual communication whereas in Delhi, you’re so used to
speaking the way - and you’re encouraged to speak properly when you’re a child, mind you.
Especially in Delhi schools. They encourage you to speak proper English, the Queen’s
English, I would say, which then changes, of course, when it becomes more and more
American as kids grow up towards their teenage. But that phenomenon is not there in the rest
of India. The rest of India, I mean, there are more social restrictions on people so people
don’t - people prefer to communicate in their own language,…whereas in Delhi, you’d find a
lot of people like me who say English is my first language because that’s the way we are. We
think in English. We were brought up speaking English. We only spoke Hindi to outsiders,
you know, where it was needed. And there are a lot of people who do that. And there are a
lot of people who want to do that as well. So as you see people, even if they don’t speak
English in the house, there are people who speak fluently outside the house, and they do so
because they wish to conform to a certain segment, which they identify themselves with so
they’d rather belong there than belong there. (m28KC 12:41-13:15)
First, Delhi, as the seat of the federal government, necessarily incorporates more English
than smaller cities. He associates this role with more frequent English prescriptive
behavior, e.g. Delhi schools focus on ‘proper’ English. Second, outside of Delhi, social
restrictions impact language preferences away from English, while in Delhi, these are
mitigated by the vast presence of English in the federal government. As he explains, at
this point, English has become habitual for affluent Delhiites. However, there are
190
downfalls to such practices—for example, reintroducing a quote, f27RG explains that
always speaking English can make Indians who are not fluent in English uncomfortable:
RG: And I see a lot of people - and now, I have to make an effort with Hindi, because I never
really learned by Hindi in school. But everyone wants to speak in English, so whether
you go to a restaurant - and as because when you go to a restaurant, if I call a waiter, I
know that the person is more comfortable in Hindi, so I’m gonna order in Hindi. But I
have to kick a few friends under the table to say, “You know, speak in Hindi. That person
is not comfortable.” We’ve - English has become a first language for most of us, and we
forget that we’re talking to people who may not be comfortable in English.
VC: So is it forgetting, or is it actually like using it to say, “Oh, I’m cooler than” RG: I think one part is that and probably it was that at one point of time, and now, I think
probably it’s, it's, you know, faded in memory. (f27RG 11:45-12:8)
English continues to divide and stratify India, because access to English is not equal, and
is especially absent in rural and lower socio-economic communities(Ramanathan 2005).
However, a different picture is painted within this upper-middle class socioeconomic
community, where English, not other languages, is used during daily walks. Here, a 73
year old Bengali housewife explains how, in the local park, where many people (who, in
my experiences at this park, range from university students to 80+ year olds) go to walk
or do yoga in the mornings, conversations are in English:
When we came here from Bengal, or Calcutta, the only thing English is - we used English to
answer question paper. Political science, economics - those are the subject we used to answer
in English. Otherwise, the talk to each other, definitely in Bengali. Here in Delhi, it’s quite
different. Sometimes I wonder. They - when I go - it’s said the Niti Bagh I go for a walk
every morning, but anybody reading me or anything, nobody speaks in Hindi or in - forget
about Bengali. Not - not many Bengalis are here. But mostly people speak in English
whenever somebody crosses. This is true by my side. I find them speaking in English.
(f73TM 3:27-34)
This population is not unified in ethnicity, religion, and caste, while they are comparable,
in terms of socioeconomics, and also, it seems, in their use of English in this domain.
IE as a Lingua Franca
IE is also viewed positively because it offers a way to communicate with Indians across
the subcontinent, with different language backgrounds—it serves as a traditional lingua
franca, and is a commodity available to fluent and non-fluent English speakers in India.
191
For example, this 62 year old former Air Force pilot describes the relative ease pilots and
air traffic controllers across India have communicating in English, despite their language
background, while he also acknowledges that some foreign pilots had extreme difficulties
in the same situations:
VC: Um-hum. And within the Air Force did you ever find it hard to communicate with
different people from different parts of India. I mean other accents you’d counter?
SCS: No, no, because the common language was English only so everybody used to speak in
English so communication wise there was no problem at all.
VC: So nobody sounded like, like could you sometimes tell when you were on the radio, oh,
this guys a Bengali speaking English or this is South Indian?
SCS: No, it is quite easy to understand if you are Indian, but of course, if you are a foreigner
for you it is difficult because of various things. I have seen foreign pilots and they don’t
understand English of Hindu controllers for a year, for one year or two years cause for
last six months I was flying with a foreign copilot and so halfway he will ask me, “So, I
don’t know what he said.” (m62SCS 8:44-9:9)
Another retired military man, m64NS, supports m62SCS by explaining that English is
used because it is a ‘common language’ across officers, while in mixed groups (noncommissioned officers and commissioned officers) multiple languages are used, though
English is again located as the common language:
VC: Yeah. When you were in the Air Force, did they typically use English for everything or
did they use Hindi, like what languages would everything be in?
VC: Uh, mostly English. When the officers cadre, mostly they communicate in English.
Even, in Air Force especially, even most of these air men as you call them, they noncommissioned officers, even they are mostly well educated and so they’re very well
understand English and so basically and the officers communication is in English.
VC: And when you’re not with the officers then it’s inNS: Then it depends on what kind of group you’re sitting. If your all Punjabi’s sitting then
you may, sometime just change over to Punjabi. If you have a mix of sitting officer from
some officer from Andhra, Tamil, or some other northeast then they co- again, you come
back to the common language and that’s English. (m64NS 9:4-14)
While a 73 year old man also sees English as a lingua franca across the nation.
VC: I wonder - because India is such a multilingual nation, do you think that’s a good thing
or a bad thing for the nation? Do you think it’s helpful?
KBS: English?
VC: Mmm.
KBS: I can not say bad, because unfortunately, Hindi maybe because of political reasons or
that reasons, has still not become a national language. So if you have to convey, you can
not do it without English today. (m73KBS 14:42-15:3)
192
The contradictions of India
I finish with two (very) extended quotes, which demonstrate the problems with trying to
interpret and ‘explain’ India—it is constructed as a series of contradictions, among which
language is but one of several factors. As the first quote demonstrates, an 85 year old
retired military man sees English as spoken ‘everywhere’ in India, but, directly
contradicting this, explains English is also ‘not spoken in very many places:’
SPB: But English is spoken everywhere in India. After all, British people ruled here for two
hundred -- more than two hundred years so everybody is talking in English everywhere.
VC: Do you think it should be a national language?
SPB: National language?
VC: Of India, do you think English should be?
SPB: No, I have not ever actually inquired about that, et cetera. Only we have been going in
in and out over here. That's all.
VC: But you were talking about how all of these people across India speak English. Do you
think it should become a national language?
SPB: India - eighty percent or seventy percent in villages, they're all illiterate people. How
can you talk to them in English unless it is spoken there in those villages also?
VC: Uh-huh.
SPB: We are very much behind in that subject.
VC: So the English speakers are all in the cities?
SPB: No, but they're not in villages also. India is mostly in villages. And Madras our capital
of the provinces if English is taken. But if I go in interior of South or in Bengal, nobody
would learn English. They will talk in their own Bengal, Bengali, or South Indian
language.
VC: Uh-huh.
SPB: You can not go and talk with them in English. They know their only language. Even in
Punjab, there are more villages. You have to speak Punjabi. If you speak to them in
English, they will not understand what you are saying.
VC: Uh-huh.
SPB: So much illiteracy is there.
VC: Uh-huh.
SPB: So you can only in big, big cities, capital, et cetera, you can talk in English, not in
remote places. (m87SPB 16 4:30)
Meanwhile, this contradictory and conflicting view of India is echoed by the following 28
year old man:
VC: So you don’t feel like India’s a unified whole?
KC: Not really. I think India’s a unified whole because English [the British] intended it to be
so. And before that, it was a set of princely states, which it still is, and the deeper
sentiments of people - of course, the brilliant thing about India is, of course, when the
need arises, they’re all Indian. So end of the day, they’re all Indian. I must grant them
that, like there - there’s no disunity like that. But if you’re Indian, and you’re within the
system - if you’re a foreigner, they’re Indian. But if you’re Indian and you belong to
something in India - obviously if you’re born in India, you do belong to some segment,
some ca-, some religion, something. You’re completely alienated from everybody else.
193
Like if I’m a North Indian, I’m fair. If I go to South India and try to enjoy temple, they
won’t let me. They tell me Oh, I - we don’t think you’re the right caste or we think
you’ve gone abroad too many times, and that’s not a good thing. Your long -- hair’s too
long or - you know, you’re not wearing the right clothes. Why are you wearing trousers?
So it’s a very different universe, and it’s very hard to connect. It’s like when I went to
step out of Delhi, I’m stepping out into a new country. And there are completely different
laws everywhere. Like we have a constitution, of course, but we know how the Indian
legal system works. We don’t need to go there. So everybody has their own interpretation
of how - what is ideal. And you will frequently clash against other people’s ideals if you
voice your opinion anywhere outside Delhi. For me, Delhi because I was born here so I
understand Delhi’s ideals, and I am Delhi’s ideals in some way. But my international
outlook, which may be considered positive in Delhi - people will say wow, you travel a
lot. You know a lot of people. You’ve done a lot of things, and they would consider that
an achievement. Down South, I’ve been told it’s a very sad thing that you had to travel
outside India to discover these things. Because their idea of discovery is completely
different from what we believe in. So it becomes and us and then situation everywhere
you go. And this is true for them when they come to Delhi because they’re dark, and they
- and they’re different. And they’re treated as bad. So if it was left for themselves, they
would love to separate India into fifty parts or maybe fifty thousand parts, if they refine
it, to them to caste. So, I mean, it’s very - India is a phenomenon. India is not a country to
me. India’s a phenomenon to me. I’ve seen it. I guess I’ve read about it. It’s been
happening with the last fifty years or since - since - since Gandhi’s time, I would say,
really. He truly is the father of the nation in that manner. And it’s a miracle this machine
works the way it does. It’s an absolute miracle. There is really no logic to it. I have never
been able to understand how these people actually managed to turn things around when
they, when they need to, which is fantastic. (m28KC 10:17-11:4)
m28KC does not see India as a consistently unified whole—just unified when the need
arises. Meanwhile, being Indian means one is also separated from most of the nation, by
caste, by religion, by region, by coloring, by how often one travels abroad, by clothing
and style, and by ideals and morals. Within such a system (if one can call it such, m28KC
prefers ‘phenomenon’), communities are united against shared threats, but divided in a
multiplicity of ways in daily life, and English in India reflects this. Specifically, he sees
English as a universal language, manifesting in a myriad of ways, but which still serves
as a communication tool:
KC: English is like this universal language which everybody speaks. Everybody kind of
speaks it their way because initials like that, it’s easy. Everybody can speak it in any way
they want, and it’s still comprehensible. Because I mean, I could say a word, which I did
not make sense in, like I told you. I said - I just messed up north VC: North, South.
KC: -- sorth and nouth [repeating earlier spoonerism]. But you still know what I’m saying,
you know? And that’s a grave error. I mean, I’ve changed the first alphabet of the letters.
And you can still understand what I’m saying, and that’s the way English works. Once
you think in English, you know. And in India, that’s - that’s - it works like that. You can
still communicate, mind you. It’s not like they can not understand what I’m saying or I
194
can not understand what they’re saying. It’s just different. It works. It works everywhere.
(m28KC 13:36-14:1)
Given what m28KC says, perhaps I have been guilty, here, of attempting to over-explain
and circumscribe English in India, by searching for similarities and differences. This
framework may not be shared by these IE speakers, who appear to be less concerned with
enumerating and describing such contrasts, and instead, are more focused on how
language functions as a communicative tool in this multilingual, pluralistic country.
Nonetheless, some common threads can be drawn out from the complex fabric of
thought, reflection, and articulated practice woven through these interviews.
Conclusion
I have drawn heavily on these speakers’ own explanations, thoughts, and reflections
within this analysis cum exposé of urban IE speaker attitudes, beliefs and ideologies
about language. We have seen how and in what ways these IE speakers are taking an
active role in interpreting, creating, and contesting their sociolinguistic identities and
contexts vis-à-vis intranational and international ideologies, and language practices, and
socio-political history. Through their thoughts addressing key terms like mother tongue,
nativity, fluency, bilingualism and the politics and their choices in naming codes, we
have seen that their personal understanding of their and surrounding IE speakers’
language are nuanced, complicated, and conflicting. These conflicts and complications
can be interpreted as both common across contexts—no situation is ever as simple as one
might hope or initially assume—, and also the direct impact of colonial and post-colonial
global ideologies about standard language, native speakers, and what constitutes a
language. These ideologies circulate in the local context, but are not simply taken up, and
195
instead, are contested, reified, and re-created to suit local needs, local language practices,
and local processes of establishing authority and through the commodification of
language.
I have attempted, within this chapter, to present these IE speakers reflections on
language, culture and identity in a linear format. However, as must be obvious by now,
they are by no means linear—one way cause and effect explanations simply don’t do
justice to how intertwined these aspects of IE sociolinguistic identity really are. We have
seen that understandings of English, IE, Hindi and Hinglish are tied to multilingual
fluency assessments, to global ideologies, to language shift and maintenance, to Indian
identity, and to economic and social needs: their language practices and sociolinguistic
identities cannot be addressed in isolation. From this, I will, in the next two chapters,
demonstrate how the social semiotics of these speakers sociolinguistic lives are a pivotal
starting point to understanding and modeling IE structure over time.
196
CHAPTER 5
POSTVOCALIC (r) IN INDIAN ENGLISH
Introduction
Varieties of English are often termed rhotic or non-rhotic, where non-rhotic behavior is
characterized as r-deletion in postvocalic coda position (e.g. [kɑ pɑk] for car park).
Postvocalic r-deletion is varyingly related to both prestigious and stigmatized forms, in
different contexts. While /r/ deletion is stigmatized and fading from use in much of
American English (AE), it is a feature of the prestige form, RP26, in the UK (Trudgill &
Hannah 2002). RP is hypothesized to have a much larger sphere of influence on World
Englishes, given England’s history of colonization, which included the introduction of
English in multilingual Asian, African and Caribbean outposts, including India (Trudgill,
Schreier, Long & Williams 2004). Widely studied in numerous English dialects,
postvocalic r-deletion ‘has been involved in a long term pattern of changes in many
English accents’ (Downes 1998: 134) and the involvement of social factors in linguistic
change has been brought to the forefront in such research. This paper seeks to explain
how rhoticity and social identity interact in a dialect of Indian English (IE) spoken in
26
Received Pronunciation, or RP, is also known as Standard Southern British English (e.g., Knight,
Dalcher & Jones 2007).
197
New Delhi, through apparent time examination of three generations of IE speakers. This
offers a lens from which to understand the present and future status of IE rhoticity and its
relationship to urban Delhi sociolinguistic identities.
Motivations/Theoretical Contribution
This project in its larger form seeks to understand how structural patterns, linguistic
ideologies and linguistic practices in a narrowly constrained population of IE speakers
collectively allow both a deeper understanding of the present and future of IE and its
speakers within an increasingly globalized society, and how they may also illuminate
larger sociolinguistic processes which are applicable in other post-colonial and bilingual
settings. Within this, three motivations are present for choosing to quantitatively model rpronunciation in particular.
First, research on IE contributes to the growing range of quantitative variationist
methodology focused on incorporating emerging post-colonial dialects, as well as
bilingual communities. It is considered problematic that ‘virtually all quantitative
sociolinguistic investigations have been carried out in standard language cultures and,
moreover, mainly in monolingual situations’ (Milroy, 2001: 546). These monolingual,
first world populations do not reflect the worldwide norms for multilingualism (though
non-standard varieties are often examined in these contexts), and variationist
methodology must attend to social and linguistic motivations in multilingual situations in
order to maintain relevancy in accounting for variation worldwide. India, by virtue of its
multilingual background with English as a colonially introduced code, and its consistent
198
label as an ESL context, regardless of the age and degree of English acquisition among
some communities, clearly can be considered an alternative marketplace (Bourdieu
1991). A nuanced model of structural variation in IE, which this research offers, will
expand and test the efficacy of variationist sociolinguistics in accounting for linguistic
variation in alternative, bi-/multilingual marketplaces.
Second, quantitatively studying (r) deletion as mediated by both social and linguistic
factors in the IE context can help towards understanding local Delhi speakers’ alignment
with various competing internationally prestigious varieties, like General American
English (AE) and RP, or conversely, alignment with a localized dialect. Indeed, the two
strongest outside social influences on India are the UK, which held much of the Indian
subcontinent as a colony, and the US, which has held global preeminence in terms of
social influence since World War II. Linguistically, these ‘prestige accents in Britain and
North America—RP and General American respectively—provide “polar norms” of nonrhotic and rhotic speech’ (Downes 1998: 136). Linking these sociolinguistic influences to
post-colonial linguistic behavior thus offers a unique opportunity to examine how local
and international influences may be visible in IE language practices.
Meyerhoff presents rhotic behavior dichotomously across English dialects as interrelated
with region and dialect history: ‘the r-fulness of North American and Bajan (Barbados)
English compared to the relative r-lessness of post-colonial varieties in the southern
hemisphere’ (2006: 186). However, the situation is arguably much more complex. This
representation ignores several English speaking regions. Meyerhoff also fails to account
199
for potential diachronic changes and age-grading changes in linguistic behavior that may
be linked to shifting ideologies (e.g., Labov 1972c; Wassink & Dyer 2004). Language
practices are not static, nor do they necessarily reflect outside norms and/or standards.
Pertinent to this, in the urban Delhi context, younger generations of IE speakers appear
resistant to notions of an external standard for their English, and instead, suggest that all
Englishes come with an ‘accent’ (V. Chand 2008). The views of these IE youth represent
an ideological change over time towards local and international English varieties. They
are accompanied by internal valorizations of IE, by Delhiites across ages, as a feature of
modern Indian identity, and as a path towards economic success. Lending further support
that IE is changing diachronically, increasingly common in popular Indian literature are
rejections of an outside standard, and support for IE as locally relevant, as one Indian
author and former UN diplomat writes:
After our chhota-pegs we sign chit-books; the next day we don our dhotis and Gandhi-topis
and do pranam when felicitating the PM at his daily darshan…As far as I’m concerned,
Indianenglish Zindabad! (Tharoor 2007: 368)27
Zindabad is an Urdu term expressing accolade, enthusiasm and approval, in this case, for
IE as the most relevant variety of English in the local Indian context. Given the
complexities of and rise in Indian globalization (Cowie 2007; V. Chand in press), it is
problematic to ignore processes of structural and ideological nativization. These are both
conscious processes, wherein speakers begin to identify IE as a commodity which serves
to index their increasingly valuable local social identity, and unconscious processes of
27
Definitions for these terms include chhota-pegs: a drink with approximately two fingers of hard alcohol;
chit books: social IOU’s or permission for entry to restricted events or clubs; dhoti(s): traditional Indian
male garment which consists of a rectangle of cloth which is wrapped around the waist and legs; Gandhitopi(s): a white cloth cap pointed in front and back with a wide band that was made popular by M.K.
Gandhi; pranam: to bow or greet with respect; felicitating: wishing well; darshan: viewing.
200
sound change where local structural features can emerge. I understand the structural
emergence of IE as interrelated with English ideological nativization and ownership.
This coupling of social, ideological and linguistic processes is challenged by Labov, who
asserts that phonological sound changes arise from features below conscious awareness
(Labov 1966; 2001). However, Woolard (2008) challenges the complete independence of
conscious indexical language practices from language change. Local identity and a rising
consciousness of distinctive features can work to encourage diachronic structural change
(Zhang 2005). Given that changes in language ideology and language ownership are
emerging in the urban IE-speaking context, and given that AE, as a social commodity,
has replaced RP on a global scale (important for the current study, the two are highly
divergent with respect to r-pronunciation), it is worth exploring whether these changes in
local sociolinguistic ideologies and the relative value of international prestige variants are
accompanied by structural change in IE. Further, while varieties of English are often
statically labeled as rhotic or non-rhotic—or, as evolving towards a stable
pronunciation—the rhoticity of IE may be in a long term state of flux, given emerging
national and (g)local—global yet local—identity in the Indian post-colonial setting.
The third overarching reason for the selection of this variable is linked to the possibility
of measuring linguistic change diachronically, rather than only predicting change through
the synchronic analysis of the practices of several generations of speakers. While
apparent time studies of variation may reflect diachronic changes, they may instead
reflect variation related to age-grading, where successive generations of speakers modify
201
their linguistic behavior at a particular stage in life (Boberg 2004; Tagliamonte & D'Arcy
2007; Wagner 2008). The choice of this structural variable and these informant
population characteristics have both been influenced by the goal of examining potential
diachronic changes in IE, given past quantitative research on r-deletion in the target
population (e.g., Agnihotri & Sahgal 1985). Comparing past results with current findings
will permit direct examination of whether current variation in r-deletion is better
understood as a change-in-progress towards a more stable pronunciation, or as evidence
of age-grading.
Past Research
Rhoticity has demonstrated strong links to both linguistic and social factors and processes
of language change in several English dialects, next explored.
(r) in IE
IE r-pronunciation is analyzed in several contrasting ways. It is considered nonexistent,
rendering IE a non-rhotic dialect (Nihalani et al. 1979: 211; Sailaja 2009: 19), present and
environmentally conditioned, rendering IE a variably rhotic dialect with linguistic
constraints considered the primary motivation for alternation (Bansal 1990; Gargesh
2004), or socially variable and indexing young educated females with more years of
English-only high school instruction (Agnihotri & Sahgal 1985; Sahgal & Agnihotri
1988; Agnihotri 1994; Trudgill & Hannah 2002: 130; Sharma 2005: 208; Wiltshire 2005:
202
282) . This third scenario suggests that IE rhotic patterns are most strongly correlated
28
with social features. In Delhi IE, rhotic behavior is a stronger socially diagnostic variable
for age and gender than other traditionally proscribed pan-IE features, e.g. alveolar stop
retroflexion (Sahgal & Agnihotri 1988). This analysis tests these claims by accounting
for both linguistic and social mediators of rhoticity.
Some studies suggest that a finer distinction, between trill, approximant or flap, and null
realization, is necessary in the Indian context (Sahgal & Agnihotri 1988; Sharma 2005).
The IE liquid /r/ is hypothesized to also manifest as trilled, both in word initial consonant
clusters, e.g. trap, drain, and in postvocalic position, e.g. car, cart (Gargesh 2004: 998).
While it is not explicitly stated, /r/’s quality as trilled is not taken to be categorical, which
suggests one area fruitful for examination in the current study.
Two quantitative analyses of (r) have been conducted on IE. The earlier of these two
studies includes the same population as the current study, Hindi/English bilinguals from
south Delhi (Agnihotri & Sahgal 1985)29 and it also examines Bengali and Tamil/English
bilinguals. The later study examines a continuum of second language learners to fluent
English speakers from a variety of Indian regional and linguistic backgrounds living
outside of India (Sharma 2003; 2005). The later study draws on Indians residing in the
Bay Area, in northern California. They also differ in their explanation for rhotic behavior.
Agnihotri and Sahgal do not attribute r-deletion to Indian L1s. They argue that an IE
28
Two dates (1988 and 1987) are regularly cited for Sahgal and Agnihotri’s paper in English World-Wide,
however, the publication date was 1988.
29
Agnihotri and Sahgal published very similar reports on rhotic behavior within two papers (1985 and
1988), with the authors reversed in the latter publication: I refer to both; however the methods and findings
are the same across both papers.
203
norm is emerging across speakers from multiple L1 backgrounds, and which is
predictable based on social factors. Sharma’s research, meanwhile, explores whether her
participants’ r-pronunciation is either merging towards IE as a ‘stable non-native variety,’
or towards American English pronunciation norms, given that they are living in the US
(2005). Neither study, thus, approaches the question of IE r-pronunciation as direct L1
influence. This study similarly is not seeking to explain r-pronunciation as caused by L1
influence, and intentionally focuses on early IE/Hindi bilinguals—that is, speakers who
acquired Hindi and English simultaneously, and are fluent in both before reaching school
age. It is very possible that IE speakers who learn English after acquiring a L1 will have
different patterns of pronunciation than those uncovered here, patterns which may validly
be attributed to L1 influence.
Agnihotri and Sahgal examine word final and postvocalic coda (r), making a binary
distinction between /r/ presence and absence (1985). They did not examine any other
internal environmental constraints, however several social factors, including age,
language background, and high school prestige are examined. They find that the older
generation has a more r-full pronunciation, while younger women with more prestigious
schooling are leading a hypothesized change: IE is becoming ‘less r-full,’ moving in the
direction of becoming a stable non-rhotic dialect (1985: 103-4).
Meanwhile, Sharma’s analysis is intended to supplement a qualitative analysis of speaker
alignment with India vs. America. It codes coda position /r/ tokens through a tripartite
division between approximant /r/, trilled and partially devoiced /r/, and /r/ absence (null
204
/r/) (Sharma 2003: 136-7). However, after creating this distinction, Sharma conflates the
null and trilled /r/ as both being indicators of an ‘Indian’ dialect in her results, which she
then contrasts with the approximant /r/, characteristic of AE (2003: 136). Given that the
current study is interested in how IE speakers may align themselves with RP’s null /r/,
AE’s variable approximant /r/, or demonstrate a localized form, e.g. the trilled /r/, and
given that Sharma finds that several of her speakers alternate between the null and trilled
realization (2003: 139), I retain this tripartite distinction within my coding.
Sharma does not examine additional internal linguistic constraints, while she does
examine several external constraints: speaker age, time of arrival in the US, and duration
of English-medium education. However, incorporating these social constraints, while
important for linking rhoticity to particular demographics, may not be enough—linguistic
constraints have played a more powerful role in mediating variation in past research. ‘We
would expect social constraints to be weaker than linguistic ones, but this is true for
virtually all variables that have been studied in any depth (e.g. Eckert, 2000; Preston,
1991) and hardly a peculiarity of new-dialect formation’ (Meyerhoff 2006: 187).
Contrasting with this and focused on black and white Bostonians, Nagy and Irwin found
that while ‘[a]ll the linguistic factors except word class proved significant… [t]he
strongest predictor of postvocalic (r) in all communities, however, was the combined
variable of age/sex, with young women leading the change in the white community, and
young men leading the change in the AA community’ (2007: 1).
205
Thus, by expanding the current coding to include linguistic factors, something not done
in past IE focused quantitative studies of rhoticity, the current study seeks to understand
how social and linguistic constraints rank and are interrelated within a hypothesized
process of new-dialect formation, which, in turn, will offer support for either social or
linguistic variables as the strongest predictors of IE (r) behavior. As well, incorporating
linguistic factors may permit this study to unravel the mystery surrounding the trill and
null realizations suggested by Sharma (2003), by uncovering social or linguistic
motivations for the alternation. While past quantitative and descriptive IE literature
guides the coding of /r/ realization, we must explore research on other dialects to form
hypotheses regarding other internal environmental factors which may correlate with (r).
Linguistic Constraints on (r)
(r) in RP
Because it is categorically considered non-rhotic, there exist no quantitative analyses of
rhotic presence in RP, southern, or eastern British English dialects (Downes 1998),
except as a hypothesized idiolectal feature, e.g. within Trudgill’s analysis of the Norwich
English dialect (1974: 77). However, the situation is in fact much more complicated:
Wales, southern or eastern British and RP English dialects are non-rhotic, while ‘large
parts of the British Isles are in fact rhotic (ScE, IrE, southwestern EngE, and much of
northern EngE)…’(Schneider 2004a). Further, several dialects are considered variably
rhotic, as this table from Downes (1998: 136) demonstrates:
206
Table 5.1 Rhoticity across English dialects
Rhotic
• General American class of
accents: midland, north
central, middle Atlantic,
etc.
• Southern mountain accents
in US, ‘hill type’ of speech
• General Canadian
• Scottish accents
• Irish accents
• Some West Indian, e.g.
Barbados
Variably Rhotic
• Local accents in the west of
England
• A few local accents in the
north of England
• New York City
• ‘Borderline’ rhotic/nonrhotic areas in the US, e.g.
South, eastern New
England, black English
vernacular
Non-Rhotic
• RP (Received
Pronunciation) in England
and Wales
• Local accents of the east
and north of England
• Most accents of Wales
and New Zealand
• Australia
• South Africa
• Black English vernacular
in US
• Some parts of eastern
New England
• Southern speech area in
US, ‘plantation’ type
• Some West Indian, e.g.
Trinidad
Given this, the linguistic constraints which may guide IE /r/ presence and quality must be
distilled from analyses of English dialects beyond RP.
(r) in AE
Examining AE analyses of (r), Labov’s Department Store study—a pilot for his larger
dissertation research on multiple variables in New York City (Labov 1966)—introduces
the systematic analysis of variable r-pronunciation (Labov 1972c). In his dissertation,
Labov examines word final and pre-consonantal coda /r/, and finds that variable rpronunciation divides the population into remarkably fine-grained strata (Labov 1966).
This analysis excludes pre-vocalic /r/, and /r/ following mid-central schwa vowel nuclei,
e.g. her, bird, while the latter are separately analyzed (Labov 1966: 50). /r/ classification
is binary, separating definite constriction from unconstricted glides or no glide, while
intermediate cases are not used in the final analysis. No additional environmental features
related to preceding or following environment, morphological status, consonant cluster
207
size, word frequency or stress are coded, given the narrow analysis of rhoticity in the
phrase ‘fourth floor,’ although formality is binarily coded (casual/formal).
Several linguistic factors demonstrate a relationship to /r/ deletion in more recent studies.
The presence of another vocalic [r] in the AE word (e.g. quarter can manifest as [kwɔtr̩]);
syllable boundary (which interacts with vowel quality) or pause; /r/’s status as syllable
final or in a consonant cluster—also termed morphological position, following Nagy and
Irwin (2007)—; syllable stress (which interacts with vowel type) wherein /r/ weakens
before an unstressed vowel (Harris 2006: 2); following word-boundary-plus-vowel—
termed ‘linking r’—which provokes r-maintenance (Downes 1998: 146); preceding
vowel quality (Harris 2006); and /r/’s status as a rhotacized schwa nucleus (e.g. in bird)
(Myhill 1988; Feagin 1990) have each been investigated. Harris (2006)’s highlighting of
/r/ weakening as motivated by both stress and vowel quality suggests that syllable stress
and vowel quality should be coded separately, and their combined effect considered.
While not quantitatively studied thus far, RP surveys have suggested that the
morphological status of a syllable that is potentially rhotic as also marking a morpheme
boundary motivates r-retention in words with a /ɜ/ nucleus. The morphological
independence of a potentially rhotic syllable as a separate syllabic morpheme, e.g. bak.er,
may also motivate /r/ retention, similar to how studies of consonant cluster reduction
have uncovered that morphological quality motivates (t,d) retention (e.g. Labov 1989).
Meanwhile, the target word’s lexical class has not demonstrated any relationship to /r/
deletion (Nagy & Irwin 2007).
208
Social Constraints on (r)
English rhoticity is a strong variable to examine in large part because of its consistent
links to social features and its involvement in larger processes of sound change distilled
from studies of numerous communities on both sides of the Atlantic. Past studies have
found the following external social factors to significantly affect the rate of /r/ deletion:
age, sex, ethnicity, degree of formality, integration into standard AE speaking
communities, regional background, socio-economic status and occupation (Labov 1972c;
Myhill 1988; Feagin 1990). These are discussed below.
Sample Population
Sociolinguistic interviews were conducted with 29 upper middle class Hindi English
early bilinguals in Delhi, India from 2007-2008: this study thus targets a sample of
linguistically, educationally, socio-economically and regionally homogeneous
informants. Data was collected from women and men from 18-87 years of age, in order to
examine apparent time variation (Bailey 2002), to restrict for potential confounding from
an overly heterogeneous population, and to permit a real time study of variation as
potentially a change-in-progress vs. age-grading, through comparison with past
quantitative work of this population (Agnihotri & Sahgal 1985). The sample is
chronologically continuous. I did not seek to create age ‘gaps,’ because such gaps might
limit the ability to interpret findings as processes of age-grading or apparent change-inprogress. Participants were recruited by exploiting already existing social networks, using
the ‘friend of a friend’ method to make initial contact within the community, and outside
of my existing Delhi social network (Milroy, 2002). In addition to the informal
209
interviews, two other types of oral data were targeted for collection immediately after the
informal interview. These include a formal reading passage (the Grandfather Passage)
and a structured retelling of the short film The Pear Story (Chafe 1975; 1980).
Collectively, these participants may be regarded as representative of modern, urban
middle and upper class Indians in several ways30. The results of this project, while clearly
not representative of the potentiality of IE dialectal variation more broadly as including
rural, lower class speakers from different language backgrounds, regions, and varying
degrees of English fluency, will provide a framework for further study of urban, globally
linked IE speakers.
Current Coding Practices
Each token realization was coded as null, trilled, or approximant based on both aural and
acoustic analysis. Seven social factor groups and four linguistic factor groups were also
coded for, to explore both internal and external motivations for rhoticity. All codes are
detailed below. To ensure reliability across coders, a factor group was created to
designate coder identity, while another factor group was created to individualize each
speaker’s results. This final factor group allowed me to examine whether any individuals’
behavior was highly divergent from their peers31.
30
Though, of course they do not fully represent modern, elite, upper middle class Indians, given the range
of cultures, ethno-linguistic backgrounds, ideologies, religions and linguistic competencies found across the
Indian sub-continent.
31
These individual codes proved important in separating out six speakers from the original sample of 35
speakers. These six were removed from the sample because they did not match the target demographics
(e.g. they had different linguistic background, were recent immigrants to Delhi, or were from a different
socio-economic class), and also evidenced very different patterns of rhoticity. The analysis presented here
is based on the remainder of the sample, a total of 29 speakers.
210
Coding Rhoticity
Rhoticity was determined through a combination of aural and visual acoustic analysis, in
Praat (Boersma & Weenick 2006). While several articulations are used to pronounce
approximant /r/, acoustic quality is relatively stable across these articulations, and
manifests as a decrease in distance between F2 and F3: the simultaneous raising of F2
and lowering of F3 (Knight et al. 2007). Tokens were examined for this convergence
formant frequency, and were also analyzed aurally. Tokens coded as non-rhotic
evidenced neither a perceptible rhotic sound nor a F2/F3 convergence. Rhotacized
schwas, which are often a source of contention for analyzing rhoticity (e.g. Yaeger-Dror,
Kendall, Foulkes, Watt, Oddie, Harrison & Kavenagh 2008), were coded as a rhotic if
there was a perceptible change in formant quality towards a F2/F3 convergence across
the vowel duration, and coded as non-rhotic if the formants remained stable and did not
converge.
Aural rhotic categorization has recently been raised as highly problematic when used as
the sole means of analysis. First, research has shown little consistency in categorization
across groups of trained listeners from disparate regions. Second, rhoticity categorization
is influenced by surrounding dialectal features (e.g. a Brooklyn pronunciation of ‘coffee’
as [kɔwfi] motivates the following word ‘bar’ to be heard as non-rhotic [ba]) (YaegerDror et al. 2008). Third, /r/ acquisition research suggests that absolute formant
frequencies should be examined in conjunction with other acoustic data, as ‘F3 lowering
on its own is only one ingredient of “correct” /r/’ (Knight et al. 2007: 1584). I suggest
that the current coding process both took into account such issues, and was relatively
211
immune to them for three reasons. First, the data is highly variable in terms of rhoticity—
there is no readily apparent default form to assume as underlying, and hence act as a
default. Second, this combination of acoustic and aural methods can counterbalance
reliance on either absolute formant values or surrounding dialectal features, and has been
suggested as a fruitful means to standardize rhoticity coding32. Third, all questionable
tokens have been verified by a second coder, and this second round of analysis was made
based on the same structured reasoning as the original coding. Tokens which could not be
reconciled through these means were excluded from analysis.
Social Constraints
Several overlapping and potentially interacting social factors were initially coded, given
that locally relevant social factors mediating rhoticity have not been uncovered for this
population. Underlying this is the assumption that these overlapping groups would be
tested in various combinations, to understand which means of categorizing social factors
offers the best ‘fit’ with the data. Social factors were developed from both traditional
social factor groups, e.g. age and gender, and from emergent social groupings evoked by
participants during the interviews. This approach was necessary, for three reasons: 1) the
lack of earlier nuanced explorations of urban Indian social groupings, 2) this group was
by design fairly homogeneous in terms of socio-economic class, location and language
background, and 3) recent compelling arguments related to participant-defined identity
that motivate a social-constructionist approach to social factor formation. These
32
This combination of aural and acoustic analyses is surprisingly rare in past analyses of postvocalic rdeletion.
212
predetermined and emergent social factor groups are next discussed (and are displayed, in
Table 5.2, along with the number of speakers in each category).
Table 5.2 Social factor coding groups
Factor Group
Conditioning
Factors
Example
Number
of
Speakers
(N=29)
Gender
Female
Male
16
13
18-20
21-30
31-40
41-50
51-60
61-70
71-80
81-90
5
5
2
1
4
5
4
4
18-24
25-38
39-59
60+
5
8
4
13
Student
Working
Modern Housewife
Traditional
Housewife
Retired from
Military
Retired from Other
Profession
Volunteer
5
11
1
1
Age, by decade
Age, by
Historical Era
Occupation
Worked until marriage/children
Never worked
Semi-retired, continued in second
profession after Military
Fully Retired, not currently working
4
Never worked, only volunteer
humanitarian (education/health) work
part time
2
Originally from West Bengal or Bengali
regions of Bangladesh
Originally from northern states of UP
and Haryana
1
5
Ethnolinguistic
Background
Bengali
UP/Haryana
Delhi
Punjabi
Mixed
Including areas now in Pakistani Punjab
e.g., 1 parent from north India, 1 from
south India
2
2
20
4
213
Table 5.2 Social factor coding groups Cont.
Factor Group
Conditioning
Factors
Example
Number
of
Speakers
(N=29)
Age &
Occupation
Student
Working
Retired
Under 20, in school
Any work experience, even if stopped
working to have children. Age is 25-52
(both men & women)
Men either retired, or semi-retired (not
working full time). Women either fully
retired, or never worked. Age: >52
(both men and women)
5
10
14
Delhi Stay
Punctuated
Continuous
16
13
Given the relative homogeneity of the target population, in comparison to past IE studies,
fewer demographic-based distinctions are possible (Paolillo 2002). For example, high
school prestige, while significant in Agnihotri and Sahgal’s (1985) research, is not coded
here because of overall similarities across the target population demographics. Coding
high school prestige was also problematic given the age range explored here. Most of the
oldest generation (65+) was schooled outside of Delhi, oftentimes in schools which no
longer exist, post-Partition. Gender, a mainstay in variationist work, is coded, however,
the lack of previous research on how some oft-used social factors can influence structural
variation in IE means that several other social code choices are experimental. Given this,
I coded several potential social factors in multiple ways. For example, age was coded
both by decade, making eight factors, and by socio-historical era, within which there
were four hypothesized groups who have lived through four chronological eras with
distinct educational and social-political periods (discussed below). Each social factor
214
group was independently examined with the rest of the factors through cross-tabulation,
to determine which factors provide the best explanation.
