Jamin R. Pelkey
DIALECTOLOGY
AS DIALECTIC
INTERPRETING PHULA VARIATION
TRENDS IN LINGUISTICS
Dialectology as
DialecticInterpreting Phula VariaDialectology as Dialectic
tion
by
Jamin R. Pelkey
Mouton de Gruyter
Berlin · New York
Trends in Linguistics
Studies and Monographs 229
Editor
Volker Gast
Founding Editor
Werner Winter
Editorial Board
Walter Bisang
Hans Henrich Hock
Matthias Schlesewsky
Niina Ning Zhang
Editor responsible for this volume
Hans Henrich Hock
De Gruyter Mouton
Dialectology as
Dialectic
Interpreting Phula Variation
by
Jamin R. Pelkey
De Gruyter Mouton
ISBN 978-3-11-024584-4
e-ISBN 978-3-11-024585-1
ISSN 1861-4302
Library of Congress Cataloging-in-Publication Data
Pelkey, Jamin R., 1974 –
Dialectology as Dialectic : interpreting Phula variation / by Jamin
R. Pelkey.
p. cm. – (Trends in linguistics: studies and monographs ; 229)
Revision of author’s (doctoral) thesis – LaTrobe University, Australia, 2008.
Includes bibliographical references and index.
ISBN 978-3-11-024584-4 (alk paper)
1. Yi language – Dialectology.
2. Yi Language – Phonology.
3. Yi (Chinese people) – Languages.
4. Yi (Chinese people) –
Ethnic identity. 5. Anthropological linguistics – China – Yunnan
Province. 6. Anthropological linguistics – Vietnam. I. Title.
PL3311.Y5P45 2011
495– dc22
2011009360
Bibliographic information published by the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;
detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.
© 2011 Walter de Gruyter GmbH & Co. KG, Berlin/New York
Printing: Hubert & Co. GmbH & Co. KG, Göttingen
̀ Printed on acid-free paper
Printed in Germany.
www.degruyter.com
To Stephanie Jill & Quynh Iris
Preface
In this work I propose a more integrative approach to dialectology and apply
the approach to the previously undefined Phula languages of China and Vietnam using original fieldwork and analysis. My hope is that the efficacy of
the argument, and hence the merit of the book, will emerge from its selfcontained integration of theory, methodology and praxis. The Phula languages are in need of definition: a task that would seem to fall to the dialectologist. If traditional approaches to dialectology are found to be inadequate
to the task, more complex approaches are called for. Such is the case. In fact,
this work adds volume to a mounting murmur that is already rising from the
halls of language variation research. Numerous recent appeals (e.g., Preston
1999a; Croft 2000; Koyama 2001; Kortmann 2004b; Bisang 2004; Nevalainen, Klemola and Laitinen 2006) suggest that mine is not a solitary voice
‘crying in the wilderness’: language variation research is currently being
called to task for re-evaluation, particularly in terms of more integrative
approaches to dialectology.
Dialectology proper has traditionally focused on the geographic distribution of language variation as an end in itself and has remained relatively
segregated from other branches of linguistic (and extra-linguistic) inquiry.
Although the field has slowly been opening up to interdisciplinary perspectives since the late 1990’s, dialectologists still tend to approach their data as
classical modern reductionists – asking exclusively synchronic or exclusively diachronic questions, using exclusively qualitative or exclusively
quantitative analysis, often neglecting insights from history, culture, sociolinguistics, language contact, historical-comparative linguistics and/or language typology. The reductive results of traditional dialectology are
intriguing and instructive in their own right, of course; but they are not wellsuited for engaging the gradient, interdependent dynamics that surround the
actual operationalization of language and dialect definitions, whether synchronically or diachronically conceived. Bringing opposing linguistic, and
extralinguistic, variables into dialectic exchange, on the other hand, results
in emergent equilibria useful for language definition at multiple levels – or
so I have found in seeking to understand Phula.
The languages affiliated with the ancient ethnonym ‘Phula’ descend
from the Ngwi (formerly Loloish) branch of Burmic in the Tibeto-Burman
family and are spoken in remote mountainous regions of southeastern
viii
Preface
Yunnan Province, China, and adjacent pockets of northwestern Vietnam, by
a scattered array of ethnic groups whose populations now total some
367,000. The data considered in this book are the results of my personal
fieldwork in 41 Phula villages gathering lexical, textual, ethnohistorical, geolinguistic and perceptual information in cooperation with numerous research and administrative units of Yunnan. Prior to the research leading up
to this publication, these language varieties were undefined – not only in
terms of synchronic identification and diachronic situation but also in terms
of degrees of separation, historical contact, status of endangerment, general
demographics, geographic distribution and dialect diversity.
As suggested above, in order to arrive at such definitions, I argue that
both data collection and data analysis should be approached from multiple
perspectives – typological-descriptive, historical-comparative and sociocognitive alike; using diagnostics that are both qualitative and quantitative;
blending insights from history, geography, ethnology, language contact and
sociolinguistics into an organic whole. The approach incorporates complexity, asserting that dialectology best flourishes as an interdependent dialectic –
a dynamic synthesis of correlative perspectives.
The overarching dialectic treated in this work is framed in terms of the
familiar ‘synchronic-diachronic’ opposition indicative of 20th century linguistic dualism. Taken as a strict dichotomy, synchrony and diachrony are,
ipso facto, irreconcilable. If we distance ourselves from the old essentialist
presuppositions and approach the actual unfolding of language use and linguistic cognition in time and space with more probing, phenomenological
attitudes, the distinction itself becomes liable to an ontological-conceptual
shift. Instead of generating artificial binary choices (e.g., between past vs.
present, history vs. typology, reconstruction vs. description), synchrony and
diachrony emerge as profoundly involved in each other’s affairs in limitless
combinations of underlying complementary tensions – tensions that might
be more aptly re-framed in interdependent biological terms (see dialogue
between Croft 2010 and Mufwane 2010 for precedence and potential problems). In place of the ‘synchronic-diachronic’ dyad, a biology-oriented triad
suggests itself, including ‘ecological’, ‘phylogenetic’ and ‘ontogenetic’ contingencies: linguistic ecology including both synchronic context and diachronic contact; linguistic phylogeny including both diachronic lineage and
synchronic inheritance; and linguistic ontogeny mediating between the two
in the form of specific, polylectal speech varieties growing through space
and time. Whatever the case, I use the traditional labels in this work if for no
other reason than to demonstrate their inadequacy as discrete categories –
and their interdependence as dialectic categories. Attempting to choose
between the two seems, at best, more and more shortsighted.
Preface
ix
Theoretical and conceptual considerations aside, the necessity of treating
the ‘synchronic-diachronic’ dialectic as central to the current study grows
out of a practical bid to disprove and refine the following hypothesis:
all synchronic languages traditionally affiliated with the Phula ethnonym
also belong to a single exclusive diachronic clade linguistically. In order to
actually make this claim falsifiable, two key sets of knowledge are needed:
1) ‘synchronic’ definitions: the number of distinct, contemporary languages
affiliated with the Phula ethnonym and 2) ‘diachronic’ definitions: the
nature of the genetic relationships they share. Understanding the nature of
either, however, is best accomplished through the lens of the other. Synchronic language definitions are shown to provide the categories necessary
for diachronic subgrouping, and diachronic subgroupings are shown to provide validation for synchronic language definitions. The two sets of
knowledge are interdependent; neither can be adequately defined in a vacuum. Numerous other analogous dynamics are illustrated. A chiastic model
emerges as a structure well suited for the dialogic task of complex
integration, and the mediation of paradox is affirmed as a fruitful mode of
scientific discovery.
My approach to fieldwork and analysis are, then, essentially hermeneutic
or ‘pragmatist’ in nature – curiosity and doubt seeking understanding, rather
than confidence and mastery seeking control. Perhaps I may even lay claim
to the ancient sign-based tradition of scientific investigation best articulated
and practiced by Charles Sanders Peirce, eminent American scientist, logician and founder of hermeneutic semiotics. According to Peirce (e.g.,
[1903] 1998), abductive inference, or hypothesis formation, is indispensible
for rational inquiry. Peirce demonstrates that deductive and inductive logic
depend on abductive inference, though they can by no means be reduced to
it; rather, the three work in tandem – abduction responding to problems
discovered through inductive testing of deductive assertions, which are
themselves habits taken up by assuming the truth of an abduction, due to the
elegant explanatory power of the latter. Under this approach, analysis is reflexive with interpretation, or riddle solving; hence the subtitle of this book:
‘interpreting Phula variation.’
The book opens with a review of the contemporary Phula varieties in
their historical context and offers a critical evaluation of background issues
in dialectology. Fieldwork methods and theoretical assumptions are made
transparent in Chapter 2, and synchronic language definitions are presented
in Chapter 3. Chapter 4 discusses the status of language vitality among the
Phula varieties while Chapter 5 sketches five comparative phonologies of
representative Phula languages. Chapters 6 and 7 demonstrate grounds for
subgrouping the Phula languages along with their next-of-kin, and the final
x
Preface
chapter reverses the perspective of the first by examining the newly defined
diachronic categories in terms of their newly defined synchronic constituents. Findings are presented in a variety of charts, tables and maps. Appendices include quantitative matrices, field note samples, and a sample
comprehension test.
The results of Phula language definition fill substantial gaps in our general knowledgebase of an immensely complex ethnolinguistic region and
add to the total number of world languages on record. Most of the Phula
languages are endangered to varying degrees; thus, the book calls attention
to the often overlooked reality that language definition must precede language preservation. Neglected diversity in the Sinosphere is also highlighted,
and research results have implications for regional ethnohistory, geolinguistics and our understanding of the interaction between language contact and
genetic inheritance. For Tibeto-Burmanists in particular, the process of subgrouping based on the newly available Phula data provides grounds for
rethinking several other Ngwi-branch relationships.
In summary, Dialectology as Dialectic demonstrates multiple grounds
for insisting on a more robust, integrative approach to dialectology while
simultaneously demonstrating grounds for defining the Phula languages: 24
synchronic languages belonging to three distinct macro-clades genetically.
In the process of interpretation, 22 of these languages are demonstrated to
descend from two exclusive clades of the Southeastern Ngwi sub-branch. I
would be pleased, of course, if others were drawn into the celebration, and
validation, of this zesty array of overlooked speech communities near the
Sino-Vietnam border. I would be doubly pleased if this work contributes, in
some miniscule way, toward encouraging the realization of a dawning ‘ontology of relations’ – a radical shift in taste, if you please, such that the interpretive mediation of complementary tensions (in language and in life)
comes to seem preferable to artificial choices between static dichotomies.
Jamin Pelkey
February 22, 2011
Fort Langley, British Columbia
Acknowledgements
This work is the culmination of a twelve year venture (or compulsion), to
define the Phula languages. Little did I know during the early stages of my
inquiry, begun in 1998, just how many individuals and organizations would
eventually assist me in the process.
From 2005–2009 I was a postgraduate student in the La Trobe University (LTU) Linguistics department, and this volume constitutes a revised version of my 2008 LTU PhD dissertation. Financial support for the La Trobe
period of research was provided by an LTU International Postgraduate Research Scholarship, an Australia International Postgraduate Research Scholarship, an LTU FHSS 2005 fieldwork grant, and an SIL-International
project funding grant – along with a number of contributors who donated
funds toward a Beijing Jeep, enabling travel to rural datapoints in Yunnan.
My two formal advisors during the La Trobe sojourn were David Bradley and Randy LaPolla – an enviable duo for any budding Tibeto-Burmanist
to study under. Of course, their guidance meant the difference between failure and success. The long hours they spent laboring through my drafts and
talking over the details of my analysis were essential for my progress and
sacrificial on their part. As my primary advisor, David’s extensive
knowledge of the Burmic group, in both breadth and depth, along with his
natural appreciation for language variation in the Sinosphere and his wealth
of experience in Asian linguistic fieldwork, made for an ideal source of inspiration. In addition to his eye for details that I had missed or mistaken David was also quick to point out practicalities, theories and implications that I
had overlooked. My appreciation goes to Randy for introducing me to
Charles Sanders Peirce during this period, for priming me with Sinitic perspectives, and for refining my philosophy of language, among many other
insights noted in the text and endnotes. Jerold Edmondson, Graham Thurgood and Harold Koch were the outside examiners of my final dissertation
draft, and each provided important affirmation and advice.
My appreciation goes to Brian Migliazza, James Matisoff, Keith Slater,
Bryan Allen, Peter Lester, Chris Stokland, Lim Chong, Jason Pounders, Nathan Davis, Sung Kim, Brandt Robbins, Andy Castro, Fraser Bennett, Alec
Coupe, Cathryn Yang, and Eric Johnson for helping me gain momentum
early in the process of research and writing. Dialogue with Cathryn and Eric,
in particular, helped define and refine my approach to research and analysis.
xii
Acknowledgements
He Lifeng and Bai Keyang were instrumental in opening the gates to official fieldwork in Yunnan Province. My primary field research partners
were Yang Liujin of Honghe University (formerly of the Honghe Prefecture
Nationalities Research Institute), Bai Bibo of the Nationalities Research Institute of Yuxi Normal University and Wang Mingfu of the Wenshan Prefecture Nationalities Research Institute. These three men coordinated
logistics, secured permissions, provided invaluable advice along the way
and became my trusted friends. Multiple other government and educational
leaders in southeastern Yunnan also granted their permission and logistical
assistance in the fieldwork effort, and hundreds of local level Phula speakers
patiently provided language data and hospitality in their mountain villages.
My data analysis can be broken down into four primary components as
will be discussed in Chapter 1. Behind each component is a key source of
influence (though, as always, I am solely responsible for remaining errors).
Ken Manson introduced me to the Neighbor-Net algorithm when I was sorely in need of a valid quantitative component for diachronic analysis, and
Noel Mann helped me think through a number of issues related to synchronic quantitative analysis. The major influence shaping my synchronic qualitative analysis was David Bradley; and, in addition to Bradley’s work on
Ngwi relationships, Randy LaPolla’s advice on subgrouping criteria had a
substantial influence on my diachronic qualitative analysis.
A number of non-academic influences helped sustain me through the
thickest stages of the analysis and drafting, such as the encouragement of
my parents, Rob and Joan Pelkey, and in-laws, Bud and Marilyn Hopkins.
The Equip Training Centre at Kangaroo Grounds provided a private office
space for a crucial three month stretch of writing. A jazz ensemble named
“Virus”, playing on Saturday evenings at the Laundry pub in Fitzroy, taught
me that live jazz, whatever else it may be, is potent therapy. The novels and
essays of Walker Percy kept me human, as did extended reflection on Colossians, Ephesians, 2Peter, 1John and other poetry (that of W.H. Auden,
Dylan Thomas and T.S. Eliot, in particular). Other sustaining influences included conversations with my friend Matt Rojahn and the birth of Quynh
Iris (a.k.a. the Seahorse).
Stephanie has been my orientation point and closest companion for nine
of the past twelve years of research and writing. She accompanied me to
most of the fieldwork datapoints discussed in this volume and helped with
video recordings, equipment maintenance, expense accounts, and a host of
other crucial details. Thank you again, Stephie.
Once again, thank you all.
Contents
Preface
Acknowledgements
Contents
List of maps
List of tables
List of figures
General abbreviations
Data source abbreviations
1. Introduction: Synchronic Phula in diachronic perspective
1.1. The Phula hypothesis
1.1.1. Falsifiability and the Phula hypothesis
1.1.2. Chapter overview
1.2. Historical background
1.2.1. Phula, Puzu, Yizu and other classifications
1.2.2. The Phula-Lolo distinction in historical records
1.2.3. Retracing Phula migration patterns from oral and
written histories
1.2.3.1. Macro-migration patterns
1.2.3.2. Micro-migration patterns
1.2.4. Previous research on the Phula languages
1.3. Field data orientation
1.3.1. Administrative orientation of Phula villages
1.3.2. Fieldwork data point locations
1.3.3. Overview of field data collection
1.3.4. External data sources
1.4. Research scope, assumptions and approach
1.4.1. Theoretical assumptions: toward an integrative
dialectology
1.4.2. On the viability of integrational standards for
language definition
1.4.3. Research scope and limitations
1.5. Argument structure and organization
vii
xi
xiii
xxii
xxiii
xxvi
xxviii
xxxi
1
1
3
4
5
7
9
11
11
15
18
21
21
28
28
29
31
31
36
40
41
xiv
Contents
2. Research background: Field methods, theory, and dialectology
2.1. Introduction
2.2. Nature and scope of the fieldwork
2.2.1. Navigating research approval and fieldwork sponsorship
2.2.2. Navigating administrative terrain
2.2.3. Navigating data point travel
2.2.4. Scope of data collection
2.2.5. Data reciprocity with language communities
2.2.6. Digital archiving and analysis
2.3. Recording equipment and use
2.3.1. Overview of recording equipment
2.3.2. Recording scenarios and setup
2.3.3. Evaluation of equipment and recording methods
2.4. Elicitation instruments and methodology
2.4.1. Sociolinguistic questionnaires
2.4.1.1. Questionnaire goals and design
2.4.1.2. Questionnaire administration
2.4.1.3. Working revisions
2.4.2. Wordlist design and elicitation
2.4.2.1. Wordlist design
2.4.2.2. Elicitation printouts
2.4.2.3. Consultant selection and screening
2.4.2.4. Elicitation framing
2.4.2.5. Cognate fishing
2.4.2.6. Use of elicitation illustrations
2.4.2.7. Non-IPA Chinese phonetic conventions
2.4.2.8. Pronunciation imitation
2.4.2.9. Working revisions
2.4.3. Natural text elicitation
2.5. Assessing identity, demography and vitality
2.5.1. Ethnic identity research
2.5.2. Demographic research
2.5.3. Geolinguistic mapping research
2.5.4. Endangerment, vitality, and language contact research
2.6. Assessing dialect intelligibility and internal contact
2.6.1. Perceptual dialectology
2.6.2. Recorded text testing and the high-intelligibility
threshold
2.6.3. Core lexical comparison and the low-intelligibility
threshold
2.7. Assessing language variation and change
44
44
45
45
45
46
47
49
50
51
51
52
54
55
55
56
57
57
58
58
60
60
61
64
65
65
67
67
68
69
69
72
73
74
78
79
80
82
84
Contents
2.7.1.
