UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Rice ears and cattle tails : a comparative study of rural economy and society in Yunnan, southwest China Guo, Xiaolin 1996

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


831-ubc_1997-195856.pdf [ 22.7MB ]
JSON: 831-1.0087936.json
JSON-LD: 831-1.0087936-ld.json
RDF/XML (Pretty): 831-1.0087936-rdf.xml
RDF/JSON: 831-1.0087936-rdf.json
Turtle: 831-1.0087936-turtle.txt
N-Triples: 831-1.0087936-rdf-ntriples.txt
Original Record: 831-1.0087936-source.json
Full Text

Full Text

RICE EARS A N D C A T T L E TAILS: A C O M P A R A T I V E S T U D Y OF R U R A L E C O N O M Y A N D SOCIETY I N Y U N N A N , SOUTFTWEST  CHINA  by XIAOLLN GUO B . A . , The Beijing Teacher's College, 1982 M . A . , S O A S , The University of London, 1987 A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES Department of Anthropology and Sociology We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH C O L U M B I A November 1996 © Xiaolin Guo,  1996  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or  by  his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  . ANTHROPOLOGY &. SOCIOLOGY  The University of British Columbia Vancouver, Canada Date  DE-6 (2/88)  December 10, 1996  ABSTRACT  This is an anthropological study of peasant economy and culture, derived from field research on patterns of social organization and production of two ethnically different rural communities (Han and Mosuo) in northwest Yunnan, China. Its aim is to explore the local contexts for understanding the changes that recent economic reforms have brought to peasant life, and the cultural as well as ecological factors that constrain peasant economic activities. Current economic reforms have been accompanied by institutional changes, of which the most important for this research is the change in political relations between local and central governments. The expansion of local autonomy has had significant implications for the management of resources. The study shows that the behavior of the two local governments has had remarkably different economic consequences. The most noteworthy policy change in the economic reforms affecting rural society has been the implementation of the household responsibility system which brought down the twenty-year old collective system and has since altered the economic landscape of the countryside. This study emphasizes how kinship systems affect the form of household organization in both Han and Mosuo communities, and how existing social relationships are manifest in economic activities. "Rice Ears" and "Cattle Tails" are images drawing attention to the culturally salient differences in the patterns of production of the two communities. Rice ears constitute a cultural image of subsistence security in the Han community; and cattle tails constitute a cultural image of prosperity and development in the Mosuo community. Apart from the ecological factors which give rise to the particular patterns of livelihood in each community, cultural values associated the particular pattern of production account for many of the economic choices of the peasants and the persistence of economic forms.  ii  T A B L E OF CONTENTS  ABSTRACT  ii  T A B L E OF CONTENTS  Mi  LIST OF P L A C E N A M E S  vi  LIST OF ILLUSTRATIONS  vii  MAPS  x  PLATES  xiii  AUTHOR'S NOTES  xviii  ACKNOWLEDGEMENTS  xix  DEDICATION  xx  INTRODUCTION  1  C H A P T E R 1. H I S T O R Y A N D S E T T L E M E N T S  21  Yunnan Province Yongsheng County Jinguan and the Three-River Basin Ninglang County Yongning and Its Peoples  C H A P T E R 2. L O C A L A D M I N I S T R A T I O N A N D E C O N O M I C A U T O N O M Y Bureaucratic System of Government Administration  hi  51  Political and Economic Autonomy of County Government The Grassroots Leadership: Township Government Village Administration: the Semi-Governmental Organization Concepts of "Public" and "State": Peasant-State Relations  C H A P T E R 3. P A T R I L I N E A L H O U S E H O L D  121  The Jia and Household Organization Social Institutions of Patrilineal Descent K i n Relations in Economic Context  C H A P T E R 4. M A T R I L I N E A L H O U S E H O L D  173  The Yidu and the Social Institutions of Matrilineal Descent Household Organization and K i n Relations Economic Matriliny  C H A P T E R 5. R I C E E A R S — C U L T U R A L I M A G E O F S U B S I S T E N C E E C O N O M Y  220  Ecosystem and Agriculture Subsistence-Security: Cultural and Economic Values of Rice Production-Consumption Loop: the Principle of Self-Sufficiency  C H A P T E R 6. C A T T L E T A I L S — C U L T U R A L A N D E C O N O M I C D I M E N S I O N S  267  General Economic Situation Crop Farming Cattle Production  CONCLUSION  311  iv  CHINESE G L O S S A R Y  325  P A T R I L I N E A L KINSHIP TERMS  334  MOSUO G L O S S A R Y  336  M A T R I L L N E A L KINSHIP TERMS  338  REFERENCES CITED  339  APPENDIX: QUESTIONNAIRE FOR HOUSEHOLD S U R V E Y  350  v  LIST OF P L A C E N A M E S  Cuihu  village in Jinguan township  D a Liangshan  Y i inhabited area in Sichuan province  Dali  second largest city in Yunnan province  Dapo  Mosuo hamlet in Yongning township  Gesawa  Mosuo hamlet in Yongning township  Guizhou  province east of Yunnan  Jinguan  township in Yongsheng county  Kunming  capital of Yunnan province  Lijiang  prefecture under which Yongsheng county and Ninglang county are administered  Neiba  Han hamlet in Yongning township  Ninglang  Y i autonomous county under which Yongning township is administered  Pijiang  Naxi hamlet in Yongning township  Sichuan  province north of Yunnan  Wengpeng  village in Jinguan township  X i a o Liangshan  Y i inhabited area in Ninglang  Yongning  township in Ninglang county  Yongsheng  county under which Jinguan township is administered  Yunnan  province  vi  LIST OF ILLUSTRATIONS  Maps 1. Yunnan Province  x  2. Yongsheng County  xi  3. Ninglang County  xii  Plates 1. Three-river basin, Jinguan  xiii  2. Cuihu village, Jinguan  xiii  3. Newly-built peasant house, Jinguan  xiv  4. Tobacco fields in Wengpeng village, Jinguan  xiv  5. Mosuo residence on hillside, Yongning  xv  6. Everyday market, Yongning  xv  7. Cattle grazing in harvested corn fields, Yongning  xvi  8. Barnyard grass fields, Yongning  xvi  9. Lamasery under construction, Yongning  xvii  10. Mosuo Daba and his family, Yongning  xvii  Figures 1. Gross national product by province (1991)  24  2. Total output value of society by province (1991)  25  3. Per capita income by province (1991)  26  vii  4. Levels of government administration  54  5. Comparison of finance situation between Yongsheng and Ninglang  63  6. Rural administration in Yongsheng County  75  7. Rural administration in Ninglang County  75  8. State-resold grain in Jinguan and Yongning  82  9. Sources of financial help in peasant household  169  10. Sources of labor help in peasant household  169  11. Matrilineal descent: Guma family  194  12. Matrilineal descent: Kuaha family  195  13. Population and land in Jinguan  223  14. Arable land and crop sown area in Jinguan  224  15. Farming seasons in Jinguan  226  16. Crop unit output in Jinguan  232  17. Seed consumption in Jinguan  235  18. Yongsheng Markets  250  19. Prices of building materials in Jinguan  ,  264  20. Sources of financial help in house building in Jinguan  265  21. Population and land in Kaiji village  268  22. Arable land in Yongning  269  23. Per capita income in Mosuo hamlets  271  24. Crop sown area in Yongning  279  25. Crops and farming seasons in Yongning  284  26. Crop unit output in Yongning  287  27. Chemical fertilizer application in Jinguan and Yongning  289  28. Seed consumption in Yongning  291  29. Livestock in Yongning  294  30. Draft animal holding in Mosuo hamlets  295  viii  31. Total crop output in Yongning  300  32. Average cattle price  305  33. Factors affecting economic strategies of local administrations  313  34. Comparison of household organizations  316  35. Comparison of basic production situations  320  ix  LEGEND o O ® —  County seat Prefecture seat Provincial capital Road 100  200  H—* f'^x  Railway River Provincial boundary National boundary  300 km ©Xiaolin Guo, 1996  X  Yongsheng County LEGEND Village Township County seat Road River  o  © ®  \  --..>,,. •i °/ \ j i Lijiang Naxi !> J Autonomous^ County / Songping hi °  >••'" o  °  o  if  o  ' y/fO I  ;  o  20 l  30 km l  ?  • >?  ...  10 l  0 1  Ninglang Yi • Autonomous V County  o )  o  Da'an ©  o  Liangguan^  -i, o tfjinsha % River  & °  ,<j  Yongbei  o  \  ^  0  /  Shunzhou  o  . Huaping County  /  ojsf  v.. I: Chenghai |o  o  N  "\  I  o  Renhe^)  oXinghu  V'  Heqing County,-;'-  x  -.