Informants’ occupation was also coded. This was done to capture any variation that may
exist between working women and housewives, and between military and private sector
professional men, who may easily have different or competing IE models based on their
daily interactions. This public/private sector occupational difference has proved
significant in Beijing Mandarin, with divergent practices of using local vs. cosmopolitanlinked phonological features (Zhang 2005). However, coding for occupation in a gender
and age delineated fashion proved problematic, in that it created structural zeros (Paolillo
2002)—some cells cannot be filled because of preconditions on the code categories.
Some structural zeros are motivated by impossible combinations, e.g., it would be
impossible for an informant to both be in their 20’s and retired. Other combinations are
conceptually possible, but were not found in the more rigidly defined Indian context, e.g.
a male housewife or volunteer. Given this, an additional factor group was created which
was gender-neutral and linked age with work status. This factor group separated younger
students, middle aged members (or former members, for women who worked until
having children) of the workforce, and retired (or, wives of retired men, who had never
worked outside the house), elderly informants. This was possible because, in my
informant pool, all women 25-52 had worked for a significant length of time. Above 52
years old, women were either housewives, or had worked as long as their husbands, and
were now retired. Admittedly this factor group fails to capture differences between, for
example, women in their 70s who had worked vs. their peers who were housewives.
215
However, the other codes mentioned above do permit this, and this particular factor
group permits an examination of occupational links to age without structural zeros.
Emergent Social Factor Groups
In the social-constructionist perspective, social factors are understood as ‘ideologically
driven processes,’ and not ‘a priori social categories’ (Woolard 2008: 439). Social
groupings thus arise and must be developed from ethnographic participant interaction,
wherein participant ideologies and local categories are fore-fronted to capture social
phenomena as experienced by the participants. This approach, advocated by numerous
sociolinguists (see examples within Woolard 2008; Levon 2009), is driven home by
Eckert’s (2004) reflections on her Jocks and Burnouts research: she suggests that her
sustained focus on social class as an independent variable nearly occluded her from
seeing and understanding the local social life as presented by her participants, which
proved critical to understanding both the social groupings, and the sociolinguistic
variation.
However, this phenomenological and nuanced social-constructionist approach to
developing and understanding social groupings and their potential links to language
practices does not always result in clean groupings or independent social factors. Real life
is far more complicated, with different identities overlapping and overlain upon each
other. In a more rigid social hierarchy, there is also less likelihood of truly divorcing
social factors, erstwhile considered independent in first world, western contexts.
Confirming this, several overlapping factors emerged in this data. For example, age was
216
coded independently by two means: by decade, and by socio-historical era. Occupation
was coded separately, however, as discussed above, this factor interacted with both age
and gender, and an additional factor group, combining age and occupation, but gender
neutral, was created, distinguishing students, workers, and retirees.
The emergent social categories allowed for two additional factor groups: ethno-linguistic
history and continued vs. punctuated stay in Delhi. My participants predominantly
identified both as Delhiites, and as from a particular ethno-linguistic background, e.g.
Bengali, Punjabi, as an explanation for their social links, cultural practices, and worldview. While Agnihotri and Sahgal (1985) mention these alignments in their population,
they do not report on any relationship between ethno-linguistic identity and rpronunciation.
Ethno-linguistic Identity
Regional ethno-linguistic identity does have potential links to other phonological features
across varieties of IE. For example, it is hypothesized to regionally segment alveolar stop
retroflexion behavior (Nihalani et al. 1979), /v/ and /w/ merged behavior (Trudgill &
Hannah 2002), vowel space (Maxwell & Fletcher 2009), and consonant cluster
simplification patterns (Bansal 1990). IE rhotic behavior has been established as socially
variable, in that it distinguishes age and gender, but past studies have not undertaken a
multivariate analysis of how regional and ethnic background, also termed ‘ethnocentrism’
(Agnihotri & Sahgal 1985), may correlate in a nuanced fashion with rhotic behavior.
217
Given that social stratification and locally significant identity are understudied in this
context, they were included here through two additional factor groups.
Ethno-linguistic history was separated into five dominant regional groups based on this
sample. Traveling northwest across India, these are: Bengali, UP/Haryanite (people from
the states of Uttar Pradesh and Haryana), Delhiite, Punjabi, and Mixed (with parents from
different regions). However, categorizing individuals based on ethno-linguistic
background is admittedly problematic. Participants will variably define their own
heritage in terms of where they themselves grew up, where their parents and grandparents
grew up, or, where their family is from ancestrally. In some cases, these three locations
coincide, but in other situations, three different locales were evoked. Complicating this,
mixed marriages are now more common (though not common overall), with parents from
different regions. Mixed parentage can then evoke, potentially, five different locales. This
research thus relied on the groupings informants provided, and reflects their ethnolinguistic alignment, while their histories may be much more complicated. In this sense,
ethno-linguistic identity is understood as an ideologically driven process. These do not
cover the range of regions in India, but instead, cover the range of regions evoked by
these speakers. There is one additional caveat to this factor group: because each speaker’s
ethno-linguistic alignment emerged within the interviews, it was not possible to pre-select
participants in a balanced fashion. As a result, there is a very uneven distribution, as
Table 5.2 demonstrates. Results pertinent to ethno-linguistic identity should be
interpreted with caution.
218
Delhi Stay
The second emergent social factor group tests whether time in Delhi, as continuous, or
punctuated by departures, can be linked to rhoticity behavior. Several participants went to
pains to assure me that they identify as Delhiites, despite having lived outside of Delhi.
There were two dominant reasons for this: 3 year military postings and pre-Partition lives
outside Delhi. Military postings, located in insulated, upper echelon English dominant
military communities, are common to this community. Excluding military families would
overly narrow the population and disregard locally defined groupings, a problematic
practice I chose to avoid. Equally common and also problematic to exclude were
participants born pre-Partition, whose childhood experiences were not in Delhi33. I thus
distinguished participants who have lived continuously in Delhi from those whose lives
in Delhi have been punctuated with departures of either sort. Subselecting only for
Delhiites who have continuously lived in Delhi would produce a population so narrow as
to not be meaningful, given the intertwined social connections that were demonstrated
between permanent Delhiites and those with departures.
Linguistic Constraints in the Current Study
Several decisions on linguistic factors to incorporate are guided by Agnihotri and
Sahgal's (1985) constraints and token exclusion choices, which allowed a real-time data
comparison with their results. For example, while Labov (1966) does not code /r/
realizations following schwa nuclei (e.g. bird), several studies—including IE studies of
postvocalic (r) (e.g Agnihotri & Sahgal 1985; Sharma 2003)—do include them, and this
33
All but one of the retired speakers were born outside of Delhi. Of these, all were affected by the upheaval
and mass migration which accompanied Partition, experiences which motivated individual and familial
relocation to Delhi.
219
environment has been targeted as a locus of r-weakening cross linguistically (Harris
2006). I thus follow these studies in including such tokens within this study. /r/ quality,
the dependent variable, separated trilled, approximant and null realizations, while four
independent variables were also coded. These include phonetic environment, syllable
stress, morphemic independence and speech formality. In total, including the dependent
variable, five linguistic factors were coded for. Details and examples of each factor group
are presented in Table 5.3.
While Sailaja suggests that Standard Indian English Pronunciation (SIEP) has seven short
vowels, seven long vowels, and six diphthongs (2009: 24-5), this is contentious. In
addition to having distinct divergences from the vowel systems of UK and American
dialects, other research suggests that no pan-IE vowel system exists (Maxwell & Fletcher
2009): ‘vowel systems vary considerably more across Indian English speakers and a basic
set of contrasts cannot be assumed’ (Sharma 2003: 136). Sailaja’s standardized vowel set
is more problematic when considered vis-à-vis her construct of SIEP—SIEP’s existence
is not endorsed by other research, and her own analysis continually explains how
speakers from various demographics deviate from the SIEP set (see Appendix B for a
detailed examination of various scholarly positions on the topic). Essentially, while
asserting a standardized vowel set, her discussion also problematizes the existence of
such, and it cannot be considered a reliable representation of this IE population’s vowels.
Thus, preceding vowel quality, while significant in Myhill’s study of (r) in Black English
Vernacular (BEV) in southern states of the US (1988), is problematic to code for in this
corpus, and was not fully distinguished within this coding. The only vowel distinction
220
made is between schwa nuclei and full-vowel nuclei contexts. Additionally,
functional/lexical word type distinctions have thus far demonstrated no significant
correlation with postvocalic (r) (Nagy & Irwin 2007) and are not examined in this study.
Word-final /r/ tokens before a vowel-initial word—that is, prevocalic (r) tokens—are
excluded in Labov (1966), but included in several other studies (Agnihotri & Sahgal
1985; Myhill 1988; Sharma 2003) because ‘[i]n many languages, final consonants which
are otherwise deleted are sometimes preserved when the following word beings with a
vowel’ (Myhill 1988:208). However, the same study finds no significant differences in (r)
deletion rates across following word-boundary-plus-vowel, consonants and glides.
Preconsonantal and prevocalic tokens—both syllable and word final—are included in this
analysis, but not coded separately, while syllable internal coda environments are coded
separately.
Syllable stress tends to demonstrate high cross-linguistic and cross-dialectal variability
(Berg 1999). Within this study, the examination of lexical stress is restricted to a tripartite
distinction between monosyllabic words, bi/multisyllabic words with primary stress on
the target rhotic syllable, and bi/multisyllabic words with primary stress not located on
the target rhotic syllable. This third category includes unstressed syllables as well as
syllables with secondary stress, while the first category, monosyllabic words, includes
both stressed and unstressed words. These are the only distinctions currently possible,
given the lack of comprehensive research on stress in IE, and this community in
particular.
221
Table 5.3. Internal factor coding groups
Factor Group
Variable Quality
Dependent
Variable
Independent
Variables
Conditioning Factors
Example
Non-Rhotic Null Realization
[ka] for car
Rhotic Approximant
Rhotic Trill
[kar]
[kaʀ]
Surrounding Phonetic Environment & Syllable Location
Full-vowel nucleus, pre-consonantal
coda position, in a CC
Full-vowel nucleus, word final, coda
position
Full-vowel nucleus, syllable final (word
internal), coda position
Schwa nucleus, with following coda
Schwa nucleus, word final position
Schwa nucleus, syllable final, word
internal
fourth
beer
sur.pris.ing
bird
her, butt.er
mur.der.er
Syllable Stress
Monosyllabic word (stress not
evaluated)
Primary stress in bi-/multisyllabic word
(stressed syllable in italics)
Non-primary stress in bi-/multisyllabic
word (anything less than primary stress
is weak)
Morphological independence
/r/ comprises an independent (bound)
syllable and morpheme
/r/ is either part of a larger syllable or
morpheme
bird, beer
mur.der.er
ans.wer, mod.ern
murder.er,
runn.er, batt.er
(one who bats)
batt.er (flour
mixture, /r/ is not
an independent
morpheme),
runn.ers (/r/ is
not an
independent
syllable)
Formality Level
Informal Speech
Medium Formality Speech
High Formality Speech
Majority of
Interview
Pear Story
Retelling
Grandfather
Passage
Formality has a demonstrated impact on /r/-realization in other English dialects (e.g.
Labov 1966). Here formality is coded through a tripartite distinction between informal
222
speech, medium formality speech (retelling the plot of a short film, The Pear Story) and
high formality speech (a reading passage, the Grandfather Passage, Appendix A). Those
contexts which encourage more attention to speech were collected at the end of each
interview.
Token Selection and Analysis Methods
Token selection was systematic: in each interview, tokens were taken starting a quarter of
the way through the interview, to uniformly handle interviews of different lengths. At this
point, the first 100 tokens were extracted for coding, with no more than three instances of
each lexical item to avoid type/token issues (Wolfram 1993). Very common individual
lexical items can have different phonological behavior (Bybee 2002; Clark & Trousdale
2009) and restricting token selection to three of any type can limit any bias their inclusion
might have on capturing overall distributions of a variable. From the Pear Story retelling,
a maximum of three /r/ tokens per lexical item were used, and all 18 /r/ tokens from the
formal reading passage were used (within which, there were not three instances of any
single lexical item).
Goldvarb X (Sankoff et al. 2005)—generically referred to as Varbrul, short for variable
rule analysis—is a multivariate analysis technique and software application designed to
model unbalanced data, i.e. naturally occurring speech. It has been successful in
determining the significance of external social and internal linguistic factors as mediators
of variation across a number of contexts, and for understanding the relationship between
and relative influence of different factor groups on realization quality (Paolillo 2002;
223
Tagliamonte 2002). There is not room here to fully explain the process of multivariate
analysis, however in addition to the discussion in Chapter 3, Bayley (2002) provides a
very useful introduction to the quantitative paradigm.
Results
The data set totaled 3813 tokens34 which were analyzed in Goldvarb X: Table 5.4 shows
the overall distribution by realization as zero, an approximant or a trill. Considered
categorical in RP, postvocalic r-deletion is clearly variable in this IE population, with less
than half of the tokens realized as null (37.6%). Trill realizations do make up a
substantial minority of the tokens, at 7.8%, and approximant realizations comprise a
majority of the tokens (54.6%)35.
Table 5.4. Overall distribution of (r)
Null
Realization (Ø)
%
N
37.6
1435
Total N
34
Approximant
Realization (r)
%
N
54.6
2082
3813
Trill
Realization (ʀ
ʀ)
%
N
7.8
296
Originally, the token set totaled over 4600 tokens, drawn from 35 speakers. However, as mentioned in
Chapter 3, a few participants fell outside my target demographics (due to their Hindi fluency, socioeconomic background, or status as visitors or recent immigrants to Delhi). Their rhotic behavior was very
different from the target population, and a decision was made to exclude them from further analysis.
35
I treat this scenario as r-deletion (with /r/ as the UR and not, alternatively, as r-insertion, with a null UR)
because it is problematic to explain how the same sound emerges across several phonetic contexts.
However, others believe that the UR would lack a /r/, and this would thus be a scenario of r-insertion. For
example, Giegerich (1999, reviewed by Orgun 2002) suggests that English does not have post-vocalic /r/ in
the UR, and rhotic dialects insert the [r] through orthographic influence. Pertinent to processes of
acquisition, this approach is problematic because children acquire rhotic vs. non-rhotic pronunciation
patterns before literacy (Knight et al. 2007). It is also problematic for explaining the behavior of speakers
who are illiterate, and thus do not have access to written forms. Finally, Orgun points out that this theory
fails to account for the behavior of epenthetic intrusive (r), e.g. idea realized as [ɑydir] (2002), which
behaves very differently than postvocalic (r)—if [r] emerges within both processes, it is unlikely to have
such divergent behavior.
224
While it would be ideal to compare overall deletion rates with those found in earlier
studies, this is not fully possible. For example, within rhoticity studies of IE samples,
there are differences in informant population. Sharma (2003) studies English learners
living in the US, while Agnihotri and Sahgal (1985) study Delhiites from three social
classes and multiple linguistic backgrounds. There are also differences in presentation of
data. Sharma (2003) conflates null and trilled realizations, comparing them with the
‘American’ variant for most of her analysis and discussion, while Agnihotri and Sahgal
(1985) do not specifically mention how trilled realizations are coded and do not present
overall rhoticity distribution separate from their interaction with social variables.
Collectively, these limit the possibility of making a direct overall comparison with earlier
studies of IE rhoticity. Further, it is impossible to derive overall deletion rates for
comparative purposes from research contrasting multiple speech communities, e.g.
Feagin (1990), with a range of 0-100% deletion across socioeconomic groups and ages,
Agnihotri & Sharma (1985) with a range of 22-80% across High School prestige level,
and Piercy (2007) with a range of 66-99% across ages, etc. Nonetheless, the overall
frequencies found here are very different than Sharma’s, where the null realization
comprises 60% of the tokens, approximants 9%, and trills 30% (2003: distilled from
Table B.8, Appendix B). Deletion rates are considerably lower in the current study.
However, this is not enough evidence to suggest any larger processes of change, given
the following three factors. First, Sharma’s participant sample is much smaller (12
speakers). Second, her sample represents different demographics in several ways, as a
continuum of non-native English speakers residing in the US for varying lengths of time.
225
Third, internal factors conditioning rhoticity in her sample are not explored. These may
have an important role in predicting (r)-realization for her sample, and, importantly, they
may not coincide with internal factors significant to this sample. Unfortunately, without
such data, it is not possible to use Sharma’s results to conduct a real-time analysis of IE
(r).
Comparing the current overall rhotic deletion rates to other contexts of variable rhotic
deletion in the US, the current overall frequency is considerably higher than the 13%
deletion rate for white speakers from New Hampshire (Nagy & Irwin 2007), similar to
the 51% deletion rate for Southern speakers—via the LAGS database, collected in the
1960’s and 70’s (Schonweitz 2001)—, yet much lower than both the 62% deletion rate
for black and white Bostonians (Nagy & Irwin 2007) and the 60% deletion rate for Black
English Vernacular speakers in Philadelphia (Myhill 1988). In New Zealand, a region
considered typically non-rhotic, a pan-New Zealand study of rural speakers demonstrates
a 91% deletion rate (E. Gordon, Campbell, Hay, Maclagan, Sudbury & Trudgill 2004).
Given that areas considered ‘non-rhotic’ have much higher deletion rates than found in
the current data, this IE sample demonstrates what we can term variable rhotic behavior.
Overall deletion rates do not, however, necessarily signify underlying grammatical
differences or similarities—it is important to also examine whether IE variable rhotic
quality is conditioned by similarly ranked linguistic and social constraints as the rankings
uncovered in earlier IE, New Zealand English, BEV and AE studies. Linguistic and social
226
factors—also analyzed within GoldvarbX—correlating with realization quality are next
examined.
Overall Constraint Ranking
Given the low number of trilled tokens, trills were conflated with approximants for the
majority of the analysis (they are, however, explored independently below). This
conflation allows a comparison of rhotic and non-rhotic realizations. Multivariate
analysis uncovered eight factor groups as significant in modeling IE rhotic behavior—
these are displayed according to their rank in Table 5.5. IE rhotic behavior is clearly a
complex phenomenon, given the number of significant factors, and the primacy of social
factors, as four of the top five influences. These factor groups are next discussed in detail.
Table 5.5. Significant factors influencing rhotic behavior
Rank
1
2
3
4
5
6
7
8
Factor
Gender
Phonetic Environment
Ethno-linguistic Identity
Age/Occupation
Delhi Stay
Morphemic independence
Formality
Syllable Stress
Type36
Social
Linguistic
Social
Social
Social
Linguistic
Linguistic
Linguistic
Linguistic Constraints on IE Rhoticity
All of the linguistic factors coded for contribute statistically significant effects for (r)
deletion in IE. Phonetic context proved to be the most significant linguistic factor.
36
I acknowledge that it is problematic to consider formality as a linguistic constraint, given that it varies
within an individual, although it is shared across speakers. Here, I follow past research in labeling it a
linguistic constraint, however, in future, I would recommend separating it into a third category, with topic,
style (Levon 2009), register (e.g. oral vs. spoken), and genre (e.g., Biber, Johansson, Leech, Conrad &
Finegan 1999). This separation has the added benefit of allowing more immediate comparisons of how data
sets are contrastive, an endeavor which has been problematic in this project.
227
Overall, schwa nuclei contexts favored deletion over full-vowel nuclei contexts. There
were also significant differences related to the following sound: coda cluster preconsonantal position (with either a schwa nucleus or a full-vowel nucleus) most strongly
favors deletion (.59, e.g. bird, fourth), while deletion rates decreased from word and
syllable final position with schwa nucleus (.52, e.g. her, murd.er.er), to word final
position with full-vowel nucleus (.45, e.g. beer), to syllable final position with full-vowel
nucleus (.36, e.g. sur.pri.sing ).
Table 5.6. Linguistic factors influencing r-deletion
Factors Considered
Phonetic Environment
Pre-consonantal, schwa or full-vowel nucleus*
Word or syllable final, schwa nucleus *
Word final, full-vowel nucleus
Syllable final, full-vowel nucleus
Morphemic Independence
Independent Morpheme & Syllable
Non-independent
Formality
High and Medium*
Low
Syllable Stress
Primary stress in bi-/multisyllabic word or monosyllabic word*
Non-primary stress in bi-/multisyllabic words
Factor
Weight
N
% realized
as [Ø]
Ø]
.59
.52
.45
.36
985
1418
897
513
46.1
38.5
33.6
26.1
.61
.49
270
3543
45.9
37.0
.55
.49
898
2915
42.2
36.2
.52
.47
Total N
2291
1522
3813
38.8
35.8
All factor groups significant; * Two factors are conflated; p = .012; Input value = 0.366;
Log likelihood = -2342.039; Chi Square/cell = 1.9345
Further, while less powerful than the social factors discussed below, morphological
independence, formality, and syllable stress were also significant (Table 5.6).
Morphologically independent tokens (.61), formal contexts (.55) and tokens with primary
stress on the syllable containing (r) (.52) favor deletion over their counterparts. In the
latter two groups, factors are conflated based on similarities in factor weight and
linguistically sound motivations—it is not appropriate to conflate factors which are
228
linguistically dissimilar or which behave differently. Bi-/multisyllabic words with
primary stress on the syllable containing (r) have been conflated with monosyllabic
words containing (r) because both factors behave identically (they had similar factor
weights), and because these two factors have a common bond. Both have primary stress
on the syllable with (r), regardless of the total number of syllables in the word, and they
stand in contrast to syllables without primary stress. High and medium formality contexts
are also conflated given similar behavior, and because they are both situations which
involve attention to speech.
However, formality, morpheme independence, and syllable stress, while each a
significant factor, were not as powerful as in other studies. Instead, in IE, social factors
rank higher, and phonetic environment stands out as the primary linguistic influence on
rhotic behavior. The latter three linguistic factors are significant, but rank below every
significant social factor in predicting rhotic behavior. The order of factor importance
suggests that r-pronunciation is largely a phonological process in IE, as opposed to a
morphological process. Further, the high ranking of social factors may be indicative of
the more rigid and complicated Indian social structure, explored in the next section.
Social Constraints on IE Rhoticity
After the overlapping social constraints were tested in various combinations, four social
factors proved consistently significant, and are next discussed.
229
The Role of Gender
Disconfirming Meyerhoff’s (2006) expectation, and Preston’s (1991) review of
monolingual variationist research, which finds that except for certain stereotypes, social
factors are always secondary to linguistic factors, social constraints demonstrated the
most powerful relationship to r-deletion in this study (Table 5.7).
Table 5.7. Social Factors favoring r-deletion
Social Factors
Gender
Female
Male
Ethno-Linguistic Identity
Delhiite
Mixed Background
Bengali
Hindi Belt (Punjabi, UP/Haryanite)*
Age/Occupation
Working
Student
Retired
Delhi Stay
Punctuated
Continuous
Factor Weight
N
% realized
as [Ø]
Ø]
.58
.40
2151
1662
44.4
28.9
.69
.59
.57
.46
252
528
138
2895
42.1
55.7
34.8
34.1
.63
.55
.39
1320
683
1810
44.2
37.0
33.0
.62
.37
Total N
2050
1763
3813
39.1
35.9
*Two factors are conflated; p = .012; Input value = 0.366; Log likelihood = -2342.039;
Chi Square/cell = 1.9345
Gender is the most significant predictor of r-deletion, with women (.58) far less rhotic
than men (.40). This coincides with formality here—formal contexts motivate less rhotic
realizations. Variants more commonly found in both women’s speech and formal speech
have been interpreted as the prestige form across several variables and many contexts.
Labov, for example, most clearly demonstrates a change-in-progress towards the prestige
form in New York City rhoticity behavior with the markedly different behavior by
middle class women in formal and informal contexts (Labov 1972c). In the Indian
context, Sahgal and Agnihotri (1988: 56) demonstrate that postvocalic (r) is more likely
230
to be unrealized by women, in more formal reading style, and by speakers from more
prestigious academic backgrounds. The current markedly different cross-gender behavior,
in conjunction with significantly less rhoticity in more formal contexts can be understood
as socially indicative—the non-rhotic realization is the more formal or prestigious form.
Diachronically, based on linguistic behavior, an r-full pronunciation was stigmatized by
Delhi IE speakers 20 years ago and this continues today.
While phonetic environment, a linguistic factor, is the second strongest factor group, each
of the other social factors discussed next (and illustrated in Table 5.7) prove more
powerful than the remainder of the linguistic factors in terms of overall significant factor
group ranking (see Table 5.5)
Ethno-linguistic Identity
Rhoticity behavior distinguishes four ethno-linguistic backgrounds to make up the third
strongest factor group: Delhiites are the least rhotic (.69), followed by mixed
backgrounds (.59) and Bengalis (.57), finishing with the Hindi Belt as the most rhotic
(.46). Hindi speakers from Punjab and UP/Haryana are collectively considered members
of the ‘Hindi Belt,’ which is a meaningful social group with specific ideological
characteristics for my participants. Interestingly, while speakers did not identify ethnolinguistically specifically as from the Hindi Belt, and instead identified as Hindi speaking
Punjabis, UP-ites, and Haryanites, there was no disagreement from participants as to
what demographics are clear members of the Hindi Belt. However, because no earlier
research has suggested that speakers from the Hindi Belt are linguistically distinct from
231
surrounding regions, and because no speaker self identified as a member of the Hindi
Belt, a conservative approach to coding was taken, and tokens by Hindi Belt speakers
were originally coded as from Punjab and UP. When quantitative analysis revealed that
these two groups’ have very similar rhotic behavior, statistical motivation, in conjunction
with the above social motivations, permitted the conflation of UP/Haryanites and
Punjabis into the Hindi Belt grouping. Contrastively, while the single Bengali speaker’s
overall rhoticity patterns very closely with the ethno-linguistically mixed speakers' rhotic
behavior, there is no justification for collapsing these two factors: Bengalis are culturally
and linguistically distinct community37, both with respect to their Bengali-speaking
background (a non-mutually intelligible cousin of Hindi), and with respect to their
English behavior, which numerous participants highlighted as different in, for example,
phonology, and intonation. Returning to the caveat offered in the initial discussion of this
factor group, the results for ethno-linguistic identity should be read with caution, given
uneven distribution and low N for some groups, in particular the Bengali and Delhiite
groupings. More data would likely flesh out this picture and provide more robust results.
As well, it is interesting to note that within these interviews, using Delhiite as an ethnolinguistic identity was limited to two men from the youngest generation. This may be an
emerging trend, wherein one’s familial and/or ancestral background are abandoned or
downplayed, and a new Delhiite identity is adopted. This would be worth re-approaching
to explore how and whether this sociolinguistic alignment develops.
37
For example, spoken English by Bengali L1 speakers has been argued to be structurally distinct from
Tamil and Hindi L1 behavior in terms of pitch accent (Pickering & Wiltshire 2000) As well, my informants
almost categorically described Bengali IE speakers as having different linguistic behavior, in particular
citing that /v/ and /w/ are pronounced as [bʰ], the IE schwa is pronounced as [o], and /s/ as [ʃ].
232
Age/Occupation
The factor group combining age and occupation demonstrates interesting links to
rhoticity. Middle aged workers are the least rhotic (.63), while their children are more
rhotic (.55) and the oldest generation—the retired parents of these workers—are the most
rhotic (.39). There are important socio-historical correlates motivating this sociolinguistic
pattern. India has undergone drastic socio-political changes across the lifespan of these
three generations and has had multiple formal and informal language policies, given that
both indigenous and externally introduced languages have been prominent on the subcontinent for over 100 years.
Exploring Indian socioeconomic and linguistic history, Indian economic self-reliance first
gained national momentum with Mahatma Gandhi and was enacted within government
policies in 1947, after India gained its independence from Great Britain. Until then,
English was the language of the government, and was spoken by a powerful, but small,
minority of the population. While it would have been convenient, in some sense, for the
newly-formed government to carry on in the same language as the colonizers, this was
not without a myriad of accompanying problems, most of which surrounded the identity
of India as a collective whole and as a newly formed nation state (T. Chand 1944). For
example, English was interpreted as a tool of colonial domination (Fabian 1986)which
should not be taken forth into India’s education system as Mahatma Gandhi wrote ‘It is
my belief that English education has bankrupted our minds…and has left us unprepared
for courageous citizenship’ (Prabhu & Rao 2003: 364, cited in Vaish 2008).
233
Starting with India’s 1950 Constitution, English was established as an official language,
while corpus based planning was enacted for Hindi (Vaish 2008), with the goal that Hindi
would become India’s official language by 1965, displacing English. However, during
this period, the Indian government recognized that issues of national identity, linguistic
and ethnic diversity would not be solved with Hindi evolving into the sole national
language. In 1963, English was permanently established as a co-official language, and
fifteen indigenous languages were chosen as official, ‘scheduled’ languages, which have
now expanded to twenty-two constitutionally recognized Dravidian and Indo-European
languages.
During the period when my workers (the least rhotic group) were growing up and
entering the workforce, India was thus grappling with how language could or should be
tied to national identity, and focused on creating nationalist links with internal languages
(Vaish 2008), while relegating English to a functional role. Educational policy was also
affected during this period. RP norms were valorized and encouraged in Indian schools
over other styles of English pronunciation (Sailaja 2009), and students’ pronunciation
was often corrected during class towards RP norms for English pronunciation (V. Chand
2008), while RP norms for English were encouraged through Indian media. Locally
produced English radio and TV programs followed national ideologies and India’s
informal language policy, by using a RP accent over other English varieties (Vaish 2008).
These workers thus grew up during in an internally focused socialist government period,
where non-rhotic RP was found in media, and promoted in school.
234
Starting in the mid 1980’s and continuing today, there has been a gradual loosening of
India’s economic borders. Major economic overhauls created during Rajiv Gandhi’s
tenure as India’s prime minister (1984-1989) specifically targeted the Indian tax code,
trade restrictions, and currency exchange, while a growing demand for skilled labor
service export and policy reforms have also been influential on Indian economics (J.
Gordon & Gupta 2004). These policy changes have been motivated in large part through
the late 1990’s increased wage-remittance by Indians working in the Gulf (Migration
Dialogue 2005) and the early 2000’s increased outsourcing and IT industries in India (J.
Gordon & Gupta 2004). Further evidence of the opening of India’s economic borders is
found in the soda market: the locally produced Campa-Cola had supplanted international
brands like Coca-Cola and Pepsi from the 1970’s until 1991, when international varieties
were again allowed access to the Indian market. India is now a free-market system, and
these economic changes clearly separate these workers from students. The worker
category captures an age group which was educated within an inward looking country
which projected RP norms.
Students have lived through a very different Indian setting. Cable TV is now a staple,
with shows from across the globe, demonstrating various accents and world-views. In
conjunction with the recent economic growth in and awareness of outsourcing, these
shows are encouraging an awareness of English dialects (Cowie 2007; V. Chand 2008).
Cable TV channels based in India, e.g. NDTV, have been influential in de-stigmatizing
various non-RP Indian accents through talk shows and other programs in IE (Sailaja
2009). There is also evidence from qualitative reflections by these participants that
235
modern schooling places significantly less emphasis on pronunciation, and on as RP as
the target (V. Chand 2008). Meanwhile, modern media display several varieties along the
rhotic continuum, including local media usage which ranges from dominantly non-rhotic
to variably rhotic. These differences in social, government and media based influences
from one generation to another are a possible motivations for the significant difference in
rhoticity across the generations: students are significantly more rhotic (.50) than their
parents (.63). The correlation between socio-historical context and rhoticity is clearly
relevant, and we will return to it after exploring how other social factors mediate rhotic
pronunciation.
Delhi Residence Length
Looking at punctuated versus continuous stay in Delhi, permanent Delhiites are
significantly more rhotic (.39) than those with punctuated stays (.62). This suggests that
past interactions with non-Delhi IE speaking communities have influenced the transient
population towards a less rhotic pronunciation. While no quantitative studies of rhoticity
exist for IE populations outside of Delhi—yet inside India—, this would be a fruitful area
for further examination.
Another possible explanation for this division is that the non-rhotic pronunciation is
identified as a Delhi feature by the transient population, even though it is not a
categorical feature of Delhi IE. These transient speakers may strive towards this
hypothesized goal within their continual identity formation to establish themselves as
Delhiites. I have uncovered no direct proof of this possibility; however, I have found two
236
tantalizing indirect leads in this direction. First, transient Delhiites are much more
vehement about their Delhiite status when asked about their travel background. Second, a
majority of my informants describe Delhi culture as more focused on appearances, as
judgmental and class conscious, and have even linked such ideologies to language
practices (underlined, below):
..more stuck up, more rude… (f18ND 17:11)
Delhi would be a lot of [HINDI punju cheap, stupid] culture and a lot of showoff and a lot
of, you know, a just one-upmanship, that’s peculiar to Delhi, so Delhi is also becoming very
glamorous, but glamorous more in a negative sense where you more, where you more just
outdo the other. (f30PG 17: 26-29)
…very class-conscious city, very class-conscious. And that reflects in the, our language..
(f27RG 11:13-14)
These quotes suggest that Delhi IE behavior may be at odds with respect to other large
cities, especially with respect to linguistic practices, like the non-rhotic pronunciation,
that are linked to prestige. As well, some speakers champion a single ‘correct’ English—
which would likely be non-rhotic, given that they highlight the worker age group as
examples of ‘good’ English. These speakers also suggest that Delhiites are not taught this
‘correct’ version in school, nor do they speak this ‘correct’ version. Collectively, these
quotes suggest that Delhi may be unconsciously identified as a non-rhotic dialect by IE
prescriptivists, but may also be a context where people are judged more harshly for
deviations from the prestige variant. This could account for the more rhotic permanent
Delhiite practice and the less rhotic transient Delhiite trend, in conjunction with the less
rhotic prestige form. However, while these links are suggestive, they are nothing more at
this point, and deserve further exploration in later research.
237
Interactions among Factor Groups
A problematic interaction arose between two of the social factor groups, namely the
occupation/age factor group, and the ethno-linguistic factor group. A cross-tabulation of
rhotic results comparing these two factor groups reflects stratified qualitative responses
from participants: students and a portion of the workers are more likely to identify
dominantly as Delhiites, while no retired speakers identify only as Delhiites, instead
always offering a regional ethnic identity. Collectively, this means that there are empty
cells and an irregular distribution. While this is problematic, statistically, it is not without
precedent (e.g., Tagliamonte, Poplack & Eze 1997; Tagliamonte 2006: 233). As well,
while oft-considered ‘basic’ social factors are often idealized as independent in the
variationist model, they have been challenged in other multilingual alternative
marketplace contexts (e.g., Rickford 1987). Given that both factor groups are significant
in their influence on rhotic behavior, neither of these factor groups can simply be
excluded from analysis. Clearly, more research is needed to determine if this coupling is
inherent to these social factors or this social context, or could be eliminated with a larger
sample.
The Delhi Prestige Form
Female, working age, transient, self-identified Delhiites are the least rhotic, overall, while
the most rhotic group is male, retired, Hindi Belt permanent Delhiites. Cross-tabulations
of each social group with formality reveals that all groups are acting as members of the
same speech community: they are all moving in the direction that they perceive as more
formal (non-rhotic) in tasks that require greater attention to speech.
238
The prestige form can be understood as non-rhotic. It is more likely in formal context, the
speech of women, and speakers who ethno-linguistically define themselves as Delhiites,
as opposed to, e.g. Punjabi or Bengali. However, it is problematic to assert the non-rhotic
pronunciation as the unequivocal Delhi prestige form within this population, when two
additional factor groups are accounted for: Age/Occupation and length of stay in Delhi.
Examining a cross-tabulation of Delhi Residence Length with Age/Occupation,
demonstrates that age and residence length are linked to rhoticity in a nuanced fashion.
There are no significant differences found by distinguishing residence patterns in the
oldest cohort, and, as already discussed, there are no students with punctuated Delhi
stays. Turning to the worker generation, the strongest factor which distinguishes rhotic
behavior is their length of stay in Delhi. Illustrated in Table 5.8, stable, working Delhiites
are far more rhotic than those whose life in Delhi has been punctuated with departures.
Table 5.8. A comparison of non-rhotic realizations for workers with continuous and punctuated Delhi stay
Workers
Continuous Stay
Punctuated Stay
%
35%
68%
Total N
N
955
365
1320
An analysis of factors motivating rhoticity in just the worker population is telling: Table
5.9 demonstrates that Delhi stay is the most significant factor: a punctuated stay
motivates an r-less pronunciation (.80) far more than continuous stay (.37). Phonetic
environment, ethno-linguistic background, syllable stress and formality also significantly
influence rhotic patterns in the worker sample. Focusing on social factors, transient
Delhiites and self-identified Delhiites are the least rhotic, while self-identified ethnolinguistically mixed, permanent Delhiites are the most rhotic.
239
Table 5.9. Non-rhotic realization for workers (27-52 years old)
Factor Group
Delhi Stay
Punctuated
Continuous
Phonetic Environment
Pre-consonantal (nucleus and non-nucleus vowel)
Word and Syllable final (nucleus vowel)
Word final (non-nucleus vowel)
Syllable final (non-nucleus vowel)
Ethno-Linguistic Background
Delhiite
Hindi Belt
Mixed
Syllable Stress
Primary Stress
Secondary Stress
Formality
High formality
Low Formality
Gender*
Morphemic Independence*
Factor Weight
N
% realized
as [Ø]
.80
.37
365
955
67.9
35.2
.62
.52
.44
.32
353
491
294
182
55.8
43.0
39.5
33.0
.73
.49
.34
124
1064
132
54.8
40.8
62.1
.53
.45
806
514
46.5
40.7
.56
.48
[ ]
[ ]
Total N
316
1004
46.5
43.5
1320
*Gender and Morphemic independence are not significant in predicting non-rhotic patterns in this
population, and their factor weights are not displayed; p = .04; Input 0.438; Log likelihood = -807.724; Chi
Square/cell = 154.5557
How can we explain this? There are two possible ways of interpreting this data. First, it
may be capturing a supra-local non-rhotic prestige form, with which speakers with
outside-of-Delhi experience are more familiar. Second, it may be that Delhi is more
rhotic than other regions of India—Delhi may not attend as closely to the nationally
prestigious non-rhotic variant. Indeed, as we saw, Delhiites are not considered ‘classy’ or
prestigious. Instead, even though national capitals are contexts typically associated with
prestige, Delhi is characterized as very unsafe city inhabited by ‘crooks and con-men.’
240
Diachronic Analysis of IE (r)
Given these apparent-time results, do the differences visible across age groups reflect
diachronic sound change, or are they more appropriately understood as age-grading?
Sahgal and Agnihotri (1988)—S&A, hereafter—compare two age groups: younger
speakers under 18 in class X and XII and speakers over 40, both in South Delhi. Given
the 22 year gap38 between that study and the current one—their younger speakers would
now be in the 36-40 range, and their older speakers would now be 62+ —these groups are
thus directly comparable with the current workers (27-52) and retirees (59+).