2.7.2.
Phonological description in light of language variation
Grammaticalization, lexicalization and variational
semantics
2.7.3. The comparative method
2.8. Assessing phylogenetic relationships
2.8.1. Clustering, subgrouping and phylogenetic situation
2.8.2. Baileyan dialectology and the dynamic wave model
2.8.3. Tree diagrams and other models
2.8.4. The Neighbor-Net algorithm and distance-based
relationships
2.9. An experiment in triadic dialectics: ecology-phylogeny-ontogeny
3. Synchronic language definitions: Identity, intelligibility, contact
3.1. Preliminaries
3.2. Phula ethnic identities
3.2.1. West-regional identities
3.2.2. South-central identities
3.2.3. North-central identities
3.2.4. East-regional identities
3.2.5. Trends and observations
3.3. Core lexical comparisons
3.3.1. West-regional comparisons
3.3.2. South-central comparisons
3.3.3. North-central comparisons
3.3.4. East-regional comparisons
3.3.5. Perspective
3.4. Recorded text testing
3.4.1. West-regional results
3.4.2. South-central results
3.4.3. North-central testing results
3.4.4. East-regional testing results
3.5. Reported dialect perceptions and intelligibility indices
3.5.1. West-regional perceptions and indices
3.5.2. South-central perceptions and indices
3.5.3. North-central perceptions and indices
3.5.4. East-regional perceptions and indices
3.6. Language contact issues
3.6.1. External contact
3.6.2. Internal contact
3.7. Integrating identity, contact, and intelligibility
3.7.1. West-regional integration
xv
84
85
87
88
90
91
93
94
94
96
96
99
101
102
110
114
118
119
120
122
124
126
126
128
128
129
132
133
133
134
136
144
149
153
154
155
159
159
xvi
Contents
3.7.2. South-central integration
3.7.3. North-central integration
3.7.4. East-regional integration
3.8. Conclusion: The synchronic Phula languages
3.8.1. Phala
[ISO 639-3: ypa]
3.8.2. Phola
[ISO 639-3: ypg]
3.8.3. Phola, Alo
[ISO 639-3: ypo]
3.8.4. Muji, Qila
[ISO 639-3: ymq]
3.8.5. Muji, Southern [ISO 639-3: ymc]
3.8.6. Muji, Northern [ISO 639-3: ymx]
3.8.7. Muzi
[ISO 639-3: ymz]
3.8.8. Bokha
[ISO 639-3: ybk]
3.8.9. Phuma
[ISO 639-3: ypm]
3.8.10. Alugu
[ISO 639-3: aub]
3.8.11. Phupa
[ISO 639-3: ypp]
3.8.12. Phupha
[ISO 639-3: yph]
3.8.13. Phuza
[ISO 639-3: ypz]
3.8.14. Phowa, Ani
[ISO 639-3: ypn]
3.8.15. Phowa, Labo [ISO 639-3: ypb]
3.8.16. Phowa, Hlepho [ISO 639-3: yhl]
3.8.17. Azha
[ISO 639-3: aza]
3.8.18. Zokhuo
[ISO 639-3: yzk]
3.8.19. Khlula
[ISO 639-3: ykl]
3.8.20. Moji
[ISO 639-3: ymi]
3.8.21. Phukha
[ISO 639-3: phh]
3.8.22. Laghuu
[ISO 639-3: lgh]
3.8.23. Pholo
[ISO 639-3: yip]
3.8.24. Thopho
[ISO 639-3: ytp]
4. Ethnolinguistic vitality: Contact, endangerment and shift
4.1. Introduction
4.2. Phola and Alo
4.2.1. Luodie and Natang
4.2.2. Adipo
4.3. Phala
4.4. Muji, Southern
4.4.1. Pujiazhai
4.4.2. Shizitou
4.5. Muji, Northern
4.5.1. Xiepo
4.5.2. Loushuidong
160
162
163
163
164
164
165
165
166
167
167
168
168
168
169
169
169
171
171
171
172
172
174
174
174
176
176
176
178
178
182
182
183
184
186
186
187
188
188
189
Contents
xvii
4.6. Muzi
4.6.1. Nuoguzhai
4.6.2. Malutang
4.7. Bokha
4.7.1. Yibaizu
4.7.2. Dixibei
4.8. Phuma
4.9. Alugu
4.10. Phupa
4.11. Phupha
4.12. Phuza
4.13. Muji, Qila
4.14. Phowa, Ani
4.15. Phowa, Labo
4.15.1. Jiajie and Lugumu
4.15.2. Wudupi
4.16. Phowa, Hlepho
4.16.1. Feizuke
4.16.2. Chekabai
4.16.3. Weibazhu/Xiaozhai
4.16.4. Suozhiwan
4.16.5. Meizichong
4.17. Azha
4.17.1. Luojiayi, Xiaopingba and Huangzhai
4.17.2. Faduke Dazhai
4.18. Zokhuo
4.19. Khlula
4.19.1. Laozhai
4.19.2. Maxi
4.20. Laghuu and Phukha
4.21. Moji
4.22. Pholo
4.22.1. Xiji
4.22.2. Fayixiazhai
4.22.3. Shangxinzhai
4.23. Thopho
4.24. Overview of Phula ethnolinguistic vitality
190
190
191
192
192
193
194
195
196
197
198
199
200
201
201
202
203
203
204
206
207
207
209
209
211
212
213
213
215
215
216
216
216
217
217
219
220
5. Phula phonologies: Five representative sketches
223
5.1. Preliminaries
5.1.1. General Phonological Features
223
224
xviii
Contents
5.1.2. Structure of the sketches
5.2. Hlepho-FZK and the Phowa Clade
5.2.1. Syllable
5.2.2. Consonant Initials
5.2.3. Vowel Finals
5.2.4. Tone
5.2.5. Phonation
5.3. Muji-PJZ and the Muji Clade
5.3.1. Syllable
5.3.2. Consonant Initials
5.3.3. Vowel Finals
5.3.4. Tone
5.4. Phuza-BJB and the Downriver Phula clade
5.4.1. Syllable
5.4.2. Consonant initials
5.4.3. Vowel Finals
5.4.4. Tone
5.5. Phola-LDC and the Upriver Phula clade
5.5.1. Syllable
5.5.2. Consonant initials
5.5.3. Vowel Finals
5.5.4. Tone
5.5.5. Phonation
5.6. Azha Phonology
5.6.1. Syllable
5.6.2. Consonant Initials
5.6.3. Vowel Finals
5.6.4. Tone
5.7. Summary
226
227
228
228
235
238
240
240
241
241
243
245
247
248
248
251
253
254
255
255
259
261
263
264
264
265
267
268
270
6. The language clades of Phula Proper: Establishing historical
subgroupings
271
6.1. Methodology for historical subgrouping
6.1.1. Overview of Ngwi tone
6.1.2. Overview of Proto-Ngwi *initial classes and guide
to notation
6.1.3. Overview of Phula tonal reflexes
6.2. A distance-based network of Phula languages
6.2.1. Neighbor-Net and distance-based phylogenetic
relationships
271
272
273
275
278
279
Contents
6.2.2.
6.3.
6.4.
6.5.
6.6.
6.7.
The Phula neighbor network and preliminary
subgroupings
The Muji meso-clade
6.3.1. Tone system innovations
6.3.1.1. Tone-class 1
6.3.1.2. Tone-class 2
6.3.1.3. Tone-class 3
6.3.1.4. The *checked tone classes
6.3.2. Other independent innovations
6.3.2.1. Phonological innovations
6.3.2.2. Morphological innovations
6.3.2.3. Lexico-semantic innovations
6.3.3. Historical dialectology
The Phowa meso-clade
6.4.1. Tone system innovations
6.4.1.1. Tone-class 1
6.4.1.2. Tone-class 2
6.4.1.3. Tone-class 3
6.4.1.4. The *checked-tone classes
6.4.2. Other Innovations
6.4.3. Historical dialectology
The Highland Phula macro-clade
6.5.1. Lexico-semantic innovations
6.5.2. Phonological innovations
The Riverine Phula macro-clade
6.6.1. Tone system developments
6.6.1.1. Tone-class *1
6.6.1.2. Tone-class *2
6.6.1.3. Tone-class *3
6.6.1.4. The *checked tone classes
6.6.2. Other Innovations
6.6.2.1. Phola-Phala (Upriver Phula)
6.6.2.2. Phupha-Alugu
6.6.2.3. Phuza-Phupa
6.6.3. Historical dialectology
Conclusion
7. Phula and Southeastern Ngwi: Sani, Axi, Azhe, Azha, Nisu, and
Phula Proper
7.1. Introduction
7.2. Expanded neighbor network
xix
281
285
285
288
289
291
293
300
301
304
305
308
313
314
316
317
319
320
322
325
330
330
334
335
336
337
338
340
340
342
342
343
345
346
350
352
352
353
xx
Contents
7.3. Southeastern Ngwi
7.3.1. Permutations of the alveolar-lateral cluster innovation
7.3.2. Sani, Axi, Azhe, Azha, Southeastern Ngwi and
Central Ngwi
7.3.3. Nisu, Southeastern Ngwi and Northern Ngwi
7.3.4. Summary of preliminary subgrouping criteria for
SE Ngwi
7.4. Azha, Azhe and Sani-Axi
7.4.1. The status of Azha
7.4.2. The Sani-Axi clade
7.4.3. Tone system developments for Sani-Axi-Azhe-Azha
7.4.3.1. Tone-class *1
7.4.3.2. Tone-class *2
7.4.3.3. Tone-class *3
7.4.3.4. The *checked tone classes
7.4.4. Working conclusion on the internal relationships of
the SA clade
7.5. Pholo
7.6. Conclusions
8. Conclusion: Diachronic Phula in synchronic perspective
8.1. Review
8.1.1. Phula language definition as chiastic dialectic
8.1.2. Folk Phula and Phula Proper
8.2. The Riverine Phula Macro-clade in synchronic perspective
8.2.1. Riverine Phula demographics and distribution
8.2.1.1. Upriver demographics and distribution
8.2.1.2. Downriver demographics and distribution
8.2.2. Riverine Phula perceptual dialectology
8.2.2.1. Upriver perceptual dialectology
8.2.2.2. Downriver perceptual dialectology
8.3. The Muji meso-clade in synchronic perspective
8.3.1. Current demographics and distribution of the Muji
meso-clade
8.3.2. Muji meso-clade perceptual dialectology by
synchronic language
8.4. The Phowa meso-clade in synchronic perspective
8.4.1. Current demographics and distribution of the Phowa
meso-clade
8.4.2. Phowa meso-clade perceptual dialectology by
language
355
356
366
368
372
375
376
378
380
383
384
386
387
388
389
391
394
394
395
398
401
401
403
404
404
406
406
407
407
411
413
413
416
Contents
8.5. Review of contributions and key findings
8.5.1. Contributions to areal language recognition
8.5.2. Contributions to research on areal language
endangerment
8.5.3. Contributions to Ngwi linguistics and regional
ethnohistory
8.5.4. Contributions to language contact research
8.5.5. Contributions to ethnic identity research
8.5.6. Contributions to tone system analysis
8.5.7. Contributions to distance-based phylogenetics
8.5.8. Contributions to intelligibility testing
8.5.9. Contributions to regional demography
8.5.10. Contributions to geolinguistics
8.5.11. Contributions to field methods in linguistic survey
8.5.12. Toward a hermeneutic dialectology
8.5.13. Toward a chiastic dialectic model of reasoning
8.6. Extra-disciplinary perspectives on neglected diversity
8.6.1. Neglected diversity and onomastic ambiguity
8.6.2. Neglected diversity, homoplasy and genetics
8.6.3. Neglected diversity and the survival of the cutest
8.7. Suggestions for future studies
8.7.1. Remaining research questions
8.7.2. New research suggestions
8.8. Coda
Appendix A: Sociolinguistic questionnaires
Appendix B: Lexical elicitation datasheet sample
Appendix C: Similarity and distance matrices
Appendix D: Sample RTT translations
Notes
References
Index
xxi
418
419
419
420
420
421
421
421
422
422
423
423
423
425
426
426
427
428
429
430
431
432
433
442
444
447
449
460
487
List of maps
1.1. Broad geographic orientation of Yunnan Province, China, and Phula 6
distribution area
1.2. Major river systems of Yunnan and ancient Phula settlement sites
13
1.3. Administrative overview of Yunnan and NW Vietnam with
22
contemporary administrative Phula regions shaded
1.4. Phula village distribution patterns by township
24
1.5. County level overview of Phula distribution marking data point
25
locations
3.1. Orientation of geographic regions and maps discussed in Chapter 3 98
3.2. General distribution of Phola, Phala, and Alo
165
3.3. Distribution of Muji, Qila and location of the three Muji, Qila
166
villages
3.4. Distribution of Southern Muji, Northern Muji, Muzi, Bokha,
170
Phuma, Phupa, Alugu, Phupha and Phuza
3.5. Distribution of Ani, Labo, Hlepho, Azha and Zokhuo
173
3.6. Distribution of Khlula, Moji, Laghuu, and Phukha
175
3.7. Distribution of Pholo and Thopho
177
4.1. Locations of Phula language endangerment maps plotted by region 181
4.2. Overview of language vitality in the core western region
185
4.3. Overview of language vitality in the core central region
208
4.4. Overview of Phula language vitality in the core eastern region
218
8.1. Spatial distribution of Phula Proper and Folk Phula delineated by 400
major genetic clade
8.2. Current distribution of Riverine Phula
403
8.3. Perceptual dialect affiliation by village and subgroup for Riverine 405
Phula
8.4. Current distribution of the Muji Clade
408
8.5. Perceptual dialect affiliation by village for Core Muji
412
8.6. Current distribution of the Phowa Meso-clade
414
8.7. Perceptual dialect affiliation by village for Phowa clade
417
(minus Phukha)
List of tables
1.1. Summary of Phula data collection linked to Map 1.5
2.1. Major stages of approval and sponsorship for Phula field research
2.2. Translated overview of elicitation frames used in lexical data
collection
3.1. West-regional onomastic identities
3.2. Onomastic identity overview for Muji and Muzi autonym groups
3.3. Other Phula onomastic identities in the south-central region
3.4. North-central onomastic identities
3.5. East-regional onomastic identities
3.6. Core lexical comparison results for the west-regional Phula
varieties
3.7. Core lexical comparison results for south-central Phula varieties
3.8. Core lexical comparison results for Phula varieties in the
north-central region
3.9. Core lexical comparison results for Phula varieties in the eastern
region
3.10. Phula intelligibility testing results for the south-central region
3.11. Results and evaluation of preliminary testing with
Muji-PJZ RTT
3.12. Intelligibility and internal contact summary for the south-central
region
4.1. Threatened language status estimates by village and language
5.1. Hlepho Phowa-FZK consonant initial phonemes
5.2. Hlepho Phowa-FZK consonants contrasts
5.3. Status of the lateral affricate series in the Phowa clade
5.4. Evidence for velar reflexes of Phula lateral clusters in Zokhuo
5.5. Reflexes of Proto-Ngwi prenasalized stops in the Phowa clade
5.6. Hlepho Phowa-FZK monophthong final contrasts
5.7. The Muji-PJZ consonant system
5.8. Muji-PJZ consonants in contrastive environments
5.9. Preliminary overview of phone reflexes for Muji lateral-cluster
series
5.10. Muji-PJZ vowels in contrastive environments
5.11. The Phuza-BJB consonant system
5.12. Phuza-BJB consonant initial contrast set one
5.13. Phuza-BJB consonant initial contrast set two
5.14. Phuza-BJB vowel final contrasts
26
46
63
101
103
106
111
115
121
123
125
127
130
131
161
221
229
230
233
233
234
235
241
242
243
245
248
249
250
252
xxiv
5.15.
5.16.
5.17.
5.18.
5.19.
5.20.
6.1.
6.2.
6.3.
6.4.
6.5.
6.6.
6.7.
6.8.
6.9.
6.10.
6.11.
6.12.
6.13.
6.14.
6.15.
6.16.
6.17.
6.18.
6.19.
6.20.
6.21.
6.22.
6.23.
6.24.
6.25.
6.26.
7.1.
7.2.