^., " J ^  r  V  it  o o  V V  Dongshan©  US;;  I Qinajgf-'  w  \  Dayao County  7T  © Dongfeng Jinsha ff % River J,'  V  ' Binchuan County  N  Pianjiao \  ©Xiaolin Guo, 1996  xi  Ninglang Yi Autonomous County /*•-  ,' o  LEGEND  /  Labo  kv  pi  s0 ^  9 o Neiba-i r, V J' 3pJ ©' O DapO V  © ®  an  ^  Yongning  '"'  o  1  )  ' —  IMLuguLake  o  V  o  Village Township County seat Road River  o  nGesawa  0  10  20 l  30 l  40 l  50 km l  /  o s  Sichuan Province  i  V  v  Hongqiao©^-''  Lijiang /< County f li  I o  1  ©Hongqi  o o  y%  Ningli ©  Daxing  © Dalaba j  V  © Xichuan o '•)° <  } © @Paomaping  © Xibuhe  o  ~ W - w - - ' ~ " ~ ' "  / o <  r  jo  J  L  ~@  z h a n h e  / ~  ) Huaping '' County  ^  \  \  o  N  Yongsheng | County )  \  }  t •"» ..of  1  i  ©Xiaolin Guo, 1996  xii  PLATES 1. Three-river basin, Jinguan  2. Cuihu village, Jinguan  xiii  3. Newly-built peasant house, Jinguan  4. Tobacco fields in Wengpeng village, Jinguan  xiv  5. Mosuo residence on hillside, Yongning  6. Everyday market, Yongning  xv  9. Lamasery under construction, Yongning  10. Mosuo Daba and his family, Yongning  xvii  AUTHOR'S NOTES  Anonymity Proper measures have been taken to disguise the identities of individual persons and households that appear in this dissertation.  Transliterations of Language Chinese terms are transliterated according to the standard Chinese phonetic alphabet (hanyu pinyin) system throughout this dissertation. In view of the absence of a written Mosuo language, Mosuo terms are transliterated according to the international phonetic alphabet and presented in a close approximation. Both Chinese and Mosuo terms are in italicized form.  Plates The photographs (Plates 1-10) included in this dissertation were taken by myself during the fieldwork, and selected for the purpose of providing ethnographic background information.  Measure Units jin = 0.5 kilogram mu = 0.0667 hectares, or 0.1647 acre  Currency Exchange (1992) l Y u a n ( R M B ) = US$0,125  xviii  ACKNOWLEDGEMENTS I would like first to express my gratitude to my research committee from whose scholarly supervision my research has benefitted immensely. I thank Prof. Martin Silverman for his stimulating guidance in the realm of economic anthropology and kinship studies; Prof. Richard Pearson for sharing with me his expert knowledge about Chinese history and the imperial state; and Prof. Stevan Harrell for expanding my research perspectives with his deep insights into Chinese society and ethnic issues. The energy and good w i l l that the three committee members infused in my work gave me support, encouragement and confidence, all of which have, finally, led to the present accomplishment of the dissertation. M y research was in part connected to a larger project titled "Rural Social Support Arrangements and the Transformation of Local Traditions in China." I wish to thank Prof. Stephan Feuchtwang for inviting me to take part in this project, and also thank the British Economic and Social Research Council for funding part of my research. The accomplishment of my work, however, would not have been possible without the help from the people in Yunnan. M y thanks go to the Institute of Rural Economy and Development of the Yunnan Academy of Social Sciences, and the Jinguan and Yongning township governments. I feel most indebted to those individuals who in different positions provided me with precious personal as well as official assistance throughout my fieldwork. They are: Chen Guoyi, Chen Yehua, Feng Lantian, Gao Zonggui, Guan Kaiqing, G u m a Luzuo, G u o Dexin, H e Shijie, He Xiuyue, H o u Zhigang, H u Quan, H u Zongyou, Kuaha Jizuo, L i Jinying, L u Guoqiang, L u Shujuan, M a Yamin, Pei X i n p i n g , Wang Jin, W e n Jianyi, W u Guochang, X i a n g Lianying, X u Yuwen, Yang Biao, Yang Bingliang, Y a n g Libang, Zhang Juhua, Zhang Y o u x i n , Zhou Haibo, and Z o u Ning. I here extend my deep gratitude to the residents of Jinguan and Yongning, who kindly and generously accepted my interviews and whose cooperation and hospitality made my fieldwork successful. The writing of my dissertation itself has benefitted from my affiliation with the Center for Pacific A s i a Studies at Stockholm University that provided me with generous use of office facilities, for which I am much obliged. I also thank the Department of Cultural Anthropology at Uppsala University for giving me opportunities to present part of my research results and for the scholarly comments made by the professors as well as the graduate students there. To other individuals who provided support and help at various stages, I thank Jean O i , K a v i n Lane and Stephan Feuchtwang, whose comments on some of my earlier draft chapters have been most helpful; I thank Charles D'Orban formerly at the library of the School of Oriental and African Studies, Nancy Hearst at the library of the Fairbank Center for East Asian Research, Harvard University, and Annie Chang at the Center for Chinese Studies Library, University of California, Berkeley, for helping me with research materials and documents in their library collections. A m o n g all and above all, Michael is the one who-apart from sharing much of my research interest as well as frustration-has, in the past years, taken upon himself a heroic role as a sympathetic listener and at the same time a critical reader, and for all of which, many, many thanks.  xix  To my parents  xx  1  INTRODUCTION  This study of rural economy concerns social and economic changes in China in the first decade of economic reform. The period covered is from 1983, when decollectivization took place in my field sites, to 1992 when the field work was carried out. Jinguan and Yongning are two townships selected for the study, both located in China's southwestern province of Yunnan. Jinguan is a community with a majority population of Han; and Yongning is a community with a majority population of Mosuo, one of China's numerous ethnic minorities. F r o m 1949 when the People's Republic was founded till today, China's rural society has undergone many changes, and a succession of political movements has had tremendous economic impact upon the life of Chinese peasants, most significantly the land reform of the late 1940s to m i d 1950s; the Great Leap Forward of the late 1950s; and the economic reforms of the 1980s. A l l these movements had specific ideological goals which, in turn, had different economic consequences. The land reform was launched soon after communist power was first established, in some places as early as 1946 (e.g. in the old revolutionary bases) and in others as late as 1956 (e.g. in the border regions and minority areas). The historical significance of the land reform was to overthrow the old economic and political relations which had persisted in the Chinese countryside for hundreds of years (Shue 1980). T o the vast number of Chinese peasants, the land reform was a movement of fanshen, an expression which literally means "to turn the  body," or "to turn over." This, in the revolutionary context, means "to stand up, to throw off the landlord yoke, to gain land, stock, implements, and houses" (Hinton 1966). The land reform movement thus succeeded in depriving the old elite of their social and economic privileges, and gave the land to the common poor. A t the end of this movement, an elementary cooperative organization of production gradually took shape in the countryside, and later expanded and grew into an advanced cooperative organization. The Great Leap Forward that followed was a disastrous nationwide mass movement in which all of Chinese society was mobilized in a radical attempt to realize the ideals of developed communism within the time span of one or more Five Year Plans. In the countryside, the movement took the form of the People's Communes (renmin gongshe)} It was a communal organization of the peasantry described in the media at the time as "the gateway to paradise," by merging small-scale agricultural cooperatives established at the end of the land reform into large cooperatives. The People's Commune was a collective economic organization characterized by the "three-tier-accounting system" of commune (gongshe), brigade (dadui) and production team (shengchan dui), operated on the basis of public ownership. The principal communist ideal of the Great Leap Forward was "From each according to his ability; to each according to his need," which at the time was interpreted as "Everyone who is able to work is able to eat his fill." The movement ended with three years of bad agricultural harvests and a catastrophic famine, yet the People's Communes survived for almost a quarter century. The recent economic reforms, in a sense, can be seen as a continuation of the Chinese revolution in modern history. But C C P policy has, to a considerable extent, been modified, and demonstrated notably favorite treatment toward the peasantry compared to the past. The economic reforms began to have their impact on rural society nationwide in 1979 under the central government's policy "to change the backward situation of China's agriculture, to  ' F o r a review of the establishment of the People's Communes, see Hudson (1960); for a critical account of Mao's ideological construct of the Great Leap Forward, see Schoenhals (1987).  