We find that yesterday’s youth—today’s workers—still lead in non-rhotic
pronunciations, though they are much more rhotic today (Table 5.10). The oldest
generation do not diverge greatly from their behavior 20 years ago—they are still far
more rhotic than the next generation. It appears that there is, across ages, currently less of
a move towards non-rhotic pronunciations in formal contexts than evidenced 20 years
ago. Unfortunately, it is impossible to determine whether these patterns are statistically
significant or not, given that S&A do not provide an overall token count (from which one
could conduct a chi square comparison).
Table 5.10. Overall percentage of non-rhotic tokens by age, comparing current results with S&A.
Informal
Formal
38
Current
Students
(17-19)
34%
47%
S&A Youth
(currently 34-38)
76%
89%
Current
Workers
(27-52)
44%
47%
S&A Elder
(currently
60+)
34%
47%
Current
Retired
(59+)
32%
37%
Here I discuss the results as presented in S&A (1988); however, their data was collected pre-1985, before
the first publication of their results with these participants. Given that this data was collected in 2007-8,
there is thus an ~22 year gap between studies.
241
Taking into consideration today’s students, we can look at three generations in real time.
It appears that there was a peak in non-rhotic behavior, which has since subsided into a
typically heterogeneous pronunciation. South Delhi IE is variably rhotic, based on both
social and linguistic factors. This peak is interesting in two ways. First, it may
demonstrate age-grading: current workers were more r-less while in high school than they
are currently. Second, it may also demonstrate diachronic change: the oldest generation
has maintained their dominantly rhotic behavior, the next generation has continued to be
far less rhotic (though more rhotic as they age), and the youngest generation is most
similar to the oldest generation, and is dominantly rhotic. Possible evidence for both of
these competing hypotheses is next explored.
Evidence of Age-Grading?
Addressing the first point, the overall rhotic behavior of the worker generation has
changed drastically over a 20 year time span. They were—and are—the peak in crossgeneration r-less behavior in both this sample and S&A’s samples. However, the worker
generation has increased in rhoticity over time. Relevant to understanding this potential
age-grading change in rhotic pronunciations, S&A have interesting divergences from the
current methodologies. They target a region—south Delhi—as their focus. From this
starting point, they ‘selected students at random from the class registers… of some
schools in South Delhi’ and for their elder population, they ‘selected informants at
random from the master-lists of some areas of South Delhi…from the local welfare
organisations’ (1988: 53). Nowhere do they inquire as to each participant’s length of time
in Delhi—however, houses change hands infrequently in India. People are less mobile,
242
there are often restrictions on who can purchase lots in particular societies and, within the
joint family system, families maintain holdings across generations. Thus, while their
sample is presented as capturing the linguistic practices of the ‘educated Delhi elite,’ it
may better reflect my subpopulation of permanent Delhiites—that is, those who have not
been posted outside of Delhi.
As well, the current worker population may diverge from S&A’s student population.
S&A select participants based on where they attend school, while I select participants
based on where they live, and these may not coincide. When the worker generation was
in high school, there were far fewer prestigious English medium public high schools in
Delhi (public schools, as in the UK, are the equivalent of US private institutions, which
charge fees), e.g., Modern, St. Columbus and DPS, each with only one location39.
Students often traveled quite far across town to attend prestigious English medium
schools. We cannot be sure that the randomly selected student population analyzed in
S&A actually reflects students who were domiciled in south Delhi. The south Delhi
public schools were, at that time, likely to reflect a student population which
encompassed a much larger region than south Delhi. Hence, they are potentially different
from the current worker population, who, when in Delhi, all grew up and continue to live
specifically in south Delhi. Comparing rhoticity across these two populations to
determine real-time diachronic changes may be counterproductive, given the potential
population differences. It is thus unclear if any significant age-grading has occurred for
39
Today there are many more prestigious public schools, and many schools have more than one location
(e.g. DPS RK Puram, DPS Mathura Road, DPS Vasant Vihar, and DPS East of Kailash). However,
students today continue to travel long distances between home and school.
243
the worker age group, and in the interests of space, detangling these possibilities will be
left to another paper.
Rhoticity across Time
Diachronic change is the second topic brought to light through the overall comparison
with S&A’s results. While their study shows a rise in r-less pronunciation over two
generations, this study demonstrates a peak in r-less pronunciation in the 27-52 age
group, followed by a decrease in r-less behavior in the youngest generation. I suggest that
this behavior can be linked to India’s colonial and post-colonial history as they relate to
media, education and ideology.
Pre-Partition India was run by Britishers from across the UK—as such there was a range
of accents, some non-rhotic, and some rhotic. All of these pronunciations were
prestigious, given their role as the colonizer’s code. Thus, Indian speakers of English had
multiple prestige targets, in terms of rhoticity. However, after Partition, the target
pronunciation in India was narrowed, and reflected non-rhotic RP through three
mediums: 1) the colonial British population was gone, and in their absence, the constant
multiple targets were also gone, 2) radio, and eventually TV media post-Partition was
dominantly BBC style (non-rhotic RP), either directly from the UK, or mimicking it
locally, and 3) the Indian school system, more structured to RP pronunciations, again
non-rhotic (Vaish 2008). Indeed, many of this worker population remember an explicit
emphasis on pronunciation during their schooling.
244
In contrast, the current youth do not feel like pronunciation was emphasized in their
schooling. Instead, they go so far as to suggest that their parents’ speech is ‘better,’ and
‘more educated.’ These youth have been educated after the opening of India’s economic
borders, within which rhotic and non-rhotic media input (through TV, movies, radio, and
the internet) is abundant, unlike the situation for their parents’ generation. Modern media,
offering multiple realizations of rhoticity, demonstrates that there is no longer a single
international media standard in terms of rhoticity. The adoption of outside norms is also
increasingly problematic for youth. They universally shun what they term ‘fake
accents’—which always manifest as mimicking RP or AE—used by schoolmates. I
suggest that the variable rhotic behavior in the youngest generation can be linked to these
changes in education (and especially attention to pronunciation), media, and ideologies,
as reflected in discourse about ‘fake accents.’
Related to their qualitative reflections on the speech of their parents’ generation, today’s
students speak like their grandparents in casual contexts. However, they speak like their
parents in formal contexts (Fig. 5.1). Their positive evaluations of and ideologies about
their parents’ speech are directly reflected in their own formal speech. Meanwhile, in
casual situations, their behavior patterns very similar to their grandparents, who can be
understood as bearers of a local or Indian culture, of which linguistic practices are just
one aspect.
245
Figure 5.1. Young people’s alternation between the casual forms of the oldest generation and the formal
forms of their parents, the middle generation
Form ality and Rhoticity
50%
45%
% of non-rhotic tokens
40%
35%
30%
Students
Workers
25%
Retirees
20%
15%
10%
5%
0%
Informal
Formal
The student population’s marked alignment with their parents’ speech in formal contexts,
and with their grandparents speech in informal contexts, may also be a result of the joint
family system. In India, it is common for multiple generations to live under the same
roof. Most often this is paternally based, that is, the younger generation tends to live with
their father’s parents. Families which live independent of other generations are a new and
marked situation. These categories—joint vs. nuclear family—are salient to this
population, and a majority of my participants have grown up in joint family systems. In
the joint family system, while parents are at work, children spend much more time with
their grandparents. The student population may thus be more influenced by their
grandparents’ rhoticity patterns, which they were surrounded by at home, while their
246
parents’ speech, markedly less rhotic, may symbolize a more formal register. This may
also be linked to the fact that the workers—their parents—are away, in more formal
working situations on a daily basis. More research is required to tease out the influence of
the joint family system on patterns of rhoticity in the youngest generation.
Trills
The trill realization has never been quantitatively studied in IE. Past research (e.g.,
Sahgal & Agnihotri 1988; Sharma 2003) has, due to low trill frequency, conflated trills
with approximant and flap realizations. However, the number of trill tokens in this data
permits analysis in conjunction with social and linguistic factors, and is next explored.
Results of Trill Analysis
Five factors have a significant influence on conditioning trill realizations40. The highest
ranked factor is phonetic environment (Table 5.11, N = 296, p = .002), within which
syllable final (word internal) position with a full-vowel nucleus motivates trills (.76),
followed by word final position (with any type of nucleus) (.61), syllable final position
with a schwa nucleus (.50), with preconsonantal position (with any type of nucleus) least
favoring the trill realization (.18). Beyond phonetic environment, no other linguistic
factors proved significant, and the results for syllable stress, morphological
independence, and formality are thus not discussed here.
40
These factor weights, interpreted as probability within variationist work, must be interpreted vis-à-vis the
input—overall, trills were very rare (7.8% of the tokens), and thus, these factor weights must be interpreted
within this context, not simply as probabilities within a given context. Environments which favor trills must
be interpreted as substantially more likely than the overall average of trilled pronunciations.
247
Table 5.11. Factors which favor trill realization
Factors
Phonetic Environment
Syllable final, full-vowel nucleus
Word final, schwa nucleus and full-vowel nucleus †
Syllable final, schwa nucleus
Preconsonantal, schwa nucleus and full-vowel nucleus †
Age/Occupation
Student
Retired
Worker
Ethno-Linguistic Identity
Delhiite
Hindi Belt/Mixed †
Bengali
Delhi Stay
Punctuated
Continuous
Gender
Male
Female
Syllable Stress*
Morphological Independence*
Formality*
Factor
Weight
N
% realized
as [R]
.76
.61
.50
.18
86
176
21
13
16.8
8.9
6.2
1.3
.64
.53
.34
50
205
41
7.3
11.3
3.1
.82
.50
.07
36
258
2
14.3
7.5
1.4
.65
.32
222
74
10.8
4.2
.56
.45
[ ]
[ ]
[ ]
Total N
175
121
10.5
5.6
296
* Three linguistic factor groups were not significant in predicting trill realizations; †Factors are conflated;
p = .002; Input = 0.041 Log likelihood = -884.966; Chi Square/cell = 1.4924
Among the significant social factor groups, Age/Occupation was the most important,
wherein students (.64) and retirees (.53) are more likely to produce trills, and workers are
far less likely to trill (.34). Retirees and students were not conflated because their
behavior is significantly different. The third strongest determining factor is ethnolinguistic identity, which is conflated into three groups: Delhiite status strongly
influences trill production, (.82) the Hindi Belt (Punjabis, UP/Haryanites and Mixed)
variably trills (.50), and the lone Bengali speaker does not trill (.07). Again, these ethnolinguistic results should be taken with a grain of salt, given the distribution. Length of
stay again proved significant, this time in locating those with punctuated stays as more
248
likely to trill (.65) than those with continuous stays (.32). The final significant factor
group is gender: men are more likely to trill (.56) than women (.45).
This patterning is interesting for several reasons. First, trill realization is conditioned
primarily by phonetic environment, and is most likely in syllable and word final position,
regardless of nucleus quality. This confirms the overall finding which suggests that
variable rhoticity behavior is a phonological, not morphological process. Indeed, Harris,
with cross-linguistic support, suggests that ‘some conditions previously attributed to
syllabic structure are better defined more locally in terms of neighboring segments or
boundaries, while others are better viewed as having a wider, suprasyllabic scope’—
within this, postvocalic (r) is ‘amenable to the more local treatment’ (Harris 2006: 20). In
conjunction with the results found here, this suggests that studies of rhoticity should
focus on immediate phonetic environment, and not on the morphemic quality of words,
which is arguably not capturing the underlying motivation for r-deletion41.
The second interesting fact which arises from analyzing trill behavior is that workers are
again separated in linguistic behavior from students and retirees, who pattern more
similarly. Clearly, the youngest generation is not behaving in alignment with either RP or
AE, with their variably rhotic and occasionally trilled patterning. The joint family system,
which encourages far more interactions between the student and retirees, may again play
a factor in the similar patterns between the two populations.
41
For example, it has been suggested that English bimorphemic words do retain a rhotic pronunciation, e.g.
furry /fɜɾɪ/, while monomorphemic words do not. (Gramley & Patzold 2003) Alternatively, this rhotic
realization may not have anything to do with morpheme structure, and instead, may have more to do with
/r/’s intervocalic position in furry.
249
The third point of discussion focuses on ethno-linguistic background: the two Delhiites
lead this trilling train pattern (.82), while the majority of the speakers (27) are conflated
into the Hindi Belt in this analysis, and pattern together, as significantly different (.50)
from Delhiites, and second in the trilling train. These behaviors, both together and
separately, support Woolard’s (2008) (among others) proposal that locally significant
alignments can demonstrate strong links to language practices, and they also provide
support for not conflating these two speakers with the Hindi Belt in the larger analysis.
Fourth, we return to Delhi residency, which again has interesting links to rhoticity.
Speakers with time spent outside of Delhi are more likely to trill, which suggests that this
feature may be more common in other areas of India, or in the Indian Military culture in
particular, given that a majority of the transient Delhiites’ outside-of-Delhi experiences
are through military postings.
Fifth, where do trills stand on the continuum of prestige variants? Men lead in trills,
formality is not significant in predicting trills, and Delhi transients are far more likely to
trill than permanent Delhi residents. As well, trills are almost as common in the oldest
generation as they are in the youngest generation, but less common in the middle
generation. These facts collectively suggest that the trill realization is not a prestige
variant, but neither is it entirely shunned. Instead, these suggest that while it is
conditioned primarily by phonetic environment, it may also hold covert prestige within
Delhi IE. Further, this may be a feature more common in other areas of India, and not a
250
particularly or uniquely Delhi IE feature—both possible explanations would benefit from
further research.
Conclusion
These rhotic results tell us much about this Delhi dialect of IE: it demonstrates ‘orderly
heterogeneity’ (Weinreich et al. 1968) in terms of rhoticity, directly challenging blanket
academic statements which frame IE as ‘wrong’ or ‘needing fixing’ (e.g., Krishnaswamy
& Burde 1998). Clearly, more structural research is needed on this and other regional IE
dialects, to understand areas of convergence and divergence, and to counter sweeping
pejorative generalizations of IE. Earlier generalizations, which devalue IE and reflect
larger societal ideologies, may motivate the marked shift found between formal and
informal situations in the youngest generations’ speech.
This data also demonstrates that the Delhi IE dialect is evolving, and is distinct from
international norms, manifesting as a variably rhotic (or semi-rhotic) dialect. Importantly,
social factors prove dominant in predicting rhoticity, which can be linked to the narrowly
circumscribed sample and the more rigid Indian social structure. Interactions amongst
these social factor groups were impossible to avoid in this study. Further analysis will
reveal if this is inherent to the more rigid Indian social hierarchy, or can be overcome
with a sufficiently large sample. This analysis demonstrates that variationist
methodologies can be successfully applied to alternative, multilingual contexts, but
clearly, more research is required to tease out locally significant social groupings, to
develop social factor groups which are relatively independent, and to determine what
251
entails a uniform vs. heterogeneous population in this context—in this study emergent
social categories, drawn from ethnographic data and qualitative reflections by
participants, were pivotal for understanding the local situation.
252
CHAPTER 6
(V) AND (W) IN INDIAN ENGLISH
What caste will English have then?... we may offend the ears of the good
Englishman by our inability to use the letter V. and W. as though they came
from the same posture of tongue and labial disposition. (Rao 1978: 421)
Introduction
This chapter explores the sociophonetic reality surrounding the IE v/w merger, a feature
that has become iconic of IE to global audiences (V. Chand in press) within which the
globally circulated understanding of the merger involves pronouncing /v/ as [w], e.g.
awailable. However, this merger has not been explored through nuanced quantitative
variationist or acoustic analysis. In keeping with the goal of understanding current
structural patterns in IE as well as diachronic change in IE practices, I frame the current
research within several related strands of research: past research on the (v/w) merger and
potential local and international historical motivations for the merger in IE, which are
discussed next. While this chapter focuses on two variables, (v) and (w), with separate
hypothesized underlying realizations and behavior, (v/w) is used as shorthand to discuss
them collectively vis-à-vis past research.
253
The IE (v/w) Phenomenon
The (v/w) phenomenon has been broadly labeled as a pan-IE feature (e.g., Trudgill &
Hannah 2002). Supporting this, the single (to my knowledge) methodical—albeit via
impressionistic auditory analysis, not acoustic analysis—IE study on aspects of this
merger focused on variable realizations of (w) as [w] versus [ʋ] . This study found the
merged phoneme [ʋ] to be stably present in the speech of non-native English speaking
upper middle class New Delhiites across two age groups (teenagers and 40+ year olds),
speech styles and language backgrounds (Sahgal & Agnihotri 1988). However, other
survey work has identified the (v/w) phenomenon as a feature exclusive of English
speakers from South India and Bengali regions (Chaturvedi 1973; Cowie 2007). Further
evidence for excluding Bengalis and South Indians is provided from personal interviews,
where IE speakers suggest that Bengalis have a different (v/w) merger, where both are
replaced with [bʰ], e.g., ‘if you talk to a Bengali and say its very, very dangerous you
know he would say its “bherry-bherry dangerous”’ (f39VD 16:14-15) and, replacing
what: ‘“bot,” they would say. The W is always pronounced as a B and, you know, little
things like that and it’s also typical’ (f59MS 32:41-42). Meanwhile, South Indians
purportedly maintain a phonetic distinction between /v/ and /w/. Thus, personal accounts
and some past research suggest that the potential IE (v/w) merger may be found solely, or
dominantly, within Hindi and Punjabi-speaking regions.
What is the nature of this merger? The single non-survey based analysis of the phonemic
status of IE (w)—that is, where dialects like RP would produce a [w] and are, underlying
254
represented as /w/—suggests that the [w] has a dominant allophone lacking velar
constriction and tongue retraction: the labio-dental approximant /ʋ/ (Sahgal & Agnihotri
1988). In total, three allophones for /w/ are suggested with hypothesized behavior—[v] is
found in more formal contexts, while [w] and [ʋ] are in free variation, and [ʋ] is more
common (1988). However, linguistic motivations beyond formality are not explored,
their quantitative analysis does not distinguish [v] production, and Agnihotri later argues
that [w] ‘appeared absent from Indian English altogether’ (1994: 243). This statement is
difficult to reconcile with his co-authored 1988 methods and results, and we are left
lacking a clear understanding of how and whether (w) variation is systematic.
Compounding this, no quantitative or acoustic research has been done on IE (v).
Literature on other non-standard42 English varieties with a similar merger has described it
as a voiced bilabial approximant (Trudgill et al. 2004). Meanwhile, a striking asymmetry
has been uncovered in other situations wherein two classes of sounds are perceived as
‘the same’, yet consistently have different acoustic signatures (Labov, Karen & Miller
1991). Given the lack of acoustic analysis thus far, it is possible that a ‘near-merger’ can
explain the (v/w) phenomenon. To this end, acoustic analysis is a necessary aspect of this
variationist study.
The historical motivations that introduce phonological variation certainly influence
development and modern day realizations—for example, variable rhoticity in New
42
By non-standard, I mean varieties that are not globally recognized as a locus of standard English—that is,
localities that are not America, the UK or Australia. There are multiple terms to capture this standard/nonstandard division (e.g. colonial/post-colonial, outer circle, peripheral, etc.), all of which are pejorative, to
varying degrees, and all of which encapsulate slightly different communities of speakers.
255
Zealand can be traced to the distinct dialects transported by British, Scottish and Irish
settlers (E. Gordon et al. 2004). The historical motivation for the (v/w) merger is thus an
important aspect towards explaining current apparent-time synchronic variation and
predicting diachronic change, and I next discuss two possible historical motivations for
the merger. These both offer a backdrop for interpreting the current apparent-time results,
and motivate the current coding practices, given that the role of internal linguistic
constraints on (v/w) production has not been examined previously.
Hypotheses for the Historical Grounding of (v/w)
Where might this merger have come from? There are two possible sources, which will be
addressed in turn. The first is a substratum contact influence from other Indian languages,
hereafter the Substratum Contact Hypothesis. The second potential source of this merger
is from the original colonial British infusion of English into India, hereafter the England
Hypothesis. Trudgill et al. (2004) suggest, with compelling evidence, that southeastern
British English had a near merger of (v/w) from the 16th through 19th centuries, which
later (post-1850) resolved itself back into two separate phonemes, /v/ and /w/.
Substratum Contact Hypothesis
Given the sociopolitical history of India as a postcolonial nation-state with ~415
indigenous languages (Registrar General 1991), the relatively recent infusion of English
in the subcontinent, and the massive population of speakers for whom English is a
second, later acquired, and domain-specific language (R.G. Gordon, Jr. 2005), IE specific
features may be the result of language contact (Sharma 2003; Herat 2005; Sharma 2005),
256
and/or may be the result of language diffusion through areal Indian linguistic features
(Emeneau 1980 [1956]).
Hindi contact is a possible influence for the (v/w) merger in Hindi/IE bilinguals
(Agnihotri 1994). While this will be explored, understanding the IE manifestation as a
direct transmission from Hindi phonology is problematic: while dominant and one of two
languages — alongside English—used across India in federal government, Hindi is not
the sole possible influence on how English is spoken on the subcontinent. It is one of 22
national languages, and one of over 415 indigenous Indian languages (R.G. Gordon, Jr.
2005). Further, Hindi is an Indo-European language, while several language families are
also represented in the region: Dravidian, Austro-Asiatic, and Tibeto-Burman are the
three other dominant families (R.G. Gordon, Jr. 2005). Given this, the structure of these
other potential substratum influences should also be examined. However, explaining a
modern (v/w) merger in terms of current-day continued substratum contact, as opposed to
a historic merger motivated by an earlier language contact situation, is problematic: this
understanding presupposes that modern IE is neither a stable nor a native dialect for any
IE speakers. While there is a large population of English L2 learners today, this does not
account for native urban Delhi or other IE speakers who are simultaneous English/Hindi
bilinguals. Delhi is one of two epicenters, with Bangalore, of the economic boom, and
thus is the focus of this study. This section examines the historical influence from various
substratum languages in order to assess the Substratum Contact Hypothesis as a plausible
motivation for the merger, and what implications it may hold for phonetic and phonemic
realizations today.
257
I focus first on Hindi as a potential contact language, given the suggested apparent
‘immunity’ of Bengali and South Indian IE speech to the (v/w) merger and past
suggestions of a contact-induced link to Hindi. In Hindi, there is no (v/w) phonemic
contrast. Instead, Hindi has one phoneme (and one grapheme, <va>43) which encompasses
[v] and [w], among other allophones. However, Chaturvedi argues that Hindi
phonetically most closely resembles the AE [v], but phonemically, encompasses AE [w]
articulations (1973: 93), while the phoneme is not consistently distinguished across the
Hindi population, with ‘uneducated’ people replacing it with [b] (Chaturvedi 1973: 46)44.
In contrast, Ohala analyzes the Hindi phoneme phonetically as [w:], with [v] as an
occasional variant articulation (1983: 6-7), while she makes no comments regarding the
phoneme’s presence (or absence, through a [b] pronunciation) across different
demographics. Others suggest that the two target sounds are subsumed within the
phoneme /β̞/: however, ‘the lower lip is so far from the upper teeth and the upper lip, it is
not possible to say whether this sound is better classified as a bilabial approximant β̞, or a
labiodental approximant ʋ’ (Ladefoged & Maddieson 1996: 324-5). There may also be a
lexical specificity to Hindi (v/w) variation: one IE/Hindi speaker commented to me on
generally having ‘a [v] dialect with [w] for certain words, e.g. pav[w] bhaji,’45 while
another said that he had to learn to pronounce all ‘w- words’ with [w], not [v]. Thus, both
[v] and [w] are possible allophones of the Hindi phoneme. The Hindi phoneme’s status as
a sonorant or an obstruent is unsettled, the presence or absence of a conditioning
43
The < > notation is used in this paper to distinguish orthographic symbols, or graphemes.
Given the earlier mention of a Bengali /v/~/b/ merger, it may be that Chaturvedi correlated ‘uneducated’
with ‘Bengali’ in his analysis.
45
Pav bhaji is a potato-based curry garnished with fresh coriander leaves and onions, and is a fast food dish
that is especially popular in Bombay.
44
258
environment for allophonic variation is also undetermined, and the range of potential
variation across different speaker populations is similarly uncertain. For the purposes of
understanding how Hindi may influence IE, the most important aspect is the lack of a
(v/w) phonemic contrast in Hindi.
Further complexity arises from the unstable status of the voiceless labio-dental fricative
/f/ in Hindi. Martinet suggests that voiced/voiceless phonemic pairs maintain and
strengthen each other’s presence within a phoneme inventory, through their shared
features and single point of contrast (1957). However, his focus is on phonemic
inventory, not phonetic: we must extend Martinet’s hypothesis, then, to suggest that a
stable /f/ in Hindi could thus encourage the presence of [v] realizations, while an unstable
or absent /f/ could act to further destabilize [v] realizations.
The grapheme <f> is a recent introduction to Hindi—this is evident through Indo-Aryan
areal analysis (Masica 1991) and in the orthography: < f. >—the grapheme is a modified
version of <f >, a grapheme representing the phoneme /pʰ/. Masica (1991) suggests that
the Hindi [f] pronunciation is maintained, in conjunction with two other fricatives
borrowed into Hindi, the /z/ and /ʃ/, however the situation is arguably more complex:
while some speakers consistently distinguish /pʰ/ and /f/, I have commonly observed
words spelled with <f> pronounced [pʰ] by other Hindi speakers in Delhi, although the
linguistic and social distribution of this variation has not been methodically examined.
259
Hindi has some voiced/voiceless fricative phonemic pairs, i.e. /s/ and /z/, /x/ and /ɣ/.
However, these are all recent borrowings into Hindi and there is evidence that these
recent introductions are only variably maintained in Hindi. Masica suggests that Hindi
speakers who do not make the contrast exist, but are marginalized: ‘[s]peakers who fail to
consistently maintain /f, z, ʃ/ (confusing them with /pʰ, j, s/) are by this definition
substandard, although these are “foreign” sounds’ while ‘Ohala also hedges by proposing
to mark all morphemes as +/- native, on the ground that even standard speakers whose
speech maintains /f, z, ʃ/ are aware that “many non-standard Hindi speakers” do not have
these segments’ (1991: 92-3). Hindi also evidences other gaps in voiced/voiceless
fricative pairings: for example, there is no voiced counterpart for /ʃ/46. The
voiced/voiceless symmetry is, however, found for all 11 oral stop minimal pairs and both
affricate minimal pairs. Importantly, within the Hindi phonemic inventory, the majority
of fricatives are recent introductions, and voiced/voiceless fricative pairs are not
consistently contrastive. Given the unstable presence of /f/ and the equally irregular
allophonic presence of [v], Martinet’s theory suggests that neither sound provides enough
scaffolding to support the other. This may, in turn, motivate the fact that past analyses of
Hindi (e.g., M. Ohala 1983; Ladefoged & Maddieson 1996) have tentatively labeled [v]
as a non-dominant—if evident—allophone. Returning to IE and Hindi’s potential
contact-induced influence with this evidence in mind, Hindi’s (v/w) merger, minimal
evidence of both a stable [v] allophone and /f/ phoneme could collectively motivate a
similar (v/w) merger in contact-induced colonial era English in India.
46
Nor is there a voiced counterpart for Hindi /h/.
260
Looking to other Indian languages for possible substratum contact influence, Indo-Aryan
languages (of which Hindi is one) typically only have a /w/, while Western Indo-Aryan
languages often have a [v] allophone which appears before front vowels (Masica 1991:
100). Dravidian languages—Tamil and Telugu, for example—also broadly follow this
pattern, and do not phonemically contrast /v/ and /w/. In addition, Tamil does not
maintain a regular voicing contrast across its phoneme inventory, and has several
inventory ‘gaps.’ Pertinent to potential non-Hindi contact-induced influence on IE (v/w),
there are no labio-dental fricatives in Tamil, although there is a labio-dental approximant,
/ʋ/, and, in intervocalic position, there is evidence that other bilabial segments weaken
into approximants (Keane 2004). Tamil, hence, does not maintain a (v/w) phonemic
contrast. These factors suggest that any Tamilian or larger Dravidian substratum
influence on IE would disfavor a (v/w) phonemic contrast, and favor a [w] pronunciation.
Importantly, this Indian linguistic data also suggests that languages are more likely than
Martinet thought to maintain stable gaps in the phonemic inventory47. Thus, the presence
of an /f/ in these various substratum languages may not be a strong precursor to the
maintenance of a [v] realization for a (v/w) phoneme, however, their collective lack of
phonemic distinction between [v] and [w] may influence their status in IE48. A (v/w)
47
Further support for this maintenance of gaps can be found in the famous case of the Arabic b/*p
phonemic gap.
48
Returning to the contact hypothesis, let us assume that historically earlier Indian speakers of various
Indo-Aryan and Dravidian languages are introduced to English—how would they interpret and then
reproduce the data, as English learners? /f/ was contrastively salient in English input, while [v] may not
have been, depending on whether the England Hypothesis (the merger was a feature of colonial era British
English, discussed below) is correct. These Indian speakers, when learning English, likely stopped /f/,
producing a [p] or [pʰ]—this stopping of fricatives (e.g. f p or pʰ, θ t or tʰ and ðd or dʰ) is a modern
feature of IE cross-dialectally (Kachru 1982; Coelho 1997), and [f] is typically found in Indo-Aryan
languages only as a borrowed and infrequently maintained phoneme (Masica 1991), while robust stop
inventories are common. Given no phonemic /f/, these contact speakers may well have also stopped /v/—
resulting in [b] or [bʰ]—or, alternatively, may have produced some sort of labial approximant, another
261
contrast, if introduced within British colonial English, could conceivably have been
collapsed, given the absence of such a phonemic contrast in the dominant Indian
languages of that era. This is my prediction of substratum contact-induced influence on
IE without the factors suggested by the England Hypothesis, discussed next.
Importantly, in the scenario set forth thus far, the proposed contact must be understood as
having occurred during the colonial period and thus currently existing as a structural
feature of IE—modern day urban IE speakers, who are the focus of this study, are not in a
contact situation, and instead, are native fluent speakers of IE. I thus offer a historical
structural motivation within the Substratum Contact Hypothesis which could produce
both modern merged (v/w) production and limited [v] production.
England Hypothesis
The second potential source of the (v/w) merger, which I term the England Hypothesis,
suggests that the merger arrived with the dialect of English originally brought to India in
the 18th and 19th centuries, the heyday of colonial British India. There is considerable
evidence that southeastern British English underwent a merger ‘with an articulation
intermediate between [v] and [w]’ (Trudgill et al. 2004: 41) which had reversed itself and
was dying out by the 1850’s (2004: 34). Examining British historical records in
conjunction with current day English varieties in seventeen ‘Lesser-Known’ English
situations—‘colonial varieties of English spoken in small communities in the North
sound common to several native Indian consonant inventories. Thus, in this earlier colonial context of
English dialect contact, we can hypothesize a scenario where /f/ is unlikely to survive—the mutually
stabilizing voiced/voiceless symmetry is absent, and [f] is not systematically contrastive within several
Indo-Aryan languages.
262
Atlantic, South Atlantic, North Pacific and South Pacific’ (2004: 26)—a (v/w) merger has
been found. These islands were all colonized pre-1850, before the British merger had
clearly reversed itself into two separate phonemes.
Further replicating the British pattern, a partial merger reverse in progress has been found
recently in young speakers from Norfolk Island, Montserrat and Anguilla (Trudgill et al.
2004: 40). Based on British historical evidence and phonetic analysis of six of the
modern island varieties, Trudgill et al. hypothesize that this merged sound is realized as a
bilabial approximant [β̞] (2004). The likelihood of each of these islands independently, or
through diverse substratum contact situations, having developed a (v/w) merger which
manifests as /β̞/ is slim—in a large-scale survey of phonemic inventories worldwide,
Maddieson found that only 1.9% of languages have a bilabial approximant (1984: 96,
cited in Trudgill et al. 2004: 40). Trudgill et al. thus reject substratum influence or
indigenous language contact as viable explanations for the same merger in seventeen
British colonial and post-colonial islands across the globe today.
Returning to the Indian context, this British merger may easily have also been transmitted
to India. Educational policy first introduced English as a supplement to existing
indigenous language courses in 1813, and English became the official language of courts
and government in 1835, via the enactment of Lord William Bentninck’s New Education
Policy (Doyle 1998). English education was then advocated as a replacement to Asian
and Oriental language education with Macaulay’s 1835 Minute (Macaulay 1920/1965).
However, English was introduced in India much earlier, circa the early 1600’s, when the
263
East India Company started trading in the region, and British Christian missionaries first
entered the subcontinent (Baldridge 2002). British English contact grew dramatically
with the 1775 institutionalization of the British East India Company as a governing
power. The British colonizers and missionaries arrived while the merger was still
common in England, and likely transported this merged speech into the Indian context,
akin to what happened in Trudgill et al’s (2004) seventeen island colony contexts.
However, Trudgill et. al’s hypothesis does not invalidate the Substratum Contact
Hypothesis—while they reasonably dismiss substratum contact as a possible motivation
in the contexts they studied, (and substratum contact would at first blush seem even less
likely as an explanation in a subcontinent with such linguistic diversity), I have shown
that several of the largest families in India, Indo-Aryan and Dravidian, in fact do have
structures that could promote (or, at least not hinder), such a merger in colonial IE. As IE
developed, there were no competing indigenous or colonial introduced phonologies with
a contrastive (v/w) distinction. Instead, it appears that both substratum and colonial
influences could have provoked and mutually reinforced this hypothesized IE merger.
Further, the historical precedent of a merger in numerous British colonies, all colonized
during a period when the British English was also merged, has important implications for
how we ideologically interpret the modern IE context. The modern IE situation—if
merged—cannot be brushed off as a case of repeated misinterpretation of input: India is
not simply a situation of a continued contact-induced merger perpetuated by generations
of English learners. Any IE merger found in native English speech must be considered an
264
artifact of a historical merger—successive generations of IE speakers are not simply
failing to acquire an aspect of their dialect. While resolving how strong each historical
influence was on the development of the suggested modern IE (v/w) merger may not be
fully possible, this quantitative analysis of (v/w) will offer the first analysis which takes
into account internal linguistic and external social influences. As well, apparent-time
analysis will illuminate potential diachronic changes in IE (v/w) and offer clues about the
future of this iconic IE feature.
There are both structural and ideological factors which may help to predict the future of
the merger in IE. British English has since unmerged (v/w) into two separate phonemes
and RP has lost prominence in India as an external standard. Following this pattern, a
similar merger reversal in progress has been uncovered in several of the merged island
dialects, while the remaining islands still have an apparently robust merger. Meanwhile,
unmerged American English (AE) has risen in prominence in India, both through media
and economic ties, while English as a code is increasingly tied to India’s economic
development and future economic prosperity (Rai 2005; Mahapatra 2006; Smerd 2007).
Clearly, this is a complicated scenario which I have set forth, one which cannot
ultimately be resolved here. However, examining three generations of IE speakers (v/w)
production acoustically and within the variationist paradigm will illuminate the
sociophonetic reality, and expose how it may be changing over time.
265
Current Study
We will now turn to the quantitative variationist analysis of IE (v/w), which was drawn
from the same population discussed throughout this dissertation, and begin by discussing
the coding choices undertaken within this analysis.
Internal Coding Choices
This section details the motivations for the internal linguistic factor groups and particular
factors. Establishing codes for acoustic realization was first addressed by examining past
acoustic research on unmerged English dialects, and cross-linguistic research on the two
sounds. Subsequently, a pilot acoustic analysis was conducted to establish whether
distinct approximant realizations could be further distinguished (V. Chand 2007). These
factor groups, in conjunction with formality, comparable through three speaking styles,
were subsequently filled out and cover all of the internal linguistic conditions that were
selected for analysis within this study, and are next discussed.
Unmerged English Dialects
In many English dialects, /v/ and /w/ are distinct phonemes with minimal pairs (e.g.
weal/veal, wine/vine, while/vial, wet/vet) and different acoustic properties (Ladefoged &
Maddieson 1996). For example, the American English voiced labio-velar approximant
[w] has the following spectrographic features: it is voiced, so one can expect to see, in the
lower frequency region of the spectrogram, vertical striations demarking a voice bar;
being a semi-vowel, the formant structure is similar to /u/ (approximately F1:300,
F2:800, F3:2600); and the F2 is lowered slightly, because of lip rounding (Ladefoged &
266
Maddieson 1996). There is, however, debate over whether the F3 also is lowered by lip
rounding, and if so, to what degree (Robert Hagiwara, personal communication).
The voiced labio-dental fricative [v] has the following acoustic features: fricatives
evidence a random noise pattern, within which [v] has a higher energy noise component
around and above 8000 Hz—this is visually perceptible with darkening in that region of
the spectrogram (Ladefoged & Maddieson 1996). A majority of the literature avoids in
depth acoustic discussions of fricative spectrographic features—they are generally
considered messy, chaotic and confusing (Ladefoged 1993; Mark Jones personal
communication). Given this, some argue that it is easier to look for the more distinctive
[w] features as distinguishing criteria. Distinguishing [v] and [w] realizations through
acoustic means, based on these different acoustic signatures, was thus done. However,
there are a range of intermediate articulations, the acoustic signatures of which are not
available from English acoustic research. As well, acoustic research on English, which
maintains a distinction, does not help us to uncover internal linguistic constraints which
may motivate different realizations. We next turn to cross-linguistic research to explore
potential acoustic features for separating these intermediate pronunciations and
determining what internal linguistic motivations to code for.
Cross-linguistic Research on (v/w)
There is cross-linguistic support for [w] being realized as a fricative allophone,
particularly in word initial position related to the relative ‘strength’ of fricatives and
approximants. In the hierarchy of sonority and openness, fricatives are stronger than
267
approximants—they are less sonorous and less open than approximants (Lass 1984: 178).
Fortition, a sound change wherein a ‘weak’ sound is changed into a ‘strong’ sound, is
subject to syllable position: strengthening is a typologically common phenomenon, where
an approximant can undergo a process of fricativization in syllable initial position
(Stephens 1988: 423; Jones 2005). Further, in strengthening environments, given that [w]
is a co-articulated, one articulation can take precedence over the other. In these cases,
‘when becoming a fricative… [w] shows itself primarily as a labial, less often as a velar’
(J.J. Ohala & Lorentz 1977: 587)49. Word and syllable initial environment and the coarticulatory properties of [w] in conjunction may thus motivate [v] realizations, and these
were incorporated into the coding. Specifically, syllable and word position was coded
with a six way distinction: word initial, syllable initial, word final, syllable final, and in
non-edge position of an onset and coda consonant clusters.