List of tables
Phola consonant initials
Phola consonants in contrastive environments
Phola vowel finals in contrastive environments
Azha-LJY consonant initial phonemes
Azha-LJY consonants in contrastive environment
Azha-PJZ vowel phonemes in contrastive environment
Guide to shorthand for major natural classes of PNg initials
Overview of Phula tonal reflexes with reference to other Ngwi
languages
Unmarked tonal reflex patterns in Phula and other Ngwi
languages
Muji clade tone correspondences for TC-1
Muji clade tone correspondences for TC-2
Muji clade tone correspondences for TC-3
Muji clade tone correspondences for *checked classes TC-L
and TC-H
*H > 33/ *[+continuant]_ in Core Muji
*H > 21/ *[-continuant]_ in Core Muji
*L > 21/ *[+continuant]_ in Core Muji
*L > 33/ *[-continuant]_ in Core Muji
*L > 21 and *H > 33 / *[-continuant] > PM *[+medial]_ for
Core Muji
An unusual rhyme reflex pattern in Core Muji
Summary of 14 Muji clade innovations
Phowa clade tone correspondences for TC-1
Phowa clade tone correspondences for TC-2
Phowa clade tone correspondences for TC-3
Phowa clade tone correspondences for TC-L and TC-H
Summary of 14 Phowa clade innovation sets
Highland Phula lexico-semantic innovation set one
Highland Phula lexico-semantic innovation set two
TC-1 tone correspondences for the Riverine Phula macro-clade
TC-2 tone correspondences for the Riverine Phula macro-clade
TC-3 tone correspondences for the Riverine Phula macro-clade
*Checked tone correspondences for the Riverine Phula
macro-clade
Summary of Riverine Phula innovation sets
*bl/y and *pl/y cluster reflexes in various Ngwi languages:
Dataset 1
*bl/y and *pl/y cluster reflexes in various Ngwi languages:
Dataset 2
255
256
260
265
266
268
274
276
277
288
290
292
293
294
295
296
297
298
301
308
316
318
320
321
325
331
332
338
339
340
341
347
358
359
List of tables
7.3. *ʔ-, *b- and *p- prefixed *resonant initial reflexes and the
exclusion of *b-l and *pr
7.4. Further lateral cluster conditioning environments and the
exclusion of *k-l syllables
7.5. Family group classifier samples from Azha and Phula Proper
7.6. ‘egg’ and ‘lay(an egg)’ in SE Ngwi, Nuosu and Nasu
7.7. ‘bat’ < ‘flyer’ in SE Ngwi
7.8. Three Azha lexical isoglosses
7.9. Exclusive devoicing of TC-1 *stop initials in Sani and Axi
7.10. Comparison of Sani tone transcription conventions
7.11. TC-1 tone correspondences for Sani, Axi, Azhe and Azha
7.12. TC-2 tone correspondences for Sani, Axi, Azhe and Azha
7.13. Problematic tone reflexes of TC-2 *s initial syllables in the
SA clade
7.14. TC-3 tone correspondences for Sani, Axi, Azhe and Azha
7.15. *Checked tone correspondences for Sani, Axi, Azhe and Azha
7.16. Pholo reflexes of SE Ngwi lateral cluster conditioning
environments
8.1. Riverine population and village distribution estimates by
language and county
8.2. Muji-clade population estimates by language and county
8.3. Muji-clade village distribution estimates by language and
county
8.4. Phowa-clade population and village distribution by language
and county
xxv
362
364
367
371
373
377
379
379
383
384
385
386
387
390
402
409
410
415
List of figures
1.1. Per-county proportion of total Phula villages
1.2. Theoretical orientation of dialectology assumed in this work:
linguistic and extra-linguistic considerations
1.3. Proposed dialogic approach to dialectology: Questions and
methods
1.4. Organization of the featured dialectic argument
2.1. Illustration of lexical elicitation recording setup
2.2. Supplementary IPA vowel conventions used by linguists in China
2.3. Bailey’s dynamic wave model of cross-lectal innovations
3.1. The Phowa dialect continuum
4.1. Summary of Phula threatened language status by village totals
5.1. The Hlepho Phowa-FZK vowel system
5.2. Hlepho-FZK pitch plots and tonemes: nasal initial environment
5.3. Hlepho-FZK pitch plots and tonemes: stop initial environment
5.4. Allophonic gradience and the PJZ vowel system
5.5. PJZ pitch plots and tonemes: Nasal initial environment
5.6. PJZ pitch plots and tonemes: Stop initial environment
5.7. The Phuza-BJB vowel system
5.8. Phuza-BJB pitch plots and tonemes: nasal initial environment
(male#1)
5.9. Phuza-BJB pitch plots and tonemes: stop initial environment
(male#2)
5.10. Phola prenasalized voiceless onsets: ɴqɑ³¹ ‘buckwheat’ and ⁿtɔ³³
‘drink’
5.11. Phola-LDC phonemic monophthong and diphthong finals
5.12. Phola pitch plots and tonemes: nasal initial environment
5.13. Phola pitch plots and tonemes: stop initial environment
5.14. Azha-LJY vowel final phonemes
5.15. LJY pitch plots and tonemes: Nasal initial environment
5.16. LJY pitch plots and tonemes: Stop initial environment
6.1. Simple network illustrating ambiguous weighted splits
6.2. Distance-based phylogenetic neighbor network phenogram of
Phula language varieties
6.3. Phula neighbor network marking languages and genetic/contact
clusters
6.4. Summary of PNg tonal reflex patterns in the nine known Muji
clade languages
23
33
34
42
53
66
92
159
220
235
239
239
244
246
246
251
253
253
257
259
261
262
267
269
269
280
282
283
286
List of figures
xxvii
6.5. Correlative redistribution of *H and *L in the Muji clade
299
6.6. Traditional isoglossic diagram of Muji clade innovations
310
summarized in Table 6.14
6.7. Simplified dynamic wave model of Muji’s core and periphery
311
6.8. Internal subgrouping hypothesis for the Muji meso-clade
312
6.9. Summary of PNg tonal reflex patterns in six Phowa clade
314
languages
6.10. Traditional wave diagram of Phowa clade innovations
327
summarized in Table 6.19
6.11. Simplified dynamic wave model of Phowa clade innovations
328
6.12. Internal subgrouping hypothesis for the Phowa clade
329
6.13. Proto-Ngwi tonal reflex patterns in the Riverine Phula languages 336
6.14. Traditional isoglossic diagram of Riverine innovations
348
summarized in Table 6.26
6.15. Internal subgrouping hypothesis for the Riverine Phula
349
macro-clade
6.16. Full Phula subgrouping hypothesis marking external contact
351
languages
7.1. Expanded Phula neighbor network incorporating Sani, Axi,
354
Azhe and Nisu
7.2. Diachronic development of *pl/y and *ʔ-kl initials in Ngwi
361
7.3. Diachronic development of lateral clusters in Southeastern
365
Ngwi
7.4. PNg tonal reflex patterns in Northern Nisu
370
7.5. PNg tonal reflex patterns in Sani-Axi-Azhe-Azha
381
7.6. Proposed internal genetic-contact composition of the
389
SA Meso-Clade
7.7. Working hypothesis on the genetic composition of
392
Southeastern Ngwi
8.1. Chiastic structure of the underlying dialectic argument
396
8.2. Dialectology as dialectic: Featuring Ontogeny as analogic
424
mediator
General abbreviations
AN
‘after next’ (e.g. ‘yearAN’= ‘year after next’; ‘yearAAN’ =
‘three years from now’, etc.)
asp
Aspirated (phonetic feature: [+asp] = aspirated, [-asp] = unaspirated)
AUG Augmentative
C
Consonant initial
C2
Consonant coda
CLF Numeral classifier (vs. NCL)
cont
Continuant (phonetic feature: [+/-cont])
CPV Completive
DEM Demonstrative
DIA
Local dialect form (e.g., from Xinan Guanhua ‘dialect’ of Mandarin)
DIM Diminutive
DRG Derogatory
DR
Downriver
EXT Existential (have ~ be)
EXC Exclusive
G
Glide segment
HON Honorific
HUM Humor (said with more jest than malice)
HP
Highland Phula (macro-clade: sister node to RP, daughter node
of PP)
ITF
Intensifier
LCh
Chinese loan
LOC Locative
Lx
Lexeme (Lx² = the second lexeme in a pair or set)
Mph Morpheme (used to refer to a specific morpheme: e.g., Mph² =
the second morpheme)
NA
Not applicable
NCL Noun classifier (vs. CLF)
NC
No cognate
ND
No data
Ng
Ngwi (formerly known as ‘Loloish’, ‘Yipho’, ‘Yiish’ and ‘Ni’)
NUM Number
(n)
Noun (denoting otherwise ambiguous part of speech for a given
lexeme)
General abbreviations
P
PB
PtLCh
PFV
PM
PW
PNg
PP
Q
REA
RP
RTT
RTP
S
SA
SE
SFX
SPV
T
TAM
TOA
TOR
UR
(v)
V
VGQ
VLQ
voc
xxix
(tense voice) Phonation
Proto-Burmic (a.k.a ‘Burmese-Lolo’, ‘Lolo-Burmese’, etc.)
Partial Chinese loan or lexeme calqued from Chinese
Perfective
Proto-Muji (meso-clade: sister node to PW, daughter node of
HP)
Proto-Phowa (meso-clade: sister node to PM, daughter node of
HP)
Proto-Ngwi
Phula Proper (perceptual clade including RP and HP)
Questionnaire
Reanalysis
Riverine Phula (macro-clade: sister node to HP, daughter node
of PP)
Recorded text test
Pilot tested RTT
Syllable (used in lexicon to refer to a specific syllable: e.g., S² =
the second syllable)
Sani-Axi-Azhe-Azha (macro-clade: sister node to HP and RP,
SE Ngwi daughter node)
Southeastern
Suffix
Superlative
Tone
Tense-Aspect-Mood/Modality
Term of address
Term of reference
Upriver
Verb (denoting the otherwise ambiguous part of speech of a
given lexeme)
Vowel final (or syllable nucleus)
Village Group Questionnaire
Village Leader Questionnaire
Voice (phonetic feature: [+voc] = voiced, [-voc] = unvoiced)
Kinship terms are abbreviated as follows but are only used in tandem
(i.e., to abbreviate a kinship relationship; e.g., FBW= ‘father’s
brother’s wife’ MOS= ‘mother’s older sister’):
B
‘brother’
F
‘father’
G
‘grand-’ (i.e., parent)
xxx
General abbreviations
H
O
M
P
S
W
Y
‘husband’
‘older’
‘mother’
‘parent’
‘sister’
‘wife’
‘younger’
Data source abbreviations
ADP (Phola): Honghe County, Menglong, Adipo
红河县勐龙乡阿底坡
BJB (Phuza): Mengzi County, Lengquan, Bujibai
蒙自县冷泉镇补鸡白
BSC (Phala): Honghe County, Baohua, Beishe
红河县宝华乡碑赊
CKB (Phowa, Hlepho, Phole): Wenshan County, Dehou, Chekabai
文山县德厚镇扯卡白
CTL (Alugu): Gejiu County (Shi), Manhao, Chongtianling
个旧市蔓耗镇冲天岭
DHN (Phowa, Ani): Mengzi County, Xibeile, Daheineng
蒙自县西 勒乡大黑能
DFC (Azha): Yanshan County, Ganhe, Faduke Dazhai (‘Dafa’)
研山县 河乡法都可大寨
DPD (Axi, YNYF 1984): Mile County, Xiyi Township, Dapingdi
弥勒县西一镇大 地
DXB (Bokha, Bokho): Jinping County, Ma’andi, Dixibei
金 县马鞍底地西
DXZ (Zokhuo): Wenshan County, Zhuiligai, Daxingzhai
文山县追栗街镇大 寨
FKC (Phupha): Gejiu County (Shi), Jiasha, Fengkou
个旧市贾沙乡风口
FNS (Phala): Honghe County, Yisa, Feinishao
红河县迤萨镇斐尼哨
FYX (Pholo): Guangnan County, Wuzhu, Fayixiazhai
广南县 铢乡法衣 寨
FZK (Phowa, Hlepho): Pingbian County, Xinhua, Feizuke
屏边县新华乡菲租克
xxxii
Data source abbreviations
GDC (Azha, Wu Z. 1996aandb): Wenshan County, Panzhihua, Gaodeng
文山县攀枝花镇高登新寨
GMD (Phupa): Mengzi County, Shuitian, Gamadi
蒙自县水田乡嘎马底
HLT (Phowa, Labo; Wang 2004): Kaiyuan, Mazheshao, Hulutang
开远市马者哨乡葫芦塘
HZC (Azha): Wenshan County, Dongshan, Huangzhai
文山县东山乡荒寨
JJC
(Phowa, Labo): Kaiyuan County (Shi), Beige, Jiaji
开远市碑格乡架吉
LCC (Moji): Xichou County, Lianhuatang, Luchaichong
西畴县莲花塘芦差冲
LDC (Phola): Yuanjiang County, Wadie, Luodie
元江县洼垤乡罗垤
LGM (Phowa, Labo): Kaiyuan County (Shi), Beige, Lugumu
开远市碑格乡鲁姑母
LJY (Azha): Wenshan County, Binglie Luojiayi
文山县秉烈乡倮家邑
LPC (Phukha, Edmondson 2002, Fried 2000): Lào Cai Province,
Vietnam
LSD (Muji, Northern): Pingbian County, Xinxian, Luoshuidong
屏边县新现镇落水洞
LZC (Khlula): Wenshan County, Liujin, Laozhai
文山县柳 乡老寨
MCD (Nisu, N., Pu et al. 2005): Shiping, Shaochong, Mocedian
石屏县哨冲镇莫测甸
MLT (Muzi): Gejiu County (Shi), Laochang, Malutang
个旧市老厂镇马鹿唐
MXC (Khlula): Maguan County, Muchang, Maxi
马 县木厂镇马西
MZC (Phowa, Hlepho): Mengzi County, Mingjiu, Meizichong
蒙自县鸣鹫镇梅子冲
Data source abbreviations
NGZ (Muzi): Gejiu County (Shi), Kafang, Nuoguzhai
个旧市卡方镇糯谷寨
NSC (Laghuu, Edmondson 2002): Nậm Sang, Sa Pa,
Lào Cai Province, Vietnam
NTC (Phola): Yuanjiang County, Lijiang, Natang
元江县澧江镇那塘
PJZ (Muji, Southern): Jinping County, Adebo, Pujiazhai
金 县阿得博乡普家寨
QLC (Muji, Qila): Jinping County, Jinshuihe, Qila
金 县金水河镇期腊
SZT (Muji, Southern, Muzi): Jinping County, Tongchang, Shizitou
金 县铜厂乡狮子头
SXZ (Pholo): Yanshan County, Ganhe, Longyong Shangxinzhai
研山县 河乡龙永 新寨
SZW (Phowa, Hlepho): Wenshan County, Baxin, Suozhiwan
文山县坝心乡所支弯
WBZ (Phowa, Hlepho): Kaiyuan, Beige, Xiaozhai (‘Weibazhu’)
开远市碑格乡尾巴猪小寨
WDP (Phowa, Labo, Asaheipho): Kaiyuan, Mazheshao, Wudupi
开远市马者哨乡乌都皮
WZC (Sani, Dai et al. 1992): Shilin County, Weize, Weize
石林县维则乡维则
XBL (Phowa, Ani): Mengzi County, Xibeile, Xibeile
蒙自县西 勒乡西 勒
XJC (Pholo): Guangnan County, Zhulin, Xiji
广南县珠林镇西吉
XPC (Muji, Northern): Mengzi County, Shuitian, Xiepo
蒙自县水田乡斜坡
XPB (Azha): Wenshan County, Binglie, Xiaopingba
文山县秉烈乡小 坝
XZC (Thopho): Guangnan County, Zhetu, Xinzhai
广南县者兔乡新寨
xxxiii
xxxiv
Data source abbreviations
YBZ (Bokha): Pingbian County, Dishuiceng, Yibaizu
屏边县滴水层乡邑佰租
YZT (Phuma): Pingbian County, Beihe, Yanzitou
屏边县白河乡岩子头
ZZC (Azhe, YNYF 1984): Mile County, Wushan, Zhongzhai
弥勒县 山乡中寨
Chapter 1
Introduction
Synchronic Phula in diachronic perspective
Thus, whether we appeal to oriental or western philosophy, to
Christian Scholasticism or to contemporary science, we find
among the deepest thinkers a resoluteness to steer clear of what
is fragmentary and to go for wholeness … a tertium quid that
provides the unifying relation that prevents these oppositions
from being antagonistic and truncated and instead makes them
complementary and mutually fulfilling.
– Charles-James N. Bailey (1982: 7)
On the Yin and Yang Nature of Language
1.1. The Phula hypothesis
The title ‘Phula’1 is a historical ethnonym with affiliates living on both sides
of the Sino-Vietnam border. The population in question currently stands at
around 367,0002 and is distributed over an area of 100,000 square kilometers – roughly the size of Arunachal Pradesh or the country of Guatemala
(see Maps 1.1 and 1.3). The ethnic title itself has traditionally been applied
to an array of related ethnolinguistic groups, consolidating them under a
common category; but, prior to the analysis and synthesis behind this book,
neither the category nor its constituency had ever been defined. In most instances of its application, the category-cum-title is accepted by in-group
members and out-group locals alike, but the range and implications of its
usage have remained unclear – and, for that matter, largely unexamined.
The Phula varieties speak languages belonging to the Ngwi3 branch of the
Burmic group in the Tibeto-Burman language family, but the actual number
of languages spoken under the Phula ethnonym and the nature of their historical relationships are two questions that have never been formally asked
or answered. In this work I attempt to define Phula by answering both questions: ‘synchronically’ speaking, how many Phula languages are there? and,
‘diachronically’ speaking, how do they fit together? To answer these questions, dialectology is called upon – and then called to task – to incorporate
2
Introduction
more complex syntheses of interdisciplinary inquiry into traditional dialect
geography.
After several years of inquiry, in the early 2000’s – after dozens of visits
to libraries great and small and scores of trips into the countryside of Yunnan Province, China – a strong, universal hypothesis slowly emerged that
provided focus for Phula data collection and analysis. The hypothesis was
explored in more tentative terms in Pelkey (2004, 2005a) and is presented in
more transparent terms in (1.1) below:
(1.1)
The Phula hypothesis:
All synchronic languages traditionally affiliated with the Phula
ethnonym also belong to a single exclusive diachronic clade
linguistically.
In scientific research, assertive hypotheses making vulnerable claims
tend to be more productive than timid hypotheses making safe claims.
Whether or not the claims of a given hypothesis are actually interesting is
another question. I personally find the Phula hypothesis to be of interest for
reasons related to undocumented diversity, undefined taxonomy, regional
ethnohistory, ethnolinguistic endangerment, intangible heritage maintenance,
linguistic human rights, cartographic experiment, dialectological reform and
general curiosity, among other interests. My assumption is that the research
question should also be of interest to others, such as Tibeto-Burmanists, areal ethnohistorians, areal language development workers, future dialectologists and Phula speakers themselves. Instead of arguing for the value of the
hypothesis in detail here, however, I will let the book speak for itself and
return to the theme intermittently.
Having asserted a transparent hypothesis of some interest in (1.1), then,
the ultimate business of this book is a thoroughgoing attempt to test it, disprove it, and refine it. By the final chapter, if I may spoil the suspense, the
Phula hypothesis is both disproved and refined; but six intervening chapters
of analysis are needed to illustrate how this is accomplished. The current
chapter is concerned with orienting the Phula research endeavor in time and
space, with some theoretical preliminaries thrown in for good measure; but
before we continue, and lest we forget, it will be useful to underscore my
intention to actually make the Phula hypothesis itself falsifiable.
The Phula hypothesis
1.1.1.
3
Falsifiability and the Phula hypothesis
In order to disprove the Phula hypothesis, it must first be falsifiable. This
tautology bears stating in order to underscore a problem: as it stands in (1.1),
the Phula hypothesis is not falsifiable; after all, ‘synchronic languages’ and
‘diachronic clades’ must be carefully defined, or ‘operationalized’, so that
their relationships can be tested and affirmed or disproved and refined. Both
must meet certain generally acceptable, sufficiently complex criteria in order to be validated (as falsifiable), and specific instances of each should be
defined through numerous layers of analysis – preferably engaging both
qualitative and quantitative measures in something of a dialectic synthesis.