3 accelerate agricultural development, to lighten the burden of the peasants, and to increase peasant income" (Xinshiqi Nongye 1992: 31). In the initial process, a new framework of agricultural production was popularized. It was known as "contracting output to the household" (baochan daohu); and later evolved into "contracting everything to the household" (baogan 2  daohu)} The establishment of the household responsibility system ended the collective management of production and distribution in existence since the establishment of the People's Communes. Under the household production responsibility system, the household became the basic unit of production-managing its own contracted land and making independent decisions. This system allows the peasants to dispose of their produce for household consumption and for market transaction once the specified quotas, taxes and levies demanded by the state and village collective have been fulfilled. In most parts of China, the household production responsibility system was implemented between 1980 and 1983. It was implemented in Yongning in 1980, and in 4  Jinguan (the Han population area) in 1983. The difference reflects a general trend in that the poorer an area was, the earlier the implementation of the household production responsibility system took place. This is the result of the flexibility of central policy at the beginning of the economic reforms towards the regions where economic development lagged behind the rest of the country. The production teams which had lived on state resold grain, state loans and state subsidies were allowed to contract land to the households as soon as this was desired by the local population (Xinshiqi Nongye 1992: 60). This policy was an attempt to provide peasants in  I n this, the production team contracts work and assigns target output to households that are responsible for production and cost. The annual output is divided between the collective and individual households with a bonus given to the households on the overfilled portion of output. 2  I n this, the individual households manage production. The output-except for the amount extracted for agricultural tax and collective funds-is retained by the contracted households. 3  A s estimated, by the end of 1980, thirty per cent of all production teams had implemented household contracts, and by the end of 1982, ninety per cent of China's production teams had embraced some form of responsibility system on a household basis (Ash 1988).  4  poor areas with incentives to in raise productivity. On the other hand, the reason for the postponed implementation of the household production responsibility system in Jinguan is that it is one of the most important grain production bases in Yongsheng county, and the county government's reluctance in implementing the household production responsibility system in Jinguan was a measure to protect the grain production in the area. The implementation of the household production responsibility system was soon followed by, i f it did not lead to, a fundamental institutional reform in rural society. This was the so-called "decollectivization," which means the dismantling of the People's Communes. Twenty-five years after the "merger of government administration and economic management" (zhengshe heyi) during the Great Leap Forward, the government reversed its policy and advocated the "separation of government administration from economic management" (zhengshe fenkai), a policy formally put forward at the end of 1982 (Xinshiqi Nongye 1992: 170). F r o m a political economy point of view, the separation of government administration from economic management was intended to ease state control over rural production and distribution, and to create a new relationship between the peasantry and the state (Walker 1988). B y 1984, the People's Communes had been dissolved throughout the Chinese countryside. A l o n g with the institutional change, there was a notable ideological change represented by a shift from "taking class-struggle as the key link" toward an advocacy of individual material interests. The ideology of class struggle was, as a matter of fact, propagated during land reform. The whole idea of the class struggle was then motivated by breaking up the old elite network and creating a new leadership in the countryside under the total control of the communist party, the economic aim of which was to "spur economic growth by eliminating much luxury consumption and make more peasant families solvent by curbing exploitation" (Shue 1980: 43). Thirty years after the land reform, the ideology of class struggle was abandoned by the central government on the assumption that "the landlord and rich peasant elements have, in the past twenty to thirty years, reformed themselves into laborers living by their own hands "  5 (Xinshiqi Nongye 1992: 12). The elirnination of class status was part of the incentive systems intended by the central government to bring every positive factor into play, which had great social impact on the internal relations of peasant society as well as on the state-peasant relations. The incentive systems also include "material interests" which ironically resemble those used by the C C P during the land reform. In both periods, such incentive systems are officially formulated as "getting rich" (zhifu). However, the motivations of the government behind the incentive systems during the land reform and current economic reform are far from alike. There are two major differences: (1) while the land reform was meant to bring social and economic equality among the population, the current economic reform was set to combat egalitarianism; and (2) while the land reform mobilized the peasantry to merge into various socialist corporations from the elementary to the advanced, the current economic reforms have encouraged peasants to break away from the collectives. Why is the idea of the peasantry getting rich so pronounced in the C C P policy? What is in the interest of the central government to let peasants get rich? These questions concern the whole importance of the economic reforms, and the bottom line is the national economic development and social stability. Because agriculture plays a crucial role in China's economic revival and the peasantry forms the majority of the Chinese population, the importance of reviving the rural economy is described by the central planners as: "The situation in the countryside is a crucial element which determines the situation of the whole country. If the countryside is stable, China w i l l be stable; i f the peasants live well, the whole country w i l l lead a good life; when the countryside is modernized, our country will be modernized" (Du 1985: 26). The urgency of improving peasant living conditions was further spelt out as:  We cannot allow tension to persist everywhere, and should begin by pacifying the peasantry. Once the peasants have grain, resolving issues like cotton, non-staple food, cooking oil, sugar and other cash crops w i l l no longer pose a problem. Having stabilized the situation among the peasants means having stabilized the situation among the majority of the population; and bringing peace to over 700 million people means bringing peace to the land under heaven... W e must relax the situation in the countryside and give the peasants some breathing space. Otherwise, I fear the peasants might rebel (Xinshiqi Nongye 1992: 6).  6  These words reveal what Susan Shirk calls "the political logic of economic reform" in that she argues that "the real challenge of economic reforms was the political one" (1993: 6). The political logic as such resembles the C C P policy during the land reform which is referred to as pragmatic in that "the Party's efforts to accommodate the day-to-day practical interests of the mass of peasants" (Shue 1980: 334). The central government's policies of the present economic reforms are indeed pragmatic by linking the political and economic interests of the nation to the individual self-interests of the peasantry. The ideology of "getting rich" has had a strong appeal to the peasants, and the household production responsibility system has proved to be the most effective means so far to accelerate rural economic growth. While the land reform is characterized as fanshen, a change of social status, the current economic reform in the countryside is characterized by fanfan, meaning doubling or multiplying output. The statistics show a remarkable growth of China's economy in the ten years between 1983 and 1992: the gross output value of agriculture in 1992 had tripled from what it was in 1983; and the rural per capita income had more than doubled since 1983 (Zhongguo Tongji Nianjian 1984; Zhongguo Tongji Nianjian  1993).  5  Still some questions remain. Have the peasants become rich? If so, how have they become rich? If not, why not? The present study is interested in finding out: when political control over the rural production has presumably been relaxed and administrative interference in peasant life has been significantly reduced, how is the peasant economy organized and what are the political, cultural and ecological forces that exist and put constraints on peasant economic choices? W i t h these questions in mind, I carried out my field research in Yunnan.  