Cross-linguistically, fricative allophones of /w/ tend to increase in noise as the following
vowel moves forward, and palatalization has been hypothesized as a motivator for
provoking fricative allophones of /w/ in 20+ languages spanning the globe—indeed,
‘palatalization, while not a necessary condition, frequently leads to the fricativization of
[w]’ (Stephens 1988: 426). This is further substantiated by Indo-Aryan allophonic [v],
which is found word initially, before front vowels (Masica 1991: 100). Voicing, place of
articulation and manner assimilation are all other possible motivations for particular
realizations. The preceding and following environments were thus each coded for seven
49
There also are several cases where ‘[m]orphophonemic alternations between /w/ and labial or velar
obstruents (or both)’ appear, including Nguna, Berber and Mende, among others (J.J. Ohala & Lorentz
1977: 581). This is relevant to the current study because morphophonemic alternations are often synchronic
reflexes of past allophonic alternations (c.f. Kiparsky 1982a; 1985).
268
possibilities: voiced/voiceless fricative, voiced/voiceless non-fricative consonant
(including approximants, laterals and rhotics), high front vowels ([i] and [ɪ]) versus other
vowels, and pause.
Pilot Acoustic Analysis
A preliminary acoustic analysis was conducted in order to determine if approximant
productions could be reliably distinguished. This pilot study focused in particular on
word initial (v/w) tokens followed by front vowels, and focused on ones which appeared,
through aural analysis, to be merged approximants. All audio was recorded and
analyzed50 in Praat (Boersma & Weenick 2006), at a sampling rate of 22,050 Hz.
Unfortunately, this analysis revealed only one clear acoustic difference useful to coding,
while several acoustic measures demonstrated a range of behaviors with no clear cut-offs
possible between ‘different’ realizations. For example, consonant length ranged from 9
ms. to 127 ms., while formant transition length and slope, as well as the length of the
following vowel each demonstrated a continuum of behaviors. However, none of these
measurements fully coincided with other measurements: in other words, tokens with
longer consonant lengths did not also have consistent formant transition lengths and
50
The acoustic realizations were measured in a number of ways. Durational measures were taken for the
target sound and the following vowel. Following Jones (2005), the cutoffs for the target sound were defined
as the beginning and end of stable structure (the structure consisting of formants or frication). The target
consonant’s length was measured from the onset of sound (and voicing in particular, but these tended to be
identical) until the formant structure began to change. Thus, consonant length did not include the transition
into the following vowel. The transition into the following vowel was measured separately, and the vowel
was only considered as the stable formant structure following the transition. Formant measurements were
made at the midpoint of the target token. The slope of the transition was taken by comparing the beginning
and endpoints of the transition period, in conjunction with the length of the transition, and captured the
period from the end of the steady formants of the target sound up until the beginning of steady formants for
the following vowel.
269
slopes or consistent following vowel lengths. These measures were hence not useful for
distinguishing different approximant realizations within a large scale analysis.
Spectrographic analysis revealed that <v> was often completely lacking in frication, with
formant structure (independent of anticipatory formants for the following vowel)
consistently visible: while F1 and F2 were present throughout the consonant, F3 tended to
resolve itself during the second half of the consonant. Figure 6.1 presents a comparison of
the word <vine> spoken in IE and AE, an unmerged dialect, to demonstrate the
difference in acoustic signatures.
Examining the approximants’ formant height, the only difference of note uncovered is
that orthographic <v> tokens sometimes started with a lower F2 than the <w> tokens,
however all of the differences were under 10% of the total formant height, this difference
was not consistent, and it would be difficult to operationalize within coding, given that
the IE vowel space is almost entirely unstudied51—such a study would need to be
undertaken first. However, some tokens demonstrated a very brief (under 10ms.) period
of frication followed by a much longer period with formants independent of the following
high vowel. I used this information to create a specific acoustic realization code of
‘mixed.’
51
E.g., Sailaja’s (2009) descriptive study of SIEP does not offer acoustic analysis of vowel space, and also
highlights several populations which deviate from the ‘standard’ she focuses on.
270
Figure 6.1. Spectrograms of vine in IE (left) and AE (right), with [w]-like formants in IE and [v]-like
frication in AE
The approximant realizations had consistent formants and no higher spectra turbulence,
narrowing the possible realizations to being bilabial, labio-velar and labio-dental
approximants. Masica suggests that in Indo-Aryan languages, the contact of the ‘[v]-like
allophone… is typically a loose one, between the upper teeth and the inside of the lower
lip’ (1991: 100). However, distinguishing between these approximant realizations, given
that they all involve a labial articulation, was not possible through acoustic analysis, and
it is not clear what means Sahgal & Agnhihotri (1988) used to reliably distinguish [w]
and [ʋ]. While they suggest that the main difference between these two sounds is that the
latter lacks lip rounding and retraction of the tongue, it is not clear how this was
271
operationalized, as these articulatory properties do not have one-to-one correlations with
acoustic or auditory signatures. Indeed, such is difficult without video data, and perhaps
unnecessary. For example, Ladefoged and Maddieson, examining Isoko, do not try to
decide whether a sound is better classified as a bilabial or a labio-dental approximant
(1996: 324). This is the nature of approximants; they are neither consonants nor truly
vowels, and their place(s) of articulation do not involve contact—though full contact
would be much easier to assess! Thus, through auditory and acoustic analysis I was
unable to find any reliable acoustic measures to distinguish between approximant
realizations beyond separating out mixed realizations, and the social and linguistic
motivations which may distinguish [w] and [ʋ] in IE shall unfortunately remain
unexplored here.
Additional Internal Factors
Given that linguistic motivations for variation in (v/w) are unstudied, additional factor
groups were created to offer a more comprehensive examination within this seminal
research. Factor groups encoding the underlying representation formality, syllable
strength and word frequency were created, while a third factor group, conflating two
other groups, was also created.
Underlying Representation
This analysis stands apart from past research, in that it is exploring whether one or two
possible underlying representations (URs) exist. Therefore, beyond each token for several
possible acoustic realizations, two URs were coded, /v/ and /w/. These were determined
272
primarily through orthography, however, there are a few exceptions to this, where a /v/ or
/w/ hypothesized to exist in the UR, but is not found within the English orthography, e.g.
words like one /wən/.
Formality
Tokens were also coded for formality, with a tripartite distinction between low, medium
and high formality, in order to determine whether attention to speech motivates different
(v/w) realizations. In particular this may prove interesting because of the assumed local
awareness of IE (v/w) and its targeting in the call center industry and in international
accent reduction courses (Cowie 2007; V. Chand in press).
Syllable Strength
Syllable stress has a demonstrated an impact on realization choice for other variables.
Notably, IE rhotic realization is mediated by syllable strength. While studies of AE are
often able to separate monosyllabic function and content words based on syllable
strength, and its effects on vowel reduction, this type of categorical distinction is not
possible in IE, which may have distinct stress patterns (please see Appendix B for a
discussion of how stress manifests in IE based on past literature). Syllable stress was thus
coded with a three-way distinction between monosyllabic words, syllables with primary
stress, and syllables with non-primary stress. Importantly, this three-way distinction
allows us to examine two different combinations: the impact of syllable weight
(monosyllabic words and multisyllabic words with primary stress on the target syllable
versus non-primary stress on the target syllable), and, the impact of total syllable count
273
within a word (multisyllabic words with primary and non-primary stress on the target
syllable versus monosyllabic words).
Word Frequency
Cross-linguistically, ‘historical sound changes often affect high frequency words first and
spread gradually to lower frequency words’ (Shi, Gick, Kanwisher & Wilson 2005: 345).
Apparent time phonological variation has also been linked to frequency for other
variables, in particular to vowel reduction, word final (t,d) deletion, and palatalization,
wherein high frequency words are more prone towards reduction and assimilation (2005).
Meanwhile, ‘conventionalized units’ (Bybee 2002: 279), that is, combinations of verbs
and particular particles, are also implicated within frequency effects, which suggest that
the diffusion of change is related to a word’s frequency within certain contexts (Bybee
2002: 278). From this, it is clear that frequency is complicated and nuanced vis-à-vis
different levels of grammar.
However, all of this aside, assessing frequency is very problematic, for two reasons: first,
there are different cutoffs—‘frequent’ does not necessarily signify any particular cutoff,
and research has used a range of cutoffs thus far. Francis and Kučera (1982, cited in
Bybee 2002) define high frequency words as occurring 35+ times per million while all
words which occur <35 times are infrequent. Bybee complicates the picture in her
explanation that, when analyzing a set of 631 non-past participle tokens for (t,d) deletion
her ‘cut-off point between high and low frequency was arbitrarily chosen to make the
number of tokens in the high and low groups approximately equal’ (2002: 265-6) (this
274
process defined low frequency words as occurring less than 100 times, while high
frequency words occur 100 or more times, essentially, distinguishing via the median).
Finally, Jurafsky, Bell, Gregory and Raymond (2001) use the 95th percentile as the cutoff
for frequent words. Frequency, is infrequently (forgive the pun) used in variationist
studies, perhaps because of the discord over what constitutes ‘frequent.’
Returning to the current issue, and starting with a surface level skim of the data, there are
several transparently frequent words in IE (and, indeed, across English varieties) which
have a target /w/, e.g. one, want, was, we, what, who, will, with and target /v/, e.g. every,
even, have, give52. It is reasonable to hypothesize that word frequency is linked to
synchronic variation, wherein high frequency words are more likely to undergo lenition,
akin to Bybee’s (2002) results. As well, word frequency may suggest a direction of
change for this variable, given that newer realizations are more likely to be found in high
frequency words for other phonological variables (Bybee 2002).
The current corpus is ~350,000 words—much smaller than those examined in the
literature discussed above—and the results of such frequency analyses may be challenged
by future research on larger corpora, but cannot be summarily ‘fixed’ here. This is
because further data cannot be collected, and it is problematic to derive word frequency
from different dialects, registers and genres, since differences within such types of data
are interrelated with defining and separating linguistic sub-systems (Biber et al. 1999).
Thus, word frequency and derived cutoff points are determined based on this corpus.
52
Several visual representations of word frequency are offered as word clouds in Appendix E, F, G and H.
275
Speakers falling outside the original participant demographics were excluded (six
speakers), given that their speech has also been excluded from these quantitative
variationist analyses. As well, each speaker’s reading of the Grandfather Passage was
removed from the corpus, because it does not reflect natural discourse word frequency53.
Finally, this corpus excludes all non IE speech—e.g., my speech, as the AE speaking
interviewer, bracketed meta comments inserted within the transcript, and Hindi speech
and English translations for Hindi speech. The remaining corpus totals 268,201 words.
Word frequency was analyzed using the concordance software program AntConc
(Anthony 2007). The cutoff for frequent words was set at 5%—that is, the most frequent
5% across the whole corpus, following Jurafsky et al. (2001). This cutoff was chosen
because the 5% cutoff is a typical point of distinction, however, as I discovered, choosing
slightly different cutoffs severely affected the significance of word frequency vis-à-vis
several factors (V. Chand & Gales m.s.).
Contractions (e.g. we’re, they’ve) and forms which are different parts of speech (e.g.
verbs worked and love versus their respective nominalizations worker and lover) were
treated as separate words, while verb conjugations (e.g. works versus working, have
versus having) and singular/plural forms (e.g. one versus ones, university versus
universities) were treated as the same word in terms of frequency codes. The tokens were
53
The ‘Pear Story’ retelling was included because, while at surface value a somewhat unnatural event with
restricted lexical class, this was not so for two reasons. First, the dominant elements of the story (pears, a
boy, a ladder, bicycles, trees, picking fruit and an older man) do not involve /v/ or /w/. Second, I was
surprised by the range of interpretations offered for this movie: to say the least, they were extremely varied,
and often strayed wildly from what I considered the ‘plot line’ and ‘dominant elements.’ Pear Story
retellings were thus included in word frequency analyses.
276
coded through a tripartite distinction separating ‘frequent’ words (the top 5% of words,
e.g. we, way, with, every, very), from two groups which both comprised the remaining
95% of the words. Within the lower 95% of words, ‘infrequent’ words (e.g. towards,
activity), were separated from compound words which contain a frequent word (e.g.
anyway, without, seventh, whenever). This distinction was made in order to test whether
the presence of a frequent word influences larger less frequent compound word
realization quality.
Function and Content Words
Word frequency has a complicated relationship to syntactic category—Bybee’s (2002)
findings, which found a frequency effect for (t,d) deletion, have been challenged by Shi
et. al. (2005), who suggest that some of Bybee’s high frequency tokens that underwent
deletion were also function words. Function and content words can be understood as
separate syntactic categories and there is evidence from several corners that speaks to an
underlying difference. Support for separately coding for syntactic category comes from
research on lexical storage and aphasia: function words are accessed more quickly
(Segalowitz & Lane 2000), and different aphasic impairments selectively impact function
and content words (Caplan 1987)—in other words, they are handled differently in
cognitive retrieval and storage. Meanwhile, word length reduction and frequency have a
complicated relationship to syntactic category: ‘content words are shorter when more
frequent, and shorter when repeated (verbs and modifiers as well as nouns), while
function words are not so affected’ with ‘shorter pronunciations, after controlling for
frequency and predictability’ (Jurafsky 2007).
277
Specifically, frequent words are more likely to be reduced and to undergo assimilation,
but variation is also influenced by syntactic category (Shi et al. 2005), wherein content
words are more likely to be reduced than function words (Jurafsky 2007). Completing the
picture, synchronic variation examining reduction in word final (t,d) and vowel reduction
to schwa have demonstrated significant correlations to syntactic category, while
assimilatory palatalizatation is not influenced by syntactic category (Jurafsky et al. 2001;
Shi et al. 2005). Thus, some variables are affected by syntactic category, but others are
not, and no satisfying overarching explanations for this divergence have been postulated
thus far. Focusing on variables affected by syntactic category, for both (t,d) and vowels,
function words are more likely to be reduced, while vowel reduction also demonstrated a
within category frequency effect, wherein high frequency function words reduce less than
low frequency ones (Shi et al. 2005). Given that word length, syllable count and word
frequency do not offer a complete picture of how reduction may be linked to word
quality, syntactic category is also examined here in order to tease apart (v/w) motivations.
Function words are different from content words on a number of levels. For example,
corpus research has highlighted several features which separate the two categories. One
corpus-based analysis of English grammar suggests eight ways of distinguishing function
and content—which they term ‘lexical’—words (Biber et al. 1999), presented in Table
6.1.
278
Table 6.1. Distinctions between Content/Lexical and Function words (reprinted from Biber et al. 1999: 55)
Feature
Frequency
Head of Phrase
Length
Lexical Meaning
Morphology
Open vs. Closed Class
Number
Stress
Lexical
Words
Low
Yes
Long
Yes
Variable
Open
Large
Strong
Functional
Words
High
No
Short
No
Invariable
Closed
Small
Weak
However, operationalizing this syntactic distinction is trickier: content and function
words cut across parts of speech (Biber et al. 1999). Really, function and content words
are better understood as fuzzy categories, with good, more prototypical examples, and
other more peripherally related examples. For example, most researchers have agreed that
determiners and auxiliary verbs are fairly ‘empty’ function words, serving primarily as
‘glue’ for binding content words together. However, a review of practices suggests
disagreement and problems operationalizing this distinction on several levels. First, there
is debate over whether any adverbs are function words. Second, few researchers even
consider how to deal with discourse markers like well (though see Biber et al. 1999 who
place them in a separate category, 'inserts,' which includes discourse markers, greetings,
interjections, attention signals, hesitations, response elicitors, etc.). Third, while
quantifiers, e.g. one, are included in some studies, they’re not in others. It is not clear if
quantifiers are simply subsumed within determiners within studies which neglect to
mention them (which they shouldn’t be), or if quantifiers are intentionally excluded (and
if so, why they are excluded). Within this, whether quantifiers include or exclude ordinals
(e.g. first, twelfth) is also uncertain. Corpus research, which regularly makes use of this
syntactic distinction, typically starts with a set list of function words, which may be
279
problematic for application here because they are based on American and British English
varieties, which may—and very likely do, given research like Schneider (2000)—vary
from IE syntactic distinctions. As well, and more importantly, these lists are not publicly
available for use here. Truly, ‘our original intuitions about the differences between
function words and content words, though undoubtedly important and correct in some
sense, are very difficult to make explicit’ (Caplan 1987: 268).
Therefore, while I draw on past distinctions to separate most words, some choices are
also made ad hoc and guided by similarities and differences in linguistic features. Within
this study I define function words as pronouns, determiners, quantifiers—which include
ordinals—, prepositions, auxiliary and modal verbs, conjunctions, some sentence level
adverbials (e.g. otherwise), and the discourse marker well. Further, some words are
polysemous, and lexical meanings are distinguished from functional ones (e.g. lexical I
have a car versus functional I have eaten) determined through the surrounding context.
Tokens are coded in a binary fashion as functional or lexical. Finally, simply because I
have been frustrated by the lack of specific word lists in past literature, below is a list of
function words found in these (v/w) tokens (Table 6.2):
280
Table 6.2. Function words found in (v/w) tokens
(w) Function Words
about, anyone, between,
down, even, everyone, few,
have, I’ve, once, ones, one,
someone(‘s), throughout,
towards, twice, twenty,
twelve, twelfth, was(n’t),
we(‘d/’ll/’re/’ve), well,
were(n’t), what(‘s), whatever,
when, whenever,
where(as/’s), wherever,
whether, which, whichever,
while, whilst, who, why, with,
within, without, won’t, would,
wouldn’t, would’ve
Polysemous (w)
Function Words
well (exclamation
or discourse
marker versus
adjective);
while (noun
versus
conjunction);
will (auxiliary or
modal verb versus
noun);
otherwise
(conjunction
versus adjective
or adverb)
(v) Function Words
above, could’ve, eleven,
eleventh, every,
everybody(‘s), everyday,
everyone(‘s), everything,
five, hav(e/en’t/ing), I’ve,
might’ve, must’ve, over,
seven, seventeen, seventh,
seventy, several,
themselves, they’ve, twelve,
vis-à-vis, we’ve, whatever,
whenever, whoever,
who’ve, would’ve, you’ve
Polysemous (v)
Function words
have (auxiliary
or modal verb
versus main
verb)
Preceding Environment + Location
Finally, a factor group was created which conflated two existing factor groups for the /w/
tokens. This was done because these groups provoked structural zeroes (Paolillo 2002)—
that is, that all factor combinations are not theoretically possible. For example, the
location factor group separates onset clusters, where (w) is preceded by a consonant in
the onset of a syllable (e.g. quick). However, the preceding phonetic environment
category allows coding for pauses, multiple consonant types, and two vowel types—
pauses and vowels will never be found in this position, hence not all factor combinations
are theoretically possible. Given that these original factor groups were independently
significant in motivating (w) realization, it was necessary to create a new conflated factor
group. Neither factor group was significant in predicting (v) realizations, thus, a conflated
factor group was not necessary for the analysis of (v).
Exploring what conflations were made, while six separate conditions were originally
coded for location within the word—word initial, word internal and syllable initial, onset
281
cluster, coda cluster, syllable final and word internal and word final—preliminary
analysis revealed (w) realization as less nuanced: word and syllable initial locations
behaved identically, and were conflated. As well, categories which comprised 1% or less
of the total data, as well as categories with categorical [w] realizations were excluded
from analysis. This excluded syllable final location (>1% and categorically [w]) and coda
cluster location (categorically realized as [w]) tokens from analysis. We are thus left with
the following three codes, and tokens: syllable initial (word initial or word internal),
onset cluster, and word final.
Further, while seven distinct conditions were originally established for preceding
phonetic environment (voiced fricative, voiceless fricative, voiced other consonant,
voiceless other consonant, high front vowels, non-high non-front vowels, and pause)
preliminary analysis revealed that (w) realization was less nuanced: preceding voiced and
voiceless fricatives behaved very similarly, and the model was not compromised by their
conflation. Preceding environment was thus conflated to six factor groups (fricative,
voiced other consonant, voiceless other consonant, high front vowels, non-high non-front
vowels, and pause).
While eighteen combinations can be made, several of these are not linguistically possible,
given English phonology (e.g. word final preceded by a high front vowel) and syllable
structure sonority rules (e.g. word final with preceding fricative). As well, some
combinations, while possible, were not found in the data. The remaining ten
combinations found in the data were coded for (Table 6.3).
282
Table 6.3. Preceding Environment + Location Conflated Factor Group—(w) analysis
Fricative (+/-voice)
Consonant (-voice)
Consonant (+ voice)
High Front Vowel
Other Vowel
Pause
Word or
syllable initial
Possible
Possible
Possible
Possible
Possible
Possible
Onset Cluster
Word Final
Possible
Possible
Possible
XX
XX
XX
XX
XX
XX
XX
Possible
XX
The conflated preceding phonetic environment plus location within the word factor group
thus contained ten factors.
Conflate Following Environment + Location?
While the above orthogonality issues also apply to code combinations within the
following environment and location within a word, an additional conflated factor group
was not necessary. A single word location code disallowed for all possible combinations
of following sounds—coda cluster, because it requires a following consonant, and
disallows for a following vowel or pause. However, this position was categorically
realized as [w], and was thus removed from (w) analysis, given the focus on motivations
for variable realizations. There are no further impossible code combinations within these
two factor groups for the (w) analysis. Meanwhile, these factor groups were left
unconflated within the (v) analysis because location was not significant in predicting (v)
realizations.
283
Current Coding Practices
Underlying /v/ tokens were analyzed separately from underlying /w/ realizations—the
UR factor group was thus used prior to analysis to separate the tokens for two
independent runs, and was then discarded as an independent variable because they were
categorical within each separate run. Beyond coding each token for one of seven acoustic
realizations—the dependent variable—, eight overlapping independent linguistic factor
groups were coded for the (w) analysis, for a total of 41 factors, and seven54 independent
linguistic factor groups were coded in the (v) analysis, for a total of 31 factors. All tokens
were also coded for an additional factor group, which separated underlying /v/ from /w/
tokens, which allowed a quick means of separating tokens for their respective separate
runs. These factors and factor groups are presented in Table 6.4, with examples.
The same social factor groups coded within the rhotic analysis (Chapter 5) were used in
this study for the (w) and (v) analyses, and are presented in abbreviated format below
(Table 6.5)55. In total they constitute six overlapping factor groups, for a total of 23
factors. Given the potential interactions amongst these factor groups, these overlapping
groups are tested in various combinations, in order to understand which means of
categorizing social factors can most productively be linked to language production. In
addition to the factor groups listed below, and in order to ensure reliability across coders,
one factor group was created to designate coder identity, while an additional factor group
was created to individualize each speaker’s results, in order to examine whether any
54
Again, the only difference in coding for (w) and (v) was the factor group that conflated preceding
environment with location, which was not applied to the (v) tokens because preliminary runs demonstrated
that location was consistently not a significant factor for determining (v) realization.
55
Age by decade was also coded for, but is not discussed here, given that other factor groups which take
age into account were significant, while age by decade was consistently not significant.
284
individuals’ behavior was highly divergent from their peers. The former did not evoke
any abnormalities, while the latter highlighted the same speakers who were excluded
from rhotic analysis based on their demographic and linguistic background, discussed
further in the results.
285
Table 6.4. Linguistic Factor Coding Groups
Factor Group
Variable Quality
Conditioning Factors
Example
Approximant with formants
[w], [ʋ] or [β̞]
[v]
[f]
Voiced labio-dental fricative
Voiceless labio-dental fricative
Mixed (fricative then approximant)
Vowel
Other Consonant
Null
[u], [o]
[b], [d]
[fai] for five
<w>
<v>
was
very
Voiced fricative
Voiceless fricative
Voiced consonant (non-fricative)
Voiceless consonant (non-fricative)
High front vowel
Other vowel, including [y]
Pause (determined acoustically)
I think is WAY, you know
That's WHY they did not
and ADVERTISE for it,
on the spot VOICE
kind of very WEIRD altogether
story NEVER look
while, VERY quiet. Ask
Voiced fricative
Voiceless fricative
Voiced consonant (non-fricative)
Voiceless consonant (non-fricative)
High front vowel
Other vowel, including [y]
Pause
it GIVES you a
journey. I'VE faced this
It's a LOVE marriage.
very FEW cars on
director MOVE into
They WON'T.
it. NOW… these
Word initial
Syllable initial (word internal)
who WAS picking
going SOME.WHERE alone,
then
that GIVE a lot of
you HAV.EN'T scored
you know, of TWELVE years
has GROWN. Corruption
Underlying
Representation
Preceding Phonetic
Environment
Following Phonetic
Environment
Location
Word final
Syllable final (word internal)
Onset cluster, with preceding consonant
Coda cluster, with following consonant
286
Table 6.4. Linguistic Factor Coding Groups Cont.
Factor Group
Preceding Phonetic
Environment
Conflated with
Location
Conditioning Factors
Example
Syllable initial (word initial or word
internal) with preceding fricative (+/voice)
Syllable initial (word initial or word
internal) with preceding voiceless
consonant (non-fricative)
Syllable initial (word initial or word
internal) with preceding voiced
consonant (non-fricative)
Syllable initial (word initial or word
internal) with preceding high front
vowel
Syllable initial (word initial or word
internal) with preceding non high front
vowel
Syllable initial (word initial or word
internal) with preceding pause
Onset Cluster with preceding fricative
(+/- voice)
Onset Cluster with preceding voiceless
consonant (non-fricative)
Onset Cluster with preceding voiced
consonant (non-fricative)
Word Final with preceding non high
front vowel
books WE’VE read
Monosyllabic word (stress not
evaluated)
Primary stress in multisyllabic word
Non-primary stress in multisyllabic
word (anything less than primary stress
is weak)
you know, WAY young age
Frequent (top 5% of words in this
corpus)
Compound words containing a frequent
word (only possible for /w/ tokens)
Infrequent
way, very
Functional words
Lexical words
would, over
wish, victory
Informal Speech
Medium Formality Speech
High Formality Speech
Majority of Interview
Pear Story Retelling
Grandfather Passage
up WHILE he
system WHICH he had
we WENT to
of the WEST about
... WHILE he’s cycling
they SWEAR by our hotel.
choose BE.TWEEN Hindi or
your
modern in his LAN.GUAGE
children GREW up and
Syllable Stress
was VER.Y scared
nod of A.PPRO.VAL really
Word Frequency
anyWAY
whilst, avenge
Syntactic Category
Formality Level
287
Table 6.5. Social Factor Coding Groups
Factor Group
Conditioning
Factors
Example
Number
of
Speakers
(N=29)
Gender
Female
Male
16
13
18-24
25-38
39-59
60+
5
8
4
13
Student
Working
Modern Housewife
Traditional
Housewife
Retired from Military
5
11
1
1
Age by
Historical Era
Occupation
Retired from Other
Profession
Volunteer
Worked until marriage/children
Never worked
Semi-retired, continued in second profession after
Military
Fully Retired, not currently working
4
Never worked, only volunteer humanitarian
(education/health) work part time
2
Originally from West Bengal or Bengali regions of
Bangladesh
Originally from northern states of Uttar Pradesh and
Haryana
1
5
Ethnolinguistic
Background
Bengali
UP/Haryana
Delhi
Punjabi
Mixed
Including areas now in Pakistani Punjab
e.g., 1 parent from north India, 1 from south India
2
2
20
4
Age &
Occupation
Student
Working
Retired
Under 20, in school
Any work experience, even if stopped working to
have children. Age is 25-52 (both men & women)
Men either retired, or semi-retired (not working full
time). Women either fully retired, or never worked.
Age: >52 (both men and women)
5
10
14
Delhi Stay
Punctuated
Continuous
16
13
Token Selection and Analysis Methods
Token selection was systematic: within each interview, tokens were taken starting a
quarter of the way through interview (this time frame, of course, varied according to the
288
length of the interview). At this point, the first 100 tokens were extracted for coding,
taking no more than three instances of each lexical item (Wolfram 1993) to avoid
type/token issues, wherein the behavior of a very common individual lexical item differs
from the generic behavior, and that item’s frequency influences the overall distribution
across realizations. From the Pear Story retelling, a maximum of three (v/w) tokens per
lexical item were used, and all 21 (v/w) tokens within the formal reading passage were
used (within which, there were not three instances of any single lexical item). Token
realization was determined through a combination of aural and visual acoustic analysis,
within Praat (Boersma & Weenick 2006), following the guidelines laid out above.
Results
Tokens with a /v/ UR were analyzed separately from tokens with a /w/ UR. Given that
the first 100 tokens of either UR were extracted and coded, the token sets were not equal
in number. In total, there were 4,671 tokens were extracted and coded, which separated
into 2,771 /w/ tokens, and 1900 /v/ tokens. These were then analyzed separately within
Goldvarb X (Sankoff et al. 2005). A few participants fell outside the target demographics
because they were identified as lacking Hindi fluency, as being from a different socioeconomic status or as visitors to Delhi—reasons for excluding these people’s speech were
already covered in the previous chapter. Further, lending linguistic credence to their
socially determined exclusion—that is, social difference as underlying linguistic
difference (Chambers 2003; Llamas 2007), their (v/w) behavior was compared to the rest
of the population and found to be very different. Importantly, this matches the differences
found within the rhotic analysis in the previous chapter, which motivated their exclusion
289
within that analysis . A decision was thus made to exclude them from further analysis
56
here. Here, I examine the results of these two analyses, by first exploring (v) results.
(v)
As discussed above, IE (v) has never been subjected to systematic quantitative or acoustic
analysis, while a persistent [w] realization of (v) is suggested by globally influenced and
influencing accent training centers, whose classes target the merger, e.g. ‘Hindi speakers
practice saying “available” instead of “awailable”’ (Gorman 2007). However, these
results demonstrate the fallacy of this belief. Systematic analysis of seven realizations
reveals that (v) is realized as [v] a majority of the time in this population of native
Hindi/English speakers from New Delhi: 82.1% of tokens are realized as a voiced labiodental fricative (Table 6.6). Although much less common, (v) is also realized as [f]
(7.0%), as [w] (3.8%), as a brief fricative [v] which transitions into a labial approximant
(2.1%), as another consonant (3.9%), as a vowel (0.8%), and as fully absent, or null
(0.3%).
Table 6.6. Overall distribution of (v)
[v]
%
82.1
56
N
1293
[w]57
[f]
%
7.0
N
110
%
N
3.8
60
Total N
Mixed
[v] [w]
%
N
2.1
33
1575
Other
consonant
%
N
3.9
62
Vowel
%
0.8
N
12
Null
[Ø]
%
0.3
N
5
This divergence across three variables also suggests that there are multiple IE dialects, which may vary in
stability, and which may stratify IE speakers in terms of region, ethno-linguistic background and socioeconomic status, beyond other factors not uncovered here. Indeed, this lends credence to my hypothesis,
posited in Chapter 2, that India, while having not necessarily completed Step 4 (endo-normative
stabilization) of Schneider’s (2007) dynamic model of post-colonial English development, may already
evidence aspects of Step 5, differentiation.
57
Here, [w] represents realizations that are labial approximants: distinctions within this category are not
examined here.
290
The above distribution demonstrates that it was worth the effort to code for multiple
possible realizations, as all were uncovered in the data—from this, we can infer that IE
/v/ does not simply have two allophones, but instead, has multiple realizations. However,
while these seven realizations were coded for, and all seven were found within the (v)
data, binomial analysis handles only two variants and multinomial analysis is not
available in current versions of Goldvarb. As is clear from Table 6.6, the distribution of
these variants is by no means equal—the [v] realization makes up ~82%—that is, 1293 of
the 1575 tokens, while at the other end, the null [ø] variant, for example, only has five
tokens. Important here, infrequent variants are not well suited for quantitative analysis
(e.g., Bayley, Lucas & Rose 2002), and this has been accounted for within this study
through three practices. First, dependent variable realizations which represent less than
1% of the total realizations are excluded from further analysis—tokens realized as null or
as a vowel were thus excluded. Second, underrepresented factors (those capturing <1% of
the total tokens) were either excluded or conflated with another factor, if linguistic or
social motivation was also present. We shall discuss conflations as they pertain to each
factor group, below. If conflation with another factor could not be justified,
underrepresented factors were excluded from analysis—one factor, preceding pause, was
eliminated in this fashion. The data was then analyzed comparing [v] realizations against
all other realizations, while motivations for the various minority realizations are
discussed later. We will now explore internal linguistic motivations and external social
motivations for the distribution.
291
Social and Linguistic Constraints on (v)
After the overlapping social and linguistic constraints were tested in various
combinations, one social factor and four linguistic factors proved significant in
motivating a [v] realization, and are next discussed in order of influence.
Following Phonetic Environment
Following phonetic environment is the most significant predictor of (v) realization: a
following voiced sound (voiced consonants, voiced fricatives, and all vowels) motivates
the most [v]-ful realization (.54), while a following voiceless fricative motivates the next
most [v]-ful pronunciation (.41), and a following pause or consonant motivates the least
[v]-ful realization (.17) (Table 6.7).
Table 6.7. Social and Linguistic Factors favoring [v] realization
Factors
Following Phonetic Environment
Following voiced sounds (consonant, vowel
or fricative)**
Following voiceless fricative
Following pause or voiceless consonant*
Syllable Count
Multisyllabic Words*
Monosyllabic Words
Word Frequency
Frequent Words
Infrequent Words*
Delhi Stay
Transient Delhiites
Permanent Delhiites
Formality
High and Medium Formality*
Low Formality
Factor
Weight
N
% realized
as [v]
.54
1350
86.7
.41
.17
79
114
73.4
46.5
.56
.34
1109
434
87.3
72.1
.58
.41
816
727
85.5
80.2
.56
.44
830
713
86.1
79.4
.58
.48
Total N
297
1246
1543
86.9
82.1
* Two factors are conflated; ** Four factors are conflated; p = .032; Input value = 0.855,
Log likelihood = -626.634; Chi Square/cell = 1.5945
292
Four conflated factors comprise the strongest motivator of a [v]-ful pronunciation: voiced
fricatives (e.g. leaves), voiced consonants (e.g. loved), high front vowels (e.g. obviously),
and non-high front vowels (e.g. five and). These were conflated because they patterned
very similarly, they all share voicing, and the model’s predictive power was not impacted
by their conflation. Most importantly, they all share a feature with the preceding [v]
which they motivate—all are voiced. Indeed, following voiced sounds is the (nearly)
neutral reference point, since a great majority of the tokens fall into this category: given
such, the ratio in this category cannot diverge greatly from the overall input percentage.
The next most influential environment, a following voiceless fricative (e.g. comparative
study), also shares a feature with the [v] realization-they are both fricatives. Meanwhile,
the two conflated factors which make up the least likely following environment for [v]ful realizations—the pause and the voiceless consonant—do not share voicing, manner or
place of articulation with [v]—it seems likely that the first two contexts motivate a more
[v]-ful realization precisely because they share features with [v]. Further, this coincides
with research on American Sign Language, which has uncovered a number of features of
the realization which are shared with the preceding and following environment (Bayley et
al. 2002).
Syllable Count
By far the most inexplicable, and also the second most significant factor motivating [v]ful realizations is syllable count, within which multisyllabic words motivate a more [v]ful realization (.56) than monosyllabic words (.34). The multisyllabic factor group
conflates syllables with primary and non-primary stress in multisyllabic words—these
293
two factors had similar factor weights, their conflation did not significantly affect the
predictive power of the model, and they also are distinguished from the third category,
monosyllabic words. In other words, syllable strength was not significant, but syllable
count was. Importantly, monosyllabic words are less likely to undergo lenition, which is
hypothesized to be related to their length-to-meaning ratio: words that are shorter are in
greater ‘danger’ of becoming unidentifiable through processes of deletion and lenition.
This is next explored.
Phonological variation and syllable count
Given phonological patterns and motivations, one might expect the reverse—that
monosyllabic words more strongly motivate a realization that is also the UR, based on
minimal size conditions of two moras or two syllables, depending on the language
(Hammond 1999)—that is, they are less likely to undergo lenition. These conditions can
lead to special behavior, where minimal words (e.g. monosyllabics) do not undergo the
same processes of deletion (which would result in words smaller than the minimum),
through a constraint which disallows this type of mora deletion (Beecher n.d.). However,
when such deletion behavior is found, it does not apply to onset positions, which are not
dominated by the mora, and instead, are thought to be dominated by the syllable (Orhan
Orgun, p.c.). This is important here because position within the word was coded for, and
not found to be significant—that is, monosyllabic words are behaving similarly,
regardless of whether the variable is in the onset or coda, which contradicts the findings
(admittedly, for a different variable) of Bybee (2002), who suggests that realization
quality is directly related to environment, for (t,d) deletion. Thus, we cannot even
294
interpret this as some sort of reverse minimality effect, wherein words which approach or
are at their minimal limit are more likely to be realized as something different from the
UR.
There is another possibility58—that higher syllable count words are more faithful to their
UR than monosyllabic words. However, this is also counterintuitive—smaller words are
likelier candidates for higher faithfulness to their UR, for two reasons. First, formally,
past literature has suggested that monosyllabics are more faithful to their UR (Beecher
n.d.). Second, there is a functional reason: smaller words have less material available for
lexical identification—current phonological analyses are taking heed of the relationship
between phonological UR, levels of production, and their relationship to listener input
and response. With less data to both produce and interpret a word, one should expect a
stronger adherence to the UR than is uncovered in this study.
There is a third possibility: many monosyllabic words are function words (e.g. auxiliary
function words was and were): several studies have shown that function words
(especially high frequency ones) are less likely to be faithful to their underlying UR (Guy
1980; 1991), and, instead have two separate URs, akin to earlier arguments by Zwicky
and Pullum. Thus far, it has been assumed that this always takes the form of more
reduction, e.g. (v) [w], rather than what is found here, the opposite, wherein (v) is more
likely to be realized as [v] over other variants. Guy’s argument (and, historically, Zwicky
58
Indeed, there is a fourth possibility—after the third one, below—, which, given space constraints, cannot
be discussed here, but is tantalizing and deserves further exploration. Words with multiple syllables versus
single syllables are equally can be considered ranked in terms of faithfulness with their UR, while one of
the two is more marked. Uncovering markedness patterns would help to develop this hypothesis.
295
and Pullams’ arguments) thus do not help to explain this situation. And, finally, there is a
fourth issue here: polysyllabic words are more likely to be ‘learned’ (e.g. more formal,
academic or ‘fancy’) words, and used in specific registers and genres (Biber et al. 1999).
However, given that formality was tested for here, as was word frequency, which is
variably realized across registers, the relationship remains murky—all three (syllable
count, formality and word frequency) are significant in predicting variant realization,
while more formal and more frequent (v) words favor [v] realizations.
Word Frequency
Originally, word frequency was coded with a three way distinction between frequent
words (the most frequent 5% of words in the corpus), infrequent words, and compound
words which did not rank in the top 5%, but do have a frequent word within them (e.g.
seventeen). Word frequency proved to be the third most significant predictor of (v),
wherein frequent words motivate a more [v]-ful realization (.58) than infrequent and
compound words (.41). The latter two were conflated because they patterned similarly,
they were both—technically—infrequent, and because the model’s predictive power
(Paolillo 2002) was not significantly impacted by this conflation.