In fact, the majority of this book is dedicated to making the premises of
the Phula hypothesis falsifiable. Falsifiability is asserted by Karl Popper
([1934] [1959] 2002) to be the touchstone for scientific inquiry – that standard which distinguishes science from other pursuits. Following up on problems pointed out by David Hume, Immanuel Kant, and C. S. Peirce, Popper
argues that purely inductive arguments, including appeals to probability, do
not constitute valid scientific claims. No quantity of positive testing can
render an empirical hypothesis verifiable. The empirical hypothesis may
still seem meaningful, or even warranted, but such claims may also be made
by non-scientific disciplines. The scientist (qua detective) hopes to learn
from experience by discarding false claims. In Popper’s own words (2002:
281), “The wrong view of science betrays itself in the craving to be right;
for it is not his possession of knowledge, of irrefutable truth, that makes the
man of science, but his persistent and recklessly critical quest for truth.”
Popper’s perspectives have gained ground among scientists and philosophers of science alike over the last 70-odd years (see discussion in Stone
1991 and Miller 2007). A falsifiable prediction provides the courtesy of a
clear scientific agenda on one hand, while discouraging the tempting fallacy
of assuming one’s ultimate conclusion, on the other. Popper’s claims also
provide a working interpretation of the natural history of science itself, 4 but
do his claims constitute grounds for an essentialist segregation, or ‘demarcation’, of scientific inquiry from other modes of inquiry? Popper’s demarcation criterion seems most plausible if we allow ‘science’ to be identified
with the pragmatist hermeneutic approach in general: the approach of the
detective, the riddle solver, the child immersed in language acquisition, the
interpreter of poems, the dialectologist puzzling over ambiguous variation –
in short, anyone genuinely seeking understanding – all who are willing to
dispense with faulty assumptions in the search for fresh truth.
Taken on his own terms, Popper’s thoroughgoing skepticism may well
be self-defeating: after all, one’s ‘recklessly critical quest for truth’ must not
4
Introduction
negate itself. C. S. Peirce ([1903] 1998) examines this and related paradoxes
in detail, noting that working hypotheses must at least be entertained, and
even assumed at some level, in order for them to be critiqued and cultivated
at all. Naturally, a certain tension emerges once a hypothesis is assumed to
be true, however tentative the assumption may be at first. Peirce provides
helpful perspective on Popper’s claims by introducing abductive inference
as the necessary grounds for both deduction and induction. The three work
in tandem in rational inquiry, with abduction (hypothesis formation) and induction (hypothesis testing) being mediated by deduction (hypothesis assumption) – inductive testing seeking to disprove and refine what abductive
guesswork has introduced before deductive certainty fossilizes and the original hypothesis is ultimately taken for granted. Thus, imagination, belief and
skepticism comprise the irreducible core of scientific logic, without any one
of which progress cannot be made. In this work, I seek to affirm all three
and, more importantly, seek to hold them in dialectic tension; thus, I hope,
pointing the way to a hermeneutic dialectology.
For the Phula hypothesis in (1.1) to qualify as a valid prediction according to Popper’s criteria, two primary courses of investigation are required: 1)
the identification and definition of synchronic language varieties putatively
affiliated with the ethnic category and 2) the identification and definition of
unique shared innovations that qualify its constituents as members of exclusive phylogenetic clades at successively higher levels. The first course of
investigation is traditionally framed as ‘synchronic’, the second as ‘diachronic’. In fact, the two are not discrete; each needs the other and both
require numerous layers of interdependent fieldwork and analysis.
1.1.2.
Chapter overview
Research stages and book content will be outlined at the end of the current
chapter. This introductory chapter provides a time-space orientation to Phula and a broad theoretical orientation to my assumptions on Phula dialectology. Having introduced the Phula hypothesis, we may move on to the heart
of the chapter in §1.2 – a discussion of the historical context of synchronic
Phula. Varieties affiliated with the Phula ethnonym are sketched out in
terms of official and folk classifications, ancient records and oral histories.
The chapter then moves on to situate the Phula varieties in space by discussing the range of their distribution in geopolitical terms and introducing
fieldwork datapoint locations in §1.3. I then discuss my approach to dialectology in §1.4 along with related assumptions and research scope. The
chapter concludes with an overview of the book itself.
Historical background
5
1.2. Historical background
In order to approach the Phula varieties in their synchronic situation, it is
helpful to understand something of their ethnic history and historical geography. Where did they originate? How did they end up where they are? How
were they identified in ancient times? How are they classified today? What
do oral histories and official records have to say about their migration routes
and relationships? Sketching answers to these questions from the available
records provides a useful backdrop for thinking about the languages by placing their speakers in a historical context. Underlying these questions is the
much more brief history of my own interest in the Phula languages.
My interest in Phula began in 1997. Fresh out of university, I was studying Chinese and slaking my wanderlust in the hidden realm of marvels that
is Yunnan Province, China (see Maps 1.1, 1.2, and 1.3). Of the 26 official
ethnic nationalities of Yunnan, the internal diversity of the Yi Nationality is
the most complex, and it had captured my imagination even before I arrived
in Chuxiong, Yunnan’s Yi Autonomous Prefecture, for Chinese studies.
After arriving in Yunnan, I started tracing the diversity of the Yi Nationality like an amateur sleuth (with more zeal than knowledge), and the clues I
followed sent me to the four corners of the province in search of understanding. Most talk of Yi that I came across in English sources at the time discussed only the Nuosu varieties of the Great Cool Mountains to the north
across the Jinsha (a.k.a. Yangtse/Changjiang) River in south-central Sichuan
Province – to the apparent neglect of Yi diversity in Yunnan. But here in
Yunnan were scores and scores of distinct groups classified as Yi who were
wearing unique ethnic dress, using unique ethnic titles, hosting distinct ethnic festivals and maintaining diverse languages and cultures. Many were
even given discrete acknowledgement in local ethnohistorical sources. Was
I just naïve or had there been some mistake? Why wasn’t there more
discussion and research on the Yi varieties of Yunnan?
One Yi title that surfaced repeatedly in local sources and in conversations with locals alike was ‘Phula’ – with written and spoken variations including ‘Pula’ 仆拉, 普拉, 濮拉, 朴喇, ‘Bola’卜拉, ‘Puzu’ 仆族 , and
‘Pulapo’ 仆拉颇. Phula seemed to be the most complicated knot in the Yi
Nationality to untangle, so I spent more and more time picking at it. There
seemed to be at least a dozen unique ethnic groups affiliated with the title.
How closely related were they? Phula was (and is) an official nationality in
Vietnam. Was it once a traditional nationality in China also? The only two
Phula groups classified by Bradley (1997a) were assigned to distinct genetic
clades linguistically. Was Phula once a single language or not?
6
Introduction
Gansu
Qinghai
Jiangsu
Henan
Shaanxi
Anhui
Sichuan
Tibet
Maps 1.2 and 1.3
Hubei
C
H
I
N
A
Zhejiang
Jiangxi
Hunan
Fujian
Guizhou
Yunnan
Maps 1.4 and 1.5
Phula
Distribution
Area
Guangdong
Guangxi
VIETNAM
LAOS
Hainan
0
200
400km
Map 1.1. Broad geographic orientation of Yunnan Province, China, and Phula distribution area
Hence, my overriding interest in dialectology from the outset has been
marked by a distinctly pragmatist, hermeneutic drive – a strong compulsion
to understand Phula that required the development of theory, methodology
and praxis adequate to the task. My early linguistic research on Phula
(Pelkey 2004) was carried out through Payap University in Chiang Mai,
Thailand, in cooperation with Honghe University in Mengzi, Yunnan. The
research focused on a polylectal variety in northern Mengzi County which I
now refer to as ‘Ani Phowa’.
The Phula category itself, I came to realize early on, also contains numerous distinct ethnolinguistic groups such as Azha, Muji, Phola, Phowa
and Zokhuo. As the years passed, I came to understand that not only is Phula embedded in Yi, but several of the groups embedded within Phula also
contain their own embedded ethnolinguistic varieties. So much ethnic embedding once seemed odd to me; especially since, generally speaking, it
does not seem odd to the individuals who are thus embedded.
Historical background
1.2.1.
7
Phula, Puzu, Yizu and other classifications
To better account for the ethnic embedding of the Phula groups, it is useful
to understand the various classifications, both official and traditional, under
which the Phula varieties are known to have been subsumed. Two key ethnic classification movements from the twentieth century that are relevant to
the discussion are the ‘Minzu Shibie’ project in China and the ‘Dân Tộc’
project in Vietnam. A contemporary local-level dichotomy between ‘Puzu’
and ‘Yizu’ is also of interest.
In the early days of my zeal for Yi diversity, I wondered if the condensed
ethnolinguistic status of the Yi Nationality did not simply represent an injustice; after all, shouldn’t diversity be described and celebrated? What I did
not understand is that this condensed status actually represents the continuation of an ancient tradition of ethnic agglomeration in the region (see discussion in Bradley 2005a: 11–12) – a tradition of agglomeration that most
regional ethnic groups happen to be fairly comfortable with.
Critical stances toward surface peculiarities are usually best tempered by
an emic understanding of underlying traditions. Take a group like PholeCKB5 of northwestern Wenshan County, for example. Even before the 1954
ethnic classification project, Phole-CKB group members would already
have been operating under five layers of embedded ethnic identity (see
§2.5.1, §3.2.3): 1) village-level: ‘Chekabai’, 2) autonym-level: ‘Phole’, 3)
endoautonym-level: ‘Hlepho’, 4) endoethnonym-level: ‘Phowa’ and 5) historical-ethnonym-level: ‘Phula’ – none of which they would have been likely to deny. With five layers of embedded identity, each incorporating a
given Phole-CKB speaker in a successively larger, more diversified group,
what difference would a sixth superordinate category make – especially if it
included the speaker in one of the largest, most celebrated Nationalities in
China? While ethnic agglomeration may lead to the neglect of ethnolinguistic diversity (see further discussion in §8.6), outside researchers should understand that the tradition is not new, and, in most cases at least, embedded
members of the Yi Nationality are not displeased with the additional layer.
The 1954 Minzu Shibie 民族识别 ethnic classification project in China is
now thoroughly documented in Mullaney (2011). One consequence of this
grand ethnic downsizing was the deep embedding of Yunnan’s Phula languages within the new agglomerate Yi Nationality. Keyes (2002) provides a
comparable description of the 1979 Dân Tộc classification project in Vietnam which classified various Phula languages along more traditional lines
under the Phu La Nationality. One unfortunate consequence of such embedding is the loss of official census statistics for specific language groups (see
further discussion in §2.5.2). However, at lower administrative levels,
8
Introduction
traditional categories remain in the collective consciousness of in-group and
out-group locals alike, even if they are not included in official census
reports.
One clue to the traditional, pre-1954, status of Phula in southeastern
Yunnan is the informal use of ‘Puzu’ 濮族 to refer to groups and individuals
affiliated with the Phula ethnonym. Since the 1950’s, the morpheme, or
character, ‘zu’ 族, which formerly meant simply ‘clan’ or ‘(ethnic) group’
when used in an ethnographic context, has taken on semantic restrictions
and is now used almost exclusively in reference to official nationalities in
China – i.e., in reference to one of the 56 official ‘minzu’ 民族. In a formal
situation, then, a reference such as ‘Puzu’ might seem to imply that Phula
has ‘minzu’ status. Since Phula is not an official nationality in China, such a
reference is formally frowned upon. In casual conversation, however, I have
noticed that even local government officials are not above using the term –
especially as a vernacular shorthand for distinguishing between a given Nisu
language (another Ngwi-branch cluster in the region also classified under
the Yi Nationality) and a given Phula language at the local level. In such a
situation, it is simply more convenient to say ‘Yizu’ and ‘Puzu’ than ‘Yizu
Nisu Zhixi’ (The Nisu Branch of the Yi Nationality)6 and ‘Yizu Pula Zhixi’,
(the Phula branch of the Yi Nationality), respectively. Locals functioning in
casual settings tend to use the Puzu title liberally in order to distinguish between various Nisu and Phula varieties, a point that is discussed further in
§1.2.2.
The title is more than a shorthand, however, as became apparent numerous times during various stages of fieldwork. A Phula official in Jinping
County, for example, reported that she hadn’t known she was officially Yizu
until she registered for high school (in the 1960’s or 1970’s). All of her
growing up years she had thought of herself as ‘Puzu’. Even official documents sometimes use ‘Puzu’ in reference to various Phula ethnic groups,
e.g., WSDZ (1988: 112), which introduces Azhabo 阿扎簸, a place name in
Xigu Township of central Wenshan County, and then explains that although
the village is now a Miao Village, it was originally an Azha village, hence
the first two syllables in the name ‘Azha’, the autonym used by the Boji
(‘winnowing basket’) Pu. The third syllable is a transliteration of the Azha
lexeme for mountain. Thus, the editor translates the village name as ‘Puzu
Mountain’.7 As mentioned in Pelkey (2005a: 52) a local tourist poster / brochure set in Mengzi County was recently recalled for editing after a Phula
group in the County was described using a ‘Puzu’ label. Thus, even though
the title is no longer officially acceptable, its continued usage suggests the
presence of an ethnic category with long historical precedence.
Historical background
9
The six Yi fangyan 方言 (usually translated, ‘dialect’) regions outlined
in Chen, Bian and Li (1985) present a more recent classification of Yi languages, slightly more oriented toward linguistic criteria, that include the
Phula varieties by matter of course. As discussed in Pelkey (2005a), different Phula languages in China are now classified as ‘Southeastern Yi’ and/or
‘Southern Yi’ depending on their respective geographic locations. In at least
one marginal case between the two regions, Chinese linguists note that a
Phula variety (which turns out to be a variety of ‘Muji’) might well be classified as either Southern or Southeastern Yi (HHYC 2002: 14). Although
some prototypical features have been outlined in support of the six Yi Fangyan hypothesis in Chen, Bian, and Li (1985), the criteria are rather more typological than diachronic; also, most boundaries fall into tidy patterns along
geographical regions, so this classification should not be thought of as
rigorously linguistic even at the typological level.
All other classifications of Phula languages at lower levels are less official and involve folk linguistics, material culture and ethnic identity issues.
Classifications at lower levels often overlap or are contradictory between
locations. Ethnonyms such as Hua Phula (Flowery Phula), Niuweiba Phula
(Cowtail Phula), Labopho (Highland Pho), Muji, and Phowa all function on
different levels and indicate different relationships, perspectives and assumptions. Discussion of this level of folk-classification is deferred until
Chapter 3, where ethnic identity issues are discussed in detail as part of the
process of synchronic language definition. For now, we turn to the PhulaLolo distinction asserted in historical records from ancient times through to
the twenty-first century.
1.2.2.
The Phula-Lolo distinction in historical records
Perhaps the best evidence for accepting Phula as an ancient ethnic category
is the long-standing distinction between Lolo and Phula asserted across
southeastern Yunnan and Northwestern Vietnam in a range of historical
records. The scope and importance of the traditional Phula-Lolo distinction
were first discussed in Pelkey (2005a).
In the contemporary local vernacular, as discussed in §1.2.1 above, the
distinction is usually represented as Yizu (Lolo) and Puzu (Phula) since the
older titles ‘Lolo’ and ‘Phula’ are now considered derogatory in many regions; but in Vietnam, where ethnic categories are slightly more specific,
the traditional Lolo-Phula distinction is trenchant enough for the two to be
assigned to distinct nationalities. In Yunnan the distinction is reinforced in
local historical documents and by the accounts of European explorers and
10
Introduction
anthropologists who traveled in Yunnan during the late nineteenth and early
twentieth centuries.
Many modern-day Chinese ethnohistorical sources of southeastern Yunnan present brief summaries of the internal diversity of the Yi Nationality at
the local level. Such sources (e.g., WSZZ 2000, WSXZ 1999, MZXZ 1995,
HHXZ 1991) frequently reference the traditional Phula-Lolo distinction
with a zhixi 支系 ‘branch’ modifier – i.e., Pula Zhixi 濮拉支系 ‘Phula
Branch’ and 倮 倮 支 系 Lolo Zhixi ‘Lolo Branch’ – and cite historical
documents as old as the Ming Dynasty (1368–1644) for precedence.
European explorers and ethnologists of the late nineteenth and early
twentieth centuries also call attention to this tradition in passing. Three examples may be noted in the writings of d’Orléans ([1898] 1999), Henry
(1903) and Davies (1909).
Prince Henri d’Orléans (1999) passed through the Phula region in early
1895 as part of an expedition surveying French interests overland from Hanoi to Calcutta. He notes many encounters with Phula (transcribed ‘Poula’)
in Mengzi and near the Honghe River in Manhao and adjacent regions during the early stages of his journey. His ethnic classifications are adopted
from casual conversation with locals, and, on these grounds, he makes a
clean distinction between Lolo and Phula.8
Linguist-anthropologist Augustine Henry (1903) affirms a Lolo-Phula
distinction to apply to the region between Yuanjiang and Mengzi (see Map
1.3). He refers to the Phula groups as ‘pigmy’ populations9 and assumes that
they have an aboriginal status compared to the Lolo.
H. R. Davies (1909), a British military surveyor and amateur linguist/
ethnologist, also makes use of the Phula-Lolo distinction in his proposed
linguistic taxonomy of the languages of Yunnan. He erroneously classifies
Phula as a subtype of Woni, however – which, in turn, he identifies as a
sub-type of Lolo (in this case represented by Nisu).
Although nineteenth century European conclusions based on the PhulaLolo distinction were not always sound, the appearance and re-appearance
of the contrast in so many historical sources provide further grounds for affirming the title as a historical ethnonym. To pick up on an earlier theme, an
important point that should be noted about the Phula historical ethnonym is
its function in history past – a function much the same as the Yizu ethnonym today: a general cover term for numerous distinct, but related, ethnolinguistic groups10. Some of the earliest of these groups are also noted in
Chinese historical records.