5  F o r a general review of agricultural change in the 1980s, see also Kueh and A s h (1993).  Fieldwork M y fieldwork was conducted during a period of nine months from A p r i l 1992 to January 1993. The reason why I selected Yunnan province was mainly due to the nature of the comparative study of economies between two ethnic communities. For the comparative purposes, it is important that the two field sites share a similar geographical situation. The two communities chosen for the present study are both located in the northwestern highland of Yunnan. The two townships, Jinguan and Yongning, respectively, belong to two counties under one and the same prefectural jurisdiction (i.e. Lijiang). Given the particular geographic situation in Yunnan, travelling is not an easy experience. A half century ago, travellers in Yunnan mostly went on horseback or on foot. Today, the extensive road networks have made travelling by motor vehicles possible and much easier, but the mountainous roads in most parts of Yunnan still present certain difficulties. From Kunming to Lijiang (the prefectural seat), it is a two-day trip with a stop-over in Dali, the capital of the ancient southwestern kingdom from 937 to 1253 and now Yunnan's busiest commercial center second only to Kunming (see M a p 1). The alternative trip is an overnight bus ride of 16 to 18 hours from Kunming directly to Lijiang. From Lijiang, it takes another three or four 6  hours by bus to get to my first field site Jinguan. M y second field site Yongning is further north, and it takes another eight hours by bus, which usually needs to break into two days to fit in the daily bus schedule. I arrived in Jinguan on a fairly warm day which happened to be a market day. I was accommodated at the township government guest house, and joined the township government staff in their communal canteen for meals. The township government seat is by the market which convenes every five days. In ordinary days, there are scores of shops on the streets  A i r links between Kunming and Dali and between Kunming and Lijiang became available in 1995.  6  supplying many articles of daily use. Post comes every day from the county seat, but is only delivered every three days. A week after I came to Jinguan, the rice transplanting season began. Peasants were busy, and I realized that I had to wait for the season to be over before I could begin my household survey. While waiting, I worked on the township statistics. After getting some idea about the basic situation in the eleven villages in the township, I selected Cuihu and Wengpeng for my household survey. The household registrations with the township public security office were not complete because household divisions and other demographic change are not frequently updated. Instead, I borrowed from the local cadres the agricultural tax registrations which are most efficiently updated. Thirty households in each village were selected by random sampling for my household survey. When the survey began in Jinguan, I went to the village every day except on market days when the streets were crowded and the peasants normally hang around at the market and are seldom home. It took me some twenty minutes to cycle from the township seat to either of the villages. Peasants generally responded to my interviews positively, and were reasonably cooperative. In the beginning they were a little concerned about why his or her household was chosen for the survey, but when I explained that their names were selected by random sampling, but not on purpose, they seemed to be more at ease. M y interviews were conducted in the peasant households. Sometimes I was received in their main room, and sometimes in the yard. I asked questions and elicited answers face to face with the interviewees, sometimes the head of the household, and other times the whole family. Being in the household enabled me to observe more carefully the situation and add more information to the questionnaire. After a dozen households or so, I became familiar with the interviewing situation and began to engage in conversations with the peasants more naturally. In the course of my interviews, I noticed that peasants had difficulty in understanding and giving direct answers to terms such as output, gross income, net income, as they calculate them in different ways. Some peasants liked to exaggerate their output, income and expenditure; and  others tended to understate. Some inaccuracies were innocent, and others were deliberate. W i t h the assistance from my local guides, these problems were easily solved. The household survey in Jinguan went smoothly; and by the end of August I had finished my work there. When the rainy season was almost over, I packed up and headed for the north to my second field site Yongning. In Yongning, I also stayed in the township government guest house. The township government has a small collective kitchen, but meals were not as regularly served as in Jinguan mainly because there were not many cadres around in the office building. Because of the poor communications in Ninglang county, post comes to Yongning every four days. The Yunnan Daily (the provincial government newspaper) usually arrives one week late. The situation became worse in the rainy seasons and in winter when the road was interrupted by rain or snow, and the post did not come at all. The poor communications make Yongning a very isolated place. It was one month before the autumn harvest when I arrived in Yongning, and I had no time to waste. I quickly went over the township population census and some other statistics as a base for my selection of hamlets. The settlement in Yongning is different from Jinguan. The hamlets are more dispersed. This, in a way, makes the village administration a loose organization as it has to cover a fairly large area. The composition of the ethnic population was crucial in the selection of hamlets for my household survey. T w o Mosuo hamlets were ideal in size, about thirty households in each. In addition, I selected one Han hamlet and one N a x i hamlet, for comparative purposes. These two hamlets were both bigger than the M o s u o ones, and I further selected the appropriate number of households through random sampling, twenty households in each hamlet. When my household survey began in Yongning, I cycled to the hamlets in the morning and returned in the afternoon. From mid-October, there was an unusual torrential rain which lasted literally non-stop for more than two weeks. It interrupted all the public roads and village tracks and abruptly brought the temperature down nearly to the freezing point. The rain washed  away parts of the river dam which led to a power failure. This was the time when most of the township cadres were back home on harvest leave, and the repair work was suspended. During the two or three weeks, even water supply became a problem. The snow came in the first week of November when the only power station of the township was out of service. M y little electric burner on which I relied for cooking and heating became useless. In spite of all the difficulties and inconveniences, my household survey went well. Towards mid December, I concluded all my work in Yongning and withdrew. O n my way back to Kunming, I stopped in Yongsheng and Lijiang for three weeks to check some data. Then, my whole fieldwork was completed just before the Chinese N e w Year. I had different experiences in the Han and the Mosuo communities. In Jinguan, the local cadres were more responsive than the peasants. The initial reluctance of the peasants in engaging in conversation was much related to their concern of possible political and economic consequences. Their worries were understandable and justifiable considering all the political upheavals the peasants have been through and the hard life they have been living in the past decades. The local officials in Jinguan township and Yongsheng county showed considerable understanding of my research and provided me with great help. They managed to spare one bicycle for me to use; they introduced me to the village cadres; and they were willing to provide assistance in any situations. Staying in the government guest house turned out to be a privilege. It gave me opportunities to make friends with the cadres. I spent most of my time outside the household survey chatting with them; and occasionally they invited me for meals. A s many of the cadres come from the villages nearby, their views of events in the villages are particularly interesting and precious. Although they live in a very isolated environment, they took much interest in the on-going events in the world which were brought to them by television. Being with the local cadres, I learned about a great deal of their value system, the implication of their official performance and the meaning of their personal sentiments. This special experience opened a window for me to observe and understand the state-peasant relations.  