It is important to note that there was not a significant relationship between the number of
syllables in a word and its frequency, unlike the case of (w), which will be discussed
below. Instead, a cross-tabulation of word frequency and morpheme count showed a
healthy distribution across cells. Word frequency thus can be interpreted as motivating a
more [v]-ful realization. As well, given that syntactic category was coded for, but
296
ultimately not a significant factor for (v) realization, we can safely say that the frequency
factor group is actually capturing frequency, and not working as a secondary effect of
syntactic category (V. Chand & Gales m.s.).
Delhi Stay
The first and only significant social factor is Delhi stay, which was the fourth most
significant factor group. This factor group separates speakers who have lived
continuously in Delhi from those whose stay in Delhi has been punctuated through job
postings or historical events like Partition, which impacted almost all of the oldest
generation. People whose stay in Delhi has been punctuated are significantly more [v]-ful
(.56) than those whose stay in Delhi has been continuous (.44).
Formality
Formality was the fifth, and final significant factor group in predicting (v) behavior. High
and medium formality (the Grandfather reading passage and the Pear Story retelling,
respectively), were conflated and compared to informal speech because they pattern
together, both stand in contrast to informal speech as contexts with attention to one’s own
speech, and because the model was not significantly impacted by their conflation. Within
this, we find that formal contexts produce a more [v]-ful pronunciation (.58) than
informal contexts (.48). This pattern was shared across speakers, which suggests that
these speakers do have a shared understanding of the prestige variant as the [v]
realization—in short these speakers are behaving as a community based on their formal
297
context realization behavior (Labov 1966). We can establish, based on all of this, that [v]
is the prestige variant in Delhi.
Taken in conjunction with Delhi stay, these results parallel the (r) findings: the prestige
variant is used more by people who have lived outside Delhi. This suggests that attention
to multiple prestige variants is lower in Delhi than in other areas of India. As discussed in
the previous chapter, Delhi is framed by its inhabitants as lacking ‘class’ and ‘prestige,’ a
city of ‘crooks and con-men.’ These evaluations can now also be correlated with Delhibased language practices, which do not adhere as closely to prestigious pronunciations.
Non-Significant Factor Groups
One social factor which was not significant stands out as initially rather surprising: age.
Unfortunately, this data can only be interpreted through synchronic apparent-time
analysis, given the lack of earlier research on the variable. However, given these
apparent-time results, the lack of significance for the age factor group suggests that the
variation is either stable diachronically or has transitioned faster because of fortition, both
of which require further research to tease apart. As well, given that gender, ethnolinguistic identity, and life-occupational stage (separating students, workers and retirees)
are all non-significant in predicting (v) realization, we can conclude that this variable
does not hold considerable power as a means of socially distinguishing speakers, and may
in fact be a case of stable variation, which further investigation can test.
298
Syntactic category was also not significant in motivating (v) realization. This neither
confirms nor denies past findings, given two important facts. First, some phonological
processes appear to be more sensitive to word frequency, e.g. lenition for (t,d) deletion
and vowel reduction, while others do not, e.g. assimilatory palatalization (Shi et al.
2005). Here, realizations were coded in a way which does not clearly separate processes
of lenition and fortition (specifically, the ‘other consonant’ category encapsulates all
consonants beyond [f], [v] and [w], and it is thus impossible to see if these processes,
when separated, are differently influenced by syntactic category). Second, this variable
stands in contrast to past analyses which are correlated with syntactic function, in that (v)
does not show a healthy distribution across realizations—(v) is realized as [v] ca. 82% of
the time. It may be that there are simply too few non-[v] realizations for syntactic
category to be influential.
Discussion of (v) Results
Overall, linguistic factors make up four of the five factor groups significant for predicting
(v) behavior. This, in conjunction with the large proportion of tokens which are realized
as [v] (82%) suggests that (v) variation is not as common in this community as past
literature and international attention (V. Chand in press)—on IE more broadly as a single
phenomenon—suggest. When it does occur, it is mediated by internal linguistic
constraints, and is not in itself significant for delineating different social identities,
alignments, age or groupings: age was not significant, by life-stage, decade, or ageoccupation, which is unsurprising given the overall percentage of tokens which are
299
realized as their UR—(v) may be relatively moribund, and may have been more variable
in earlier generations or in different demographics.
Importantly, the attention paid to this variable within educational materials, accent
reduction courses, and call center training (V. Chand in press) stands in contrast to its
proportion of appearances within this community (althought it does come closer to the
SIE Pronunciation Sailaja 2009 envisions and/or describes)59. However, having just
explained the non [v] realizations as uncommon, they do make up almost 20% of the
tokens, a percentage which is easily above the level of perception for listeners—the
unexpected can be quite salient. Thus, even though it is not as ‘common’ as it is
ideologically portrayed to be, its presence is in itself, under 20% of the time, may
motivate these ideologies.
As well, as I have been at pains to point out in the previous chapters (e.g. Chapters 2-4),
India is a highly heterogeneous setting, and the Indians studied here, as upper-middle
class native English speakers, may not represent or even resemble pronunciation norms in
other regions, other linguistic communities or those of non-native English speakers as
more ‘average Indian’s vis-à-vis (v) pronunciation (Sailaja 2009: 20). Thus, the question
of how relevant—or, indeed, necessary—training that focuses on (v) pronunciation really
is cannot be answered simply by looking at this population and (v)60, although it does
59
Unfortunately, these results are highly problematic for Sailaja’s (2009) documentation of Standard Indian
English Pronunciation (SIEP)—if this set of speakers, as native IE speakers from upper-middle class
backgrounds, does not match her descriptions, this begs the question of who actually does, and whether
SIEP exists.
60
Ideally, this should be tested through a conjunction of perceptual and attitudinal tests, akin to the
methods used in Chand (2006).
300
offer a data point for later comparative analysis across Indian populations. I now turn to
the analysis of (w), to explore the other side of the coin, after which I will explore how
the cumulative results can be reconciled with the historical evidence and respective
theories for the merger.
(w)
Despite the widespread disagreement in past research concerning the acoustic realization
of the /w/ phoneme in IE, these quantitative results demonstrate that (w) is realized as a
voiced labiodental approximant a majority of the time in this population: 89.5% of tokens
are realized as a labial approximant (Table 6.8). Although clearly in the minority, (w) is
also realized fully as [v] (2.5%), partially as a brief [v] which transitions into an
approximant (5.1%), never as the voiceless labiodental fricative [f] (0.0%), as another
consonant (0.8%), or as a vowel (1.2%). By comparing (v) and (w) results in a 2-way
Chi-Squared test, we see that speakers are significantly more likely to hit the target with
(w) than (v) (Chi Square = 47.546, p < .0001) and that in comparison to the above
analysis of (v) realizations, (w) realized as [w] is more frequent than (v) realized as [v].
Table 6.8. Overall distribution of (w)
[w]
%
89.5
N
2064
[v]
%
2.5
N
57
mixed
[v] + [w]
%
N
5.7
131
Total N
[f]
%
0.0
2298
N
0
Other
consonant
%
N
0.8 19
Other
vowel
%
N
1.2
27
This data can only be interpreted through apparent-time synchronic analysis, because the
single previous quantitative study, Sahgal & Agnihotri (1988), only distinguished
between [w] and [ʋ], although they suggested that [v] realizations were more likely in
301
formal contexts. Their data is not comparable for three reasons: I could not uncover any
reliable acoustic or auditory means of distinguishing the two labial approximant
realizations—nor do they explain how they did so—, they do not separate out [v]
realizations, and they do not offer token counts for their realizations, which would permit
statistical comparison.
While seven realizations were coded for, and all seven were found within the (w) data,
VARBRUL only handles two dependent realizations for binomial logistic regression. As
is clear from Table 6.8, the distribution of these variants is by no means equal—the [w]
realization makes up almost 90% of the 2298 tokens, while at the other end, the [f]
variant, for example, has zero tokens. Important here, infrequent variants are not well
suited for quantitative analysis (e.g., Bayley et al. 2002), and this has been accounted for
within this study through three practices. First, non-[w] realizations are conflated in the
rest of this analysis, second, variants which represent <1% of the tokens are excluded
from this binomial analysis (e.g. tokens with the [f] realization or the other consonant
realization), and third, some factors are conflated, if both statistical and theoretical
motivations are present—these are discussed below. Linguistic and social factors which
correlate with realization quality are next examined.
Social and Linguistic Constraints on (w)
After the overlapping social and linguistic constraints were tested in various
combinations, one social factor and three linguistic factors proved significant in
motivating a [w] realization.
302
Age/Occupation
Further affirming the results of the accompanying rhotic analysis (Chapter 5), which
found that a social factor—gender—was the primary motivation for (r) realization
quality, a social factor also has the most powerful relationship in this analysis of (w) in
IE. The age/occupation factor group was the most significant predictor of (w) realization,
and demonstrates a linear progression wherein the youngest generation has the most [w]ful pronunciation (.78), their parents, the worker generation, has the next most [w]-ful
pronunciation (.60), and the retired generation, the students’ grandparents, has the least
[w]-ful pronunciation (.33) (Table 6.9).
Table 6.9. Social and linguistic factors favoring [w] realization
Factors
Age/Occupation
Student
Working
Retired
Location + Preceding Environment
Onset cluster preceding voiceless
consonant
Syllable initial with preceding voiceless
consonant*
Syllable initial with preceding fricative*
Syllable initial with preceding voiced
non-fricative (consonant or vowel)**
Syllable initial with preceding pause*
Syllable Weight
Monosyllabic words
Multisyllabic words*
Following Environment
Following non high, non front vowel
Following high front vowel
Factor
Weight
N
% realized
as [v]
.78
.60
.33
354
735
1005
97.5
93.9
84.7
.89
111
98.2
.64
291
94.8
.48
.45
207
1250
90.3
89.0
.35
235
85.5
.55
.39
1499
595
91.1
87.4
.52
1668
90.7
.42
426
87.6
Total N
2094***
* Two factors are conflated; ** Three+ factors are conflated; ***knockouts were excluded which resulted
in a smaller N; p =.020; Input value = 0.926, Log likelihood = -617.951; Chi Square/cell = 1.3574
Age was coded in three different ways—by decade, by historical era, and by generation.
The former two categories were perhaps too nuanced to capture variation which is
303
transitioning into moribundity, while the three way distinction of the age/occupation
factor group did successfully distinguish three groups by generation. [w] is realized a
majority of the time, and this is increasing in the younger generations. The importance of
this factor group will be discussed further below, after we contextualize the motivations
for (w) realization by exploring the secondary significant factor groups, all of which are
linguistic factors.
Location + Preceding Environment
[w]-realization distinguished five environments conflating various word positions and
preceding sounds to make up the second most significant factor group. The variable was
not common in word final position and all but one onset combination. Several factor
groups were thus excluded due to low token count positions (specifically, the following
were eliminated from analysis: onset position with preceding fricative, onset position
with preceding voiced consonant and word final with preceding non-high front vowel).
As well, several factors behaved similarly enough for conflation without affecting the
model’s predictive power, and were conflated, given linguistic motivation. For example,
word and syllable initial positions behaved identically, are both the initial position within
a syllable, and were thus conflated. Preceeding voiced consonants and vowels (sans
fricatives) provoked near identical (w) behavior in word and syllable initial position, and
share voicing quality, and were conflated.
Overall, preceding voiceless sounds motivated more [w]-ful realizations than preceding
fricatives, voiced sounds or pauses, while fricatives provoked less [w]-ful realizations
304
than voiceless consonants, and a more [w]-ful realization than voiced consonants and
pauses—in other words, fricatives split other voiced and voiceless sounds. Specifically,
onset clusters with preceding voiceless consonants (e.g. quick) are separated as provoking
the most [w]-ful realization (.89). Decreasing from there, syllable initial (word internal
and word initial) positions have a hierarchy, wherein preceding voiceless consonants
provoke more [w]-ful realizations (.64) than preceding fricatives (.48), which, in turn are
more [w]-ful than preceding voiced non fricative consonants and vowels (.45). The least
likely environment for [w]-ful realizations (still with enough tokens to study
quantitatively) is the syllable initial position with a preceding pause (.35).
Given that almost all tokens were realized in syllable initial and onset positions, statistical
analysis comparing syllable initial position with non-strengthening environments beyond
onset clusters, e.g. coda cluster, word and syllable final other environments, was
impossible. However, this data is consonant with processes of strengthening in word and
syllable initial positions mentioned above. The onset cluster environment is significantly
more likely to motivate a [w] realization than syllable initial (word internal or word
initial) position.
Syllable Count
The third most significant motivator of [w]-ful realizations separated monosyllabic words
from multisyllabic words, wherein monosyllabic words motivate more [w]-ful
realizations (.55) than multisyllabic words (.39). This factor group was also able to test
syllable stress, by separating monosyllabic words from primary and secondarily stressed
305
syllables in multisyllabic words, e.g. one might expect monosyllabic tokens and stressed
syllables in multisyllabic tokens to pattern against multisyllabic tokens with non-primary
stress on the target syllable. Conflating multisyllabic words into a single category did not
affect the predictive power of the model, and they patterned similarly, so I conflated these
two groups, and the factor group is reinterpreted in lieu of this. However, one concern
arose: Is the monosyllabic category simply capturing and separating word frequency
differences or function/content differences, given that many frequent /w/ words are
monosyllabic and functional?
Word Frequency
As discussed above, an additional factor group was created to deal with the first
possibility addressed above: word frequency separated tokens as frequent, infrequent, and
compounds. The top 5% of all words (not just words with /w/) within the larger corpus
were treated as frequent words, while two other factors were created: this second factor
captures less-frequent words which incorporate a frequent word (e.g. the lexical item way
was in the top 5% of words used, while anyway was not)—these were labeled as
compounds, while the third factor was reserved for all remaining infrequent words. The
word frequency factor group did not prove significant in the final analysis, and thus, we
can say that the syllable type factor group is actually capturing that monosyllabic words
(e.g. quick) are more likely realized as [w] than multisyllabic words (e.g. quicker),
regardless of where stress is located.
306
Function vs. Content Words
Testing the second possibility, tokens were separated into two groups: function and
content words, detailed above. However, this factor group ultimately was also not
significant in predicting realization quality. Thus, the syllable count factor group, as it has
been reformulated, must be interpreted as directly capturing variation vis-à-vis syllable
count, and not as secondarily capturing function/content distinctions.
Following Environment
Past cross-linguistic research has suggested that palatalization provokes fricativization in
the preceding sound (Stephens 1988: 426). This data shows that a following high front
vowel does motivate less [w]-like realizations, however, it is not clear whether the
following high front vowel environment truly represents palatalization—a process which
encourages high front vowels—, or simply captures underlying high front vowels. The IE
vowel space has not been charted, and it is thus not clear if the high front vowels found
here are underlyingly present, or are themselves the surface representation of a process of
palatalization. Based on this data, I cannot comment on how palatalization as a process
affects (w) realization, but this data does demonstrate that following non high front
vowels do significantly motivate more [w]-ful realizations (.52) than following high front
vowels (.42).
Thus, following high front vowels, whether they are the result of a palatalization process
which also affects the preceding (w) or whether they are underlyingly present, do
influence (w) away from an approximant realization. However, they are not as influential
307
as the higher ranked preceding environment and word location and total syllable count.
Future research on the relative strength of preceding versus following phonetic
environments would be profitable in that it may help explain why preceding environment
is more influential, within this data, towards predicting (w) realization.
Discussion of (w) Results
The number of significant factors is necessarily smaller with so many tokens
categorically realized—there is less room for conscious, indexical, stylistic variation
(e.g., Levon 2009) within such a context. However, the features which were significant
merit discussion here.
Social Factor Groups
Several factor groups were not significant in predicting (w) realization in this study, and
their lack of significance is important in understanding IE (w) realization diachronically.
Neither ethno-linguistic identity nor Delhi stay as permanent versus punctuated was
significant in predicting (w) realization. This suggests that the evolution of (w) towards a
categorical [w] realization is not confined to narrower social groupings or areas of
India—this process may be a pan-Indian change, and further research in other language
groups would help in assessing this.
As well, gender was not significant—men and women are behaving the same. Given the
collective three significant linguistic factors are phonological, this first suggests that (w)
realization is largely a phonological process. Second, given that the age/occupation factor
308
was significant, and demonstrates that speakers are moving in a linear direction towards a
more w-ful realization, I can say that (w) is diachronically moving towards a more
categorical [w] realization. Gender may have been significant 20+ years ago (e.g. S&A),
but currently, it appears that variable (w) realizations are almost entirely absent,
encompassing only ca. 10% of the tokens. Social factors beyond age are not significant in
predicting variable behavior, while several linguistic factors do still influence variation.
This process, whereby social factors beyond age fail to significantly influence variation,
may be common in contexts of moribund variables, and merits further exploration with
other moribund variables. This theory has larger impacts—it suggests that social
motivations require more healthy variation, while almost categorical variation may be
more commonly influenced almost entirely by linguistic factors.
Linguistic Factor Groups
Of the eight hypothesized internal linguistic influences on (w) realization, two were
conflated into a single factor group—preceding environment and word location—while
two others were not significant. One non-significant factor group is word frequency,
which was discussed above. Next is formality: speakers do not significantly change their
pronunciation across the three contexts. This lends credence to the suggestion that (w)
variation is moribund—speakers are not aware of the variation, consciously or
unconsciously—and this is reflected in their static behavior across different formality
contexts. This is also in line with the globalized attention to the ‘merger,’ which focuses
on alternate realizations of (v), not (w) (V. Chand in press). Further, given that word
309
frequency was also not significant, this suggests that the variable, if ever influenced by
word frequency, is no longer, and it has transitioned past this point.
However, this leads to an interesting conundrum—if formality doesn’t influence variation
and the only social variable is chronologically linear, then are we really looking at
sociolinguistic variation within a community or are we instead looking at a sound change
which is either independent of social influence and simply receding or only apparent by
comparing across demographics? The apparent-time analysis demonstrated that social
factors beyond age are not significant in this group, and further research with additional
populations is necessary to explore the whether this variable separates different IE
populations.
Comparison of (v) and (w)
Returning to our original question on the nature and reality of the much discussed (v/w)
merger, we have explored the results of variationist analysis upon both (v) and (w), and
can now compare them.
The reality of this ‘merger’
Clearly, the overall statistics from each variable demonstrate that there is no widespread
merger. Briefly, I would like to explore how often each variable was realized as the
other—that is, the merged behavior of (v) realized as [w] and/or of (w) as [v], as is
hypothesized in the literature.
310
Before tokens were excluded because they fell within factors which represented less than
1% of the tokens or were knockouts, (v) was realized as [v] 82% of the time, while (w)
was realized as [w] 89% of the time. Meanwhile, ‘merged’ behavior manifested in the
following way: (v) was realized as [w] or as mixed (which involved a prolonged
approximant after a brief fricative) 5.9% of the time (93 tokens out of 1575 total), and
(w) was realized as [v] or as mixed (which involved frication, albeit briefly) 8.2% of the
time (188 tokens out of 2773 total). However, examining the other realizations
demonstrates an interesting difference: the mixed [vw] realization is the second most
common realization for (w), following [w], while the [f] realization supersedes both the
[w] and mixed [vw] realizations for (v) (Table 6.10).
Table 6.10. A comparison of the distribution of realizations for both variables, with mixed realizations
respectively conflated as non-dominant factors (i.e. for (w), the [v] category includes merged realizations,
and for (v), the [w] category includes merged realizations)
[v]
(v)
(w)
%
82.4
8.2
N
1566
188
[w]
%
N
6.0 114
89.5 2064
[f]
%
6.7
0.0
N
127
0
consonant
%
N
3.5
66
0.8 19
vowel
%
N
0.7 14
1.2 27
null
%
N
0.7
13
0.0
0
Merged behavior, thus, was not often apparent—it was 6% of the time for (v) and 8.2%
of the time for (w)—, and (v) realization as [f] and [w] are not statistically significant,
through 2-way chi-square tests. However, the sets of significantly influential factors and
the ordering of these significant factors for each variable collectively suggest that (v) and
(w) are not working in collusion, but instead, are undergoing distinct processes with
distinct motivations. Further, given that both variables were coded based on the same
principles, and both used a combination of acoustic and aural analysis, it is clear that this
is not a near-merger situation (c.f. Labov et al. 1991), wherein two sounds are
consistently perceived as the same, but have different acoustic signatures.
311
Conclusion
Returning to the historical English and substratum contact motivations presented above,
these speakers do maintain a phonemic distinction between (v) and (w) a significant
majority of the time. When they produce another realization, a merged pronunciation—
with /v/ realized as [w] or /w/ as [v]—is only one of several possibilities. The strongest
predictor of (w) realization is the age/occupation factor group, which is linearly linked to
age: younger speakers are significantly more [w]-ful than older speakers. This is a strong
indicator that earlier generations had a more variable pronunciation of (w), yet the
number of non-target tokens is too small to perform a separate test on. In short, we can
say that IE used to have a more variable (w) pronunciation, but we can not say what the
nature of that pronunciation was. Without a larger token count it is also impossible to say
whether the seven realizations uncovered here were all equally present in a historically
earlier setting, or whether some or all of them are secondary changes as IE moves
towards increasing homogeneity in (w) pronunciation.
Meanwhile, (v)—which, based on percentages is more variable—does not correlate to
age, and hence, to diachronic change. Instead, linguistic factors most strongly motivate
(v) realizations, and Delhi is more variable than other Indian locales, given the difference
uncovered between formality, non-permanent Delhiites and a more [v]-ful realization.
This process, if indeed, it is a change-in-progress, may have moved ‘faster’ than (w)
change over time, and have transitioned from socially variable to only linguistically
variable. Moribundity may not equate in a linear fashion with increasing categorical (v)
behavior—it may be that variables lose their social diagnostic predictive power at
312
different stages of categoricity. More research is needed in this vein to answer this
question.
Thus, neither of the historical motivations—the Substratum Contact Hypothesis and the
England Hypothesis—are a significant impact on modern day IE (v/w) as spoken by
native IE/Hindi bilinguals in Delhi, nor do we see much residue of variability in their
current realizations. Importantly, the current behavior also excludes continued contact
with non-native English speakers within this Indian community—who may be more
influenced by their native language phonologies—because, as detailed above, such would
also likely motivate a higher percentage of merged realizations than is found here.
Unfortunately, we cannot determine, from this data, whether either of these hypothesized
historical processes impacted earlier versions of IE, for three reasons: first, age was not a
significant motivation for (v) realization, second there is no earlier comparable data on
this community to compare with these results, and third, the number of merged tokens
uncovered here is too small to conduct a separate variationist analysis on. Thus, while IE
may at one point have had a merger of some sort, one which is now almost entirely gone,
we cannot test this hypothesis. Instead, what we see, from these results, is that internal
linguistic constraints are the dominant influence on (v) and (w) realization—this, in
conjunction with the rhotic results discussed in the last chapter, offers significant
evidence that IE is a natively spoken dialect with its own internal grammar.
313
CHAPTER 7
CONCLUSIONS
Several sociolinguistic aspects of these speakers’ lives, as upper-middle class native IE
speakers living in New Delhi, have been woven together here to create a tapestry whose
threads are these overlapping, conflicting, empowering and authentic attitudes, ideologies
and language practices. As the weaver (if readers will permit me to extend this
metaphor), I have organized these threads to flow from thematic concept to concept, but I
have not re-dyed nor consciously excluded any threads which were the ‘wrong’ color.
Instead, my interpretation is that I have been handed a set of apparently distinct threads—
these IE speakers voices—which, by drawing on theories, methods and findings from a
range of disciplines, I have organized and woven together into an overarching patterned
tapestry. By drawing so heavily on my speakers’ voices, even when these mar otherwise
‘clean’ findings, I have attempted to stay true to their intent, their ideologies, and their
sociolinguistic lives.
Each of the data and analysis chapters has offered conclusions, which I will not reiterate
here. Instead, I will return to the original research questions in light of the findings
presented over this dissertation, in order to address what I have—and haven’t—been able
314
to examine here, and the results and their broader implications of the questions I have
been able to explore in this data through these quantitative and qualitative methods. My
original research questions were:
1.
What are speaker attitudes towards and understandings of IE as a variety of
English, and how does their IE awareness correlate with their actual linguistic
repertoire?
2. What social factors, alliances and categorizations do these IE speakers bring up
when discussing IE, and how are these related to IE with respect to the ideology
of standard languages with native speakers?
3. How do speaker linguistic plans, domains of language use, and the perceived
function of languages within their multilingual linguistic repertoire interact both
with each other and with IE cross-generationally and across genders?
4. How do previously cited features of IE manifest in actual oral conversation?
Are they variable, what motivates this variation, and can they themselves be
quantitatively modeled?
5. How is IE structure changing diachronically (or, contrarily, what sort of age
grading is apparent)? Are globalization and increased access to RP and
American English (AE) media influencing IE towards assimilation with one of
the two, or towards increased distance from these dominant English varieties?
315
Research Question 1
Chapter 4 has demonstrated that speaker attitudes and interpretations of IE are varied,
multifaceted and at times even internally conflicted vis-à-vis global ideologies about IE
intelligibility, ‘correctness,’ and native speakerhood status, while the ‘fake accent’
discourse offer a lens for interpreting and understanding the authenticity, value, and
legitimacy IE offers in this local context. Addressing the second half of this question, IE
awareness correlates in a positive manner with to linguistic repertoire. The two IE
speakers (f39VD and f25GS) who explicitly disavow IE as a separate English dialect and
emphatically condemn IE structure as ‘wrong’ also evidenced different patterns for (r)
and (v/w) from the rest of the sample, and can be interpreted as socially and linguistically
distinct, while the majority of the sample are more positive about their English, and
behave as a community in terms of ‘orderly heterogeneity’ (Weinreich et al. 1968) for
three structural variables.
Research Question 2
As Chapters 3 and 4 demonstrate, several social factors are evoked when discussing IE
that relate to educational and linguistic background, socioeconomic status, fluency,
urbanity, and bilingual and biliterate proficiency (e.g. vernacular students were framed as
the ‘other,’ and are also interpreted as from lower socioeconomic strata, rural areas, less
fluent in English, and holding more advanced literacy skills in Hindi than this
population).
316
Further, there are age-related differences in how these speakers interpret their own
nativity—and, fundamentally, the potential nativity of IE, e.g. is it possible to have IE as
a native tongue? IE is explicitly disavowed as a native dialect by some, while standard
and native language ideologies are evoked, challenged, and dismissed by others. IE as a
standard is valued for its neutrality, simplicity and international intelligibility by some,
while it is compared to ‘English’ (e.g. ‘Language’ with a capital L, Blommaert 2003) and
found lacking by others. Some speakers refuse to even discuss IE as a concept, explaining
that there are more differences than similarities within the population of English speakers
in India: several ‘types’ of IE are distinguished as impacted by regional and ethnolinguistic background.
Collectively, these speak to and negotiate global language ideologies: by discussing how
‘other’ Indian populations speak English in various ways that are further from global
norms, how international varieties are also ‘dialects’ which are variably intelligible in the
local context, and by offering their own speech as a ‘neutral’ comparison to
internationally prestigious varieties, these speakers re-frame their IE as having economic,
social, and lingua franca value on the global stage. Native speakerhood holds less import
in this setting as bestowing social authority, which is instead determined through
intelligibility, communicative competence and adherence to local practices.
Correlations with Research Question 4
Importantly, as Chapters 5 and 6 demonstrate, the social alliances which speakers evoke
as locally salient positively correlate with and stratify speakers in terms of their language
317
practices. For example, ethno-linguistic identity and local interpretations of life-stage as
separating three generations into students, workers and retirees were both significant
factors for understanding (r).
Research Question 3
These speakers’ linguistic plans, domains of use, and their perceptions of the functions of
IE and Hindi are intertwined, and vary across generations, and genders. Younger
speakers with less professed Hindi fluency and literacy have vaguer long term
transmission plans, and place a heavier emphasis on IE as part of their urban uppermiddle class Indian identity than their parents and grandparents do. Older speakers use
Hindi in more domains, have more advanced Hindi literacy skills, and have passed—to
varying degrees of success—Hindi on to their children. These older speakers see Hindi as
having more functions—though this is also variable, within this community—, while
younger speakers see Hindi as restricted to the marketplace and for communicating with
domestic servants. For youth, Hinglish is encroaching on intimate domains earlier
reserved for Hindi. Youth also have more conscious monitoring skills, as their almost
entirely English monolingual interviews, in comparison to later generations’ more Hindi
and Hinglish infused interviews, demonstrate.
While females typically assume the burden of maintaining home languages and are
carriers for cultural transmission and continuity (Zentella 1987; Giles & Coupland 1991),
Hindi transmission has a more complicated relationship to gender. Younger women are
slightly more focused towards passing an Indian language on to their children, as part of
318
their cultural heritage, while their male counterparts are more concerned that their
children show ‘respect,’ regardless of their heritage language fluency—these are both
hypothetical scenarios, as these youth haven’t had children yet. Older generations of men
and women are more unified in their opinions that indigenous language fluency is an
important part of ‘being Indian’ and reflects one’s cultural familiarity. However, across
generations and genders, the social and economic value of English fluency overshadows
that of indigenous languages, and IE’s increasing value as an aspect of modern Indian
identity also factors in to language maintenance plans. While in the older generation the
men were and are all using English at work, their wives have brought English into the
home, and raised their children to be English speakers.
Research Question 4
While academic literature on IE has offered many interpretations of IE structure,
Chapters 5 and 6 stand apart as the first quantitative analysis of natively spoken IE, and
offer a far more complicated explanation of IE structure and variation than past IE
literature. Drawing on over 50 hours of sociolinguistic interviews, quantitative analysis of
systematic phonetic and phonological variation for three variables—(r), (v) and (w)—in
conjunction with the results of qualitative analysis of local social alignments has exposed
significant links between language practices and social identity. Gender, age, ethnolinguistic background, and domestic mobility within India have each proven to be
foundational elements of individual sociolinguistic identity and practices, and collectively
are significant factors for understanding systematic phonological variation through
319
synchronic and diachronic analyses. The variables examined here were deliberately
chosen and are discussed next.
(r)
(r)-pronunciation was modeled in Chapter 5 for four reasons: 1) past IE surveys have
suggested distinctive IE behavior (null, approximant and trill realization); 2) two
dominant international English standards have very different rhotic behavior (dominantly
non-rhotic UK vs. dominantly rhotic US); 3) (r)-pronunciation has consistent social links
across Englishes; 4) it is often involved in larger processes of sound change. In this postcolonial, multilingual, alternative marketplace, social factors proved even more
significant and complicated than the traditional age/gender/class based distinctions
typically used in English monolingual dominant marketplace studies. This demonstrates
that variationist techniques can be successfully applied to alternative multilingual
marketplaces, but also challenges the assumption that oft-considered ‘basic’ social
factors, widely used in variationist studies, are adequate in accounting for alternative or
multilingual contexts. Further, this research presents a serious counterargument to the
underlying theory in the variationist model—that social factors are idealized as
independent—given the more rigid social structures found in the Indian context.
(v/w)
The v/w merger, meanwhile, has become iconic as a feature of ‘non-standard’ (and
specifically Indian) English, in ways which make it a target of social sanctions in the
struggle over who gets to decide what counts as ‘good’ English. However, this analysis
320
revealed that younger IE speakers are aware of and proud of this IE feature—the local
and international values of IE thus are highly divergent. It is becoming further
stigmatized internationally, and yet is an in-group identity marker locally. This
divergence is made more complicated by comparing local and international salience with
the results of quantitative modeling of (v) and (w), Chapter 6. While both are infrequently
merged across age groups, their hypothesized variability is nonetheless used
internationally to index India (e.g., in the speech of Apu, from The Simpsons V. Chand in
press), and features prominently in youth local awareness. Theoretically, these results
offer evidence that these local indexical features are not engaged in larger processes of
language change.
The variability of variables
Considering (r) and (v/w) in conjunction, this dissertation is in the relatively unique
position of having explored different types of variation. (r) is a ‘healthy’ variable, with a
lot of variability which is linked to both social and linguistic factors, and is likely to be
changing over time, through apparent-time extrapolations across the three generations.
Meanwhile (v) and (w) have far less variability—that is, they are realized as their
underlying form ca. 82% and 89% of the time, respectively. These moribund variables
demonstrate what may be a larger phenomenon, that when the range of variation is
narrower, linguistic factors are more influential to realization quality than social factors.
Both types of variables offer insights about this sociolinguistic setting and IE structure,
while they also both contribute to understanding and theorizing language change and
variation.
321
Research Question 5
IE appears, through apparent-time analysis across three age groups, to be evolving, as all
languages do. This supports my original hypothesis that IE is a natively spoken and ‘real’
dialect with internal structure and trajectories that can be explored and explained without
reference to contact influences and India’s massive ESL population. Chapter 5 explored
diachronic changes in IE rhoticity through comparison with an earlier study, however, the
conclusions were tentative, given methodological, population and coding differences, and
missing token totals. Further research on this population will illuminate whether the
projected apparent time findings are realized through diachronic analysis.
Importantly, Chapter 5 demonstrates a pattern of rhoticity which is not moving towards
globally prestigious AE or RP—increased access to AE, and the renewed influence of RP
cannot be interpreted as influencing IE towards assimilation and dialect leveling, and
instead, IE—in terms of rhoticity—is becoming increasingly distinct, as evidenced in the
youngest generation of speakers. However, this process is not linear—the results
demonstrated a curvilinear (V-shaped) peak in practices which separated the worker
generation from their parents and their children. This pattern can be explained by drawing
on socio-political, educational and ideological periods in India’s development, and also
through the traditional joint family system, where children are raised by their
grandparents.
322
Chapter 6 shows a different process—IE (v) and (w) may have been more variable in past
generations, but at this point, they come closer to AE and RP pronunciation norms. This
may have to do with focused international attention on this ‘merger’ as ‘wrong.’
However, it may not—earlier generations of IE speakers may never have been any more
variable than they are now, and instead, more variable realizations may be a Indian
feature of other regions, other fluencies, or other socio-economically distinct groups.
Earlier—or future—data on this population vis-à-vis these two variables will answer this
question.
Broader Implications
Broadly, examining alternative, multilingual settings raises important theoretical
questions about the application of variationist methods in new contexts, and the
relationship between language change, shifts in linguistic ideologies, sociolinguistic
identity, and the long term impact of linguistic colonialism. IE, as a case study, verifies
that a distinct dialect of English is natively spoken in India, while it also offers clues
towards structural and sociolinguistic patterns in other post-colonial English varieties,
given shared historical and linguistic patterns of inequality, while it also, through the
qualitative methods used here, suggests that they could have very different trajectories
given different global and local ideologies, socio-political events, educational agendas,
and social alignments. Local interpretations of identity, authenticity, and legitimacy and
how these dialogically interact with global ideologies and assessments are foundational to
interpreting these results, and must be a part of future studies on language development
and change.
323
APPENDIX A
GRANDFATHER PASSAGE
You wish to know all about my grandfather. Well, he is nearly ninety-three years old; yet
he still thinks as swiftly as ever. He dresses himself in an old black frock coat, usually
several buttons missing. A long beard clings to his chin, giving those who observe him a
pronounced feeling of the utmost respect. When he speaks his voice is just a bit cracked
and quivers a trifle. Twice each day he plays skillfully and with zest upon our small
organ. Except in the winter when the snow or ice prevents, he slowly takes a short walk
in the open air each day. We have often urged him to walk more and smoke less, but he
always answers, “Banana oil!” Grandfather likes to be modern in his language.
324
APPENDIX B
IE FEATURE CHART
Feature
Example
Reference
All bracketed labels (e.g. [S])
attached to references refer to
divergent opinions coindexed in the parallel
Feature column—labels are
not standardized across
different rows.
These features are arrived
at from multiple and variable
sources; if unlabeled,
references is a descriptive
work with no cited source.
Otherwise, data source is
provided after divergent
opinions, i.e. (Name Year:
Page [OpinionLabel],
DataSource) and labeled with
the following conventions:
Speaker Background
AIE= Anglo-Indian
informants (small community
researched by Coelho)
BIR= Butler IE speakers
(servants in South India,
researched by Hosali)
US-IE= IE speakers living in
US
L2= speakers labeled as
English learning
Language Form
ES= elicited speech,
NS= natural speech,
NLJ=Native listener
judgments,
WE= written essays (school)
WF= written fiction
WG= written English usage
guides
WK= written: Kolhapur
Corpus of Indian English
WP= written print
(newspapers, government
writings)
WU= written, genre
unspecified
Analysis
Q= quantitative or corpusbased analysis
325
PHONETIC AND PHONOLOGICAL FEATURES
Feature
reduced vowel inventory
Example
VOWELS
11-12 pure vowels, 6-9
diphthongs
Tense vowels not
consistently long (unlike AE,
RP)
Open vowel in place of RP
/ɔ://o:/
/ɔ:/, negatively correlated
four, bored, course
with prestige use in formal
situations
Monopthongs for RP
diphthongs
/ey/ /e:/
/əu/ /o:/
Word initial front vowels
eight /je:t/
preceded by /j/ , back vowels own /wo:n/
preceded by /w/ (South
India)
CONSONANTS
Consonant inventory size
22 phonemes
22-24 phonemes
Voiceless stops /p, t, k/
unaspirated in stressed word
initial position, only
occasionally aspirated in
word-final position [WF]
Reference
(Nihalani et al. 1979:
210; Bansal 1990;
Agnihotri 1994: 237;
Trudgill & Hannah
2002)
(Bansal 1990)
(Sahgal & Agnihotri
1987, cited in
Agnihotri 1994;
Agnihotri 1994; Coelho
1997, AIE NS Q)
(Nihalani et al. 1979:
211; Coelho 1997, AIE
NS Q; Trudgill &
Hannah 2002: 130;
Sharma 2005: 210, USIE L2 NS Q)
(Trudgill & Hannah
2002: 130)
(Bansal 1990;
Agnihotri 1994)
(Bansal 1990;
Agnihotri 1994; Coelho
1997 [WF], AIE NS Q;
Trudgill & Hannah
2002: 130; Sharma
2005: 208, US-IE L2
NS Q)
326
Feature
Example
Alveolar stop /t, d/
/t/, /d/ [ʈ], [ɖ] / #___
retroflexion.
Debate regarding whether
retroflexion is socially
conditioned (SC),
completely replacing
alveolar (RA), in free
variation [FV],
environmentally conditioned
[EC], or regionally variable
[R]; Trudgill & Hannah
(2002) argue that /s, l, z/ also
retroflexed.
/ʧ, ʤ/ lack lip rounding
typical in RP; /ʃ, ʓ/ also
included by (Nihalani et al.