Historical background
1.2.3.
11
Retracing Phula migration patterns from oral and written histories
No Phula language is known to have had a written form prior to 1987.11 All
non-linguistic details about Phula history must rely on passing mention in
official Chinese records and oral traditions passed down between generations. Such details are not always historically reliable, but they at least provide a good starting place for further inquiry. Official ethnohistorical
documents in rural Yunnan have long recorded selected details on the origin
and status of ethnic groups under the purview of a given administrative region. Such information often includes oral traditions of ancient migration
routes. During the fieldwork component of this research project I also elicited a small collection of oral traditions concerning migration patterns. Combining this information and tracing some of the major themes enables a
tentative sketch of Phula history useful for approaching synchronic varieties
and forming initial hypotheses about their relationships. Reported patterns
may be sub-divided into ancient macro-geographical movements and more
recent micro-movements.
1.2.3.1. Macro-migration patterns
The earliest records of reported migration routes involve the ancestors of at
least six contemporary groups affiliated with the Phula ethnonym: Azha,
Phowa/Muji, Zokhuo and Phola/Phala. Two locations of origin and four migration routes can be traced for these five groups. Azha and the Phowa/Muji
groups reportedly trace their ancestry back to the shores of Dianchi Lake
near modern-day Kunming, while most Zokhuo and Phola reportedly trace
their ancestry back to the Dali region in northwestern Yunnan, as do the
Sani of Shilin County. These locations are marked on Map 1.2.
Based on genetic linguistic findings discussed in Chapters 6 and 7, we
may assume that one of these Phula settlements was prior to the other. The
best candidate for this proto-homeland would seem to be the Dali region
since speakers of two widely distinct Phula languages, Zokhuo and Phala,
both trace their ultimate ancestry to the region. The region near Dali was a
center of political power for two successive kingdoms which were independent of Chinese rule for a combined total of more than 500 years: the
Nanzhao Kingdom (718–902) and the Dali Kingdom (937–1253)12. Given
this scenario, the ancestors of Sani-Axi-Azhe-Azha and Highland Phula
may well have moved (or been relocated) to the shores of Dianchi Lake
from Dali along with the ancestors of the Nisu sometime during the
Nanzhao Kingdom (see also MLXZ 1987: 698).
12
Introduction
In support of this theory, Backus (1981: 63–66) describes the events surrounding the Nanzhao conquest of the divided Cuan 爨 kingdom during the
740’s which would have precipitated such an emigration from Dali. He
notes that in the year 748 more than 200,000 Cuan families (the majority of
the Cuan population) were forcibly removed from the area that is now
northeastern Yunnan, including the shores of Dianchi Lake, and resettled into western and southwestern Yunnan (Backus 1981: 66). This resettlement
enabled the Nanzhao kingdom to gain control of a region that was slowly
yielding to the influence of the Tang Dynasty. After deporting the local
population and cutting their leadership off from further negotiations and alliances with Tang rulers, Nanzhao authorities would have naturally repopulated the region with their own loyal citizens and soldiers. This historical
scenario provides a plausible rationale for the relocation of the ancestors of
modern-day language groups such as Phowa, Muji, Azha and Sani from the
Dali region to the shores of Dianchi Lake sometime during after the middle
of the eighth century.
Some 150 years later, in the waning years of the Nanzhao, near the beginning of the tenth century and sometime before the end of the Tang Dynasty in 907, the ancestors of modern-day Azha are said to have begun
moving into the region which is now northeastern Wenshan County (WSZZ
2000: 389; WSXZ 1999: 184). To this day Azha speakers consider Milewan
迷勒湾 in Binglie Township13 to be the home of their ancestors’ spirits –
this being the place they send the spirits of their deceased, according to
custom (WSZZ 2000: 389) as discussed in Pelkey, Wang and Johnson
(2005).
The early ancestors of the Phola and Phala languages, which I dub ‘Riverine Phula’ in Chapter 6, reportedly migrated into Yuanjiang, Shiping and
Honghe Counties along the Honghe River from their ancestral homeland in
the Diancang Mountains, just west of Erhai Lake (HHYC 2002: 53, 83). The
headwaters of the Honghe begin just south of Dali. HHYC (2002: 83) records a composite oral history from Phola and Phala speakers in Menglong
勐龙 Dayangjie 大羊街, Baohua 宝华 and Yisa 迤萨 that traces the Riverine Phula migration route along the Honghe river basin from modern-day
Weishan County through Shuangbai, Xinping and Yuanjiang Counties into
Honghe. The time of this migration is unknown but is unlikely to have been
later than the end of the Dali Kingdom in 1253 at which time numerous Bai
and Ngwi-speaking inhabitants in the area fled from Mongolian invaders.14
In support of this hypothesis, the middle waters of the Honghe River were
known as Pu Shui 濮水 15 in ancient times, presumably named after the
Historical background
13
Phula inhabitants whose villages lined its banks (HHYC 2002: 53, 83;
WSZZ 2000: 388).
TIBET
SICHUAN
GUIZHOU
Erhai
Dali
Y
U N N A N
Kunming
Dianchi
Fuxian
Yuanjiang
Honghe
Huaning
Kaiyuan
Mengzi
Wenshan
VIET NAM
MYANMAR
0 50 100 km
GUANGXI
LAOS
Map 1.2. Major river systems of Yunnan and ancient Phula settlement sites
The ancestors of modern-day Phowa and Muji groups, which I collectively dub ‘Highland Phula’ in Chapter 6, reportedly dwelled along the
shores of Dianchi Lake for several hundred years until the middle of the
thirteenth century. Speakers of the Highland Phula ancestor language are
said to have begun migrating south sometime between the fall of the Dali
Kingdom in 1253 and the advent of the Yuan Dynasty in 1271. Chinese historical records and oral histories (HHYC 2002: 43, 50, 83; MZXZ 1995:
131) hold that this wave of Phula migration passed into the fertile basin that
lies between the Mengzi and Kaiyuan County seats via what is now Huaning County – a slow migration that continued into the Ming Dynasty (1368–
1644). In support of this migration route, Huaning County historical records
14
Introduction
(HNMZ 1990: 46) note the presence of Phula in ancient times (location and
time period unspecified) with an ethnonym transliterated Pula 扑拉 and a
possible exonym transcribed Heipu 黑噗 ‘Black Phu’, but the source notes
these former inhabitants left no trace of their presence in the contemporary
local Ngwi population – neither in terms of identity nor in terms of oral
traditions.
The earliest known mention of the ethnonym ‘Muji’ shows up in local
Ming Dynasty records. The title is transliterated 母基 (HHYC 2002: 48).
During the Ming and Qing Dynasties, what is now northern Mengzi County
(see Map 1.3) was part of a distinct administrative region known as ‘Ami
Zhou’, whose capital, Ami, corresponds with modern-day Kaiyuan City (see
You 1994: 273 and Map 1.2). HHYC (2002: 50) and MZXZ (1995: 131)
record that in the year 1573 the Muji of Mengzi and the Phowa of Ami
joined forces to revolt against the oppression of a local monarch. They appointed their own king and rallied the mountain peoples into a revolt. The
conflict lasted three years, but was finally unsuccessful (HHYC 2002: 50).
Thus, by the middle of the sixteenth century speakers of the Phowa and Muji ancestor languages had already developed somewhat distinct identities. In
the case of Muji, the identity was loconym-based (see §3.2.2 for further discussion). After the revolt failed, however, Muji descendants moved further
south into the mountains of Gejiu, Mengzi, Pingbian and Jinping Counties
(HHYC 2002: 50, 65–66, 83), among other places. Thus, a likely MujiPhowa split can at least be traced back to the middle of the sixteenth century,
as MZXZ (1995: 131) concurs.
A final ancient migration route that can be traced with relative certainty
is that of the ancestors of the Zokhuo. This overland migration route from
Dali to Wenshan seems to have been established during the Nanzhao Kingdom period when soldiers were sent southeast into what is now Wenshan.
Zokhuo speakers report that their ancestors came into the area over a period
of four hundred years for military and commercial purposes alike. Details of
the various stages are somewhat hazy, but they are recorded in multiple
sources such as WSZZ (2000: 389), WSXZ (1999: 184), WSJZ (1990: 117),
Lu (1987: 5–6) and YNSJ (2000: 97–99). Clearly, the most dramatic event
that led, in turn, to the most massive migration, occurred in the year 1253
when the Dali Kingdom fell to Mongolian invaders. The memory was apparently vivid since many Zokhuo speakers are still able to recount the specific details and stages of the journey which passed from Dali through what
is now Yuxi Prefecture and northern Honghe Prefecture into Wenshan – a
route that would have potentially reunited them with their (not-so-)long-lost
compatriots, the Highland Phula, along the way.
Historical background
15
Further macro-migration details related to Phula are available in Chinese
sources, but the above movements constitute a sketch of the major, most
widely cited, themes. As will become clear in Chapters 6 and 7, these details
do not account for all Phula groups or historical relationships, nor can they
all be accepted as historical fact, but they present us with a starting place for
understanding the diachronic background of the Phula milieu in its modern,
or ‘synchronic’ context.
1.2.3.2. Micro-migration patterns
Micro-migration patterns reported by the inhabitants of single villages are
also sometimes revealing, and the reported age of a given village site provides a further gauge of time depth useful for understanding population
movements. These questions were included in interviews with locals at most
of the datapoints researched during the fieldwork component of this project.
The oldest reported village sites will be reviewed here along with a more detailed treatment of Muji-affiliated migration patterns which are by far the
most complex in the Phula milieu. See Map 1.6 for county and datapoint
locations.
Phupa residents of Gamadi (GMD) in Shuitian Township of southern
Mengzi County estimate that the old upper-village site has been there for
over 500 years. Locals reports that their ancestors arrived at this site by
passing through Jianshui, presumably by following the Honghe River, down
to Hekou and then back up to their current location.
Phola residents of Luodie (LDC) in Wadie Township of Yuanjiang
County claim their village site is anywhere from 200–500 years old. Local
ethnohistorians, including those of Nisu descent, affirm that the village was
ethnically Phula at the time of its founding. Across the Honghe River and
slightly upstream, Phola residents of Adipo village, Lijiang Township,
Yuanjiang County report their village site to be more than 300 years old.
Phala residents of Feinishao (FNS) note that their village site was originally located in the heart of Yisa Zhen – the area that is now home to the
downtown shops and streets of the Honghe County seat. Feinishao residents
moved to their current location on the outskirts of town some 300 years ago,
but other Phala residents still live in pockets closer to downtown Yisa. This
micro-migration pattern lends evidence to the general folk knowledge in the
area that Honghe County seat was founded by Phula.
Phowa residents of Weibazhu (WBZ, more recently known as ‘Xiaozhai’)
in southeastern Beige Township of Kaiyuan County, claim their village has
been established for some 300–400 years, but most of the inhabitants are
16
Introduction
reported to have descended from a Phowa speaking family, with the surname Wang 王,16 who moved in to the village four generations (100–120
years) ago from Xin’ansuo Township in Mengzi County.
Phowa residents of Feizuke (FZK) village, in Xinhua Township of
northern Pingbian County report that their village was established some
eleven generations ago, making the village between 200–300 years old.
Some elderly Khlula residents of Maxi in Muchang Township of
Maguan County estimate that their forefathers immigrated to their current
location over 10 generations ago from the area that is now Wenshan County.
Alugu residents of Chongtianling and Alonggu villages in Huangcaoba
Township of southern Gejiu County report an interesting micro-migration
scenario. The Phula variety spoken in these villages is apparently a mixture
of closely related lects whose speakers converged on the village site. Speakers with the surname Li 李 are said to have moved in upriver from the east,
speakers with the surname Shi 石 are said to have moved in downriver from
the west. Speakers from the two clans maintain different pronunciations
and/or different forms for certain basic lexical items to this day such as
those listed in (1.2):
(1.2)
Clan-based lexical/pronunciation distinctions in Alugu
‘paddy field’
Li: tjɛ⁵⁵ŋə⁵⁵
Shi: tjɛ⁵⁵mi³³
‘demon’
Li: ʰɔ²¹a²¹ma²¹
Shi: ʰɛ²¹a²¹ma²¹
‘scatter’
Li: ɕɛ³³
Shi: ɕi³³
‘old’
Li: mɔ²¹
Shi: mɛ²¹
The /ɔ/~/ɛ/ variation in ‘old’ and the first syllable of ‘demon’ are both reflexes of the PNg *aŋ rhyme (i.e., PNg#535 *maŋ² and PNg#156 *tsaŋ¹,
‘person’, respectively). Based on rhyme correspondences in neighboring
languages, the Li clan seems to have once been a variety of Phupa, which,
together with Phuza, forms a closely related sister node to Alugu and Phupha – all four of which belong to the ‘Downriver Phula’ meso-clade of
‘Riverine Phula’ as defined in Chapter 6. Accordingly, Phupa is situated
downriver from the Alugu distribution.
Of all Phula migration patterns, those of the Muji-affiliated varieties
seem to range the widest. Since the defeat of their ancestors in 1573 (see
discussion in §1.2.3.1 above), Muji migration patterns have become more
and more intricate. Several Muji-affiliated dialect communities have migrated into Vietnam since the sixteenth century; and, as discussed in Pelkey
(2006), more than one such community has migrated back into Yunnan
from Vietnam.
Historical background
17
The last known villagers from the Muji-affiliated Bokho population of
Vietnam are said to have migrated back across the border into Yunnan in the
year 1979 (perhaps due to unrest associated with the Sino-Vietnam war)
where they were resettled in two adjacent villages of central Mengqiao
Township of eastern Jinping County. Bokho inhabitants of Dixibei in western Mengqiao Township report that their forefathers established the village
more than ten generations ago, but they report that five Bokho families
moved in from Vietnam sometime during the Qing Dynasty (from villages
named Xiaotongchang and Phulazhai), so Bokho migration has been fluid
for centuries in some cases but relatively stable in others.
The Muji inhabitants of Qila (QLC) in Jinshuihe Township of Jinping
County, on the other hand, returned to Yunnan from Vietnam 5 generations
ago and report that speakers of their unique lect still live in at least one village in Vietnam. They report being descended from Muji in Mengong and
Mengla villages of Lai Chau Province, Vietnam – situated to the southwest
of QLC.17 Originally, however, QLC speakers maintain that their ancestors
started off in China.
Geographic place name gazetteers at the county-level also record village
names that show evidence of previous Muji occupation in regions where
Muji speakers no longer dwell. Muji distribution apparently once extended
to Huangmaoling and Huangcaoling Townships of southeastern Yuanyang
County, for example, as is evidenced by several place names which are now
inhabited by Hani and Nisu residents (YYDZ 1992:95, 101, 104). Two villages named Mujizhai 姆基寨 in Huangmaoling are now inhabited by Nisu
and Hani and a third, named simply Muji 姆基 is also now a Hani village.
Also in central Leshichong Township of western Wenshan County is a
village named Mujichong 母鸡冲 which is now inhabited by Han and Miao.
A Qing Dynasty (1644–1912) document, Kaihua Fuzhi 开 府志, written in 1828, further attests to the historical presence of Muji (transliterated
母鸡) in the western Wenshan region and lists 192 villages under the supervision of Kaihua Fu,18 inhabited by Muji at the time (Tang et al. [1828]
2004: 54, 66–67). Kaihua Fu was a Qing Dynasty administrative region that
included much of modern-day Wenshan Prefecture and parts of far southeastern Honghe Prefecture (Tan 1996: 48–49). Muji villages are listed in the
source in general regions that correspond with modern-day Pingbian and
Hekou counties, the southern third of Wenshan County, the far eastern tip of
Jinping County and southeastern border areas of Mengzi County. The number of Muji villages listed in the source is roughly 15% higher than the total
number of modern-day Muji-affiliated villages that are in the geographic
18
Introduction
area corresponding to the old Kaihua Fu administrative region – a fact that
may indicate obsolescence, out-migration or both.19
A handful of early Muji speakers apparently migrated back into Huaning
County sometime during the Ming Dynasty according to historical records,
but they have since been assimilated into the local population (HNMZ 1990:
46). Oral tradition in Huaning County holds that they were the previous
inhabitants of Dalila and Xiaolila villages of Tonghongdian Township.
Understanding the wide-ranging migration tendencies of various Muji
lects helps to explain the surprising genetic status of a group using the autonym ‘Thopho’ as discussed in Chapters 3, 4 and 6. The variety is spoken in
two villages of eastern Guangnan County and turns out to be the most
distantly separated Muji-affiliate of all.
Not all Muji have ranged so broadly, of course. The Muji population
center is eastern Gejiu, southern Mengzi, western Pingbian and northern
Jinping Counties. The ancestors of Muji inhabitants in these areas did not
move far from their original homeland, and many villages such as Shizitou
(SZT) that lie in pockets just beyond the population center have been stable
for hundreds of years. SZT locals use Muzi as an autonym and claim to have
been in their current location for a total of seven generations. SZT elders
report that the original inhabitants of the village moved there from Mengzi.
1.2.4.
Previous research on the Phula languages
Linguistic research on the Phula languages was very sparse until the late
twentieth century, and this book constitutes the first comprehensive, published attempt to define these languages, but previous research on various
Phula varieties has been crucial in paving the way. Research done by Chinese scholars such as Yang Liujin, Bai Bibo, Wu Zili and Wang Chengyou
on various Phula varieties of China and research carried out by Jerold Edmondson on various Phula varieties in Vietnam have been particularly helpful. A summary of previous research on the Phula languages was presented
in Pelkey (2005a). I will review that discussion here and add further entries
published or discovered since that time.
Until the 1990’s, wordlist collection (mostly short wordlists) and sparse
comparisons with other Tibeto-Burman languages were the extent of linguistic research carried out on any of the Phula varieties. In addition to the
brief wordlists elicited by d’Orléans and Davies in the late nineteenth century (see discussion in §1.2.2, above), Fu ([1950] 1997) reports that F. S. A.
Bourne elicited two 38-item wordlists from Phula varieties in 1888 –
Historical background
19
perhaps the first linguistic fieldwork in history to be carried out on a Phula
variety.