11 In Yongning, my relationship with the local cadres and the ordinary peasants turned out to be the other way round. The Mosuo peasants were warm, hospitable and friendly from the beginning to the end. Most of the time, I was treated as a guest and offered butter tea and other home-made snacks. To most of my questions, the Mosuo peasants gave frank answers. They were cooperative and spontaneous in conversations, and seemed not worried about exposing their economic situations. The village and hamlet leaders were also friendly and helpful. The township cadres were reasonably cooperative in many matters, although most of them were often not at work. There was a clear expression of reservation among the cadres in the beginning. The situation was somewhat improved later as they learned that I was there studying "economy" but not "matriarchy." I could see that we gradually came to terms of understanding each other. Their reservations seemed to have derived from some kind of local exclusionism-the same sentiments shown towards the commercial traders and peddlers on the market. A s the traders are seen flagrantly making money out of local community by taking their produce (e.g; mushrooms, herbs, etc.), the "cultural people" (which are collectively called as wenhua ren by the locals, including journalists, novelists, film-makers, anthropologists and alike) are equally regarded as making a fortune out of the local population by writing books or shooting pictures about the local life. The local cadres (many of them are Mosuo) are sensitive at the mention of "matrilineal society" (muxi shehui) by accidental tourists, as it has been perceived by Chinese society as something "exotic" and "primitive." The conflict of cultural and economic values 7  between the locals and the non-locals and between the Mosuo society and the larger Chinese society, is observable; and an outsider needs to show a considerable understanding in the situation as such.  M u c h of the resentment resulted from the publications of Chinese ethnographies on the Yongning Mosuo in the 1980s, in which the Mosuo are perceived as "living fossils" of the ancient society.  7  12 In both field sites, my household survey was assisted by local guides, one from each village as in Jinguan, and one from each hamlet as in Yongning. M y guides worked both as my interpreters (especially in Yongning) and as liaison between me and the peasant households. They introduced me to every household and assisted me with interviews. Apart from occasionally chasing away fierce guard dogs and clarifying misunderstandings, they provided me with valuable information about the village, hamlet and individual households. They also helped me duplicate village maps, work out the farming calendar and inform me of local customs. The local status of the guide is important. In Jinguan, an ordinary peasant was better qualified to be a guide than a village leader because the peasants in general are suspicious about cadres. In Yongning, the hamlet leaders turned out to be a better choice of guide because they are authoritative and respectable in the eyes of their fellow hamlet residents. Understanding local dialects was also a challenge in my fieldwork. When I first arrived in Kunming, I could hardly understand a word when people spoke at a regular speed. In Jinguan, the local speech was even more difficult to understand. But, after the first three or four weeks, I got used to the local accent, and was able to talk to the peasants on my own. In Yongning, the Han Chinese language {Hanhua) is commonly used in the government institutions and on the market because population there is multi-ethnic. Most of the young and middle-aged Mosuo speak Hanhua fluently. A l l my Mosuo and N a x i guides were very efficient in interpreting and translating conversations between the interviewees and myself. After a while, some Mosuo terms in frequent use became familiar to me, which to certain extent, helped smooth the procedure of interviews. I have benefitted greatly from my different experiences in the two field sites. They have taught me how to observe different social and economic realities; how to appreciate the people of different backgrounds; and how to compare their life with ethnographic open-mindedness.  Primary Sources of Data One difficulty with this study is the limited literature available. There have been some previous studies of rural Yunnan between the 1920s and 1940s, but most of the works are based on the peasant life in the central-eastern part of Yunnan (e.g. Fei and Chang 1945; H s u 1949). Systematic ethnographic studies of the rural society in northwest Yunnan are almost non-existent. Yongsheng county to which Jinguan township is administratively subordinate, for example, "has not been studied by anybody," as the local officials put it. The major sources of reference are two editions of a county history: one compiled in the Qing Dynasty titled Yongbei Fuzhi (1992, reprint of 1765), and the other compiled a few years ago titled Yongsheng Xianzhi(l989)  which was constructed on the basis of Yongsheng Lishi Ziliao Huiji 1-12  (1962), a collection of historical materials held at the Lijiang prefectural government archives. Other written materials include three volumes of Yongsheng Wenshi Ziliao Xuanbian  I, II, III  (1989; 1990; 1991); Yunnansheng Yongshengxian Diming Zhi (1988) and one volume of Zhonggong Yongsheng Difangshi (1990). In addition, there is a small collection of the old county government's documents in the county archives, a pile of some 100 rolls of papers which, as I was told, have not been sorted since 1948 when the county magistracy court was set on fire by bandits. The literature on the Yongning site is comparatively rich. There have been many Chinese ethnographic monographs and reports, in addition to bits and pieces in the Chinese historical records. The county history Ninglang Yizu Zizhixian Zhi (1993) was available in manuscript version at the time of my fieldwork, and published a year later. The Chinese ethnographies and reports are mainly based on the research from the 1960s to the 1980s. The two major monographs are titled: Yongning Naxizu de Azhu Hunyin he Muxi Mating (Zhan et al., 1980); and Yongning Naxizu de Muxizhi (Yan and Song, 1983). The richest materials come from three volumes of research reports titled: (1) Ninglang Yizu Zizhixian Naxizu Shehui ji Mating Xingtai Diaocha I (1986); (2) Ninglang Yizu Zizhixian Yongning Naxizu Shehui jiqi  Muxizhi Diaocha II (1986); and (3) Ninglang Yizu Zizhixian Yongning Naxizu  Shehuijiqi  Muxizhi Diaocha III (1986). These volumes contain extensive materials about the political 8  system, religious practice and kinship relationship of Mosuo society. The existing English language literature includes notes of Western explorers and ethnographies. The most notable Western ethnographer in the early period, who has contributed the most interesting and detailed ethnographies about the native peoples of northwest Yunnan, was an American botanist named Joseph Rock. He first came to Yunnan on an assignment to work for the United States Department of Agriculture, and later for the National Geographic Society, but his stay in Yunnan extended beyond his official assignments. Between 1920s and 1930s, Rock travelled extensively and collected massive materials about the nature and society of N a x i (or in Rock's term: Na-khi) and Mosuo (or in Rock's term: Hli-khin). H i s ethnographic studies on the history, political systems, folklore and religions were later published in a series of research volumes. The Ancient Na-khi Kingdom of Southwest China, Volume I and II (1947) consists of over 400 pages on the geography of northwest Yunnan, history of the N a x i (Nakhi) society, and the livelihood of the Mosuo (Hli-kin). Rock's other writings include The Na-khi Naga Cult and Related Ceremonies (1952) which is about the N a x i religion (and its B o n origin) and ceremonies; and The Zhi ma Funeral Ceremony of the Na-khi of Southwest China (1955) which is a collection and descriptions of N a x i funeral rituals.  9  The more recent English language literature on Mosuo society are two doctoral dissertations (Shih 1993; Weng 1993), which only became available a year after I had returned from my field work. Both studies of the Mosuo are focused on kinship, gender relations and myth, but neither of them deals with the economic life of the people in the society.  N a x i is the Chinese official ethnic identification for Mosuo, used in the government documents and many academic works. 8  I n about the same period, Peter Goullart wrote Forgotten Kingdom (1955) which contains some ethnographic materials in this part of Yunnan.  