1979)
/ ʧ, ʤ, ʃ, ʓ/ pronounced with
blade of tongue making
contact, tongue tip down
lack of /ʓ/, merger with /ʃ /
/l/ is always clear (no dark l)
Reference
(Nihalani et al. 1979
[R]; Kachru 1982: 359
[RA]; Sahgal &
Agnihotri 1987 [SC],
cited in Agnihotri
1994; Bansal 1990
[SC]; Agnihotri 1994
[SC]; Coelho 1997
[FV] AIE NS Q;
Baldridge 2002 [RA]
US-IE NS; Trudgill &
Hannah 2002: 130
[RA]; Sharma 2005:
210 [SC] US-IE L2 NS
Q; Sailaja 2009: 21
[EC], [RV] and [SC])
(Nihalani et al. 1979:
212; Bansal 1990;
Agnihotri 1994)
(Nihalani et al. 1979:
212)
(Nihalani et al. 1979:
212; but see Sailaja
2009: 22, who directly
disagrees with this)
(Kachru 1982: 359
WU; Bansal 1990;
Agnihotri 1994; Coelho
1997AIE NS Q;
Sharma 2005: 208, USIE L2 NS Q)
327
Feature
IE /ʋ/ subsumes RP /w, v/.
Debate regarding whether
this is stable [S], regionally
variable [R],
environmentally conditioned
[EC], free variation [FV], or
socially variable [SV] with
‘non-standard’ speakers
merging and SIEP speakers
maintaining a phonemic
distinction, wherein /v/ has
less frication than RP, but
still has frication
Geminate consonant
pronunciation in words like
upper and rubber
Consonant cluster
simplification word-initially
via initial /i:/ insertion,
resyllabification
OR
via [ə] insertion within CC
(for Punjabi speakers, within
CC’s where first C is /s/ or
second C is /l/)
word-final CCC reduction:
medial or final C dropped
Consonant cluster
simplification word-finally
(/t,d/ deletion) conditioned
by environment with gender,
age and style correlation:
simplified in women, youth
and spoken style61
61
Example
EC:
/w/ [v] / C__
[w] / V[+back]__
[ʋ] / elsewhere
sp-, st-, sk- i:
sp is.p…
Reference
(Nihalani et al. 1979
[S]; Sahgal &
Agnihotri 1987 [S];
Bansal 1990 [EC];
Agnihotri 1994 [S];
Coelho 1997 [FV] AIE
NS Q; Trudgill &
Hannah 2002: 130 [R];
Sharma 2005: 218 [FV]
US-IE L2 NS Q;
Sailaja 2009: 20 [SV])
(Krishnamurti 1978,
cited in Agnihotri
1994)
(Kachru 1982: 359;
Bansal 1990; Agnihotri
1994; Trudgill &
Hannah 2002: 130,
limited to north India)
school [səku:l]
glass [gəla:s]
(Bansal 1990)
/-kts, -sts, -skt]
(Bansal 1990)
/t,d/ Ø
(Kahn 1991)
These social correlations aren’t evident in the current study, perhaps because her population was
nonnative Muslim IE speakers.
328
Feature
r-pronunciation—present
[P], environmentally
conditioned [EC] or socially
variable [SV], diagnostic for
age and gender: r-lessness
for young females, as well
for high school’s prestige
(high) and years of Englishonly instruction in HS
Example
/ɹ / [ɹ] / __V
Ø / elsewhere
OR
/ɹ / Ø / ___#C and ___C
62
(Sailaja 2009:19 [EC])
[ɹ] / elsewhere
OR
/ɹ / Ø e.g., feather
[fɛdɜ]
/r/ flapping and possible
retroflexion
of dominantly pronounced
with [f]
voiced interdental fricative,
typically interdental stop,
instead manifests alveolar
(and aspirated), or as a
dentalized voiced alveolar
stop
stop/fricative merger [M],
free variation between dental
and alveolar/retroflex [FV],
or fricatives manifest as
dentalized alveolar
consonants [D] distinct from
alveolar C’s; aspiration
conditioned by regionallinguistic background [RL]
Reference
(Nihalani et al. 1979:
211 [P]; Bansal 1990
[EC])
(Agnihotri & Sahgal
1985 [SV]; Sahgal &
Agnihotri 1987 [SV];
Agnihotri 1994 [SV];
Coelho 1997 [SV] AIE
NS Q; Trudgill &
Hannah 2002: 130
[SV]; Sharma 2005:
208 [SV] US-IE NS Q;
Wiltshire 2005: 282
[SV], citing Singh
2004) 6263
(Trudgill & Hannah
2002: 130)
(Bansal 1990)
/ð/ [d]
[s]
(Bansal 1990; Coelho
1997 AIE NS Q;
Baldridge 2002 NS)
/p/ & /f/; /t/ & /Ɵ/; /d/ &
(Nihalani et al. 1979:
212 [D]; Kachru 1982:
359 [M]; Coelho 1997
[FV] AIE NS Q;
Trudgill & Hannah
2002: 130 [M] and
[RL]; Sailaja 2009: 21
[RL])
/ð/; /s/ & /ʃ/
/Ɵ/, /ð/ [t͍], [d͍]
(Coelho 1997) found complete r-lessness in her sample population of Anglo-Indian English speakers.
Sharma (2005)only looked at rhoticity in coda position, following Labov (1966), finding that IE may be
completely non-rhotic or may manifest as a partially devoiced, trilled /r/. However, generalizing past her
findings is problematic because the study focuses on non-native Indian English learners within the US Bay
Area (important because it is outside of India).
63
329
Feature
/z/ replacing /s/, /gz/
replacing /ks/ for some
speakers
Medial /ŋ/ never reduced;
always followed by /k/ or /g/
/n/ retroflexed by south
Indians
h-dropping (thus far only in
Anglo IE speakers, socially
variable)
h-insertion (thus far only in
Anglo IE speakers, socially
variable and less common
than h-dropping)
Resyllabification based on
spelling
Syllabic /n/ and /l/ absent,
replaced with [ən] and [əl]
by most speakers
Syllable reduction
Increased palatalization and
nasalization
64
Example
loose [z], purpose [z],
maximum [gz]
Reference
(Bansal 1990)
singer [sɪŋgəɹ]
(Bansal 1990)
mo[ɳ]ey
(Sailaja 2009: 22)64
His leg is hurt?
[ʔiz lɜ́ɡ ɪ̆̆z ɜ́:ʈ]
(Coelho 1997: 574 AIE
NS Q)
the person who injured
[ɖɜ́ pɜ́:sn ʊ̈ hɪ́ndʒɜ́:ɖ]
(Coelho 1997: 576 AIE
NS Q)
typically ty.pic.a.lly
(4 syllables)
button [bətən]
(Baldridge 2002: 4 NS)
government [ɡɒrmɛnt]
(Bansal 1990)
(Bansal 1990)
(Labru 1984: 50)
The remainder of Sailaja’s Standard Indian English Pronunciation (SIEP) features are not included in this
table, because her study—descriptive in nature—continually switches between what SIEP speakers do and
what actual social, regional and linguistic communities in India do (2009). Entering her views on SIEP, and
the range of variation here would necessarily make this chart as long as her book, and instead, I refer
speakers to that source directly. I do not, however, endorse her views, because they are descriptively
derived and ideologically constructed. It is not clear who—if anyone—is a SIEP speaker. Her list of
features does serve as a useful starting point for picking variables for variationist or acoustic analysis.
330
MORPHOLOGICAL FEATURES
Feature
Past tense marker –ed
uniformly pronounced one
way, without
environmentally conditioned
allomorphs
OR
/d/ used instead of /t/ in past
tense pronunciation
OR
-ed past tense marker
dropped after /t, k/
Present tense marker –s
uniformly used by many
without environmentally
conditioned allomorphs
Plural marker uniformly
used by many without
environmentally conditioned
allomorphs
a morpheme replaces an
before vowel-initial words
Lower use of contracted
forms in IE as compared to
RP & AE
English affixes productively
applied to Indian terms
Possessive ‘s absent
Example
GRAMMATICAL AFFIXES
Reference
(Krishnamurti 1978,
cited in Agnihotri
1994)
asked [a:skd], laughed
[la:fd], traced [tre:sd]
(Bansal 1990)
walked walkØ
(Baldridge 2002 NS)
runs [rʌns], not [rʌnz]
(Krishnamurti 1978,
cited in Agnihotri
1994; Bansal 1990)
dogs [-s], falls [-s], keys [s], matters [-s]
(Bansal 1990)
an a / __#V
(Labru 1984: 69 WP;
Baldridge 2002 NS)
(Sharma 2001: 350
WK Q)
will vs. ‘ll
have vs. ‘ve
not vs. n’t
-fy, –ism, pre- attached
productively
I living next to my
memsahib sisterØ house.
COMPOUND NOUN FORMATION
Extensive compound
English-speaking, conventformation;
going,
Form: N1 + for N2 N2N1 chalk-piece (piece of
AND
chalk), key-bunch (bunch of
Form: noun + deverbal noun keys), salt giver (giver of
labeled Rank-reduction,
salt),
explained as
age barred, pindrop
‘modifier+head,’ replacing
silence, time-pass,
schoolgoer, immoral
‘head+qualifier’ by (Labru
charges (for charges of
1984)
immorality in 24 persons
were arrested on immoral
charges.)
(Baldridge 2002: 3 NS)
(Hosali & Atchison
1986: 72 BIE NS Q)
(Kachru 1982: 361;
Labru 1984: 46, 76-77
WP; Baldridge 2002: 3
NS; Trudgill & Hannah
2002: 131; Mehrotra
2003)
331
Feature
Example
Reference
WORD CLASS CHANGES RELATED TO PLURALIZATION
Nominal form used instead
colourØ pencils (replacing
(Trudgill & Hannah
of participial when
coloured)
2002: 131)
functioning as adjective
scheduleØ flight (replacing
scheduled)
Mass nouns pluralized
litters, furnitures, woods
(Hosali & Atchison
My son has got three
1986: 71 BIE NS Q;
childrens, madam, two
Chelliah 2001: 171
daughter and one son.
WG; Baldridge 2002
NS; Trudgill &
Hannah 2002: 130)
Singular nouns used in
She’s a tease, she loves to
(Sahgal & Agnihotri
generic sense pluralized
pull your legs.
1985 NLJ Q; Hosali &
In his hearts of hearts he
Atchison 1986: 71 BIE
knows he shouldn’t have
NS Q)
done it.
I been work lot of
memsahibØ. (memsahib is
respectful singular term for
female employers, lady of
the house)
Count nouns not pluralized
(Chelliah 2001: 171
One of my relativeØ …
WG; Baldridge 2002
NS)
(Hosali & Atchison
all as incipient plural
All the soup I know. And
1986: 71 BIE NS Q)
marker65 (variable)
tomato soup, and vegetable
soup and clear soup. (I
know how to make all kinds
of soup, e.g.. tomato soup,
vegetable soup, clear soup.)
All the girls I had. (My
children were all girls.)
OTHER WORD CLASS CHANGES
Bare nouns for RP partitive alphabets (letters of the
(Trudgill & Hannah
phrases
alphabet)
2002: 131)
a chalk (a piece of chalk)
clothes (items of clothing)
toasts (pieces/slices of
toast)
65
This use of all to mark plural may be related to bare nouns for partitive, next feature (my opinion, not
suggested by (Hosali & Atchison 1986).
332
Shortening of words,
syntactic class then opened
Feature
Indian lexical borrowings
already plural also get
English pluralization
Reduplication for emphasis
at lexical and clause level
enthusiasm (n) enthu (n,
adj),
fundamentals (n)
fundas/u (n, adj)
Example
roti (Hindi: bread.pl) roti
(Eng: bread.sg) and rotis
(Eng: bread.pl)
Come on! sit sit!
hot hot water
(Baldridge 2002: 3 NS)
Reference
(Baldridge 2002: 3 NS)
(Kachru 1982: 361;
Baldridge 2002: 6 NS;
Chelliah 2006 WF Q,
citing Kachru 1986:40)
SYNTACTIC FEATURES
Feature
SOV word order (Dravidian
‘interference’)
Example
And hot water take, and tea
put. (Take some hot water,
and put tea into it.)
PREPOSITIONAL PHRASE
Optional (or limited) use of
Optional (or limited) use of
prepositions
prepositions
NOUN PHRASES
Article presence/absence
We are going to Ø temple.
different from AE/RP for
Orissa face an innings
indefinite [I], definite [D] and defeat at Ø hands of Bihar.
zero [Z] article (mass/count
For you what is a nonsense
noun properties different from of Masses it is something
AE/RP).
different.
66
Reference
(Hosali & Atchison
1986: 73 BIE NS Q)
Optional (or limited)
use of prepositions
(Kachru 1982: 360
[D, I, Z]; Labru
1984: 46, 68 [D,I, Z]
WP; Williams 1987;
Shastri 1992: article
use is broadly
defined as different
WK Q; Chelliah
2001: 171 [I] WG;
Baldridge 2002: 6
[C] NS; Sharma
2005: 199 [D, I] USIE L2 NS Q;
Chelliah 2006 [I]
WF Q) 66
Sharma found that definite articles have a high rate of null realization, however, her informants are
labeled contrastively as IE language learners at the beginning of the study, and as stable non-native IE
speakers given results which show a remarkable stability across multiple levels of fluency—her participants
are never considered native IE speakers (2005: 202).
333
Feature
Bare noun (in sg. or plural) as
generic noun, problematized
by ‘generic vacillation,’
where number of generic
noun repeatedly shifts
between sg & pl.
Subject and Object omission
(pro-drop); suggested that
pronoun omission more likely
than full NP omission [P]
Initial is accompanying
subject pro-drop, functioning
like existential There is, there
are, as well as I am, he is, she
is, it is.
Invariant negatives no and
not67
Dummy subject it not
required
67
Example
Before it was thought Ø
woman’s work is only in
between four walls.
So father thinks that
daughter will be Ø liability
to him son is Ø asset.
1. He played cricket all day
today—and now (pro.subj)
does not want to work on his
homework..
2. It is simple: take a dollar
bill, and insert (pro.obj) in
the machine, face up, and
you get four quarters.
3. You got tickets? No,
(pro.subj) sold (pro.obj)
already.
Is working in the mess. (My
father worked in the mess.)
Is only one father and one
brother. (I have only my
father and one brother.)
Is everyday working some
bearers. (Everyday I work
as a bearer.)
Wednesday not eating meat.
(On Wednesdays we don’t
eat meat.)
I no go Jesus. (I won’t go to
Jesus.)
One girl no married. (One
of my daughters isn’t
married.)
No, is not help you. (No, she
doesn’t help you.)
Here (pro.subj) is not safe to
wait.
It is unclear if this feature is syntactic or lexical.
Reference
(S.N. Sridhar 1996:
63-4 L2 WE)
(Hosali & Atchison
1986: 72 [P] BIE NS
Q; Bhatt 2000: 79;
Chelliah 2006)
(Hosali & Atchison
1986: 68-0 BIE NS
Q)
(Hosali & Atchison
1986: 69 BIE NS Q)
(Bhatt 2000: 81)
334
Disagreement in number
between subject and verb,
esp. with respect to mass
nouns; It is alternatively
suggested that this is a lack of
inflection for tense/aspect
rather than explicit
disagreement [L].
Preposed adverb preference:
‘adverb+verb’ preferred over
‘verb+adverb’
Transitive verbs used
intransitively
Intransitive verbs used
transitively
VERB PHRASE
Simla, Dalhousie and KuluManali are attractive spots
and diversity of atmosphere
in different places heightenØ
their glamour and charm.
The outgoing calls from the
Bangalore system also is
being routed through this
line with effect from 6 p.m.
today.
He quickly ran away. (vs.
ran away quickly)
One should first work hard
and then ask for a reward.
(vs. work hard first)
reach, waive, enjoy exert
preside, dissent
Mr. Mishra wished the
conference a success and
assured that…
Copula omission (variable)
I Ø the dressing boy.
I Ø poor boy.
PARTICLE VERB CONSTRUCTIONS
Particle verbs (verb +
Common (WF): care for,
preposition/adverb):
come from, burst into, ball
low rate of particle verbs in
into live with, laugh at, look
IE, when used, found in
after
formal IE writing, as
compared to high rate in
Uncommon Domain: talk to,
informal spoken RP. Rogers
speak to, think about,
(2002: 196 Q) suggests that
depend on
IE written fiction has highest
rate of prepositional verbs.
Semantic domains of
communication, existence and
mental processes are clarified
as rare [SD]
68
(Labru 1984: 71 WP;
Hosali & Atchison
1986 [L] BIE NS Q;
Chelliah 2006, citing
Hosali 1982 [L] BIE
NS Q)
(Labru 1984: 47)
(Labru 1984: 48
quoting Kindersley
1938)
(Labru 1984: 48
quoting Kindersley
1938, 75 WP)68
(Hosali & Atchison
1986: 65 BIE NS Q)
(Rogers 2002: 196
[SD] WF, WP, WU
Q; Schneider 2004b
WK Q)
This doesn’t seem to be a transitivity issue, and instead might better be characterized as applying an
article to a mass noun; however the original author’s labeling has been maintained.
335
Particle and preposition
insertion,
omission and ‘incorrect’ use;
Rogers (2002: 196 Q) finds
that to is rare in IE, but
common in RP and AE, while
from is more frequent in IE
than RP and AE
Differences in complement
structure of certain verbs
Complementation used
differently from AE, in
particular talk, do, sell, wear,
burden, stay and be occur
with the –ing form
69
We tried to spot out the
filmstars. (out insertion)
He made me to go. (to
insertion)
Police sources said that he
was holding to telephone
receiver in his hand when he
was discovered dead. (to
insertion)
Please convey Ø him my
best wishes. (to deletion)
The meeting was presided Ø
by Mr. A.G. Kher. (over
deletion)
Applications accompanied
with testimonials are invited
for five temporary
posts…(with replacing by:
‘incorrect’ use)
We are involved to collect
poems. (vs. involved in)
She was prevented to go.
(vs. prevented from)
I would like that you come.
(vs. like for)
They want that you should
leave. (vs. want…to)
to + V.inf replaced by prep
V+ing
Among those who like
talking politics…vs. like to
talk)
I won’t like going to Italy.
(vs. to go)
…for doing… (vs. to do)
…for writing purposes…
(vs. to write with)
(Labru 1984: 75, 978 WP; Sahgal &
Agnihotri 1985: 126
NLJ Q; Trudgill &
Hannah 2002: 131;
Chelliah 2006, citing
Verma 1982)
(Trudgill & Hannah
2002: 134)
(Shastri 1992: 266-7
WK Q)69
(Labru 1984: 48
quoting Kindersley
1938)
Variation in complementation was originally hypothesized for Pakistani English, though IE appears to
display behavior that is not statistically significant in its difference from AE (Shastri 1992: 266-7).
336
Lack of reflexive objects with
some typically reflexive verbs
You are invited to avail Ø
this golden opportunity. (vs.
avail yourself of)
(absent, acquit, apply, avail,
avert, betake, distinguish,
drink, enjoy, express, overreach, pride, resign, revenge,
seat, set)
QUESTION FORMATION
Frozen (uninflected/invariant) no?
or
isn’t it?
tag question complete [C] or
There is so many people no
socially variable [S]—Sahgal madam?
& Agnihotri argue that no? is There is so many people,
not used, and isn’t it is used
isn’t it, madam?
(1985).
Lack of subject-auxiliary verb
inversion in direct questions
complete [C] or socially
variable [S]
Where you are coming
from? (vs. are you)
Who you will come with?
(vs. will you)
Subject-auxiliary verb
inversion in indirect questions
(distinction between
embedded and non embedded
interrogatives neutralized)
I asked Hari where does he
work. (vs. he works)
(Labru 1984: 48
quoting Kindersley
1938; Chelliah 2001:
169 WG)
(Kachru 1982: 360
[C]; Verma 1980
cited in and refuted
by Sahgal &
Agnihotri 1985: 123
[S]; Hosali &
Atchison 1986: 70
BIE NS Q; Coelho
1997 [S] AIE NS Q;
Chelliah 2001: 170
WG; Baldridge
2002: 6 [C] NS;
Trudgill & Hannah
2002: 134)
(Kachru 1982: 360
[C]; Sahgal &
Agnihotri 1985 [S]
NLJ Q; Coelho 1997
[S] AIE NS Q; Bhatt
2000: 74 [S];
Chelliah 2001: 170
WG; Baldridge
2002: 7 [C] NS;
Trudgill & Hannah
2002: 134)
(Verma 1980, cited
and disproved in
Sahgal & Agnihotri
1985: 122; Trudgill
& Hannah 2002:
134)
337
Question formation via
auxiliary movement vs.
intonation: IE speakers use
intonation type bare, while
AE speakers typically add
discourse markers, e.g. so
Direct/Indirect question
proportions: indirect
questions with modals more
frequent for upwardly mobile
Westernized, direct
imperatives more common for
traditional, culturally tied
students; also depends on
status of interlocutor and
context.70
Q-fragment
No ‘do’ support
Auxiliary Movement:
Are you coming?
Intonation:
You are coming?
(Coelho 1997 AIE
NS Q)
Direct:
Give me a glass of water.
Indirect:
Can I have a glass of water?
(Sridhar 1991: 311
L2 WE)
Going then?
(Chelliah 2006 WF
Q)
(Bhatt 2000: 75)
Why Ø you look so worried?
TENSE, ASPECT AND AUXILIARIES
Deviations in underlying
I am doing it often. (vs. do)
Tense/Aspect of verbs:
I am having many sarees.
(vs. have)
Progressive morpheme for
stative verbs [S], habitual and The surface is feeling rough.
completed actions [A], or to
(vs. feels)
signal imperfectivity (I)71.
They are having vindictive
motives. (vs. have)
I was doing. (vs. I
(constantly) did.)
70
(Kachru 1982: 360-1
[S]; Labru 1984: 48
quoting Kindersley
1938 [I]; Sahgal &
Agnihotri 1985
[SAI] NLJ Q; Hosali
& Atchison 1986: 61
BIE NS Q; Chelliah
2001: 169 [S], 171
[I] WG; Sharma
2001: 368 [I] WK Q;
Baldridge 2002:6 [I],
7 [SA] NS; Rogers
2002: 192 disproves
[S], 193 supports [A]
WF, WP, WU Q;
Trudgill & Hannah
2002: 132 [SA];
Chelliah 2006 [I]
WF Q)72
Static form of will you used (rather than distinguishing between can, could, will, would, will, shall, may,
etc.) by traditional students (Sridhar 1991: 315).
71
Sharma points out that “Williams (1987) noted that the use of progressives with stative verbs occurs in
West African varieties of English, and Harris (1984: 56) cited certain Celtic dialects as also exhibiting this
usage” (2001: 368).
338
Auxiliary absent with
progressive73
I Ø doing for domestic
line—butler. (vs. am doing)
When small I Ø working to
the British. (vs. was working
or worked)
Extensive use of pluperfect
78% of IE had V+ed used
(had V+ed) for present perfect for pluperfect, 12% for
or simple past 74,75
present perfect, 10% for
preterit meaning
--had V+ed with present
meaning
--had V+ed with remote past
meaning
--had V+ed with sustained set
of affairs and explicit
extended time adverbial
--had V+ed for covertly
reported speech
Present tense with durational
phrases instead of present
perfect
72
(Hosali & Atchison
1986: 64-5 BIE NS
Q)
(Shastri 1992: tense
is broadly
characterized as
stylistically different
WK Q; Sharma
2001: 356 WK Q;
Rogers 2002: 194-5
disproves 'extensive
use' in WP, supports
in WF and email Q)
It is done. (vs. has been)
(Labru 1984: 48
Never before in the
quoting Kindersley
Capital’s history these
1938)
colonies had faced such a
flood threat. (vs. have faced) (Sharma 2001: 352
The capsule had been silent WK Q)
about the role of the early
(Sharma 2001: 353
pioneers in the freedom
WK Q)
struggle. (vs. was silent)
Similar concessions had
been in force for years in the
southern States. (vs. have
been)
The agitators had also
disrupted road traffic. (vs.
disrupted)
I am here since two o’clock.
(vs. have been)
I am reading this book for
two hours. (vs. have been)
(Sharma 2001: 363
WK Q)
(Sharma 2001: 363
WK Q)
(Trudgill & Hannah
2002: 133)
Hosali and Atchison (1986) suggest that the progressive is used 55% of the time, compared to other verb
types, for younger speakers, 34% of the time for older speakers. Additionally, they argue that its use is
haphazard, given that both bare verb stems and the progressive were used by the same speakers in the same
context (1986: 62).
73
This is difficult to accurately characterize, as the progressive was sometimes used sans auxiliary to
reference simple past (e.g. the second example is glossed as When I was young I worked for the British.).
74
Sharma, following Shastri (2001: 362), considers this an ‘opaque’ feature, which, in contrast to
‘transparent’ features, like code-switching and lexical borrowing, is not readily observable because “it is
perhaps not the form that is at variance but the function’ .(2001: 351)
75
Sharma also highlights increased non-standard pluperfect use in regional press and bureaucratic texts, as
compared to national texts in both genres (2001: 361)
339
Future form used for temporal When you will arrive, please
and conditional, instead of
visit me. (vs. when you
present
arrive)
If I will come, I will see you.
(vs. if I come)
Relaxed tense restriction on
I asked Hari where does he
embedded sentences
work. (vs. he works)
(Labru 1984: 48
quoting Kindersley
1938; Trudgill &
Hannah 2002: 133)
(Verma 1980, cited
and disproved in
Sahgal & Agnihotri
1985: 122)
Absence of sequence-of-tense When I saw him last week,
(Trudgill & Hannah
constraints
he told me that he is coming. 2002: 133)
(vs. was coming)
When had +Ved is used with Similar concessions had
(Sharma 2001: 363
time marking adverbials
been in force for years in the WK Q)
(TMA), meaning is present
southern States. (vs. have
perfect
been)
would and could TMA’s to
He would be coming to
(Sharma 2001: 367
mark simple future76
receive you on 15th and he
WK Q)
OR
would be bringing the other
ticket to. (vs. will come, will
bring)
would and could replacing
We hope that you could join (Trudgill & Hannah
present tense will, can
us.(vs. can)
2002: 132)
because past tense felt to be
The lecture would begin at
more tentative, therefore more 2:00.(vs. will)
polite
Copula favored over modals
be used in 70% of the
(Hosali & Atchison
constructions, modals
1986: 66 BIE NS Q)
comprise less than 30%, for
both young and old
speakers, auxiliary have
completely absent
Present perfective instead of
We have known, four years
(Sharma 2001: 367
simple past
ago, that weightlifting was
WK Q; Trudgill &
going to be an Olympic
Hannah 2002: 134)
sport. (vs. knew)
Imperfective to express
He used to do… (vs. did)
(Labru 1984: 49
perfective
The train was halting at the
citing Goffin 1934,
platform and another train
96 WP)
rammed into in from behind.
(vs. was halted)
76
Sharma (2001: 269) notes a similar parallel in Trinidadian English, referencing Winford (1993: 172), and
in Singapore English, referencing Foley (1998: 142).
340
Feature
Example
SYNTACTIC GENRE-SPECIFIC FEATURES
Increased pluperfect in
Type
AE RP
IE
reported speech
Reported
39% 41% 44%
Non-reported 61% 59% 56%
Increased paraphrasing of
Mahadevan sahib told me
direct quotes (in newspaper
don’t ride the cycle. (Dr.
genre), as compared to RP & Mahadevan has forbidden
AE [P]. Conversely, in oral
me to ride a bicycle.)
speech, direct quotes directly
reported [D]
AE RP IE
Longer sentences (in
Press
(report)
7.9 8.7 10.6
newspaper genre) as
Press (editorial) 8.1 9.2 10.8
compared to RP & AE
Skills, trades, hobbies
Reference
(Sharma 2001: 359
WK Q)
(Hosali & Atchison
1986: 73 [D] BIE
NS Q; Sharma 2001:
350 [P] WK Q)
(Sharma 2001: 351
WK Q)
8.3 9.9 11.1
Scholarly/scientific
8.3 9.8 11.5
Detective Fiction
8.8 10.1 11.2
‘Failure’ of voice harmony,
where nominal groups calling
for active verbs instead
followed by passive verbs.
Variation in placement of
also—typically at end of S.
only and itself:
Emphasis on time/place
OR
Simple right edge focus
marker and
nonquantificational
(presentational) focus
OR
Emphatic postmodifier
particle for NP’s and larger
constituents
Fronting for focus/topic
OTHER FEATURES
I had aversion for posting in
hill stations and, as such,
Simla was ruled out.
(Labru 1984: 78, 86
WP)
We never even used Hindi
word also.
I was in Toledo only.
Can we meet tomorrow
itself?
(Baldridge 2002: 7
NS)
(Baldridge 2002: 6
NS; Trudgill &
Hannah 2002: 132)
He will buy over there tickets
only.
(Bhatt 2000: 77-78,
and references
therein)
…but the woman then were
not allowed to go out itself.
(S.N. Sridhar 1996:
59-60 L2 WE)
Two is of course your
(Chelliah 2006 WF
choice.
Q)
341
Feature
Conjoined verb
Dislocation (dominantly to
left, occasionally to right)
Complex sentences with lots
of embedding
Absence of to in number
range expression
‘So’ complementizer
Postposed existential ‘there’
in conjunction, adverbial
‘there’ replaces dummy
‘there’
Relative pronouns distanced
from referent noun phrase
Example
Soon I’ll have to wake up
and work.
L: This chicken curry I don’t
know madam. (I don’t know
how to make chicken curry.)
R: There she is working, my
daughter. (That’s where my
daughter works.)
Reference
(Chelliah 2006 WF
Q)
(Hosali & Atchison
1986: 74 BIE NS
Q)77
(Kachru 1982: 360)
two Ø three hundred
I think so you are in
Amayrica now?
That is of course there.
I’m sure an explanation is
there.
which, who, that not directly
following NP’s they refer to
The impression has been
created which can not be
removed.
(Baldridge 2002: 6
NS)
(Chelliah 2006 WF
Q)
(Trudgill & Hannah
2002: 132; Chelliah
2006 WF Q)
(Labru 1984: 79-81
WP)
LEXICAL AND SEMANTIC FEATURES78
77
Feature
Compounds created to mark
gender
Example
cousin-brother, cousinsister
English affixes productively
applied to Indian terms
pre- -fy, –ism
Hindi/Vernacular affixes
productively applied to
English terms
policewala (one who
polices)
Reference
(Kachru 1982: 361;
Labru 1984: 48
quoting Kindersley
1938; Kachru 1986;
Baldridge 2002: 3
NS)
(Kachru 1982: 363;
Baldridge 2002: 3
NS)
(Kachru 1982: 363)
I suspect that these are simply conversation strategies for emphasis and focus, although corpus research
is required to verify this.
78
Sometimes labeled collocations (e.g., Kachru 1982)
342
Feature
New lexical items via
acronyms
New lexical items via
abbreviation (different from
abbreviations in IE because
pronounced as abbreviation
Borrowing of Hindi terms
into IE where (near)
equivalents are available in
English
Borrowing of Hindi terms
into IE where English
equivalents are unavailable
(with the genres of religion,
fine arts, music, dance and
drama, folk art and
mythology, administration,
flora and fauna, society/caste
terms)
English word replaced with
semantically similar, but
distinct word
one as indefinite pronoun
substitution
79
Example
FOC (free of charge),
ABCD (American Born
Confused Deshi)
Jan (January), Gen. Sec. or
G. Sec [ʤiɛk] (General
Secretary)
Reference
(Baldridge 2002: 4
NS)
achchaa (good), beta
(child), bus (stop), lekhin
(but), hartal (strike), bhook
(hunger), maida (flour)
(Kachru 1982: 362-3;
Labru 1984: 105 WP;
Kachru 1986; Shastri
1992: 264 WK Q;
Baldridge 2002: 6
NS; Trudgill &
Hannah 2002: 135) 79
(Shastri 1992: 265
WK Q; Trudgill &
Hannah 2002: 135)
dharmic (religious), sarod
(stringed instrument),
kathak (Indian dance style),
Diwali (fall festival of
lights), lakh (unit of
measurement for 100,000),
sari (female dress made of
one long cloth),
shri/shrimati (Mr./Mrs.)
…if we see only in the
direction of art and
literature. (vs. look)
Women came out of the
house for studying. (vs.
home). Even if they do not
encourage it, they stay
indifferent. (vs. remain).
foot (vs. leg), hand (vs.
arm), pull on (vs. get on
with), side (vs. direction),
best (vs. worst with
indefinite article), loves (vs.
likes in My teacher ___ me),
got (vs. have, e.g. I got
money, but not enough.
one used for a: And one
black lady…
(Baldridge 2002: 4
NS)
(S.N. Sridhar 1996:
58-9 L2 WE)
(Labru 1984: 49
citing Kindersley
1938)
(Hosali & Atchison
1986: 63 BIE NS Q)
(Baldridge 2002: 6
NS)
Shastri considers these ‘transparent’ features, because they are ‘obvious’ (1992: 266)
343
Feature
‘Redundant’ marking of
inclusion via even… also
Example
In our country, even women
also encourage.
Even India is also changing.
Semantic nativization80,
wheatish (to describe
where localized meanings are complexion),
developed for English words sick/ ill (archaic AE
meaning retained)81,
dark/black82, biodata
(curriculum vitae), almirah
(chest of drawers), cobrother (wife’s sister’s
husband), furlong (1/8 of a
mile; archaic), bearer
(room-servant)
Frequent use of archaic
serpent, ox, thrice
forms and idioms
Verb choice influence from
Hindi
Preposition choice
Collocational restriction on
semantic pairs and root/affix
extended
Register change within
sentence
80
Reference
(S.N. Sridhar 1996:
60 L2 WE)
(Kachru 1982: 364-5;
Hosali & Atchison
1986: 74 BIE NS Q;
Kachru 1986; Shastri
1992: 269 WK Q;
Trudgill & Hannah
2002: 135; Mehrotra
2003)
(Labru 1984: 48
quoting Kindersley
1938 and Goffin
1934; Sharma 2001:
350 WK Q)
open/close the light (vs. turn (Kachru 1982: 361;
on/off)
Shastri 1992: 269
giving a test (from student’s WK Q; Baldridge
perspective),
2002: 6 NS)
taking tea (vs. having tea),
keep the ball there (vs.
put)83
since (vs. for), as (vs. like)
(Chelliah 2006 WF Q,
I have been writing this
citing Verma 1982:
essay since two hours.
183)
This tea is too light for me.
(Sahgal & Agnihotri
I am sorry for the
1985: 125 NLJ Q)
bother+ation I have caused
you.
Armed man nabbed.
(Chelliah 2006WF Q,
citing Dixon 1991:
440)
Semantic nativization is also documented for Aboriginal Australian English speakers (Sharifan 2005).
IE use of sick aligns with AE, not RP, in that the feature [+nausea] is optional, while it is obligatory in
RP (Shastri 1992: 270).
82
In IE, black ‘has acquired an additional component [+indian] (caste marker) which is defined by the
universe of discourse belonging to low caste,’ systematically contrasts with white when used to describe
hair color, while dark ‘has acquired an additional component [-norm] (Indian) as applied to complexion’
(Shastri 1992: 269-70).
83
Shastri argues that IE keep suppresses the aspect feature [+durative] and introduces the feature [by
movement] (1992).
81
344
Feature
could replacing was able to
Example
He could just only finish it
before we left.
Modal auxiliaries used
differently in conditional
constructions and indirect
speech
Emphatic reflexive84, 85
This I myself told them was
wrong.
Emphatic nothing
Nothing. (as response to
question if informant had
ever experienced any sad
incident in his life.)
Lexical hybrids (1 vernacular tiffin carrier (one who
word, 1 English word)
carries lunches packaged in
metal tiffins)
British sarkar (British
government)
lathi-charged (cane/baton
charges used by police to
quell riots)
like this, likewise, and totally Like this, the position of
as discourse adverb signaling women has been changing.
summary or conclusion in
Totally, women is equal to
narrative
God.
Quotative ‘as’ preceding
Indian woman was
quoted material
considered as a machine.
Some of them have called
her as “…”
They were called as prepuberty marriages.
Culture bound simile and
In olden times, woman just
metaphor use, literal
worked like a bullock.
translation of (Kannada)
flower bed (nuptial bed),
idioms
dead womb (a barren
woman)
Disproportionate use of
saddest event, brutal
ecomiums86 and
unabashed aggression,
condemnation
charmingly beautiful, the
happiest man on earth
84
86
Reference
(Trudgill & Hannah
2002: 132)
(Shastri 1992: 266
WK Q)
(Chelliah 2006 WF
Q)
(Hosali & Atchison
1986: 70 BIE NS Q)
(Kachru 1982: 363;
Labru 1984: 130 WP;
Kachru 1986;
Mehrotra 2003)
(S.N. Sridhar 1996:
61 L2 WE)
(S.N. Sridhar 1996:
60 L2 WE)
(S.N. Sridhar 1996:
58 L2 WE)
(Labru 1984: 58 WP)
(Labru 1984: 57 WP)
Confirmed in written IE novel genre, not yet studied in spoken IE.
I have no idea what an ‘ecomium’ is, but am guessing that it is the opposite of a condemnation based on
the examples Labru offered.
345
Feature
IE abuses most often curses
invoking death or disease
Honorifics to address elders
Kinship terms to address
elders: socially variable,
marking Westernized,
upwardly mobile middle
class
Example
You of evil stars
May the vessel of your life
never float in the sea of
existence.
You cock-eyed son of a bowlegged scorpion.
POLITENESS
name+Sahib: Mohan sahib
name+ji: Ghandiji
Aunty/Uncle (replacing
Mr./Mrs.)
Frequent use of formal style
markers
Please grant me…
Would you like to give me…
Use of please in requests
correlated with Western,
upwardly mobile student
population, who use it
consistently, while less
westernized, more traditional
students used with social
superiors, and did not with
social inferiors
Can I please have a glass of
water?
Reference
(Labru 1984: 58-9)
(Labru 1984: 60)
(Sridhar 1991 L2
WE)
(Sridhar 1991: 313
WK Q; Sharma 2001:
350 WK Q)
(Sridhar 1991: 312 L2
WE)
PROSODIC/SUPRASEGMENTAL FEATURES
Feature
Equal stress in each
syllable, creating syllabletimed rhythm, possibly
socially variable [SV]
Example
Equal stress results in
unreduced vowels
e.g. in function words and
suffixes
Suffixes receive stress
Reference
(Nihalani et al. 1979:
220; Bansal 1983, cited
in Agnihotri 1994;
Shackle 1987, cited in
Agnihotri 1994; Bansal
1990; Coelho 1997 [SV]
AIE NS Q; Trudgill &
Hannah 2002: 130)
(Bansal 1990; Coelho
1997AIE NS Q; Trudgill
& Hannah 2002: 130)
(Trudgill & Hannah
2002: 130)
346
Feature
Tone group divisions differ
from RP
Nucleus location differs
from RP
Example
Word accent different from
AE, RP
IE accent assignment is
phonetic, on penultimate
heavy syllable, while other
native Englishes are
morpho-phonemically
assigned (syntactic category
relevant)
IE is pitch accent, not
stress-accent system.
Stressed syllables fall in
pitch, fall and low-rise pitch
replace stress typical of
RP87
on 1st syllable in IE:
another, correct, degree
on 2nd syllable in IE:
character, energy,
fortnight, governor,
minister
Smaller cues for stress in
IE, and different location of
stressed syllable, as
compared to AE.89
AE pitch & stress occur
together, while in IE, they
are relatively independent.