According to items listed in Huffman (1986), the Phula varieties of Vietnam were first researched in 1906 through the comparative work of
Etienne Edmond Lunet de Lajonquière. Claudius Madrolle and Alfred Liétard followed up on this work in 1908 and 1909, respectively (Huffman
1986). These treatments include a small data sample from a Phula-affiliated
dialect as supporting evidence in more comprehensive comparisons between
various Tibeto-Burman languages.
Some Phula varieties were likely considered by Chinese researchers engaged in the 1954 Minzu Shibie project, but no data is known to be published from this era of research. Robert Shafer (1952) was apparently the
next linguist to publish research using Phula data. The data is from a variety
listed as ‘Phupha’ from Alfred Liétard and Claudius Madrolle. Shafer (1974)
also includes this Phupha data and makes reference to a Ngwi language
Tśökö (1974: 4) which likely corresponds with the glossonym Zokhuo in this
book.
According to Edmondson and Lama (1999), research on the Phula varieties of Vietnam resumed in 1975 when Vietnamese linguist Nguyễn Văn
Huy published a 13-page article claiming that out of three major ‘dialects’
classified as Phula in Vietnam, most share lexical similarity of 50% or less
and none share greater than 65%.
In 1984 (YNYF 1984) the Yunnan Institute of Nationalities (now the
Yunnan University of Nationalities) published a comparative wordlist of Yi
varieties, each listing over 2,700 lexical items. This lexical compendium
included Zokhuo, a Phula variety spoken in southeastern Wenshan County.
Various ethnohistorical sources in southeastern Yunnan have reported
lexicostatistic counts between a handful of Phula varieties. Criteria for determining similarity are not discussed, nor is the number of items compared
in several cases, but the presence of such reports demonstrates that some
degree of linguistic analysis has been undertaken. Such reports may be
found in WSZZ (2000: 394) which reports Tula (represented herein as
Khlula) to share 58% lexical similarity with Phowa (presumably Hlepho
Phowa) and 54% lexical similarity with Pholo. MZXZ (1995: 131) reports
that Phowa (presumably Ani Phowa) and Muji (presumably Northern Muji)
share 50% similarity (though lexical comparison is not necessarily implied).
GJMZ (1990) reports 70% lexical similarity between Muji and Phola.
ZSKY (1994: 636) reports that language varieties spoken by the Yi Nationality in Pingbian County (including four to five distinct Phula varieties and a
Nisu variety) share 40% lexical similarity on average. Finally, WSZZ (2000:
394) claims that Zuoke and Azha lexical similarity stands between 30% and
20
Introduction
60%.20 The source of the discrepancy in this final case is unknown. In many
cases these findings correspond with my own. Lexical comparison will be
discussed further in Chapters 2 and 3.
The most helpful and thorough linguistic research on the Phula languages has been published since the mid-to-late 1990’s. Bai Bibo pioneered
linguistic research among a remote Phula affiliated group called ‘La’ou’21
with fieldwork carried out in far southeastern Mengzi County in late 1988
and early 1989. Published in Bai (1994), his elegant sketches of La’ou phonology and grammar are the first known descriptive work carried out on a
Phula language. The next year, David Bradley (1995) included lexical and
morphological research on a Yunnan Phula variety, based on his own data
collection, as part of the analysis of a Ngwi-branch paradigm involving statives of dimensional extent. Two years later, Bradley (1997a) used data
from this Phula variety to support its preliminary genetic classification in
the Ngwi branch. Wu Zili’s (1996a, 1996b) 13-page phonological and
grammatical sketch of Azha constitutes the second published description of
a Phula-affiliated language. Wu Z. (1997) went on to provide a comparative
account of lateral clusters in minority languages of southwest China which
included data from six Phula varieties in Yunnan – a topic he had earlier
treated in Wu Z. (1992).
A few years later, Edmondson and Lama (1999) introduced the Vietnam
Phula language, ‘Xá Phó’, or ‘Laghuu’, to the English linguistic literature.
The article provides a phonological sketch of Laghuu and a brief word list.
The next year Fried (2000) used more of Edmondson’s Vietnam data to introduce another language of the Vietnam Phu La Nationality: ‘Phukha’.
Fried’s MA thesis includes a phonological sketch of Phukha and a 573 item
wordlist comparing Phukha and Xá Phó.
A few years later the editorial committee of the Honghe Yizu encyclopedic dictionary, HHYC (2002), published brief phonological inventories of
two previously undescribed Phula languages along with historical facts and
demographic figures related to several Phula varieties, largely based on
fieldwork carried out by Yang Liujin. The same year, Edmondson (2002)
published his comparative lexicon of seven minority languages of northern
Vietnam including Phukha and Laghuu data, and Bradley (2002) proposed
that the Phula varieties should be assigned to a distinct genetic subgroup of
Ngwi, ‘Southeastern Ngwi’. The next year, Edmondson (2003) discussed relationships between Phukha and Laghuu and compared lateral clusters between these and a number of other Phula languages in Yunnan using
supplementary data from Wu Z. (1997). The same year Wang Chengyou
(2003a) published a brief description (32 pp.) of a Phowa variety spoken in
Mazheshao Township of central Kaiyuan County. His work includes a
Field data orientation
21
grammatical and phonological sketch of the variety which he later reworked
and republished as (Wang 2004) with the addition of a 685-item wordlist.
Long (2004) published a similar sketch of ‘Bola’ phonology, syntax and
lexicon based data collected from a Phala variety in Honghe County.
My own published and presented research on the Phula languages includes Pelkey (2003, 2004, 2005a, 2005b, 2006, 2007a, 2007b), Pelkey et al.
(2005) and the current volume, which is a revision of my PhD dissertation
(Pelkey 2008). My MA thesis (Pelkey 2004) provides phonology and
grammar sketches of Ani Phowa, a Phula variety spoken primarily in Xibeile Township of northern Mengzi County, along with an analysis of
Phowa verbal semantics, an interlinearized narrative and a 927-item wordlist. Pelkey (2005a) constitutes my early attempt to consolidate the Phula
languages, and Pelkey (2006 and 2007b) constitute my early attempts to
define the Muji clade – the most diversified subgroup in the Phula complex.
1.3. Field data orientation
Modern-day Phula affiliates inhabit expansive pockets of southeastern Yunnan and also live in scattered villages of northwestern Vietnam (see Maps
1.4 and 1.5). An understanding of regional geo-political administration is
key for approaching data collection among the Phula languages. This section provides an overview of relevant administrative regions, population
pockets and other factors influencing the selection of village-level data
points for Phula fieldwork. The section also illustrates the locations of these
datapoints, provides a summary of the data collected and discusses a
handful of supplementary data sources considered in my analysis.
1.3.1.
Administrative orientation of Phula villages
Total Phula village counts and population estimates were unavailable at the
beginning of this research project (see further discussion in §2.5.2). I offered an early population estimate of 400,000 in Pelkey (2005a) based on information compiled from numerous Chinese ethnohistorical sources. Further
on-site fieldwork has enabled a more refined understanding of Phula demographics. I now place the ethnic population at 367,000 living in some
1,156 villages of southeastern Yunnan and northwestern Vietnam. Of this
number, 388 (34%) villages are reported to be mixed with other ethnic
groups.
22
Introduction
TIBET
SICHUAN
Diqing
Nujiang
Zhaotong
Lijiang
Dali
GUIZHOU
Chuxiong
Baoshan
Y
Dehong
International
Provincial
Prefectural
County
0
50
100 km
Simao
Xishuangbanna
MYANMAR
Kunming
GUANGXI
U N N A N
Lincang
Border Key
Qujing
Yuxi
Honghe
Mengzi
LAI CHAU
LAOS
Wenshan
Wenshan
HA CAO
GIANG
BANG
LAO CAI TUYEN BAC
QUANG KAN
YEN BAI
SON LA
VIET NAM
Map 1.3. Administrative overview of Yunnan and NW Vietnam with contemporary
administrative Phula regions shaded
More than 97% of the Phula population is located in China, where I estimate Phula affiliates inhabit some 1,125 villages in 105 townships of 18
counties. As Map 1.4 illustrates, Yunnan Province is divided into 16 prefectures (州/市/地区) and 130 counties (县/市/区). Below the county level lie
the township (乡/镇), community (村民委员会)22 and village (自然村/村民
小 组 ) administrative levels, respectively. Phula-inhabited counties are
located in three prefectures: Honghe, Wenshan and Yuxi.
By contrast, only about 3% of Phula-affiliated villages are located in
northwestern Vietnam, where they are scattered through four contiguous
provinces. As can be noted in Map 1.3, the land area of one Vietnam province is roughly equivalent to the land area of two or three counties in China.
The relative scale between administrative regions in the two countries is not
equivalent and should not be confused.
Field data orientation
23
Figure 1.1 illustrates the per-county breakdown of Phula villages.
Wenshan, Yanshan and Mengzi counties are home to the largest density of
Phula villages followed by Kaiyuan, Jinping, Maguan, Pingbian and Gejiu.
Village counts in other counties are sparse by contrast.
~1,156 TOTAL PHULA VILLAGES
VIETNAM, 31, 3%
Xichou, 1, 0.1%
Yuanjiang, 29, 3%
Shiping, 21, 2%
Honghe, 23, 2%
Guangnan, 22, 2%
Jianshui, 1, 0.1%
Malipo, 7, 1%
Yuanyang, 15, 1%
Maguan, 81, 7%
Gejiu, 58, 5%
Jinping, 99, 9%
Wenshan, 249, 22%
Mengzi, 148, 13%
Kaiyuan, 106, 9%
Yanshan, 165, 14%
Qiubei, 18, 2%
Pingbian, 76, 7%
Hekou, 6, 1%
Figure 1.1. Per-county proportion of total Phula villages
Map 1.4 illustrates the cluster density of Phula village distribution by
township. Comparing these patterns with Maps 1.2 and 1.3, we may note
that village clusters roughly pattern into four geographic regions (illustrated
in Map 3.1; cf. Map 1.5 for county locations):
1. Western: near the confluence of the Yuanjiang (Honghe) and Xiaohedi rivers in
Yuanjiang, Shiping and Honghe Counties.
2. South-central: especially along the common borders of Gejiu, Mengzi,
Yuanyang and Jinping Counties
3. North-central: especially along the common borders of Mengzi, Kaiyuan,
northwestern Wenshan and western Yanshan Counties.
4. Eastern: especially in Maguan, Wenshan, and eastern Yanshan Counties.
24
Introduction
These general geographic patterns are not necessarily meaningful in
themselves, but they provide a good starting place for understanding Phula
variety and relationships. Surface cluster patterns are especially useful for
understanding identity and language contact issues. Accordingly, this fourway geographical division will be picked up again in Chapters 3 and 4,
where it is interpreted, and tested, as a Sprachbund hypothesis.
0
25
50km
Many village locations in Vietnam are uncertain
cf. Map 1.3 and 1.5 for administrative names
Map 1.4. Phula village distribution patterns by township
Village cluster patterns provide one gauge for understanding Phula distribution. Other factors that should be considered for the definition of distribution polygons include intervening villages of other ethnic groups, villages
containing ethnic groups mixed with Phula, topography, river systems, and
administrative boundaries. Related issues are discussed in more detail in
§2.5.3. Phula distribution polygons rendered according to these guidelines
are compiled in Map 1.5 and elsewhere in the book.
39-XZC
Mile
Xinping
Qiubei
(MCD)
Shiping
Jianshui
21-NTC
31-DHN
34-XBL
20-LDC
19-ADP
18-FNS
Gejiu
14-MLT
17-BSC
Honghe
28-MZC
Mengzi
15-NGZ
22-CKB 5-XPB
4-LJY
3-HZC
33-SZW (GDC)
40-DXZ
26-FZK
38-BJB
1-CTL 13-XPC 10-LSDPingbian
35-YZT
36-GMD
37-FKC
Yuanyang
16-SZT
Wenshan
8-LZC
Maguan
6-YBZ
11-PJZ
Lüchun
24-SXZ
2-DFC
Yanshan
32-LGM
Kaiyuan
27-JJC
29-WBZ
Yuanjiang
25-XJC
23-FYX
(HLT)
30-WDP
9-MXC
12-QLC
N
0
25
(LPC)
VIETNAM
50km
Xichou
41-LCC
Malipo
Lao Cai
(NSC)
Lai Chau
25
Map 1.5. County level overview of Phula distribution marking data point locations
KEY
#- (e.g., 8-LZC): See Table 1.1
ABC: Primary datapoint
ABC: Supplementary datapoint
(ABC): External data source
Field data orientation
7-DXB
Jiangcheng
LAOS
Guangnan
Ha Giang
Hekou
Jinping
CHINA
Datapoints at which the full 1,200-item wordlist was elicited are
considered ‘primary’ datapoints. All other datapoints are considered
‘supplementary’. Datapoint importance was anticipated ahead of time,
but village stays were adjusted as necessary during the course of the
research, and some locations such as Alugu-CTL were visited more
than once in order to acquire sufficient data.
(ZZC)
Map# County
1. Gejiu
2. Yanshan
Township
Data Collected
Village
Huangcaoba Chongtianling
Dafa
Ganhe
Identity
Texts
Q
RTP
RTT
1200
250
3
0
2
2
1
0
Lx
Autonym
Shorthand
3
0
a⁵⁵lɯ³³ɡɯ³³
a⁵⁵tʂa³³
Alugu-CTL
Azha-DFC
3. Wenshan Dongshan
4. Wenshan Binglie
Huangzhai
Luojiayi
250
1200
5
5
2
2
0
1
0
0
a³³ a²¹
a³³ a²¹
Azha-HZC
Azha-LJY
5. Wenshan Binglie
Xiaopingba
500
0
2
0
0
a³³ a²¹
Azha-XPB
6. Pingbian
Dishuiceng
Yibaizu
1200
4
2
1
0
pʌ²¹kʰɑ⁵⁵
Bokha-YBZ
7. Jinping
Ma’andi
Dixibei
1200
2
2
1
5
po²¹kʰo⁵⁵
Bokho-DXB
Laozhai
Maxi
1200
250
3
0
2
2
1
0
0
0
kɬʰu²¹la³³
kɬʰu²¹la³³
Khlula-LZC
Khlula-MXC
8. Wenshan Liujing
9. Maguan Muchang
10. Pingbian
Xinxian
Luoshui
20
0
1
0
0
m̩ ²¹ʨi³³
Muji-LSD
11. Jinping
12. Jinping
13. Mengzi
Adipo
Jinshuihe
Shuitian
Pujiazhai
Qila
Xiepo
1200
1200
1200
5
5
2
2
2
2
1
0
1
0
1
1
m̩ ¹³ʥi³³
14. Gejiu
Laochang
Malutang
1200
3
2
0
2
mɯ²¹ʥi³³
m̩ ¹³ʣɿ³³
Muji-PJZ
Muji-QLC
Muji-XPC
15. Gejiu
Kafang
Nuoguzhai
16. Jinping
Tongchang
Shizitou
17. Honghe
Baohua
Beishe
18. Honghe
Yisa
Feinishao
19. Honghe
Menglong
Adipo
m̩ ¹³ʥi³³
Muzi-MLT
250
0
2
0
0
m̩ ¹³ʣɿ³³
Muzi-NGZ
1200
3
2
0
1
m̩ ¹³ʣɿ³³
Muzi-SZT
85
0
2
0
1
pʰa³¹la³³
Phala-BSC
1200
7
2
1
0
pʰa³¹la³³
Phala-FNS
500
4
2
0
1
pʰo³¹la⁵⁵
Phola-ADP
Introduction
Location
26
Table 1.1. Summary of Phula data collection linked to Map 1.5
Yuanjiang
Yuanjiang
Wenshan
Guangnan
Yanshan
Guangnan
Pingbian
Kaiyuan
Mengzi
Kaiyuan
Kaiyuan
Mengzi
Kaiyuan
Wenshan
Mengzi
Pingbian
Mengzi
Gejiu
Mengzi
Guangnan
Wenshan
41. Xichou
Wadie
Lijiang
Dehou
Wuzhu
Ganhe
Zhulin
Xinhua
Beige
Mingjiu
Beige
Mazheshao
Xibeile
Beige
Baxin
Xibeile
Baihe
Shuitian
Jiasha
Lengquan
Zhetu
Zhuiligai
1200
250
1200
10
250
1200
1200
1200
250
500
250
1500
100
250
110
1200
1200
1200
1200
1200
1200
5
0
1
0
0
4
5
9
0
4
1
5
3
0
0
4
6
4
4
2
6
2
2
2
1
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
1
0
0
0
0
1
1
1
0
1
0
0
0
0
0
1
1
1
1
0
1
0
0
2
0
0
0
0
0
1
1
3
0
0
0
0
0
1
0
1
0
0
pʰo³¹la⁵⁵
pʰo³¹la⁵⁵
pʰo²¹lə³³
pʰo⁵⁵lo⁵⁵
pʰo⁵⁵lo⁵⁵
pʰo⁵⁵lo⁵⁵
pʰo²¹va³³
pʰo²¹(va⁵⁵)
pʰo²¹va³³
pʰo²¹(va³³)
pʰo²¹va⁵⁵
pʰo²¹(va³³)
pʰo²¹(va⁵⁵)
pʰo²¹wa³³
pʰo²¹va³³
pʰɯ⁵⁵ma²¹
pʰɯ³³pa²¹
pʰɯ⁵⁵pʰa³³
pʰɯ⁵⁵za³¹
tʰo²¹pʰo³³
o²¹kʰuo⁵⁵
Phola-LDC
Phola-NTC
Phole-CKB
Pholo-FYX
Pholo-SXZ
Pholo-XJC
Phowa-FZK
Phowa-JJC
Phowa-MZC
Phowa-WBZ
Phowa-WDP
Phowa-DHN
Phowa-LGM
Phowa-SZW
Phowa-XBL
Phuma-YZT
Phupa-GMD
Phupha-FKC
Phuza-BJB
Thopho-XZC
Zokhuo-DXZ
Lianhuatang Luchaichong
250
0
2
0
0
Unspecified
Moji-LCC
114
79
18
24
32000
27
Luodie
Natang
Chekabai
Fayixiazhai
Shangxinzhai
Xiji
Feizuke
Jiaji
Meizichong
Weibazhu
Wudupi
Daheineng
Lugumu
Suozhiwan
Xibeile
Yanzitou
Gamadi
Fengkou
Bujibai
Xinzhaicun
Daxingzhai
Field data orientation
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
28
Introduction
1.3.2.