9  Aside from the publications and manuscripts, local statistics provided a useful source of information, and helped in the selection of a desirable field site. M y major concerns in selecting townships were the population composition and economic situation; and the local statistics at county-, township- and village-levels have the figures necessary for this purpose. Many researchers are skeptical about the reliability of government statistics and they have good reasons to be so. The main problem with the statistics emerging from my fieldwork was that certain statistical categories may not represent what they appear to be. The reason for this is often related to administrative constraints. The local cadres know how the statistics are compiled, and are often sarcastic about the work they do themselves under the circumstances. They refer to their compilation of statistics as "market-supply" responding to "market-demand" by higher authorities. Therefore, ambiguities are sometimes made deliberately for administrative convenience. The usefulness of local statistics, however, overrides such problems, i f one knows how the figures are compiled. It is fair to say that all statistics are based on something, and one only needs to be alert and find out what exactly the category includes, and for what purpose it is included. Some dubious information can be easily tested in the household survey and during interviews with the local cadres and peasants. Without the local statistics it would have taken me longer to understand the general demographic and economic situations. The most substantial sources of my research come from my household survey which consists of over one thousand pages of household records, and an equal quantity of field notes. The household survey includes information on household organization, land contracts, crops, output, income, expenditure, investment, social network, social welfare, and so on. It was conducted through prepared questionnaires and spontaneous interviews. The questionnaire contains over one hundred questions (see Appendix). Half of them are open questions which enabled me to elicit a great deal more information than what the questions might superficially suggest. The advantage of using the questionnaire was that they saved me from the initial awkwardness of not knowing what to ask and losing directions.  s  But I did not distribute the questionnaire to the peasants and later collect the answer sheets from them, because no matter how simple the vocabulary one uses in designing the questionnaire, one has a lot to explain in the real situation. Otherwise, many misunderstandings w i l l arise and lead to confusion and inaccuracies. The records of household surveys conducted with the help of the questionnaire provide the contents of the major chapters of this dissertation.  Economic Anthropology and Peasant Studies The present study is intended as a contribution to economic anthropology. Economy is perceived by anthropologists as a group or groups of people utilizing a common set of resources. It concerns a process of production and the circulation of material goods. Because the material wealth of societies is produced by human beings, production and distribution do not only involve individuals and activities, but also a set of social relations that people establish in the process of production and reproduction. In this approach, economy is a "total social phenomenon where religion, politics and the like find simultaneous expression" (Gregory and Altman 1989: 198), which is what economic anthropologists refer to as "embeddedness" of economic activities. In this "embeddedness" of economic anthropology, kinship and political-religious relations constitute the "invisible part" of the mode of production (Godelier 1977). The importance of kinship relations lies in their function as a social condition of the abstract appropriation of nature and its resources; as the basis of the social organization and cooperation of individuals and groups in the various concrete processes of material exploitation of resources; and as the framework for their distribution (Godelier 1986: 139). In this respect, economic anthropology is committed to the task of studying two aspects of one issue: social relations of economy and economic dimension of social relations. The significance of the association of social relationships with economic activities is, as argued by Firth that: "most social relations have an economic co-efficient; many social relations are primarily concerned with economic values" (Firth 1951: 122). In this dissertation, the focus of the study of social  organization on local government and households is intended to highlight economic relations between peasants and the state, and economic relations within the kinship system. It compares the mechanisms of how these political and social relations function in the economic activities in the two communities. Peasant society and peasant economic behavior are the center of the study.  10  The  peasantry has been broadly defined in terms of its residence, type of production, relation to the market and class segment (Cancian 1989, Shanin 1988). B y residence, peasants are rural; by type of production, peasants are mainly subsistence cultivators employing simple technology; by relation to the market, peasants are incorporated into a local community where social relationships are established through transactions of local produce surplus; and by class segment, peasants form a subordinate group in relation to the urban population and the government. Peasants are sometimes distinguished from other groups in society by their cultural status which is a combination of world-view, value system and life style (e.g. 1989; E . W o l f 1957; Foster 1965; Fei and Chang 1945). Sometimes they are distinguished by their economic status which is "occupational," indicating that peasants are subsistence-oriented agricultural producers (e.g. Firth 1951; Dalton 1969). They are also distinguished by their political status which means that they have a subordinate relationship with the larger society (e.g. Shanin 1988). However, none of these categories is so clear-cut in that one excludes another. Rather, as Geertz argues, the dimensions are interrelated in demarcating the boundaries of "both the part-society and part-culture of the peasantry" and in delineating the "relation of the peasantry to the wider civilization within which it exists" (Geertz 1961: 4).  T h e term "peasant" used here refers to the rural residents-regardless of their specific occupations (e.g. farmer, artisan, building contractor, peddler, etc.)-whose productive activities constitute part of household economy. In this context, the controversy about whether "peasant" should be understood as a more derogatory term than "farmer" (see for example, M y r o n Cohen, 1993) is not relevant. 10  The traditional Chinese peasantry has been perceived as "a way of living, a complex of formal organization, individual behaviour, and social attitudes, closely knit together for the purpose of husbanding land with simple tools and human labour" (Fei 1988: 57). In a general sense, this is still the picture of Chinese peasantry today. In this study, the peasantry is presented as an economic category as well as a social category. A s an economic category, the primary distinctive trait of peasantry is its form of livelihood;  11  and as a social category, the  peasantry shares certain culture, economic life and polity with the larger society, but at the same time retains its own tradition and social organizations.  Outline of the Study The dissertation embodies studies of social organization and patterns of production. They incorporate culture and economy into one which illustrates how the peasants live under the given social and economic conditions, and why they live the way they do. The result of the research w i l l show that the two communities-the Mosuo and the Han-have different priorities and choices in their economic activities largely due to differences in social institutions, economic environments and cultural values. The main body of the dissertation consists of six chapters. Chapter one presents the local history, natural and cultural environments. The descriptions of topography, geography and communications of the field sites reveal the grim situations of frontier life in the past and the nature of the economies at present in this part of highland Yunnan. The 600 hundred years of garrison history of the Han peasants in Jinguan and the political and religious systems of Mosuo society provide a cultural background for our understanding of the peasant economy in the two communities.  ^t has been noted that in the Chinese context, "a fair proportion of the population designated as rural obtains its livelihood not from agriculture but from the extensive range of rural industries, social services and government administration" (Hussain and Feuchtwang 1988: 36). 