‘angry’ or ‘aggressive’
perception of IE speakers
because of pitch range
‘sing song’ intonation,
more pitch accents used,
with correlation between L1
and kinds of pitch accents
used
87
Reference
(Bansal 1983, cited in
Agnihotri 1994; 1990)
(Nihalani et al. 1979:
225; Bansal 1983, cited
in Agnihotri 1994; 1990)
(Nihalani et al. 1979;
Bansal 1990; P. Pandey
1994: 199; Wiltshire
2005: 295 TIE ES NS)
(Pickering & Wiltshire
2000 ES NS) 88
(Contradicting this,
Wiltshire & Moon 2003:
291 argue that stress, or
loudness, differs in
magnitude but not in
direction from AE, ES
NLJ)
(Wiltshire & Moon 2003:
300 ES NLJ)
(Wiltshire & Moon 2003:
301 ES NLJ)
(Nihalani et al. 1979:
223)
(Kachru 1986; Wiltshire
2005: 296 TIE ES NS)
It is suggested that ‘drop in pitch and lack of amplitude change on accented syllables is a transfer from’
Indian languages and is an Indian areal feature (Pickering & Wiltshire 2000: 174).
88
He hypothesizes that IE speakers assign accent based only on phonetic composition of word, not by this
AND syntactic category.
89
The analysis presented by (Wiltshire & Moon 2003) is very confusing; I am not sure if I have correctly
represented their findings on pitch, syllable length and stress.
347
GLOBAL FEATURES
Feature
Earnest moral tone
Stylistically ornate (heavy
Latinate vocabulary instead
of Germanic)
Lengthy or wordy prose
style
Bookish English and
‘excessive’ use of idioms
Free, unaffected
exaggeration (via Indian
idealism, contemplative
outlook, effusiveness and
gaudiness)
‘Excessive’ politeness
Example
We are fumbling in our
pockets for this key today. If
we do not recover it in good
time, we may lose it for ever.
We need this key to throw
open to our youngsters the
window that will give them a
world perspective of the
problems of humanity and
their possible solution.
coming almost after a week
(vs. come back), inhabit (vs.
lives), demise (vs. death)
In his attempt to resolve this
conflict while between the
two, Labov tries at times to
hunt with the hounds while
running with the hares.
The affliction was real
enough. Miss Shahbano, the
Parsee ayah whom Bilquis
had employed on her return
to Karachi, complained on
her first day that when she
gave Sufiya Zinobia a bath
the water had scalded her
hands, having been brought
close to boiling point by a red
flame of embarrassment that
spread from the roots of the
damaged girl’s hair to the
tops of her curling toes.
Thank you very much indeed.
I am extremely grateful
At the feet of our parents…
…beg to submit…
Reference
(Labru 1984: 57 WP)
(Chelliah 2001: 170
WG)
(Shastri 1992: 271
WK Q; Sharma 2001:
350 WK Q)
(Chelliah 2006 WF Q,
citing Dixon
1991:440)
(Rushdie 1983 cited in
Labru 1984: 125-6
WF)
(Labru 1984: 57 WP)
348
READING/WRITING CROSSOVER
Feature
Example
Overfaithful to written form pronunciation of silent
within pronunciation
letters in bomb, lamb
Geminate consonant
pronunciation from written
CC
Medial /ŋ/ never reduced;
always followed by /k/ or
/g/
Syllabic /n/ and /l/ absent,
replaced with [ən] and [əl]
by most speakers
Codeswitching into Hindi
in writing conveys
informality, while
codeswitching in oral
speech changes the tone of
conversation, excludes
people, or signals the
introduction of a more
formal subject
singer [sɪŋgəɹ]
Reference
(Krishnamurti 1978,
cited in Agnihotri 1994;
Labru 1984: 62; Shackle
1987, cited in Agnihotri
1994; Bansal 1990)
(Krishnamurti 1978,
cited in Agnihotri 1994;
Bansal 1990)
(Bansal 1990)
button [bətən]
(Bansal 1990)
The janta (public) was also
curious to see in what
andaaz (style) the girls
dressed. Lara came in
glamourously clas, while
Priyanka stuck to her new
dress-code of showing little
skin.
(Viswamohan 2004: 35
WP)
upper and rubber
IE REFERENCES NOT COVERED HERE
•
(Bhatia & Ritchie 2001): Generative approach to codeswitching, not descriptive
for IE, based on relatively few examples.
•
(Bhatt 2005): Discussion of alternate meanings being exploited through IE
features, no new features cited.
•
(Chaudhary 2001): A monologue in IE/Hindi, with full RP translation on right.
Demonstrates many of the features found above, lends credence to the notion that
IE is typically mixed with Hindi, akin to how urban-Wolof is defined as involving
English (Swigart 2001), but no generalization of features
349
•
(D'souza 1997): A thought piece on myths associated with IE, no documentation
of actual features.
•
(Hohenthal 1998): Research based on an online survey of IE language practices,
ideologies and attitudes, not on IE features.
•
(Hosali 2005): Examines Butler English (BE), which is arguably distinct from IE.
As well, does not characterize BE, just offers multiple examples.
•
(Jenkins 2000): No specific characterizations of IE, beyond proposal that L2
English varieties differ most noticeably in their pronunciation.
•
(Kandiah 1991): A book introduction which summarizes Kahn (1991) and Sahgal
(1991); offers interesting original critique of past IE literature and nativity, but no
new IE features.
•
(Krishnaswamy & Burde 1998): A critique of written IE through different time
periods as non-native, wrong and deviant; IE not characterized, and simply
located as ‘wrong.’
•
(Leitner 1990): IE features are compared to Australian English; IE features pulled
from Hosali’s (2005) Butler English, hence not useful here.
•
(Leitner 1994): A corpus based study of syntactic, aspectual and semantic range
of two words (begin and start) across three dialects, including IE. Not applicable
for discussion of generic IE features here.
•
(Mehrotra 1998): A presentation of IE texts, no analysis.
•
(Mehrotra 2002): A presentation of examples of IE letter writing genre. Short,
lacks generalizations of how to circumscribe this genre, and pejorative in its view
of IE features as incorrect and nonnative.
350
•
(Mesthrie 2006): No IE features are listed—discussion of types of Englishes
(Kachru’s circle paradigm) and historical background of English.
•
(Nihalani et al. 1979): Most of this book is not included here because it is a ~200
page, word by word, alphabetized IE Usage Guide. It lists words, shows what
would be ‘preferable’ in RP, but fails to generalize across instances.
Phonetic/Phonological/Suprasegmental grammar points at the end are included
here.
•
(Sailaja 2009): An entire comprehensive book on SIEP features, and what
communities do not use them. Too large to include here, and problematic, in that
it is unclear who a SIEP speaker would be.
•
(Sahgal 1991): A survey of domains of language use, no documentation of IE
features.
•
(Schneider 2000): A corpus based analysis of IE, RP and AE vis-à-vis three nonIE particular features, which finds that IE falls between the other two for these
features.
•
(Sharma 2005): An analysis of non-native English learners; she is cited for null
article use, because she acknowledges that it is a feature mentioned previously in
IE literature. However, the rest of the features she examines are not IE features,
and instead are treated as ESL features within a cline of bilingualism. These
features (null copula, non-standard S-V agreement, non-standard past marking)
may prove to be features of IE, or may be simply features of ESL speech.
•
(K.K. Sridhar 1996) Very similar to (Sridhar 1991).
351
•
(Williams 1987): A discussion of language acquisition for non-native varieties of
English, with no discussion of IE specific features.
•
(Wiltshire 2005): I skipped the majority of features (phonetic, phonological and
suprasegmental) listed in this article because it draws on CIEFL (1972) for IE
features and focuses on contrasts between those and features from IE speakers
with Tibeto-Burman L1, arguing both for the influence/transfer from TibetoBurman languages and overarching Indian areal features.
352
APPENDIX C
INTERVIEW MODULES
•
Introduction
o I’m in school in the US and am doing a project on language and life in
Delhi. I will be talking to over thirty people from 18-70 years old, to see
how different people see Delhi and what their lives are like. If it is ok, I’d
like to talk for about an hour.
o Because this is a school project, I need you to sign this consent form that
you are OK with being interviewed. You’re not giving up any rights, just
acknowledging that you are allowing this interview. Is this ok?
o Is it ok if I record this, so i don’t forget anything important that you’ve
said? It’s very small and won’t get in your way.
•
Personal Background
o Where were you born? Have you lived here throughout your life? (If not,
where did you live and how long?)
o Where were your parents born? Did they always live there? If not, where
else and how long?
o What's your occupation? Have you always had this job? If not, what else
and when?
o Your parents' occupation? If too numerous to list, when did they have
what job?
o What was your childhood like?
353
o Where did you go to school?
•
Marriage
o Are you married?
If no •
What do you think of the marriage practices in India?
If yes •
Where is your spouse from?
•
How did you meet your spouse?
•
How did you get married?
•
What was the wedding like?
o Do you think there is a right and wrong age for marriage?
o What makes a good marriage?
•
Family
o Tell me about your family. How many children do you have?
o And what about the family you were born in?
•
Family Background and influence of other English varieties
o Do you have family abroad?
o Do you travel abroad? Where do you go?
o Do you have any ethnically mixed marriages in your family? Regionally
mixed?
If so, what languages do their children speak? What do you think
of that?
354
o Has anyone in your family married someone from a different state in
India?
Do the two share an Indian language, or is English the only
common language?
If they have kids, what languages do the kids speak?
o Do your family members that are abroad now sound different when they
visit? Can you describe how?
•
Specialized area
o Ask about link through which i met them—birders, family/friends,
gardening, etc.
o How did you get into ___?
o What do you like about it?
•
Language plans/ past language choices
o Did you/will you teach your children English at home? Why?
o Did you/will you teach your children Hindi at home? Why?
o What about local and vernacular languages—if you had lived in Mysore,
for example, would you have wanted your child to learn Kannada?
•
Would this be in addition to or instead of Hindi?
Would this be in addition to or instead of English?
Personal language background
o what languages learned
o at what age
o at home vs. school vs. community
355
•
Domains of Language Use
o What languages do you use when you…
read the paper
read novels
talk on the phone
IM
email
talk to friends
talk to family
talk to servants
talk to shopkeepers when you are shopping
•
Are there any markets where you always speak English?
•
Where do you use language X? (School, friends, work, TV,
music, internet, phone, servants, marketplace, etc.)
•
Delhi
o Have you always lived in South Delhi?
o Do you like Delhi?
o What are your favorite shopping areas in Delhi?
o What do you think of the traffic in Delhi? the metro?
o What do you think of the pollution in Delhi?
o Do you think that the upcoming Commonwealth Games will be good for
Delhi?
o How has Delhi changed over time?
356
o Do you think that people’s attitudes towards school have changed over
time?
•
Language Ideologies/Attitudes
o Was your school English medium? And college?
o Did your teachers at school consider nice English speech important?
o What rules did they stress often?
o (For people born outside Delhi) When you moved/started commuting to
Delhi did you get comments about your accent from locals? What did they
say? Were they right? Why?
o Where do people speak nice English? Why?
o (For informants who lived in the south for a long stretch of time)
When you lived in the south did you get comments from locals like you
speak in a funny way? I once heard a story about a schoolgirl from Delhi
who moved to Bangalore and her classmates told her she was putting on
airs because of how she talked. Has anything like this happened to you?
What exactly?
o What's the English used in the South like? And that of the east, in
Calcutta? And north, in the hill stations? And west, in Rajasthan or
Gujarat?
o
Who do you think speak nice English of the following people?
leading politicians
elementary school teachers
shop assistants
357
teenagers
radio and television announcers
priests
o Have you got any book on language or linguistics?
o Have you got any dictionaries? Bilingual dictionary? Which one? Do you
use/read them?
o Do you listen to or watch the programmes on language cultivation on the
radio and television?
•
Personal Assessment of linguistic ability (repeat for each language collected
in personal language background)
o How do you feel about your level of language X?
o Are you more comfortable reading, writing, speaking or hearing, or are
you equally comfortable in each domain?
•
Language Policy and Diversity in India
o How do you feel about the number of languages spoken in India?
Do you think it is helpful or makes it more difficult to govern the
whole country?
o Do you think a lot of people across India speak English? Hindi? How
many?
o How do you feel about it when politicians make speeches in different
languages?
o Do you think India would be better off if it had one language that
everyone spoke? What language would you choose? English? Hindi?
358
•
Language attitudes and IE
o How do you feel about the English spoken in Delhi? Is it different from
the English spoken in other countries? in other cities?
o What do you like about it? Can you give me some examples?
o What do you dislike about it? Can you give me some examples?
o Do you think that it is different from how they speak in England?
America? Which one do you prefer? Does any one sound more polite or
smart or cultured than the others?
o What kinds of differences are there?
o Do you think that the younger generation sounds different from the older
generation? How so? Is it better or worse?
•
Pear Story
o I’d like to show you a short video called the pear story and talk about it
afterwards, is that ok?
o AFTER can you tell me, in your own words, the plot of the story?
o Did you like the story? Did it remind you of anything?
•
Grandfather clause
o Do you mind if we turn off the fan/AC for a minute? I’d like you to read
this short passage, and i don’t want the fan noise to interfere.
o Grandfather passage
•
Participant’s Choice
o Is there anything that I have not asked you about but you would have liked
to talk about?
359
•
Closing
o Thank you for your time. It’s been very nice to talk to you. Is it ok if I
contact you again if i have any other questions? I can also give you my
phone number here or my email address in the States if you would like to
hear about how this study turns out.
360
APPENDIX D
QUALITATIVE CODES
These codes were used in Atlas.ti, and were developed from emergent themes within the
interviews as well as outside concepts which have proved useful in earlier research. The
codes are organized alphabetically by Code Family, because this is how files are
organized within Atlas.ti.
Background: BG
1. Culture/Lg links—groups who do maintain ‘their’ language: whoMaintainsLg
2. Family Education (parents as doctors vs. illiterate): scalar: famEdu %
a. (0=uneducated; 1=high school; 2=college, 3=post-graduation, 4=PhD )
3. English Acquisition: home vs. school vs. other
a. learnEngHome
b. learnEngSchool
c. learnEngOther
4. Type of schooling (convent, gov’t):
a. Public, non-government School: schoolPub
b. Government school: schoolGov
c. Boarding School (live away): schoolBoard
5. English vs. Hindi Medium schooling
a. schoolEngMed
b. schoolHindiMed
361
c. schoolOtherMed
6. English Education-why and justification of how it would help/hinder them to have
Eng/Hindi medium school: engSchoolingBenefits
7. Lg rules at school: disallowed from speaking X at school: oneMediumBetter
8. Ditched L1/family L1 (only code if yes): l1Loss
9. Education level
a. Primary: schoolLevPrimary
b. High school: schoolLevHS
c. College: schoolLevCollege
d. Post-grad: schoolLevPostGrad
e. DR: schoolLevelDr
10. Other languages spoken by informant, be they fluent or otherwise.: L3
11. Parents bad at English: parentsDontSpeakEngWell
Language and codeswitching: CS
1. Reflections on codeswitching: why do people switch, do they think of it as
modern, as lazy, as youth thing, or what?
a. whoCS
b. whyCS
c. attTowardsCS
2. Hinglish: within codeswitching/mixing, explicit references to Hinglish: hinglish
3. Understanding of it as mixing vs. switching vs. another thing: whatIsCSing
362
4. Domain appropriateness: Feelings towards language x in domain y as appropriate/
inappropriate—if too many, divide up by language.
a. whenLgAppropriate
b. whenLgInappropriate
5. Interviewer=Interviewee: Codemixing within interview or shared knowledge w/in
interview: sharedKBwVC
Language and cultural transmission: CultTran
1. Cultural maintenance: Importance of Hindi (or other lg):
a. +HindiLinks2Culture
b. -HindiLinks2Culture
c. +OtherLgLinks2Culture
d. -OtherLgLinks2Culture
2. Other cultural practices that they’re up to (Indian dance, etc): culturalPractices
3. Abandoned native lg (e.g. did not teach kids Punjabi, etc.): ditchL1
4. Lg and future children: comfort level for passing on Hindi, etc. and motivations
for passing this on
a. easilyPassHindiToG2
b. hardPassHindiToG2
5. Language wishes: wish I knew lg x b/c of Y: wishLearnedLgX
6. Mixed marriage (w/in India, w/ foreigner)—anyone informant mentions
a. marriedForeigner
b. marriedDiffCommunity
363
Demographic Coding: DG
1. Sex
a. Male: genderM
b. Female: genderF
2. Ageage (1, 2, 3, 4)
a. 18-24: age1
b. 25-38: age2
c. 39-59: age3
d. 60+: age4
3. Cultural/Ethnic Background
a. Delhi: ethD
b. Punjabi: ethP
c. Bengali: ethB
d. UP: ethU
e. Bihar/East: ethE
f. Bombay: ethBom
g. Pakistan: ethPak
h. Rajasthan: ethR
i. Mixed: ethM
j. Other: ethO
4. Nuclear Family vs. joint family (current state)
a. famN
b. famJ
364
5. Special oddball mark for my few who fall outside the original demographic lines
(m73ks, f25gs, f39gs, f73tm)
a. oddball
6. Religion
a. Hindu: religH
b. Sikh: religS
c. Muslim: religM
d. Buddhist: religB
e. Atheist: religA
f. Generic religousness: religionGeneric
7. Career: career (R, SR, NA, b4kids, Gov't, Priv, Edu)
a. Retired: careerR
b. Semi-Retired (now private, was public): careerSR
c. stay at home, never worked: careerNA
d. stay at home, worked till marriage/kids: careerb4kids
e. working: Nationalized work (gov’t service, armed forces): careerGovt
f. working: Private Sector: careerPriv
g. in school/student: careerEdu
h. career volunteer:careerVolunteer
8. future plans (work, marriage, etc.) futurePlans
9. Familial/Social resistance to future plans: resistanceToFuturePlans
365
Gender and links to life choices, practices or behavior: Gender
1. Gendered choices for life steps: studied x b/c girls study that, got married b/c
dangerous to be unmarried girl, etc. :
a. bcGirlDidX
b. bcBoyDidX
2. Gendered choices for language steps: learned lg x b/c in-laws spoke:
a. husbandFamilyLgAcq
b. wifeFamilyLgAcq
3. Modernity and Changes in Lifestyle or practices related to gender. Modern
4. Boys speak more Hindi, girls more English: boyHindiGirlEng
Indian English: IE
1. English in India as separate from other englishes: ieExists
2. Influence of other Eng on IE.
a. AEinfluencesIE
b. RPinfluencesIE
3. IE speech: positive/negative examples
a. +IEdesc
b. –IEdesc
c. neutralIEdesc
4. Future of IE in India: IEfuture
5. All Eng accents neutral: allEngAccNeutral
6. Eng accents can be ranked: canRankAccents
366
Key concepts: KW
1. ‘Mother tongue’ references: motherTongue
2. native: native
3. fluent: fluent
4. ‘thinking’ in lg x as indication of nativity: thinkInLgX
5. World Englishes: worldEng
6. Language Terms—
a. Standard English: standardEng
b. Hinglish: hinglish
c. Indian English: indianEng
d. ‘good english’: goodEng
e. ‘accented English’: accentedEng
f. ‘neutral English’ : neutralEng
Language attitudes: LgAtt
1. Lg learning/multilingualism as a natural process:multilingNormal
2. Lg practices/descriptions/ideologies attributed to age groups
a. youngPplLg
b. oldPplLg
3. Hindi Class-stopped taking Hindi when allowed to quit it and why:
stopHindiClass
4. Hindi Class- continued in Hindi class after it was no longer compulsory:
continuedHindiClass
367
5. Feelings about Hindi: hindiFeelings
6. Feelings about English: englishFeelings
7. Hindi fluency as important: hindiFluencyImpt
8. Hindi fluency as unimportant: hindiFluencyUnimpt
9. English literacy- any discussion of how its used, learned: englishLiteracy
10. Other lg fluency impt: otherLgFluencyImpt
11. Other lg fluency unimpt: otherLgFluencyUnimpt
12. Hindi literacy as important: hindiLiteracyImpt
13. Hindi literacy unimportant: hindiLiteracyUnimpt
14. other vernacular literacy as important: otherLgLiteracyImpt
15. other vernacular literacy unimportant: otherLgLiteracyUnimpt
Language Domains: LgDom
(Eng/Vernacular/Mixed): LgDom: these are all scalar, w/ possible answer of E, H, M O P
(English, Hindi, Mixed, OtherLg, Punjabi)
1. Lg used at home/intimate setting: scalar: homeLg %
a. homeLg %E
b. homeLg %H
c. homeLg %M
d. homeLg %O
e. homeLg %P (…., and did same for each item in lgDom through 16)
2. Lg w/ mother: wMomLg %
3. Lg w/ father: wDadLg %
368
4. Lg w/ siblings: wSiblings %
5. Lg w/ children: wKids %
6. Lg w/ spouse: wSpouse %
7. Lg used at work/school wkSklLg %
8. Lg used w/ friends: wFriendsLg %
9. Lg used w/ extended (older) family: wOlderFam %
10. Lg used w/ NRI family: wNRIfam %
11. Lg used for formal: formalSettingLg %
12. informal setting: informalSettingLg %
13. Lg used for joking, swearing (separate?): jokingSwearingLg %
14. Lg used when get angry or upset: emotionalLg %
15. Lg used w/ servants, wServantsLg %
16. in market: inMarketLg %
17. in particular region: regionalLg %
18. Lg used on internet: internetLg %
19. Lg used for reading : reading %(E,H,O)
20. Other languages used in home (but not by informant): otherLgsSpokenInHome
21. When codeswitching, Eng dominant: codeswitchEngDom
22. vs. Hindi dominant: codeswitchHindiDom
23. Interactions w/ non-hindi, non-Eng community: do they have a place for other Lg
usage? Do they speak vernacular w/ anyone?: nonHindiEngInteractions
24. other domains that are English focused: english
25. other domains that are hindi focused: hindi
369
Language and globalization: LgGlob
1. English= world/universal lg: engIsWorldLg
2. English=success: explicit links between English and job opportunity, success,
prosperity, confidence, etc.: engEqualsSuccess
3. Hindi =/= success; as a limiting factor, if only monolingual hindi speaker:
hindiNotSuccess
4. Education (medium irrelevant) equals success, more money.:
educationEqualsSuccess
5. Pride in being an Indian, expressions that link to how ppl should be proud of
being Indian, having own way of talking, own culture. indianPride
6. Colonial References: colonial
7. English Media Influence/Input:
a. local IE media: domIEmedia
b. AE : domAEmedia
c. RP: domRPmedia
8. Changes across lifespan in media and languages : changesInMediaAccess
9. changes across lifespan in educational practices: changesInEducation
10. Media type
a. Newspaper: newspaperMediaEng
b. Newspaper: newspaperMediaHindi
c. Tv: tvMediaEng
d. TV: tvMediaHindi
e. Radio: radioMediaEng
370
f. Radio: radioMediaHindi
g. Internet: internetMedia
11. ‘aping the west’ references through fake accents: apingWest
12. +/- opinions of the western influence on India today, etc.
a. +WesternInfluence
b. -WesternInfluence
13. Modernity, liberalness and Changes in Lifestyle or practices that are unrelated to
gender: modernLiberal
14. References to traditional, ‘backwards’ or conservative factions that are not
directly related to gender: traditionalConservative
Language and nationhood: LgNation
1. Indian nationality: explicit descriptions of Indian nationality and how it is or isn’t
tied to Hindi lg/culture: India=Hindi
2. Where English is appropriate at national/gov’t level: engInGovt
3. Regional differences w/ respect to language preferences: regionalLgPreferences
4. Lg diversity of India: indianLgDiversity
5. English as national lg?: engNationalLg
6. Class distinctions for lg practices: lgAndClass
7. Urban/rural dichotomy w/ respect to Eng. Knowledge/fluency: urbanRural
8. S. Indians all know Hindi: southIndiansKnowHindi
9. S. Indians know more Eng: southIndiansKnowEng
10. Partition: references (personal or familial) to partition: partition
371
11. Partition motivation: references to Partition as motivation for doing x or y,
learning x or y, caring about x or y. : partitionAsMotivation
12. Future migration plans: futureMigrationPlans
Language Now: LgNow
1. English level/quality (self-described): personalEngLevel
2. English style (self described or self reported descriptions made by 3rd party): e.g.,
accented, Punjabified, neutral, etc. engStyleDesc
3. Hindi style (accented, punjabified, English accent in Hindi): hindiStyleDesc
4. Hindi level (personal assessment of hindi level: personalHindiLevel
5. Shuddh Hindi vs. informant’s Hindi (how correctly they feel they speak Hindi):
shuddhHindi
6. Changes in language practice over lifetime: lgChoiceLifespanChanges
7. Labels: capturing distinctions between what one is called and what one considers
oneself: personalOutsideAssesmentDiff
8. Literacy levels in various languages
a. literacyEng
b. literacyHindi
c. literacyOther
9. Own accent as malleable for identity or location: personalAccentMalleable
10. direct comments on pronunciation.: pronunciation
372
Diaspora reflections: NRI
1. Have family living outside India. familyOut
2. NRIs: accents, habits, how are they understood-as locals or foreigners: behavior
3. NRI loss of vernacular lg/failure to learn vernacular lg.: noVernacularLgSkills
4. No loss of vernacular language skills from living abroad (anyone):
vernacularLgSkillsGood
5. NRI’s (and their kids) should learn vernacular, even if living abroad:
shouldLearnVernacular
6. NRI/foreign-returned ppl discussion:
a. seenAsForeign
b. seenAsLocal
7. NRI accents/habits as “more Indian than locals” and what that may mean for later
studies (not so related to this study): moreIndThanLocals
8. Mistaken for NRI b/c of lg practices/attitudes: mistaken4NRI
Delhiites and local/international travel: RegAtt
1. Delhiite definitions, characteristics, who comes to Delhi, etc.: Delhiites
2. Born/brought up outside Delhi- if so, discussions of why migrated, and how they
feel about it : whyShiftedToDelhi
3. Discussion of traveling within India-what its like, whether it feels ‘similar’ or
united, what lg one needs for such travel: domesticTravel
4. Regular visits outside India: why they go, where they go, what that says about
their understanding of local dialects): intlTravel
373
Attitudes towards regional language variation: RegAttLg
1. Imitations/descriptions of different intra-Indian English accents
a. delhiAccent
b. southAccent
c. westAccent
d. centralAccent (Maharashtra, Madhya Pradesh)
e. bombayAccent
f. punjabiAccent
g. bengaliAccent
h. gujaratiAccent
i. chennaiAccent
j. jNk (jammu and Kashmir)
k. otherDomesticAccent
2. Attitudes linking verbal style w/ group characteristics (e.g. Bengalis are ‘X’
because their speech is drawn out or S. Indians are ‘Y’, because you hear that in
their speech): personalityLgLinks
16. Imitations/descriptions of different non-Indian English accents:
a. AEaccent
b. RPaccent
c. otherIntlAccent
17. Hindi Belt references-both linguistic and attitudes towards it: hindiBelt
18. English accent jokes (e.g. if a __ says ‘m’ it means __):
a. IEaccentJokes
374
b. intlEngAccentJokes
19. Special slang related to particular region or group: slang
20. imitations of how groups speak Hindi: hindiImmitations
Random Coding: Rndm
1. Explicit grammar references: grammarRef
2. Lexical Fillers: lexicalFillers
3. stative+progressive instances: progressive
4. discussion of rhotics: rhotic
5. discussion of w/v: w.v
6. Life events w/ respect to normal schedule (she got married on time): lifestage
7. Intelligibility breakdowns, cross-dialectally: intelligibility
8. Instances of VC/informant comprehension breakdowns: huhVC
9. references to anglo-indians.: angloIndian
10. references to call centres: callCentre
11. BE LIKE or BE ALL quotatives: quotative
Inter-interview difference in speaker stance: SpkrStnc
1. Discussion of traditional/normal vs. transgression/rebellion/alternative life choice:
a. traditionalChoice
b. transgressionChoice
2. Indian vs. narrower ethnic identity, e.g. Gujarati, Punjabi, discussion:
a. indianIdentity
375
b. delhiIdentity
c. northIdentity
d. gujaratiIdentity
e. punjabiIdentity
f. bombayIdentity
g. bengaliIdentity
h. modernIdentity
i. traditionalIdentity
j. casteIdentity
3. Parent vs. child understandings/framings of story:
a. parentPerspective
b. childPerspective
4. Work vs. family framing:
a. workPerspective
b. homePerspective
376
APPENDIX E
WORD CLOUD DEMONSTRATING IE WORD FREQUENCY
VISUALLY
377
APPENDIX F
WORD CLOUD DEMONSTRATING IE (r) WORD FREQUENCY
VISUALLY
378
APPENDIX G
WORD CLOUD DEMONSTRATING IE (V) WORD FREQUENCY
VISUALLY
379
APPENDIX H
WORD CLOUD DEMONSTRATING IE (W) WORD FREQUENCY
VISUALLY
380
REFERENCES
Accent Reduction Institute, The. 2008. Accent Reduction Institute website main page.
Accessed February 15, 2008. http://www.lessaccent.com/.
Acrolect Technologies, Pvt Ltd. 2009. Acrolect Technologies Website. Accessed 17
April, 2009. http://www.acrolect.co.in/.
Agnihotri, Rama Kant. 1994. Sound patterns of Indian English: A sociolinguistic
perspective. In Rama Kant Agnihotri and A.L. Khanna (eds.), Second language
acquisition: Socio-cultural and linguistic aspects of English in India. 235-246.
New Delhi: Sage Publications.
Agnihotri, Rama Kant and Sahgal, Anju. 1985. Is Indian English retroflexed and r-full?
Indian Journal of Applied Linguistics 11: 97-109.
Allbritten, Rachael. 2006. The anatomy of a sociolinguistic interview. Presented at New
Ways of Analyzing Variation (NWAV 35): Columbus, OH.
Annamalai, E. 1998. Nativity of language. In Rajendra Singh (ed.), The native speaker:
Multilingual perspectives. New Delhi: Sage Publications.
Annamalai, E. 2004. Nativization of English in India and its effect on multilingualism.
Journal of Language and Politics 3(1): 151-162.
Anthony, Laurence. 2007. AntConc. 3.2.1w.
http://www.antlab.sci.waseda.ac.jp/antconc_index.html.
Ash, S. 2002. Social class. In J. K. Chambers, Peter Trudgill and Natalie Schilling-Estes
(eds.), The handbook of language variation and change. 402-422. Oxford:
Blackwell.
381
Bailey, Guy. 2002. Real and apparent time. In J. K. Chambers, Peter Trudgill and Natalie
Schilling-Estes (eds.), The handbook of language variation and change. 312-332.
Oxford: Blackwell.
Bakhtin, Mikhail M. 1981. The dialogic imagination. Austin: University of Texas Press.
Baldridge, Jason. 2002. Linguistic and social characteristics of Indian English. Language
in India. Accessed December 15, 2006.
http://www.languageinindia.com/junjul2002/baldridgeindianenglish.html.
Bansal, R.K. 1983. Studies in phonetics and spoken English. Hyderabad: CIEFL.
Bansal, R.K. 1990. The pronunciation of English in India. In Susan Ramsaran (ed.),
Studies in the pronunciation of English. 219-233. London: Routledge.
Basu, Indrajit. 2004. The faces of globalization: A dilemma for India. United Press
International. Calcutta. March 12, 2004.
Bauman, Richard and Briggs, Charles L. 1990. Poetics and performance as critical
perspectives on language and social life. Annual Review of Anthropology 19: 5988.
Bayard, Donn, Gallois, Cynthia, Weatherall, Ann and Pittam, Jeffery. 2001. Pax
Americana? Accent attitudinal evaluations in New Zealand, Australia and
America. Journal of Sociolinguistics 5(1): 22-49.
Bayley, Robert. 2002. The quantitative paradigm. In J. K. Chambers, Peter Trudgill and
Natalie Schilling-Estes (eds.), The handbook of language variation and change.
117-141. Oxford: Blackwell.
382
Bayley, Robert, Lucas, Ceil and Rose, Mary. 2002. Phonological variation in American
Sign Language: The case of 1 handshape. Language variation and change 22(1):
19-53.
Baynes, Kathleen, Farias, Sarah, Chand, Vineeta and Bonnici, Lisa. m.s. Degradation of
Idea Density in Alzheimers narratives: Methods for analysis and conclusions.
Davis.
Beecher, Henry. n.d. Stem faithfulness in the derivational morphology of Italian.
Linguistics Department San Diego: UC San Diego.
Berg, Thomas. 1999. Stress variation in British and American English. World Englishes
18(2): 123-143.
Bhatia, Tej K. 2006. Super-heroes to super languages: American popular culture through
South Asian language comics. World Englishes 25(2): 279-297.
Bhatia, Tej K. and Ritchie, William C. 2001. Language mixing, typology, and second
language acquisition. In P. Bhaskararao and K.V. Subbarao (eds.), The yearbook
of South Asian languages and linguistics. 37-62. New Delhi: Sage Publications.
Bhatt, Rakesh M. 2000. Optimal expressions in Indian English. English Language and
Linguistics 4(1): 69-95.
Bhatt, Rakesh M. 2005. Expert discourses, local practices, and hybridity: The case of
Indian Englishes. In A. Suresh Canagarajah (ed.), Reclaiming the local in
language policy and practice. 25-54. Mahwah, NJ: Lawrence Erlbaum
Associates.
383
Biber, Douglas, Johansson, Stig, Leech, Geoffrey, Conrad, Susan and Finegan, Edward.
1999. Longman grammar of spoken and written English. Harlow: Longman
Publications Group.
Blackburn, Mollie V. 2005. Agency in borderland discourses: examining language use in
a community center with black queer youth. Teachers College Record 107(1): 89113.
Blake, Renee and Josey, Meredith. 2003. The /ay/ diphthong in a Martha's Vineyard
community: What can we say 40 years after Labov? Language in Society 32: 451485.
Blommaert, Jan. 2001. Investigating narrative inequality: African asylum seekers' stories
in Belgium. Discourse & Society 12: 413-449.
Blommaert, Jan. 2003. Commentary: A sociolinguistics of globalization. Journal of
Sociolinguistics 7(4): 607-623.
Bloomfield, Leonard. 1933. Language. New York: Henry Holt.
Boas, Franz. 1911. Introduction. In Franz Boas (ed.), Handbook of American Indian
Languages. Washington: Government Print Office (Smithsonian Institution,
Bureau of American Ethnology).
Boberg, Charles. 2004. Real and apparent time in language change: Late adoption of
changes in Montreal English. American Speech 79(3): 250-269.
Boersma, Paul and Weenick, David. 2006. Praat: doing phonetics by computer. 4.0.30.
http://www.praat.org.
Bonasera, Fausto and Maffi, Anna. 2004. Mother Tongue. Terralingua.
384
Bonnici, Lisa and Chand, Vineeta. 2007. When talk becomes .doc: Issues in the clinical
analysis of Alzheimer’s speech. Presented at UC Davis Symposium: The Curious
Lives of Documents: UC Davis, CA.
Bourdieu, Pierre. 1991. Language and symbolic power. Cambridge: Polity Press.
Briggs, Charles L. 1986. Learning how to ask: A sociolinguistic appraisal of the role of
the interview in social science research. Cambridge, UK: Cambridge University
Press.
Bucholtz, Mary. 1999. "Why be normal?" Language and identity practices in a
community of nerd girls. Language in Society 28: 203-223.
Bucholtz, Mary. 2003. Sociolinguistic nostalgia and the authentication of identity.
Journal of Sociolinguistics 7(3): 398-416.
Bucholtz, Mary and Hall, Kira. 2008. All of the above: New coalitions in sociocultural
linguistics. Journal of Sociolinguistics 12(4): 401-431.
Bybee, Joan. 2002. Word frequency and context of use in the lexical diffusion of
phonetically conditioned sound change. Language variation and change 14: 26190.
Cameron, Deborah. 2000. Styling the worker: Gender and the commodification of
language in the globalized service economy. Journal of Sociolinguistics 4(3):
323-347.
Cameron, Deborah. 2001. Working with spoken discourse. London: Sage Publications.
Canagarajah, A. Suresh. 2006. Negotiating the local in English as a Lingua Franca.
Annual Review of Anthropology 26: 197-218.
385
Caplan, David. 1987. Neurolinguistics and linguistic aphasiology. Cambridge:
Cambridge University Press.
Chafe, Wallace. 1975. The Pear Story. Accessed March 1, 2007.
http://www.pearstories.org.
Chafe, Wallace (ed.) 1980. The pear stories: Cognitive, cultural and linguistic aspects of
narrative production. Norwood: Ablex Publishing.
Chafe, Wallace. 1985. Linguistic differences produced by differences between speaking
and writing. In D. R. Olson, A. Hildyard and N. Torrance (eds.), Literacy,
language, and learning. 105-123. Cambridge: Cambridge University Press.
Chafe, Wallace. 1994. Discourse, consciousness, and time. Chicago: University of
Chicago Press.
Chafe, Wallace. 2007. The Pear Film. Accessed March 1, 2007,
http://www.linguistics.ucsb.edu/faculty/chafe/pearfilm.htm.
Chafe, Wallace and Tannen, Deborah. 1987. The relation between written and spoken
language. Annual Review of Anthropology 16: 383-407.
Chambers, Jack K. 2003. Sociolinguistic theory: Linguistic variation and its social
significance. Oxford: Blackwell.
Chand, Tara. 1944. The problem of a common language for India. In Tara Chand (ed.),
The Problem of Hindustani. 13-40. Allahabad: Indian Periodicals Ltd.
Chand, Vineeta. 2007. The acoustics and cross-dialect intelligibility of [v] and [w]
Qualifying Research Paper. Linguistics, UC Davis: Davis
386
Chand, Vineeta. 2008. Hinglish identity: Changes in language practices and attitudes in
urban India. Presented at The 14th Conference of the International Association for
World Englishes: City University of Hong Kong.
Chand, Vineeta. in press. [v]at is going on? Local and global ideologies about Indian
English. Language in Society 38(4)
Chand, Vineeta and Gales, Tammy. m.s. The importance of word frequency: From
variationist problems to corpus-based solutions. Department of Linguistics Davis:
University of California, Davis.
Chaturvedi, M.G. . 1973. A contrastive study of Hindi-English phonology. Delhi:
National Publishing House.
Chaudhary, Rakesh. 2001. 'Aati kyaa New York?' - a Nampally monologue. English
Today 65(17/1): 27-30.
Chelliah, Shobhana L. 2001. Constructs of Indian English in language 'guidebooks'.
World Englishes 20(2): 161-177.
Chelliah, Shobhana L. 2006. The representation of Indian English in Indian English
novels. Presented at South Asian Languages Analysis (SALA) 26: Kannada
University & Central Institute of Indian Studies, Mysore.