Fieldwork data point locations
The decision-making process for selecting data points factored in major
known population pockets, reported ethnic identity and/or dialect distinctions, geophysical features (i.e. mountain ranges and riverbeds) and geographic spread. Initial preferences on datapoint placement were revised as
necessary during the course of the research in order to factor in new insights
from locals, such as previously undocumented reports of diversity, and in
order to adjust for unexpected obstacles, such as impassable roads.
In the end, with the oversight of government administrators (discussed
further in §2.2.1 and §2.2.2), I carried out fieldwork in 41 village-level
datapoints located in 35 townships of 13 counties in three prefectures. Field
methods, data archiving and data analysis are discussed in §2.2. For now, I
draw attention to the distribution of the Phula data points considered in the
book which are plotted in Map 1.6. Datapoint numbers and acronyms are
linked to the Table 1.1 data elicitation summary.
Overall, the data-point spread, placement and quantity satisfied my
qualitative criteria for researching spatial distribution as widely as possible
and investigating reported variety in as much detail as possible within the
13-month time frame allotted for fieldwork. Due to time constraints, research priority and available lexical data from two varieties (see §1.3.4), no
fieldwork was carried out in Vietnam.
Some outlying pockets remain unresearched and may turn up important
distinctions which I have not accounted for in this book. Several of these
possibilities are reviewed in §8.7. More datapoints were originally planned
for Maguan and Malipo Counties in particular, but newly discovered diversity in Guangnan (39-XZC), Jinping (12-QLC), Gejiu (1-CTL) and Mengzi
(38-BJB) Counties took precedence.
1.3.3.
Overview of field data collection
Table 1.1 presents an overview of data collected during the fieldwork component of this research project.23 Datapoints are arranged in alphabetical order sorting autonym romanizations first followed by village names. Each
data point is assigned a number in the left hand column linked to its geographic location in Map 1.5. The three-letter abbreviations used in this chart
serve as identifiers of the corresponding language variety in the remainder
of the book. Chinese character versions of administrative locations are provided in the Data Source Abbreviation reference table listed in the front
matter of this book on pages xxxi-xxxiv.
Field data orientation
29
In most cases, the lexeme counts listed under ‘Lx’ in Table 1.1 are round
estimates including (for the primary data points) a basic wordlist of 1,127
items plus additional semantic distinctions, numeral classifiers, sentence final particles, and other items collected through elicitation frames or in order
to clarify various lexical issues. Thus, the total number of elicited items is
rounded off to 1,200. Research methods for lexical data elicitation are
discussed in §2.4.2.
Generally speaking, if a variety turned out to exhibit more variation than
expected in terms of lexicon, phonology, ethnic identity, folk linguistics and
other surface indicators during the course of the research, the research
schedule was adjusted to spend more time gathering data in the village, and
vice-versa for varieties that turned out to exhibit less variation than expected.
Thus, many of the primary datapoints turned out to represent unique synchronic languages; though this was not always the case (e.g., Phole-CKB,
Muzi-SZT and Bokho-DXB), as will become clear in Chapter 3.
The Q column in Table 1.1 represents the administration of sociolinguistic questionnaires as discussed in §2.4.1. The RTT column represents the
administration of recorded text tests using pilot tested (RTP) RTT’s from
other locations as a qualitative gauge of intelligibility. The RTP column represents the development and pilot-testing of a recorded text test in the village in-question. RTT and RTP research methods are discussed in more
detail in §2.6.2.
Oral text collection was carried out for recorded text testing, language
documentation and other purposes as discussed in §2.3, §2.4.3 and §2.6.2.
1.3.4.
External data sources
In order to make valid phylogenetic claims, a large corpus of data is needed
for comparison. Thankfully, recent decades have seen a substantial amount
of newly published lexical and phonological data on Ngwi-branch languages
spoken in China. External data sources consulted in this book are summarized in (1.3) listed in alphabetical order of language name with varieties
that are Phula-affiliated being listed in bold:
(1.3)
External data sources considered in this book
1. Axi (DPD): Mile County, Xiyi Township, Dapingdi (YNYF 1984)
2. Azha (GDC): Wenshan County, Gaodeng (Wu Z. 1996a, 1996b).
3. Azhe (ZZC): Mile County, Wushan Township, Zhongzhai24 (YNYF
1984); supplementary data from Wu Z. (1992), two varieties from an
unspecified location in Huaning County.
4. Bisu: Laopinzhai, Menghai County (Li 2004).
30
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Introduction
Hani: Luchun County, Dazhai (Sun 1991); Supplementary data from Dai and
Huang (1992), Mojiang and Luchun lects.
Jinuo: Jinghong County, Jinuo Township, Manka (YNSZ 1998).
Laghuu (NSC): Lao Cai Province, Sa Pa Township, Nam Sang (Edmondson
2002).
Lahu: Lancang County, Shofumenglaba (Sun 1991)
Lalo: Weishan County, Wuyin Township, Baiwudi (Dai and Huang 1992) with
supplementary data from Björverud (1998), Weishan County, Longjie Township, and Sun (1991), Nanjian County, Zhonghua Township.
Laomian: Lancang County, Zhutang Township, Laomiandazhai (Zhang 2004).
Lisu, Northern: Bradley (1994), Nujiang Prefecture; some forms also from
(Sun 1991), Fugong County, Jiakedi Township, Chada and YNSZ (1998),
Lushui County, Gudeng Township.
Lisu, Southern: Bradley (2006), northern Thailand.
Nasu: Luquan County, Sayingpan Township (YNSZ 1998); supplementary data
from Chen, Bian and Li (1985)
Nisu, Northern (MCD): Shiping County, Shaochong Township, Mocedian (Pu,
Kong and Pu 2005); supplementary data from YNSZ (1998), Xinping County,
Laochang Township.
Nisu, Southern: Zhu (2005), Wang (2003b), Yang (2009).
Nuosu: Sichuan Province, Xide County (Dai and Huang 1992); supplementary
data from Chen, Bian and Li (1985).
Phala: Honghe County, village unspecified (Long 2004).
Phowa (HLT): Kaiyuan County, Mazheshao Township, Hulutang (Wang
2004).25
Phukha (LPC): Vietnam, Lao Cai Province, Bảo Yên Township, 30 km southeast of Lao Cai city along the Honghe River (Edmondson 2002, Fried 2000).
Proto-Burmic: Matisoff (1972a, 2003)
Proto-Ngwi: Bradley (1979)
Proto-Tibeto-Burman: Matisoff (2003)
Sani (WZC): Shilin County, Weize Township (Dai and Huang 1992); Supplementary data from Wu, Ang and Huang (1984).
Nyisu: Shilin County, Zhuqing Township, Gaohanshan Community (Wu Z.
1997; YNYF 1984).
In total, Dialectology as Dialectic considers data from more than 70 language varieties, 41 from my own fieldwork on Phula and 30 or so from external sources, including three proto-languages, five additional Phula
varieties and 25 or so additional dialects of various Ngwi languages not affiliated with Phula. Other Tibeto-Burman languages are also referenced for
comparative purposes as noted in the text.
Research scope, assumptions and approach
31
1.4. Research scope, assumptions and approach
In terms of approach to fieldwork and analysis, the primary focus of this
work is dialectology. Dialectology is concerned with the discovery of variation patterns at the intersection of language usage and human geography –
whether socially, spatially or temporally conceived. I do not wish to pursue
dialectology simply for dialectology’s sake, however. The empirical discovery of isoglossic boundaries in geographic space should also attempt to operationalize – to interpret, to signify (see Bailey’s 1996d: 62-64 critique of
Labovian positivism in this regard). In short, I am faced with the necessity
of a more hermeneutic, integrative dialectology useful for making the Phula
hypothesis falsifiable in terms of its antecedent (‘synchronic’ languages)
and its consequent (‘diachronic’ relationships). This section introduces my
theoretical assumptions on dialectology and language definition and attempts to define the necessary boundaries that have delimited my fieldwork
and analysis.
1.4.1.
Theoretical assumptions: toward an integrative dialectology
The traditional focus of dialectology proper has been the geographic (traditionally spatial, more recently also social) distribution of language variation
as an end in itself. Chambers and Trudgill (1998) provide a standard account
of the traditional approach – a field of inquiry which has primarily been limited to dialect geography, isoglossic mapping and selective insights from sociolinguistics. Labov’s (1972) classic quantitative work on the social
motivations of language variation breathed new life into the field in the late
20th century and continue to inspire practitioners (cf. Labov 1994, 2010, in
press). Change is slow, but, as Chambers and Trudgill (1998: 189) note, traditional dialectology is continually opening up to include more and more
interdisciplinary perspectives.
One emerging branch is now known as ‘perceptual’ dialectology. Preston
(1989, 1999) presents the standard account of this approach by outlining
ways in which folk knowledge and language attitudes can be useful for understanding the significance of geographic variation. The importance of this
approach is discussed further in §2.6.1.
Other approaches to dialectology emphasize the overlapping significance
of historical linguistics and language contact with dialect geography. Although most practicing dialectologists tend to incorporate doctrines of the
regularity of sound change into their assumptions, traditional dialectology
has focused on synchronic surface variation, to the neglect of underlying
32
Introduction
diachronic processes or historical-comparative interpretations of this variation. Strains of dialectology such as Labov (e.g., 1994, 2001, 2010) and Bailey (e.g., 1973, 1982a, 1996a, 1996b) both emphasize the intersection of
comparative linguistics and dialectology. Both also highlight the gradual
gradience of language change, and both insist on the importance of investigating language contact for adequate dialectology. Bailey, however, is skeptical of the usefulness of traditional isoglossic representations to convey
anything substantial about the nature of language or linguistic relationships
(1996a, 1996d). He is also unwilling to accept the antitheoretical Labovian
‘inventory mentality’ that seems to find only chance statistical norms (see
Bailey 1996d: 64); instead, in order to better interpret the competing ambiguities of phylogenetic divergence and ecological convergence, Bailey seeks
to establish a theory that blends natural internal change, normative external
change and the implicational hierarchies that mediate between the two. His
approach is discussed further in §2.8.
Another strain of dialectology emphasizes intelligibility testing as a
means of discovering dialect divergence. Labov (2010: 19-86), Casad
([1974] 1987) and B. Grimes (1988, 1994) are examples of dialectologists
who work under the assumption that intelligibility issues are closely linked
to socio-cognitive consequences of language variation. The latter two, at
least, argue that comprehension testing via carefully prepared recorded text
tests should be actively incorporated into dialect definition. This approach is
discussed further in §2.6.2.
In defining the Phula languages, I am not interested in choosing between
these various strains of dialectology; rather, I wish to incorporate insights
promoted by each insofar as they are useful for making the Phula hypothesis
falsifiable. Two key theoretical assumptions underlying this work are that
dialectology should be useful for defining actual languages and dialects and
that, in the pursuit of such definitions, dialectology should be approached
dialectically. To the degree that perspectives which have traditionally fallen
outside the purview of dialectology proper are actually useful or relevant for
defining languages and dialects, they should, on these grounds, be incorporated into dialectology. I will argue for the viability of operationalizing
language definitions in §1.4.3 and §2.6. For now, let me simply assert that
major language varieties should be defined according to some adequate,
consistent set of criteria and that one job of the dialectologist should be to
discover and apply such criteria.
Figure 1.2 presents an illustration of the orientation I assume dialectology to take within linguistics proper (solid lines) and some of the most basic
extra-linguistic realities with which linguistics is concerned (dotted lines).
Research scope, assumptions and approach
Space
&
Geography
Historical-Comparative
•Reconstruction
•Subgrouping
•Semantic Shift
•Language Contact
•Grammaticalization
•etc.
Socio-Cognitive
•Demography
•Intelligibility
•Identity
•Vitality
•Usage
•etc.
Dialectology
33
Time
&
History
Typological-Descriptive
•Phonetics
•Phonology
•Lexicology
•Morphology
•Syntax
•etc.
Society & Culture
Figure 1.2. Theoretical orientation of dialectology assumed in this work: Linguistic
and extra-linguistic considerations
In addition to the innovation and maintenance of unique linguistic forms,
human languages are dependent on time, space, society and other extralinguistic variables. Presumably, then, an adequate study of human language
would seek to gain insights from descriptive-typological linguistics, sociocognitive linguistics and historical-comparative linguistics, within linguistics proper, and would also be well advised to ensure that such insights are
informed by historical, political and ethnographic contingencies lying beyond the scope of linguistics. If the job of the dialectologist is to discover
major patterns of language variation, dialectology itself would be best carried out at the overlapping margins of these three approaches to language
and would also be interested in their outside influences.
If dialectology is to be carried out at the intersection of comparative linguistics, descriptive linguistics and sociolinguistics for the identification of
languages and dialects, research methods and models should be designed
that foster insights from all three perspectives (a topic discussed more fully
in Chapter 2). Nor is there any apparent need to limit the dialectological
34
Introduction
analysis to procedures that are either purely quantitative or purely qualitative. Figure 1.3 summarizes some key practical ways in which dialectology
might be carried out as a fusion of ‘synchronic-diachronic’ oppositions on
the one hand and ‘qualitative-quantitative’ oppositions on the other.
core lexical comparisons,
acoustic measurements,
speaker demographics
U A L I T A T I V E
onomastic identity,
perceptual dialectology,
intelligibility testing
descriptive typology,
culture, ethnohistory
distance-based cladistics,
implicational hierarchies
comparative method,
subgrouping,
dynamic wave analysis
contact
distribution,
geolinguistics
DI
ACHRONIC
Q
U A N T I T A T I V E
endangerment
SYNCHRONIC
Q
Figure 1.3. Proposed dialogic approach to dialectology: Questions and methods
Research methods strongly affiliated with each of the four possible combinations in Figure 1.3 are listed in one of the four corresponding boxes.
Other relevant research considerations are listed along the horizontal and
vertical axes which function as transition zones between the four methodological oppositions. Adequate language definitions should, for example, accommodate contact-induced phenomena into their criteria. Contact is a
diachronic phenomenon, but its presence may be detected by research results in any of the four domains. Diachronic contact leads to synchronic
endangerment, but research methods on endangerment need not be restricted
to qualitative or quantitative analyses (see §2.5–§2.7 and Chapter 4).
As is discussed further in §8.1.1, building on research methods and theory outlined in Chapter 2 and implemented in the remainder of the book, this
approach to dialectology provides research means that are not only integral
but also organic. In such a system, rather than competing claims cancelling
each other out, different insights from different approaches are called on to
inform and/or correct each other. At a macro-level, this process of dynamic
Research scope, assumptions and approach
35
synthesis proceeds in such a way that synchronic language definitions
(which are themselves already based on multiple interdependent criteria)
may be grouped together genetically via adequate diachronic analysis. With
this mode of inquiry well under way, the diachronic analysis may then be
called on to re-interpret the original assumptions about various synchronic
phenomena. This is a condensed summary of my approach to Phula
language definition, to be more fully realized in the coming chapters.
Human languages are complex, dynamic and gradient – contingent on
time, space, linguistic structures and extralinguistic variables alike. Such realities may have engendered defeatist positions on the part of some dialectologists in the past regarding the possibility of language and dialect
definition, but I see no compelling reason that the definition of language varieties should be a reductionist task carried out in a theoretical vacuum. If
anything, the task of language definition should seek to incorporate complexity and competing claims. The process of hermeneutic synthesis enables
inclusivity without yielding to relativism and yields to strong theoretical
claims without defaulting to foundationalism.
Such an approach proceeds in the spirit of interdisciplinary pragmatism
rather than seeking to align itself with any particular reductionist paradigm.
Pragmatism’s general affirmation of organic, integrational system building
welcomes multiple insights from numerous traditions and, thus, seems better
suited for approaching actual language candidates, which are themselves
composed of far more complexity than a single linguistic tradition, or
dialectological perspective, can adequately account for.
The idea of incorporating multiple perspectives into language definition
and/or linguistic description is certainly not novel; but rigorous, broadscale
applications are lacking. Croft began issuing calls in the 1990’s for shifts
toward a more ‘dynamic paradigm’ in language variation research. He calls
for a “… paradigm in which the study of all types of linguistic variation –
cross-linguistic (typology), intralinguistic (sociolinguistics and language acquisition) and diachronic (historical linguistics) – are unified” (1990: 258–
259).26 Kortmann (2004a) constitutes perhaps the clearest concerted effort
to-date to articulate the possibilities of such an approach. Therein Bisang
(2004: 39) issues a clarion call for the “integration of typology, dialectology,
and sociolinguistic models of diffusion plus findings from contact linguistics” in order to better interpret both micro-level variation between individual speakers and macro-level variation between the extant distribution of
typological features in the languages of the world.
With some recent exceptions such as Labov 2010, ethnohistory, cognition and culture do not yet appear to weigh in heavily in the emerging conversation related to integrational dialectology. A cluster of closely related
36
Introduction
issues were also recently reviewed by Eira and Stebbins (2007) who call attention to ‘lineages of authenticity’, or traditional assumptions about what
counts as a valid perspective on the status of a given language variety or
linguistic form. Eira and Stebbins point out that different viewpoints on authenticity can be contradictory but are frequently negotiable. The two authors also remind us that linguists and language speakers both approach
such definitions within their own cultural frameworks. They conclude their
review of the problem by suggesting that those engaged in language definition and linguistic documentation should seek to incorporate and affirm
competing claims – not by treating them as if they were separate strands in
dissonance, but, rather, by inquiring at what interval in the lineage of the
language variety in question such a viewpoint or form came to be part of the
overall narrative or lineage. This is a particularly useful insight for dealing
with competing claims – one that is reinforced by the outcomes of this
research.
1.4.2.
On the viability of integrational standards for language definition
One might expect that a discipline taking human language as its focus
would have a time-honored set of complex criteria for defining what does
and does not constitute a basic subtype of its own set. On the contrary, the
very possibility of language definition is itself still unsettled in linguistics.
