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