Chomsky, Noam. 1955. Transformational analysis. Ph.D. dissertation. Department of
Linguistics, University of Pennsylvania: Philadelphia
Chomsky, Noam. 1957. Syntactic structures. The Hague: Mouton.
Chun, Elaine. 2007. 'Oh my god!': Stereotypical words at the intersection of sound,
practice, and social meaning. Presented at New Ways of Analyzing Variation
(NWAV 36): Philadelphia, PA.
387
Clark, Lynn and Trousdale, Graeme. 2009. Exploring the role of token frequency in
phonological change: Evidence from TH-fronting in east-central Scotland.
English Language and Linguistics 13(1): 33-55.
Coelho, Gail M. 1997. Anglo-Indian English: A nativized variety of Indian English.
Language in Society 26(4): 561-589.
Coupland, Nikolas, Bishop, Hywel, Evans, Betsy and Garrett, Peter. 2006. Imagining
Wales and the Welsh language. Journal of Language and Social Psychology
25(4): 351-376.
Cowie, Claire. 2007. The accents of outsourcing: the meanings of "neutral" in the Indian
call centre industry. World Englishes 26(3): 316-330.
Crystal, David. 2005. Indian English. Radio National: Lingua Franca. Accessed
February 7, 2007. http://www.abc.net.au/rn/arts/ling/stories/s1298284.htm.
Cukor-Avila, Patricia and Bailey, Guy. 2001. The effects of the race of the interviewer on
sociolinguistic fieldwork (Research Notes). Journal of Sociolinguistics 5(2): 254270.
D'souza, Jean. 1997. Indian English: Some myths, some realities. English World-Wide
18(1): 91-105.
Das, Anupam and Subbarao, K.V. 2008. Echo-word and politeness: A socio-pragmatic
aspect of South Asia. Presented at New Ways of Analyzing Variation (NWAV 37):
Rice University.
Detaramani, Champa and Lock, Graham. 2003. Multilingualism in decline: Language
repertoire, use and shift in two Hong Kong Indian communities. Journal of
Multilingual and Multicultural Development 24(4): 249-273.
388
Downes, William. 1998. Language in Society. Cambridge: Cambridge University Press.
Doyle, Tricia. 1998. Western Education in Nineteenth-Century India. Modern Literary
Studies, Queen's University of Belfast: Belfast
Du Bois, John, Schuetze-Coburn, Stephan, Cumming, Susanna and Paolino, Danae. 1993.
Outline of Discourse Transcription. In Jane A. Edwards and Martin D. Lampert
(eds.), Talking Data: Transcription and coding in discourse research. 45-89.
Hillsdale, New Jersey: Lawrence Erlbaum.
Dubois, Sylvie and Horvath, Barbara. 2000. When the music changes, you change too:
Gender and language change in Cajun English. Language variation and change
11: 287-313.
Dudley, Brier. 2004. Microsoft's call-center business in India gets an American accent.
Seattle Times. Seattle.
Eckert, Penelope. 2004. Variation and a sense of place. In Carmen Fought (ed.),
Sociolinguistic variation: Critical reflections. 107-118. Oxford, U.K.: Blackwell.
Eckert, Penelope. 2008. Variation and the indexical field. Journal of Sociolinguistics
12(4): 453-476.
Edwards, Jane A. 1993. Principles and contrasting systems of discourse transcription. In
Jane A. Edwards and Martin D. Lampert (eds.), Talking data: Transcription and
coding in discourse research. 3-31. Hillsdale: Lawrence Erlbaum Associates.
Edwards, John R. 1982. Language attitudes and their implications among English
speakers. In Ellen Bouchard Ryan and Howard Giles (eds.), Attitudes towards
language variation. 20-33. London: Arnold.
389
Emeneau, Murray B. 1980 [1956]. India as a Linguistic Area. In Anwar S. Dil (ed.),
Language and linguistic area: essays by Murray B. Emeneau. Stanford: Stanford
University Press.
Erbaugh, Mary S. 2001. The Chinese Pear Stories--Narratives across seven Chinese
dialects. Accessed March 1, 2007. http://pearstories.org/.
Fabian, Johannes. 1986. Language and colonial power. Berkeley: University of
California Press.
Feagin, Crawford. 1990. The dynamics of a sound change in Southern States English:
From r-less to r-ful in three generations. In Jerold A. Edmondson, Crawford
Feagin and Peter Muhlhausler (eds.), Development and diversity: Language
variation across time and space. A festschrift for Charles-James N. Bailey. 129146. Arlington: The Summer Institute of Linguistics & The University of Texas at
Arlington.
Feagin, Crawford. 2002. Entering the community: Fieldwork. In J. K. Chambers, Peter
Trudgill and Natalie Schilling-Estes (eds.), The handbook of language variation
and change. Oxford: Blackwell.
Figueroa, N.L. 2004. 'U.S. and 'them': A study of the language attitudes of speakers of
high- and low-prestige varieties of Spanish toward 'World Spanishes'. Ann
Arbor: University of Michigan.
Fishman, Joshua A. 1965. Yiddish in America: Socio-linguistic description and analysis.
Bloomington: Indiana University Press.
Fox, Aaron A. 2004. Real country: Music and language in working-class culture.
Durham: Duke University Press.
390
French, Brigittine M. 2001. The symbolic capital of social identities: The genre of
bargaining in an urban Guatemalan market. Journal of Linguistic Anthropology
10(2): 155-189.
Gal, Susan and Irvine, Judith T. 1995. The boundaries of languages and disciplines: How
ideologies construct difference. Social Research 62(4): 967-1001.
Gargesh, Ravinder. 2004. Indian English: phonology. In Edgar W. Schneider, Kate
Burridge, Bernd Kortmann, Rajend Mesthrie and Clive Upton (eds.), A handbook
of varieties of English: Phonology. 992-1002. Berlin: Mouton de Gruyter.
Giles, Howard and Coupland, Nikolas. 1991. Language, contexts & consequences.
Buckingham: Open University Press.
Gippert, Jost. 2002. Languages of India. Accessed November 28, 2007, 2007.
http://titus.uni-frankfurt.de/didact/karten/indi/indiega.jpg.
Glaser, Barney G. and Strauss, Anselm. 1967. Discovery of Grounded Theory: Strategies
for qualitative research Chicago: Aldine Publishing Company.
Gordon, Elizabeth, Campbell, Lyle, Hay, Jennifer, Maclagan, Margaret, Sudbury, Andrea
and Trudgill, Peter. 2004. New Zealand English: Its origins and evolution.
Cambridge: Cambridge University Press.
Gordon, James and Gupta, Poonam. 2004. Understanding India's services revolution.
International Monetary Fund: New Delhi.
http://imf.org/external/pubs/ft/wp/2004/wp04171.pdf
Gordon, Raymond G., Jr. (ed.) 2005. Ethnologue: Languages of the world. Dallas: SIL
International.
391
Gorman, Anna. 2007. Accentuating the 'American' in their speech. Los Angeles Times.
Los Angeles. October 23, 2007.
Gramley, Stephan and Patzold, Kurt-Michael. 2003. A survey of modern English.
London: Routledge.
Gumperz, John. 1961. Types of linguistic communities. Anthropological Linguistics 4(1):
28-40.
Guy, Gregory R. 1980. Variation in the group and in the individual. In William Labov
(ed.), Locating language in time and space. 1-36. New York: Academic Press.
Guy, Gregory R. 1991. Contextual conditioning in variable lexical phonology: An
exponential model of morphological constraints. Language variation and change
3: 223-239.
Hammond, Michael. 1999. The phonology of English: A prosodic optimality-theoretic
approach. Oxford: Oxford University Press.
Harris, John. 2006. Wide-domain r-effects in English. UCLA Working Papers in
Linguistics 18: 1-24.
Heath, Jeffrey. 1979. Diffusional linguistics in Australia: Problems and prospects. In S.A.
Wurm (ed.), Australian linguistic studies. Canberra: Department of Linguistics,
Research School of Pacific Studies.
Heath, Jeffrey. 1984. Language contact and language change. Annual Review of
Anthropology 13: 367-384.
Heller, Monica. 2003. Globalization, the new economy, and the commodification of
language and identity. Journal of Sociolinguistics 7(4): 473-492.
392
Heller, Monica. 2008. Language and the nation-state: Challenges to sociolinguistic theory
and practice. Journal of Sociolinguistics 12(4): 504-524.
Herat, Manel. 2005. BE variation in Sri Lankan English. Language variation and change
17: 181-208.
Hoare, Rachel. 2001. An integrative approach to language attitudes and identity in
Brittany (Research Notes). Journal of Sociolinguistics 5(1): 73-84.
Hohenthal, Annika. 1998. English in India: Loyalty and attitudes. M.A. Thesis. Language
In India, University of Turku, Finland: Turku
Horvath, B and Sankoff, D. 1986. Delimiting the Sydney speech community. Language
in Society 16: 179-204.
Hosali, Priya. 2005. Butler English. English Today 81(21/1): 34-39.
Hosali, Priya and Atchison, Jean. 1986. Butler English: A minimal pidgin? Journal of
Pidgin and Creole Languages 1(1): 51-79.
Hymes, Dell. 1974. Foundations in sociolinguistics: An ethnographic approach.
Philadelphia: University of Pennsylvania Press.
Ito, Rika. 2001. Belief, attitudes, and linguistic accommodation: A case of urban sound
change in rural Michigan. University of Pennsylvania Working Papers in
Linguistics 7(3): 129-143.
Jenkins, Jennifer. 2000. The phonology of English as an international language. Oxford:
Oxford University Press.
John, Binoo K. 2007. Entry from backside only: Hazaar fundas of Indian-English. New
Delhi: Penguin Books.
393
Johnstone, Barbara, Andrus, Jennifer and Danielson, Andrew. 2006. Mobility,
indexicality, and the enregisterments of "Pittsburghese". Journal of English
Linguistics 34(2): 77-104.
Jones, Mark J. 2005. An acoustic study of labiodental /r/ in British English. Cambridge
Occasional Papers in Linguistics 3: 100-109.
Jurafsky, Dan. 2007. Predictability effects on content versus function word pronunciation
in conversational English Presented at Workshop on Variation, Gradience and
Frequency in Phonology: Stanford University, Stanford, CA.
Jurafsky, Dan, Bell, Alan, Gregory, Michelle and Raymond, William D. 2001.
Probabalistic relations between words: Evidence from reduction in lexical
production. In Joan Bybee and Paul Hopper (eds.), Frequency and the emergence of
linguistic structure. Amsterdam: John Benjamins.
Kachru, Braj B. 1982. South Asian English. In R.W. Bailey and M. Gorlach (eds.),
English as a world language. 353-383. Ann Arbor: University of Michigan Press.
Kachru, Braj B. 1983. The Indianization of English. Oxford: Oxford University Press.
Kachru, Braj B. 1985. Standards, codification and sociolinguistic realism: The English
language in the outer circle. In R. Quirk and H. Widdowson (eds.), English in the
world: Teaching and learning the language and literatures. 11-36. Cambridge:
Cambridge University Press.
Kachru, Braj B. 1986. The Indianization of English. English Today 6: 31-34.
Kachru, Braj B. 1986b. The alchemy of English. The spread, functions and models of
non-native Englishes. Oxford: Pergamon Press.
394
Kahn, Farhat. 1991. Final consonant cluster simplification in a variety of Indian English.
In Jenny Cheshire (ed.), English around the world: Sociolinguistic perspectives.
288-298. Cambridge: Cambridge University Press.
Kandiah, Thiru. 1991. South Asia. In Jenny Cheshire (ed.), English around the world:
Sociolinguistic perspectives. 271-287. Cambridge: Cambridge University Press.
Keane, Elinor. 2004. Tamil: Illustrations of the IPA. Journal of the International
Phonetic Association 34(1): 111-116.
Kemper, Susan, Greiner, Lydia H., Marquis, Janet G., Prenovost, Katherine and Mitzner,
Tracy L. . 2001. Language decline across the lifespan: Findings from the Nun
Study. Psychology and Aging 16: 227-239.
Kemper, Susan, Kynette, Donna, Rash, Shannon, O'brien, Kevin and Sprott, Richard.
1989. Life-span changes to adults' language: Effect of memory and genre. Applied
Psycholinguistics 10: 49-66.
Kennedy, Miranda. 2007. India Calling. American Public Media.
Kerswill, Paul and Wright, Susan. 1990. The validity of phonetic transcription:
Limitations of a sociolinguistic research tool. Language variation and change 2:
255-275.
Kiesling, Scott. F. . 2001. “Now I gotta watch what I say”: Shifting constructions of
masculinity in discourse. Journal of Linguistic Anthropology 11(2): 250-273.
Knight, Rachael-Anne, Dalcher, Christina Villafana and Jones, Mark J. 2007. A real-time
case study of rhotic acquisition in Southern British English. In Jürgen Trouvain
and William J. Barry (eds.), 16th International Congress of Phonetic Sciences
1581-4. Saarland University, Saarbrücken, Germany.
395
Krishnamurti, Bh. 1978. Spelling pronunciation in Indian English. In R. Mohan (ed.),
Indian Writing in English. 129-139. Delhi: Orient Longman.
Krishnaswamy, N. and Burde, Archana S. 1998. The politics of Indians’ English:
Linguistic colonialism and the expanding English empire. Delhi: Oxford
University Press.
Kristiansen, Tore. 2003. The youth and the gatekeepers: Reproduction and change in
language norm and variation. In Jannis K. Androutsopoulos and Alexandra
Georgakopoulou (eds.), Discourse Constructions of Youth Identities. 279–302.
John Benjamins.
Ktims. 2006. Map of Punjab. Accessed November 28, 2007, 2007.
http://infinity.gotroot.ca/~ktims/punjab_map.png.
Labov, William. 1966. The social stratification of English in New York City. Washington
DC: Center for Applied Linguistics.
Labov, William. 1972a. The design of a sociolinguistic research project. Report of the
Sociolinguistics Workshop held by the Central Institute of Indian Languages.
Central Institute of Indian Languages: Mysore.
Labov, William. 1972b. Language in the inner city: Studies in the Black English
Vernacular. Philadelphia: University of Pennsylvania Press.
Labov, William. 1972c. Sociolinguistic patterns. Philadelphia: University of
Pennsylvania Press.
Labov, William. 1982. Objectivity and commitment in linguistic science: The case of the
Black English trial in Ann Arbor. Language in Society 11: 165-201.
396
Labov, William. 1989. The child as linguistic historian. Language variation and change
1: 85-97.
Labov, William. 2001. Principles of linguistic change. Oxford: Basil Blackwell.
Labov, William, Karen, Mark and Miller, Corey. 1991. Near-mergers and the suspension
of phonemic contrast. Language variation and change 3: 33-74.
Labru, G.L. 1984. Indian newspaper English. Delhi: B.R. Publishing Corporation.
Ladefoged, Peter. 1993. A course in phonetics. Orlando: Harcourt Brace and Company.
Ladefoged, Peter and Maddieson, Ian. 1996. The sounds of the world’s languages.
Oxford: Blackwell Publishers Ltd.
Lass, Roger. 1984. Phonology. Cambridge: Cambridge University Press.
Lavandera, Beatrice. 1978. Where does the sociolinguistic variable stop? Language in
Society 7: 171-183.
Lee, Jamie Shinhee. 2006. Linguistic constructions of modernity: English mixing in
Korean television commercials. Language in Society 35: 59-91.
Leech, Geoffrey. 2005. Adding linguistic annotation. In Martin Wynne (ed.), Developing
linguistic corpora: A guide to good practice. 17-29. Oxford: Oxbow Books.
Leitner, Gerhard. 1990. Divergence and similarity in English--Australian English in
contrast with Indian English. In Jerold A. Edmondson, Crawford Feagin and Peter
Mulhausler (eds.), Development and diversity: Language variation across time
and space. 323-349. Arlington, TX: Summer Institute of Linguistics
The University of Texas at Arlington.
Leitner, Gerhard. 1994. Begin and start in British, American and Indian English. Hermes
13: 99-122.
397
Levon, Ezra. 2009. Dimensions of style: Context, politics and motivation in gay Israeli
speech. Journal of Sociolinguistics 13(1): 29-58.
Lippi-Green, Rosina. 1997. English with an accent: Language, ideology, and
discrimination in the United States. London: Routledge.
Liu, Lydia H. 1999. The question of meaning-value in the political economy of the sign.
In Lydia H. Liu (ed.), Tokens of exchange: The problem of translation in global
circulations. 13-41. Durham: Duke University Press.
Llamas, Carmen. 2007. "A place between places": Language and identities in a border
town. Language in Society 36: 579-604.
Macaulay, Thomas Babington. 1920/1965. Minute on Education by the Hon'ble T. B.
Macaulay, dated the 2nd February 1835. In Delhi: National Archives of India
(ed.), 107-117. Calcutta: Superintendent, Government Printing.
Machin, David and van Leeuwen, Theo. 2003. Global schemas and local discourses in
Cosmopolitan. Journal of Sociolinguistics 7(4): 493-512.
Mahapatra, Rajesh. 2006. IBM's technology to boost English skills at call centers. IndiaWest. California. November 3, 2006.
Malkani, Gautam. 2006. Londonstani. Great Britain: Fourth Estate.
Mallikarjun, B. 2001. Language of India according to the 1991 Census. Language in
India 1(7)
Martinet, Andre. 1957. Phonetics and linguistic evolution. In L. Kaiser (ed.), Manual of
phonetics. 252-273. Amsterdam: North-Holland Publishing Company.
Masica, Colin P. 1991. The Indo-Aryan languages. Cambridge: Cambridge University
Press.
398
Maxwell, Olga and Fletcher, Janet. 2009. Acoustic and durational properties of Indian
English vowels. World Englishes 28(1): 52-69.
McGory, Julie. 2006. Using ToBI to analyze intonation patterns. Presented at New Ways
of Analyzing Variation (NWAV 35): Columbus, Ohio.
McPhate, Mike. 2005. Outsourcing outrage: Indian call-center workers suffer abuse. San
Francisco Chronicle. Thursday, November 17, 2005.
Mehrotra, Raja Ram. 1998. Indian English: Texts and Interpretation. Amsterdam: John
Benjamins.
Mehrotra, Raja Ram. 2002. English in private letters in India. English Today 72(18/4):
39-44.
Mehrotra, Raja Ram. 2003. A British response to some Indian English usages. English
Today 75(19(3)): 19-25.
Mendoza-Denton, Norma. 2002. Language and identity. In J. K. Chambers, Peter
Trudgill and Natalie Schilling-Estes (eds.), The handbook of language variation
and change. 475-499. Oxford: Blackwell.
Mendoza-Denton, Norma. 2008. Homegirls: Language and cultural practice among
Latina youth gangs. London: Wiley-Blackwell.
Mesthrie, Rajend. 2006. World Englishes and the multilingual history of English. World
Englishes 25(3/4): 381-390.
Meyerhoff, Miriam. 2006. Linguistic change, sociohistorical context and theory-building
in variationist linguistics: New dialect formation in New Zealand. English
Language and Linguistics 10(1): 173-194.
399
Meyerhoff, Miriam and Niedzielski, Nancy. 2003. The globalisation of vernacular
variation. Journal of Sociolinguistics 7(4): 534-555.
Migration Dialogue, The. 2005. India: Remittances, high-tech. Migration News Online
12(2). Accessed March 13, 2005.
http://migration.ucdavis.edu/mn/more.php?id=3104_0_3_0.
Milroy, James. 2001. Language ideologies and the consequences of standardization.
Journal of Sociolinguistics 5(4): 530-555.
Milroy, Lesley. 2002. Social networks. In J. K. Chambers, Peter Trudgill and Natalie
Schilling-Estes (eds.), The handbook of language variation and change. 549-571.
Oxford: Blackwell.
Milroy, Lesley and Gordon, Matthew. 2003. Sociolinguistics: Method and interpretation.
Oxford: Blackwell.
Mufwene, Salikoko S. 2001. The ecology of language evolution. Cambridge: Cambridge
University Press.
Muhr, Thomas. 2004. User's manual for ATLAS.ti 5.0. Berlin: ATLAS.ti Scientific
Software Development GmbH. http://www.atlasti.com/.
Myhill, John. 1988. Postvocalic /r/ as an index of integration into the BEV speech
community. American Speech 63(3): 203-213.
Nagy, Naomi and Irwin, Patricia. 2007. The return of R. Presented at New Ways of
Analyzing Variation (NWAV 36): University of Pennsylvania, Philadelphia.
Newbrook, Mark M. 1998. The attitudes and beliefs of some educated Malaysians with
respect to grammatical and lexical features of Malaysian English. Te Reo 41: 163177.
400
Ngom, Fallou. 2003. The social status of Arabic, French and English in the Senegalese
speech community. Language variation and change 15: 351-368.
Nichols, John. 2004. Voters in India say no to globalization. Madison Capital Times.
Madison, WI. May 18, 2004.
Nihalani, Paroo, Tongue, R.K. and Hosali, Priya (eds.) 1979. Indian and British English:
A handbook of usage and pronunciation. Delhi: Oxford University Press.
Ohala, John J. and Lorentz, James. 1977. The story of [w]: An exercise in the phonetic
explanation for sound patterns. Berkeley Linguistics Society 577 - 599. Berkeley.
Ohala, Manjari. 1983. Aspects of Hindi phonology. Delhi: Motilal Banarsidass.
Orgun, C. Orhan. 2002. Review of Giergerich, Heinz. 1999. Lexical strata in English:
Morphological causes, phonological effects. Cambridge: Cambridge University
Press. Journal of Germanic Linguistics 14(2): 177-85.
Pandey, Anita. 2004. Culture, gender, and identity in cross-cultural personals and
matrimonials. World Englishes 23(3): 403-427.
Pandey, Pramod. 1994. On a description of the phonology of Indian English. In R.K.
Agnihotri and A.L. Khanna (eds.), Second language acquisition: Socio-cultural
and linguistic aspects of English in India. 198-207. New Delhi: Sage Publications.
Paolillo, John C. 2002. Analyzing linguistic variation: Statistical models and methods.
Stanford: CSLI Publications.
Park, Joseph Sung-Yul and Wee, Lionel. 2008. Appropriating the language of the other:
Performativity in autonomous and unified markets. Language & Communication
28: 242-257.
401
Pattanayak, D.P. 1998. Mother tongue: An Indian context. In Rajendra Singh (ed.), The
native speaker: Multilingual perspectives. New Delhi: Sage Publications.
Pennycook, Alastair. 2007. Global Englishes and transcultural flows. London:
Routledge.
Pickering, Lucy and Wiltshire, Caroline. 2000. Pitch accent in Indian-English teaching
discourse. World Englishes 19(2): 173-183.
Piercy, Caroline. 2007. A quantitative analysis of rhoticity in Dorset: Evidence from four
locations of a urban to rural hierarchy of change. Presented at CamLing 2007
University of Cambridge.
Preston, Dennis R. 1991. Sorting out the variables in sociolinguistic theory. American
Speech 66: 33-56.
Preston, Dennis R. (ed.) 1999. Handbook of perceptual dialectology. Amsterdam: John
Benjamins.
Preston, Dennis R. 2002. Language with an attitude. In J. K. Chambers, Peter Trudgill
and Natalie Schilling-Estes (eds.), The handbook of language variation and
change. 40-66. Oxford: Blackwell.
Preston, Dennis R. under review. The South: Still different. In Michael Picone and
Catherine Davies (eds.), Language and variety in the South III. Tuscaloosa:
University of Alabama Press.
Purnell, Thomas, Idsardi, William and Baugh, John. 1999. Perceptual and phonetic
experiments on American English dialect identification. Journal of Social
Psychology 18(1): 10-30.
402
Rahman, Tariq. 2009. Language ideology, identity and the commodification of language
in the call centers of Pakistan. Language in Society 38(3): 233-58.
Rai, Saritha. 2005. India's boom spreads to smaller cities. The New York Times. New
York. January 4, 2005.
Raj, Dhooleka S. 2003. Where are you From? Middle-class migrants in the modern
world. Berkeley: University of California Press.
Ramanathan, Vaidehi. 2005. The English Vernacular divide: Postcolonial language
politics and practice. Clevedon: Multilingual Matters Ltd.
Rampton, Ben. 1995. Crossing: Language and ethnicity among adolescents. London:
Longman.
Randall, Andrew. 2007. The British Empire: India. Accessed November 28, 2007, 2007.
http://freespace.virgin.net/andrew.randall1/india.htm.
Rao, Raja. 1978. The caste of English. In C. D. Narasimhaiah (ed.), Awakened
conscience: Studies in commonwealth literature. New Jersey: Humanities Press.
Registrar General, Office of the. 1991. Census of India. Accessed February 10, 2007.
http://www.censusindia.net/cendat/language/lang1.html;
http://www.censusindia.net/.
Reynolds, Mike. 2005. Panjabi/Urdu in Sheffield: A case study of language maintenance
and language loss. In Janet Cotterill and Anne Ife (eds.), Language across
boundaries. London: Continuum.
Rickford, John R. 1987. Social class grouping in sociolinguistic research. American
Speech 62: 281-285.
403
Rickford, John R. 1988. Connections between sociolinguistics and pidgin-creole studies.
International Journal of the Sociology of Language 71: 51-57.
Rickford, John R. 1997. Unequal partnership: Sociolinguistics and the African American
speech community. Language in Society 26: 161-197.
Rogers, Chandrika K. (formerly Balasubramanian). 2002. Syntactic features of Indian
English: An examination of written Indian English. In Randi Reppen, Susan M.
Fitzmaurice and Douglas Biber (eds.), Using corpora to explore linguistic
variation. 187-202. Amsterdam: John Benjamins.
Roy, Arundhati. 1997. The god of small things. New York: HarperCollins.
Rubel, Paula and Rosman, Abraham. 2003. Introduction: Translation and anthropology.
In Paula Rubel and Abraham Rosman (eds.), Translating cultures: Perspectives
on translation in anthropology. 1-22. Oxford: Berg.
Rushdie, Salman. 1983. Shame. Picador.
Ryan, Ellen Bouchard, Giles, Howard and Sebastian, Richard J. 1982. An integrative
perspective for the study of attitudes toward language variation. In Ellen
Bouchard Ryan and Howard Giles (eds.), Attitudes toward language variation. 119. London: Arnold.
Sahgal, Anju. 1991. Patterns of language use in a bilingual setting in India. In Jenny
Cheshire (ed.), English around the world: Sociolinguistic perspectives.
Cambridge: Cambridge University Press.
Sahgal, Anju and Agnihotri, R.K. 1987. Indian English phonology: a sociolinguistic
perspective. English World-Wide 9(1): 51-64.
404
Sahgal, Anju and Agnihotri, Rama Kant. 1985. Syntax: The common bond. Acceptability
of syntactic deviances in Indian English. English World-Wide 6: 117-129.
Sahgal, Anju and Agnihotri, Rama Kant. 1988. Indian English phonology: a
sociolinguistic perspective. English World-Wide 9(1): 51-64.
Sailaja, Pingali. 2009. Indian English. Edinburgh: Edinburgh University Press.
Sankoff, David, Tagliamonte, Sali and Smith, Eric. 2005. Goldvarb X: A variable rule
application for Macintosh and Windows. University of Toronto.
http://individual.utoronto.ca/tagliamonte/Goldvarb/GV_index.htm.
Santa Ana, Otto. 1993. Chicano English and the nature of the Chicano language setting.
Hispanic Journal of Behavioral Sciences 15(1): 3-35.
Santa Ana, Otto and Bayley, Robert. 2004. Chicano English: phonology. In Edgar W.
Schneider, Kate Burridge, Bernd Kortmann, Rajend Mesthrie and Clive Upton
(eds.), A handbook of varieties of English: Phonology. 417-434. Berlin: Mouton
de Gruyter.
Saussure, Ferdinand de. 1983 [1916]. Course in general linguistics. London: Duckworth.
Schecter, Sandra R. and Bayley, Robert. 2002. Language as cultural practices:
Mexicanos en el Norte. Mahwah: Lawrence Erlbaum.
Schieffelin, Bambi, Woolard, Kathryn A. and Kroskrity, Paul (eds.) 1998. Language
ideologies: Practice and theory. New York: Oxford University Press.
Schiffrin, Deborah. 1994. Approaches to discourse analysis. Oxford: Blackwell.
Schneider, Edgar W. 2000. Corpus linguistics in the Asian context: Exemplary analyses
of the Kolhapur Corpus of Indian English. In Teodoro A. Llamzon Lourdes S.
Bautista, Bonifacio P. Sibayan (ed.), Parangal cang brother Andrew : Festschrift
405
for Andrew Gonzalez on his sixtieth birthday. Manila: Linguistic Society of the
Philippines.
Schneider, Edgar W. 2004a. Global synopsis: Phonetic and phonological variation in
English world-wide. In Edgar W. Schneider, Kate Burridge, Bernd Kortmann,
Rajend Mesthrie and Clive Upton (eds.), A handbook of varieties of English:
Phonology. 1111-1137. Berlin: Mouton de Gruyter.
Schneider, Edgar W. 2004b. How to trace structural nativization: particle verbs in world
Englishes. World Englishes 23(2): 227-249.
Schneider, Edgar W. 2007. Postcolonial English: Varieties around the world.
Cambridge: Cambridge University Press.
Schonweitz, Thomas. 2001. Gender and postvocalic /r/ in the American South: A detailed
socioregional analysis. American Speech 76(3): 259-285.
Seale, Jennifer Marie. 2007. An analysis of the syntactic and lexical features of Indian
English oral narrative: A Pear Story study. Linguistics, University of North
Texas: Denton
Segalowitz, Sidney J. and Lane, Korri C. 2000. Lexical access of function versus content
words. Brain & Language 75: 376-389.
Shackle, C. 1987. Speakers of Indian language. In M. Swan and Bernard Smith (eds.),
Learner English: A teacher's guide to interference and other problems. CUP.
Shankar, Shalini. 2004. Reel to real: Desi teens' linguistic engagement with Bollywood.
Pragmatics 14(2/3): 317-335.
Shankar, Shalini. 2008. Desiland: Teen culture, class, and success in Silicon Valley.
Durham: Duke University Press.
406
Sharifan, Farzad. 2005. Cultural conceptualisations in English Words: A study of
Aboriginal children in Perth. Language and Education 19(1): 74-98.
Sharma, Devyani. 2001. The pluperfect in native and non-native English: A comparative
corpus study. Language variation and change 13: 343-373.
Sharma, Devyani. 2003. Structural and social constraints on non-native varieties of
English. Linguistics, Stanford University: Palo Alto
Sharma, Devyani. 2005. Dialect stabilization and speaker awareness in non-native
varieties of English. Journal of Sociolinguistics 9(2): 194-224.
Shastri, S.V. 1992. Opaque and transparent features of Indian English. In Gerhard Leitner
(ed.), New directions in English language corpora: Methodology, results,
software developments. 263-275. Berlin: Mouton de Gruyter.
Shenk, Petra Scott. 2007. 'I'm Mexican, remember?' Constructing ethnic identities via
authenticating discourse. Journal of Sociolinguistics 11(2): 194-220.
Shi, Rushen, Gick, Bryan, Kanwisher, Dara and Wilson, Ian. 2005. Frequency and
category factors in the reduction and assimilation of function words: EPG and
acoustic measures. Journal of Psycholinguistic Research 34(4): 341-364.
Shuck, Gail. 2004. Conversational performance and the poetic construction of an
ideology. Language in Society 33: 195-222.
Silverstein, Michael. 1979. Language structure and linguistic ideology. In Paul R. Clyne,
William F. Hanks and Carol L. Hofbauer (eds.), The elements: A parasession on
linguistic units and levels. 193-247. Chicago: Chicago Linguistic Society.
Silverstein, Michael. 1985. Language and the culture of gender: At the intersection of
structure, usage, and ideology. In Elizabeth Mertz and Richard J. Parmentier
407
(eds.), Semiotic mediation: Sociocultural and psychological perspectives. 219259. Orlando: Academic Press.
Silverstein, Michael. 1996. Monoglot "standard" in America: Standardization and
metaphors of linguistic hegemony. In Donald Brenneis and Ronald K.S.
Macaulay (eds.), The matrix of language. 284-306. Boulder: Westview Press.
Singh, Rajendra (ed.) 1998. The native speaker: Multilingual perspectives. New Delhi:
Sage Publications.
Singh, Udaya Narayana. 2006. The new Linguistic Survey of India. Presented at South
Asian Languages Analysis (SALA) 26: Kannada University & Central Institute of
Indian Studies, Mysore.
Smerd, Jeremy. 2007. ArcelorMittal: Optimas award winner for global outlook.
Workforce management. Accessed 19 April, 2009.
http://www.workforce.com/section/11/feature/25/88/56/index.html.
Sridhar, Kamal K. 1991. Speech acts in an indigenised variety: sociocultural values and
language variation. In Jenny Cheshire (ed.), English around the world:
Sociolinguistic perspectives. 308-318. Cambridge: Cambridge University Press.
Sridhar, Kamal K. 1996. The pragmatics of South Asian English. In Robert J.
Baumgardner (ed.), South Asian English: Structure, use and users. 141-157.
Urbana: University of Illinois Press.
Sridhar, S. N. 1996. Towards a syntax of South Asian English: Defining the lectal range.
In Robert J. Baumgardner (ed.), South Asian English: Structure, use and users.
55-69. Urbana: University of Illinois Press.
408
Stephens, Laurence D. 1988. The role of palatalization in the Latin sound change /w/ >
/β/. Transactions of the American Philological Association 118: 421-432.
Swigart, Leigh. 2001. The limits of legitimacy: Language ideology and shift in
contemporary Senegal. Journal of Linguistic Anthropology 10(1): 90-130.
Tagliamonte, Sali. 2002. Comparative sociolinguistics. In J. K. Chambers, Peter Trudgill
and Natalie Schilling-Estes (eds.), The handbook of language variation and
change. 729-63. Oxford: Blackwell.
Tagliamonte, Sali. 2006. Analyzing sociolinguistic variation. Cambridge: Cambridge
University Press.
Tagliamonte, Sali and D'Arcy, Alex. 2007. To peak or not to peak: Exploring the
incrementation of linguistic change. Presented at New Ways of Analyzing
Variation (NWAV 36): University of Pennsylvania.
Tagliamonte, Sali, Poplack, Shana and Eze, Ejike. 1997. Plural marking patterns in
Nigerian Pidgin English. Journal of Pidgin and Creole Languages 12(1): 103129.
Tharoor, Shashi. 2007. The elephant, the tiger, & the cell phone. India: The emerging
21st-century power New York: Arcade Publishing.
Tourism, Indian. 2006. Indian States. Accessed November 28, 2007, 2007.
http://www.indian-tourism.us/india-map.jpg.
Trudgill, Peter. 1974. The social differentiation of English in Norwich. Cambridge:
University Press.
Trudgill, Peter and Hannah, Jean. 2002. International English: A guide to varieties of
Standard English. New York: Oxford University Press.
409
Trudgill, Peter, Schreier, Daniel, Long, Daniel and Williams, Jeffrey P. 2004. On the
reversibility of mergers: /W/, /V/ and evidence from lesser-known Englishes.
Folia Linguistica Historica 24(1-2): 23-45.
Vaid, D.D. 1977. Improve your English. Delhi: Orient Paperbacks.
Vaish, Viniti. 2008. Biliteracy and globalization: English language education in India.
Clevedon: Multilingual Matters Ltd.
Viswamohan, Aysha. 2004. Code Mixing with a difference. English Today 20(3/79): 3436.
Wagner, Suzanne. 2008. Gettin' ahead: Social class and the transition to linguistic
adulthood. Presented at American Anthropological Association annual meeting:
San Francisco.
Ward, Nigel. 2006. Non-lexical conversational sounds in American English. Pragmatics
and Cognition 14(1): 129-182.
Wassink, Alicia Beckford and Dyer, Judy. 2004. Language ideology and the transmission
of phonological change. Journal of English Linguistics 32(1): 3-30.
Weinreich, Uriel, Labov, William and Herzog, Marvin. 1968. Empirical foundations for a
theory of language change. In Winifred P. Lehmann and Yalkov Malkiel (eds.),
Directions for historical linguistics. 95-189. Austin: University of Texas Press.
Whorf, Benjamin Lee. 1941. The relation of habitual thought and behavior to language.
In John B. Carroll (ed.), Language, thought and reality: Selected writings of
Benjamin Lee Whorf. 134-159. New York: Technology Press of M.I.T.
Williams, Jessica. 1987. Non-native varieties of English: A special case of language
acquisition. English World-Wide 8: 161-189.
410
Wiltshire, Caroline. 2005. The “Indian English” of Tibeto-Burman language speakers.
English World-Wide 26(3): 275-300.
Wiltshire, Caroline and Moon, Russell. 2003. Phonetic stress in Indian English vs.
American English. World Englishes 22(3): 291-303.
Wolfram, Walt. 1993. Identifying and interpreting variables. In Dennis Preston (ed.),
American dialect research. 193-221. Amsterdam: Benjamins.
Woolard, Kathryn A. 1986. Language variation and cultural hegemony: toward an
integration of Sociolinguistic and social theory. American Ethnologist 12(4): 738748.
Woolard, Kathryn A. 1998. Language Ideology as a Field of Inquiry. In Bambi
Schieffelin, Kathryn A. Woolard and Paul Kroskrity (eds.), Language ideologies:
Practice and theory. 3-47. New York: Oxford University Press.
Woolard, Kathryn A. 2008. Why dat now?: Linguistic-anthropological contributions to
the explanation of sociolinguistic icons and change. Journal of Sociolinguistics
12(4): 432-452.
Woolard, Kathryn A. and Gahng, Tae-Joong. 1990. Changing language policies and
attitudes in autonomous Catalonia. Language in Society 19: 311-330.
Woolard, Kathryn A. and Schieffelin, Bambi. 1994. Language ideology. Annual Review
of Anthropology 23: 55-82.
World Bank, The. 2006. Purchasing power parity: GDP. The World Bank Group:
Washington D.C.
http://siteresources.worldbank.org/DATASTATISTICS/Resources/GDP_PPP.pdf
411
Yaeger-Dror, Malcah, Kendall, Tyler, Foulkes, Paul, Watt, Dominic, Oddie, Jillian,
Harrison, Phil and Kavenagh, Colleen. 2008. Perception of r-fulness by trained
listeners. Presented at New Ways of Analyzing Variation (NWAV 37): Houston,
Texas.
Young, Richard and Bayley, Robert. 1996. VARBRUL Analysis for second language
acquisition research. In Robert Bayley and Dennis R. Preston (eds.), Second
language acquisition and linguistic variation. 253-306. Amsterdam: John
Benjamins.
Zentella, Ana Celia. 1987. Language and female identity in the Puerto Rican community.
In Joyce Penfield (ed.), Women and language in transition. 167-179. New York:
State University of New York Press.
Zhang, Qing. 2005. A Chinese yuppie in Beijing: Phonological variation and the
construction of a new professional identity. Language in Society 32: 431-466.