In part, of course, this is due to the interdisciplinary nature of the task in
what has been a remarkably segregated field of essentialist inquiry. In the
words of Chambers and Trudgill (1998: 4–5):
… paradoxically enough, a ‘language’ is not a particularly linguistic notion
at all. Linguistic features come into it, but it is clear that we consider Norwegian, Swedish, Danish and German to be single languages for reasons that
are as much political, geographical, historical, sociological and cultural as
linguistic.
Here we would do well to note that the same may be said of anything we
attempt to define – especially superordinate categories with gradient, processual underpinnings, inclusive of ecological and phylogenetic contingencies: A ‘plant’ is not a particularly botanical notion, for example; nor is an
‘animal’ a particularly zoological notion; a ‘nation’ is not a particularly political notion; an ‘economy’ is not a particularly economic notion. Classical
modern thought since Kant has been confronted with the assertion that everything we know and categorize (excluding, in his view, the a priori analytic) must come to us empirically through various conceptually constructive
Research scope, assumptions and approach
37
filters such as space and time. Less metaphysical filters, such as interaction
and convention, also apply. 27 In other words, any given domain of inquiry is
naturally implicated in (frequently neglected) domain-external dynamics
such as geography, history, sociology and culture.
It would be a mistake to assume on these grounds, however, that linguists are ill-equipped, or no better equipped than non-linguists, to set forth
complex criteria for operationalizing the subtypes of their own set – any
more than we should assume a skilled botanist is ill-equipped to classify and
identify plant species. Naturally, linguists and laypersons alike tend to refer
to the world’s ‘languages’ as distinct entities. Unlike the layperson, however,
the linguist should be more acutely aware of the dynamics, complexities and
interdependent contingencies underlying linguistic variation. Thus, even if
language definitions must range beyond linguistics proper, e.g., into geography, history, sociology, cognition, political administration and culture, the
lot for defining a ‘language’ would still seem to fall to the linguist – or,
more particularly, the dialectologist.
The difference between an epistemologist and a layperson, to draw a further analogy, is not that the former is endowed with an understanding of
epistemology while the latter has none. One need not study epistemology in
order to have beliefs and knowledge. One need not be an epistemologist to
classify systems of thought. The difference between an epistemologist and a
layperson is that the former is more acutely aware of how and why s/he does
and does not know what s/he does and does not know. The epistemologist is
uniquely trained to identify the limits of reason and classify systems of
thought. If s/he is led beyond the bounds of philosophy proper, say into the
realm of neuroscience or cultural anthropology, in order to develop specific,
consistent criteria for the definition of the boundaries of reason or systems
of thought, then so be it. This should not seem odd.
In both cases, the person who has been trained to understand the complexities, contingencies and limits of her field of study is uniquely qualified
to interpret and incorporate criteria external to her field of study for defining
the subtypes of her own set. In neither case should we assume the task to be
impossible or unnecessary or, worse, equally suitable to the abilities of the
experienced investigator and the layperson alike.
Language varieties, like systems of thought, happen to be fluid, complex
and gradient. Since any given lect constitutes a dynamic phenomenon with
underpinnings that are linked to contact, intelligibility, socio-history, geopolitics, culture, cognition and structure alike, the dialectologist should
simply require that a given language or dialect definition attempt to be
equally complex in its application.
38
Introduction
This is often not the case in Sinosphere linguistics, however – nor in
many language regions around the world. Languages are frequently asserted
or assumed to exist without clear criteria – even by linguists. Contrast, for
example, two statements from van Driem’s (2001) otherwise breathtaking
overview of language contact in the Indoshpere and Sinosphere. Upon discussing various Tibeto-Burman varieties of southwest China he notes,
“There are two Naxi or Moso languages … Eastern Naxi consists of heterogenous, mutually unintelligible dialects, whereas Western Naxi is a single
fairly homogenous language” (2001: 443). A few pages later he states the
following: “The rGyal-rong dialects are diverse enough to warrant treating
some of the dialects as separate languages …” (2001: 446). Careful dialectology does not appear to underlie such definitions, and the author’s own ad
hoc criteria seem to be in conflict between the two sets of languages. This
will not seem unusual to most, however, since rigor and consistency in
definition of languages and dialects seem to be lacking worldwide.
In an attempt to address such problems, the International Organization
for Standardization has, since the advent of the ISO639-1 codes in 2002,
called on linguists to define clear criteria for accepting one or more linguistic varieties as a single synchronic ‘language’. Five years later, the ISO6393 code guidelines (ISOL 2007) propose specific but flexible standards for
language definition. I present the ISO639-3 guidelines in (1.4):
(1.4)
ISO639-3 language code guidelines (ISOL 2007)
a. Two related varieties are normally considered varieties of the same
language if speakers of each variety have inherent understanding of
the other variety (that is, can understand based on knowledge of their
own variety without needing to learn the other variety) at a functional
level.
b. Where spoken intelligibility between varieties is marginal, the existence of a common literature or of a common ethnolinguistic identity
with a central variety that both understand can be strong indicators
that they should nevertheless be considered varieties of the same
language.
c. Where there is enough intelligibility between varieties to enable
communication, the existence of well-established distinct ethnolinguistic identities can be a strong indicator that they should
nevertheless be considered to be different languages.
Note that criterion (1.4a) primarily involves dialect intelligibility and
language contact, both of which overlap substantially with descriptive linguistics, historical linguistics and sociolinguistics. Criteria (1.4b) and (1.4c)
Research scope, assumptions and approach
39
primarily involve ethnic identity – although, literacy, education, culture,
political administration and other contingencies may be implicated as well.
In terms of writing systems, a distinction must be made between politically motivated orthographic cohesion, i.e., cohesion enforced on a group
from the outside using orthography as an ideological rationale for unity, and
identity motivated orthographic cohesion, i.e., cohesion arising from within
a given ethnolinguistic group using orthography as an extension of common
ethnic identity. The ISO639-3 language code guidelines are not simply allowing for the former possibility in criteria (1.4b); otherwise, in such cases,
the presence of a distinct identity could be invoked in spite of a common orthography, and the language definition process would default to a slightly
modified interpretation of criteria (1.4c). Bradley (2006: xv-xvi) provides an
example of two otherwise distinct synchronic languages, Northern and
Southern Lisu, that have recently begun to merge due to identity-motivated
orthographic cohesion. In a case such as this, intelligibility cannot be woodenly invoked as the ultimate criteria for language definition. In both cases,
ethnic identity is the key issue at stake, not politics or ideology. In this way,
languages may be justifiably defined without capitulating to official definitions – taking as much care as possible to define the language from the point
of view of its speakers.
As discussed above, identity issues related to orthography do not factor
in to Phula language definition. The key components of Phula language definition, then, seem to be threefold: identity, intelligibility and contact. In
practice, however, we find that diachronic considerations are inextricably
enmeshed in the definition process. Insights from historical dialectology,
subgrouping, internal reconstruction and comparative reconstruction are also essential for understanding distinctions between individual varieties. This
point constitutes a serious critique of the ISOL (2007) criteria that should be
incorporated into future versions. These themes will be considered in more
detail in Chapters 2 and 3. Methodology for researching contact, phylogeny
identity and intelligibility will be discussed more fully in Chapter 2.
As is outlined in Harris (1990), and elsewhere, the definition of language-internal dialects leads us ever deeper in the fog of abstraction where
idiolect, sociolect, isolect, and minilect seem to cavort and interchange at
will; nevertheless, in the words of Charles-James N. Bailey (1996a: 146),
“If the notion of a dialect is an abstraction … it is an abstraction that is
clearly a psychologically real one for the users of a language.” Dialectology
as Dialectic will attempt to define both Phula languages and dialects – both
synchronically and diachronically. ‘Synchronic’ dialect definitions are (out
of necessity) more preliminary, and lower in priority, than historical dialect
definitions, however. Synchronic language definitions also take precedence
40
Introduction
over synchronic dialect definition. Furthermore, the latter will be defined
with more reference to perceptual dialectology (see §2.6.1) than the former.
1.4.3.
Research scope and limitations
Approaching dialectology integrationally with the intention to define actual
language varieties is, as one would expect, labor intensive. Dialogic or not,
the dialectologist must admit to limitations. Here I sketch the major boundaries that mark my task and note some questions that will be left for later
research.
First of all, the work is bounded by the constraints of my proposed approach to integrational dialectology sketched out in Figure 1.3 above. I will
outline this approach further in the next chapter in terms of methodology. In
attempting to disprove and refine the Phula hypothesis, I will be unable to
engage in extended discussions of acoustic phonetics, morphosyntax, or textual analysis. My forays into (variationally-informed) phonology will be
somewhat more thorough, but they may also be found lacking in their attention to detail if considered in isolation. Comparative morphosyntax has the
potential to further refine the conclusions of this research, but due to the
breadth and geographical range of language varieties researched, I have, in
terms of descriptive-typological linguistics, necessarily limited myself to
phonological analysis and a smattering of morpho-semantic discussions.
In terms of ‘synchronic’ dialectology, I place primary limitations on intelligibility testing and the definition of sub-dialects. I adopt the former
limitation largely because comprehensive testing proved unnecessary (see
§8.5.8) and time-prohibitive. The latter limitation I have discussed above in
§1.4.2. Further intelligibility testing may be deemed useful in some cases,
but the incorporation of multiple perspectives, including quantitative
measures and diachronic subgroupings, tends to provide satisfactory answers to questions originally intended for intelligibility testing in most cases.
In terms of diachronic dialectology, my research emphasis does not require a full phono-lexical reconstruction. The hybrid internal/comparative
reconstruction plus historical dialectology analysis carried out in this work
is intended to identify embedded historical relationships. In fact, subgrouping (vs. the mere identification of regular sound correspondences) is considered to be the primary goal of the comparative method by historical
linguistics as diverse as Nichols (1996) and Ruhlen (2005). This work, then,
provides a proper foundation for more comprehensive phonological and
lexical reconstruction of Proto-Phula and/or any of its sub-branches in the
future.
Argument structure and organization
41
In addition to these constraints, natural limitations of time such as those
that marked the data collection schedule, will be noted, when relevant, in the
remainder of the book.
1.5. Argument structure and organization
Thus far in this introductory chapter, I have presented the Phula hypothesis,
discussed the historical background of the Phula milieu and argued for a
more interpretive approach to dialectology. In the next chapter, I will further
develop the practical and theoretical implications of my proposed approach
to dialectology and summarize the fieldwork and analysis that underlie the
project. These two introductory chapters set up the core dialectic tensions
that I rely on for Phula language definition. Allow me to summarize the
general weave of the argument here.
As it stands in (1.1), the Phula hypothesis is not falsifiable. For its logical
antecedent to function as an argument in a valid scientific prediction, synchronic dialectology must be able to identify and define the actual languages
affiliated with the Phula ethnonym. This pursuit will be the focus of Chapter
3. For the logical consequent of the Phula hypothesis to function in a valid
prediction, diachronic dialectology must be able to identify and define the
distinct genetic clades to which the synchronic languages belong. This pursuit will be the focus of Chapter 6. The analysis presented in Chapter 6 is
dependent on the prior analysis of Chapter 3, both of which are dependent
on the fieldwork, theory and background outlined in Chapters 1 and 2. Thus,
the arguments of Chapters 1, 2, 3 and 6 essentially flow in reverse chronological order from synchronic to diachronic. Chapter 7 pushes this trend further back in time by exploring the broader genetic position of the Phula
subgroups identified in Chapter 6. The very process of making the Phula
hypothesis falsifiable determines whether or not it is indeed false and, if so,
how it should be refined. By the end of Chapter 7, both the antecedent and
consequent arguments of the Phula hypothesis are provided with ample
definition to prove the Phula hypothesis false (but only partially so).
In Chapter 8, the Phula hypothesis is refreshed and refined, and the direction of the dialectic argument is reversed. Building on the diachronic definitions in Chapters 6 and 7, Chapter 8 is able to review and reinterpret the
results presented in the first half of the book. While the initial flow of the
work moves from synchronic to diachronic in order to establish the arguments of the Phula hypothesis, the final ascent of the discussion moves from
diachronic to synchronic in order to reorganize insights gained from Phula
‘synchronic’ dialectology into embedded ‘diachronic’ clades. Thus, in
42
Introduction
addition to refining the Phula hypothesis, the end result of this work provides a historical reinterpretation of the demographics, distribution, identity,
contact, intelligibility, and folk linguistics applied to Phula definitions in
earlier chapters. The final chapter also examines some possible factors related to neglected diversity in the Sinosphere and reviews possibilities for
future research and analysis.
At the heart of the book lie two practical applications of hermeneutic dialectology. Chapter 4 is primarily an application of dialect ecology, examining the status of endangerment that currently characterizes the Phula
languages, with an eye to predicting future decline and/or maintenance. This
is accomplished, in part, by establishing a typology of threatened statuses.
Chapter 5 anticipates the reverse chronology of Chapter 8 by using insights
from dialect phylogeny to select representative Phula phonologies for analysis. Five representative languages are examined from five meso-clade
subgroups affiliated with the Phula ethnonym.
My argument is organized in something of a figure-eight pattern or chiastic structure in which the top half complements the thematic content of the
bottom half and the flow of the argument is reversed once it has run its
course, allowing an inversion of antecedent and consequent, as is illustrated
in Figure 1.4.
Chapter 1: Hypothesis
Chapter 2: Theory
SYNCHRONIC
Chapter 3: Variety
Synchronic Diachronic
Synchronic research background
Synchronic language definitions
Synchronic application to future
Diachronic application to present
Chapter 6: Phylogeny
Chapter 7: Consanguinity
Chapter 8: Reinterpretation
Diachronic language definitions
DIACHRONIC
Diachronic research background
Diachronic Synchronic
Figure 1.4. Organization of the dialectic argument featured in this work
The argument may also be understood in terms of dialectic interpretation,
such that the first chapter presents a thesis, the final chapter presents its
Argument structure and organization
43
antithesis and the middle chapters comprise an organic synthesis connecting
the two extremes. Such terminology must not be confused with the static holism of Hegelian dialectics. ‘Synthesis’ in this work is intended to signify an
organic sense of growth—the growth enabled by opposing forces being
brought into complementary relationship and held in tension. The meaningful ontogenetic gestalt that emerges from such dynamics I hold to be the
best evidence for ‘synchronic’ and ‘diachronic’ (or better, ecological and
phylogenetic) definitions. These points are discussed more thoroughly in
§8.1.1 in order to better draw on insights in the intervening chapters of
analysis.
An approach that celebrates paradox and incorporates complexity may
seem peculiar, if not fundamentally flawed, to theorists functioning under
reductive presuppositions. Givón, in this connection, prompts us to reconsider “… our Western aversion to paradoxes, our inherent recoil from logical contradiction, as an analogical stand-in for our well-documented
penchant for reductionism” (2005: 248–249). In his work on the complex
emergence of communication pragmatics from the evolution of other complex systems ranging from biology to culture, Givón argues that language
theorists should seek to forge “… adaptive compromises … whereby the
complex design of the whole responds to and integrates together conflicting
adaptive demands” (2005: 249). The application of dialectology to the task
of language definition should be no different.
Instead of assuming polarized perspectives to be antagonistic or simply
unrelated, I assume in this work that opposing viewpoints should be called
upon to inform and correct each other in unifying relationships via a tertium
quid or triadic intermediary. Here, the structure of the intermediary is chiastic (compare Figures 1.4 and 8.1). The argument of the intermediary is dialectic. The method of the intermediary is, in this case, a hermeneutic
dialectology – dialect geography beyond the inductive warehouse, historical
dialectology released from deductive confinement, social dialectology beyond the statistical vacuum – an abductive, integrational dialectology gathering to itself sufficient means whereby the Phula hypothesis itself may be
made falsifiable.
Chapter 2
Research background
Field methods, theory, and dialectology
There is a case in which a disciple of the noble ones notices:
When this is; that is.
From the arising of this comes the arising of that.
When this isn't; that isn't.
From the ceasing of this comes the ceasing of that.
– Siddhartha Gautama
Anguttara Nikaya 10.92
2.1. Introduction
From its beginning stages, this project has yielded to an overriding principle:
the necessary symbiosis of its various research components. Without a carefully crafted wordlist, for example, eliciting the basic ‘lexicon’ of a given
Phula speech variety would have been time and cost prohibitive, but the full
development of the wordlist itself depended on lessons learned during local
elicitation sessions (see §2.4.2.8). Eliciting the developing wordlist from
multiple related speech varieties provided comparative lexical data needed
for language definitions, but the lexical data itself was often best elicited by
using lexemes and morphemes collected from other Phula speech varieties
to fish for cognates (see §2.4.2.4). Counting apparent cognates between
Phula speech varieties provided ways of predicting both low intelligibility
(see §2.6.3) and phylogenetic affiliation, but the ultimate determination of
cognacy and lineage alike could only be established diachronically using the
comparative method. Using the comparative method for phonological and
lexical reconstruction enabled the definition of genetic relationships between Phula languages, but the validity of diachronic sound correspondences depended on the quality of synchronic phonological analysis (and viceversa!). Peter Ladefoged (2003: 1) articulates a further step in this interdependent series: “The phonology has to be clear before you can make a
meaningful description of the phonetics; and without a description of the
sounds, you cannot get very far with the phonology. The two kinds of
investigation have to advance hand in hand.”
In fact, each component necessary for interpreting Phula variation have
had to advance ‘hand in hand’ with other components – including such
Undertaking a broadscale experiment in theory and praxis, this book
demonstrates grounds for insisting on a more integrational approach
to dialectology while simultaneously demonstrating grounds for defining
the hidden Phula languages of the Sino-Vietnam borderlands. The modern
languages and their ancestral lineage are defined dialectically, through
dynamic syntheses of competing perspectives.
TRENDS IN LINGUISTICS. STUDIES AND MONOGRAPHS
Trends in Linguistics is a series of books that publishes state-of-the-art work
on core areas of linguistics across theoretical frameworks as well as studies
that provide new insights by building bridges to neighbouring fields such as
neuroscience and cognitive science. The series considers itself a forum for
cutting-edge research based on solid empirical data on language in its various
manifestations, including sign languages. It regards linguistic variation in
its synchronic and diachronic dimensions as well as in its social contexts
as important sources of insight for a better understanding of the design of
linguistic systems and the ecology and evolution of language.
www.degruyter.com
ISBN 978-3-11-024584-4
ISSN 1861-4302