1  The central part of the dissertation begins with social institutions that have undergone dramatic changes along with the economic reforms. The social institutions that have had direct effect on peasant economy are the administration of local government and the form of the peasant household (patrilineal in the Han community and matrilineal in the Mosuo community). The study of these social organizations is focused on economic relations between peasants and the state, and economic relations between the members of the peasant household. Chapter two mainly deals with functions of the local administrations at county and township levels. It focuses on how the economic reforms have strengthened government autonomy at local levels, and how this has affected the local management of economic resources. The pursuit of these questions leads to an understanding of the role that the local government play in deciding the course of local economy, and the effect of ethnic status and poverty on local politics and peasant life. The behavior of the local government has impact on the state-peasant relations which are dealt with on two levels. One is the relationship between the central and local governments; and the other is the relationship between local cadres and peasants. In this context, the peasant concept of the state is dealt with in terms of moral economy. The village administration here is presented as an semi-governmental and semicollective organization; and the change of its functions also has impact on the state-peasant relations. The household is the center of peasant life. It is an economic entity and at the same time a descent organization. Household organization is associated with a set of principles regarding descent and affinal relations, and influences the local people's idea of social institutions such as marriage and adoption. In the different household organizations of the two communities, and by different descent rules (patrilineal and matrilineal), the particular kinship systems embody economic obligations and responsibilities. Chapter three is about the household organization of the Han peasants in Jinguan. It centers on the patrilineal concepts reflected in the descent system and institutions of marriage and adoption. The discussion revolves around the relationships between parents and children,  siblings, husband and wife. The analysis of the social obligations and responsibilities entailed by these social relationships reveals a network of economic relations in and outside the peasant household. Chapter four concerns the household organization of the Mosuo peasants in Yongning. It focuses on the descent system along the female line, the institution of "walking marriage," and the special sibling relationships. The particular form of matrilineal household in Mosuo society is characterized by the exclusion of affinal relations and the interdependence of sisters and brothers. This particular household organization has a direct influence on the pattern of production in Mosuo society. The differing patterns of production in the Han and Mosuo communities are dealt with in chapters five and six respectively indicating how the pattern of production in Jinguan concentrates on rice farming, while the pattern of production in Yongning concentrates on cattle breeding. "Rice Ears""and "Cattle Tails" are used in this dissertation to point out to different models of production for the two communities. They are anthropological models which are constructed on the basis of participant observation, and represent the ethnographer's understanding of the world of the local people (see Holland and Quinn 1987; Lakoff 1987; Gudeman 1986). Each model reflects an image of a particular peasant economy and denotes a value system. While the image of Rice Ears is associated with the cultural cognition of subsistence security and the production-consumption loop, the image of Cattle Tails is rooted in the cultural cognition of wealth and development. The contrast of the peasants' enthusiasm between rice cultivation and cattle production illustrates different economic forms in that both economic and cultural values are accommodated.  21  CHAPTER 1 HISTORY A N D SETTLEMENTS  This chapter outlines the natural and cultural environments of the societies under study and aims to provide background information for the reader to understand the social organization and patterns of production dealt with in the chapters to follow. A general account of topography and a graphic display of the economic status of Yunnan sketch the economic situation of the field sites. A n account of the garrison history and the Han immigration to Yunnan indicates how the Three-River Basin came into being as we see it today, and why rice farming predominates in the economy in Jinguan. The picture of Yongning differs in population and economic environment. There the old political and religious systems offer important clues to understanding the particular social organization and economy dominated by cattle breeding in Mosuo society. The immigration history of the Yongning basin not only has impact on ethnic relations, but also has implications in economic development in the area.  Yunnan Province  Yunnan is a province situated in southwest China. It borders Sichuan province in the north (an area that used to be known as Xikang, or Hsi-k'ang), Guizhou and Guangxi provinces i n the  22 east, Tibet in the northwest, Burma in the west and south, and Laos and Vietnam in the south (see M a p 1). Yunnan has an area of 394,000 square kilometers of which ninety-four per cent is mountainous. The highest point is in the northwest, with an elevation of 6,740 meters; and the lowest area is the river valley in the southeast which is only 74.6 meters above the sea level. Three major rivers flow right through this undulating land in the southwest. They are the upper reaches of the Yangtze (known as Chang Jiang in Chinese, and this particular part of Chang Jiang flowing in Yunnan is called Jinsha Jiang, or "Gold-Sand River") from the west to the east, the Mekong (Lancang Jiang) and the Salween (Nu Jiang) from the north to the south. Yunnan literally means a place "south of the clouds," which gives an impression of its geographical remoteness from the central China, and at the same time, signifies the mysterious nature of the place. Almost all Chinese writings about Yunnan favor the expression "threedimensional" (liti) because the topography creates diversities in climate, flora and fauna distributions and levels of economic development. The mountainous topography gives Yunnan difficult access to the outside world. It was not long ago that horseback was still the major means of transportation in Yunnan. The only railroad from before 1949 was built by the French in the first decade of the twentieth century. It is a 464 kilometer long narrow gauge railway (1,000 mm) from Kunming to Hekou on the border with Vietnam, and still in use today. The earliest public highway in Yunnan was 34 1  kilometers long, built in 1928 around the capital Kunming. The Second W o r l d War promoted some expansion of communications in Yunnan for the province was the military base used by the Allies. A t that time, the Yunnan-Burma road was built. It was 955 kilometers long. The conditions of public transport in Yunnan have significantly improved since the 1950s, owing to the forestry development. Since the 1980s, many "state standard highways" (guodao ) have  Yunnan is the only province in China to have both standard and narrow gauge railways in operation (I owe this information to Thomas Kampen who has done extensive research on Chinese railway system). Along the 464 k m long railway, there are 182 bridges and 171 tunnels (Yunnan Shengzhi V o l . 34, Tiedao Zhi, 1994). 1  23 been built. They are mainly sightseeing routes around and between the capital Kunming and other major cities. Entering the 1990s, several air links between cities were established to ease the travelling difficulties within the province. Probably due to the geographic situation as such, Yunnan was almost out of the reach of the Central Court of China before the 13th century. The first military invasion to Yunnan was launched by the Yuan Emperor. In 1252, Kublai Khan came from the North to conquer the Dali K i n g d o m . Henceforth Yunnan became one of China's provinces. Following this military 2  conquest, a century later, the M i n g court began to send garrison troops to Yunnan from China's central plains (mainly Hunan, Hubei, Jiangxi and Jiangsu provinces), which began the history of Han immigration to Yunnan. The garrison troops were ordered to support themselves by farming while guarding the borders. A s estimated, the M i n g troops and their dependents soon brought the number of Han immigrants to nearly one million cultivating 1.3 million mu of land which was about 42 per cent of the total arable land in the province at the time (Dangdai Zhongguo de Yunnan I 1991). Today, Yunnan has a population of 38.3 million (1992). One third of it is composed of peoples of scores of ethnic groups other than the Han. O f the total population, 87.5 per cent is agricultural. The economic development in Yunnan is, among a few other hinterland provinces, lagging behind other parts of China. The general economic status Yunnan is illustrated in Figures 1-3.  T h e Dali Kingdom in 937 A D was preceded by the Nanzhao Kingdom set up in 649 A D by the peoples known in the Chinese literature as the "black barbarians " (perhaps including the ancestors of Y i , Lisu, N a x i and Mosuo) and "white barbarians" (i.e'. the Bai) in the territory of today's Yunnan province, the south part of Sichuan province and the west part of Guizhou province. 2  Billion Yuan o b o  o b o  o b o  a> o b o  o o b o  00 o b o  Guangdong Shandong Jiangsu Sichuan Liaoning  ..raMfc>.r.tt  Henan Zhejiang Hebei Hubei Shanghai  -  1  Hunan  Anhui Beijing Fujian Jiangxi Guangxi Yunnan Shanxi Shaanxi Jilin  =3  Tianjin Inner Mongolia Xinjiang  I  Guizhou Gansu Hainan  Ningxia Tibet  1 ZJ  Heilongjiang  Qinghai  ]  I  •  o b o  o b o  a> o b o  00 o b o  Billion Yuan —L  o b o  o b o  o o b o  o o b o  o b o  Ol  o b o  CO  o o b o  CO Ol  o b o  4^  o o b o  Jiangsu Shandong  i  Guangdong Sichuan  ZZ3  Liaoning  Zi  Zhejiang Shanghai  1  Henan Hebei Hubei  I I  Heilongjiang Hunan Anhui Beijing  I  Z2  Fujian Tianjin Jilin Jiangxi Shanxi Shaanxi  I  Guangxi Yunnan Inner Mongolia Xinjiang Gansu Guizhou Hainan Ningxia Qinghai Tibet  ZZ3  o c c c  Yuan 01  CO  o o o  b o o  "o o o  o o o  b o o  b o o  Shanghai t Beijing Tianjin Liaoning Guangdong  >  Zhejiang Jiangsu I  ~1  Heilongjiang »  "1  Xinjiang P 1  Shandong Fujian Jilin Hainan Hubei Hebei Inner Mongolia Tibet  I—  Qinghai Shanxi Ningxia Hunan Shaanxi o c  Jiangxi Yunnan  a  Gansu  c o  Sichuan Henan Anhui Guangxi Guizhou  1  , 3  22. S5  a. S' 3  The Yangtze River is sometimes analogized with a dragon image to refer to China's economic development, in that the east and southern coastal regions form the dragon head and the western hinterlands compose the tail. A s shown in the figures above, though the economic status of Yunnan is not quite at the tip of the dragon's tail, 102 out of 127 counties in the province are financially subsidized by the state (Dangdai Zhongguo de Yunnan 11991: 709). In 1992, Yunnan had forty-one state designated poverty-stricken counties; and in 1994, thirty-two more counties were added. The 73 state poverty-stricken counties in Yunnan province 3  constitute 12 per cent of the total number of state poverty-stricken counties in the country; and the 7 million population living below the poverty line constitutes 8 per cent of the total population living below the poverty line in the whole country (Wanxiang Shenzhou 1994). Both Yongsheng county (in which Jinguan township is included) and Ninglang county (in which Yongning township is included) are in northwest Yunnan located in Lijiang prefecture (see M a p 1). Yongsheng is in the category of Yunnan's 102 state subsidized counties, and Ninglang is one of Yunnan's 73 state poverty-stricken counties. The primary feature of the economy in these two counties is strictly subsistence agriculture. Yongsheng had the highest economic development and living standard among four counties in Lijiang prefecture with a rural per capita income of 380 Yuan (1991). Ninglang was the poorest county in Lijiang with a rural per capita income of 259 Yuan (1991) which was one third lower than in Yongsheng.  O n e of the reasons for the increase of the state poverty-stricken counties is that the state raised the poverty line. In 1986, China classified 328 state poverty-stricken counties as in need for state priority support, and the official poverty line at the time was 150 Yuan (net) per capita income in general, and 200 Yuan (net) in minority areas and the old revolutionary bases (Zhongguo Pinkun Jingji Kaifa Gaiyaol9S9). In 1994, the poverty line was redefined and became 400 Yuan (net) per capita income (the elevation of the poverty line should reflect the inflation rate over the years), and the total number of state poverty-stricken counties reached to 592 (according to the Office of the Leading Group of Economic Development in Poor Areas under the State Council. 3  Yongsheng County  Yongsheng is some 600 kilometers northwest of Kunming (see M a p 1). In the old days, when draft animals were the exclusive means of long distance transportation, it took 18 days to get to Kunming on horse back. The people in Yongsheng were so cut off from the outside world that they had not much idea about what was going on when seeing the American and British planes flying over their villages (during W o r l d War II), nor did they understand what Liberation was about when they were liberated, as the locals say. The first public road in Yongsheng was built in the early 1940s when the Republican government retreated to Chongqing city, Sichuan province. From 1958 to the present, several major public highways have been built leading to Lijiang, Binchuan (south of Yongsheng, between Yongsheng and Dali), Huaping (east of Yongsheng, between Yongsheng and Panzhihua, a industrial center in the south of Sichuan province) and Ninglang (north of Yongsheng). W i t h the same military settlements in other parts of Yunnan in the late 14th century, M i n g troops came to be stationed in Yongsheng. The military unit at the time was called Lancang Wei: Lancang as the name of the military district, and wei as designation of the military unit (one of forty in Yunnan). The garrison settlement in Yongsheng consisted of three thousand soldiers and officers and some six thousand dependants (Yongsheng Xianzhi 1989). They brought seeds and tools with them from home in the central plains, and turned some 36,000 mu of shrub land into paddy fields. Thus began the rice farming history in the area (ibid.). Most of the troops in Yongsheng came from Changsha F u (prefecture) in Hunan province. Even today, the people in Yongsheng still speak with a particular Changsha accent which is remarkably different from other places in Yunnan. Yongsheng county today has a population of 357,515 (1991). Seventy per cent of it is Han. The rest of the population is composed of ethnic minority groups including: Y i , Lisu, N a x i , B a i , D a i , Shuitian, Taliu, Liude and H u i . The Han and H u i generally inhabit the basin  29 area. Dai, B a i , Shuitian, Taliu and Liude live in the river valley. The Y i , L i s u and N a x i dwell in the mountains. The Yongsheng county seat is at an elevation of 2,140 meters. The county has an area of 5,000 square kilometers of which 23 per cent is basin, a strip of paddy fields stretching from north to south between two mountain ranges. Yongsheng is one of the most important grain production bases in Yunnan. There are three agricultural zones featured by varieties of produce: grain crops, cash crops and animal husbandry and logging. The grain producing area is the basin and the main crops are rice and broad beans. The.cash crop producing area is in the south, and the main crops are sugar cane, tobacco, cotton and peanuts. Animal husbandry and logging are located in the mountains. The grain cropping area has two harvests a year, and the cash cropping area has three or more. The development of industry in Yongsheng is not significant. It only includes some small scale machine repairing; grain, sugar and tobacco processing; power and timber plants; mining, porcelain, garment and building material manufacture. Its output value constitutes only 20 per cent of the total output value of society (1991).  Jinguan and the Three-River Basin  Jinguan is one of eighteen townships under the jurisdiction of Yongsheng county. The township is twenty-eight kilometers northwest of the the county seat (see M a p 2). Jinguan is one of the biggest market centers in Yongsheng county, and situated on the major communication link between Yongsheng and Lijiang. It has an area of 205 square kilometers and occupies the middle and upper parts of the Three River Basin (see Plate 1). The basin derives its name from the three rivers named Meng, H u i and Ji flowing through it. The Three-River basin is probably the most fertile land in northwest Yunnan province, and has the highest rice yield in Yongsheng county. The basin has 63,418 mu of arable land  supporting a population of more than 50,000 in two townships: Jinguan and Liangguan. The basin itself is 1,550 meters above the sea level. The annual rain fall is 821 millimeter and the frost season is 96 days (Yongsheng Xianzhi 1989). The primary livelihood of the people in this area is rice farming. The 1991 rural per capita income was 543 Yuan which was the highest of all townships i n Yongsheng. Jinguan has 8,165 households with a population of 34,402 (1991). It administers eleven villages and 84 hamlets. Ten of the eleven villages in the township have access to electricity and public roads. The evidence of the history of the M i n g military settlement is still reflected in place names. The name of the township Jinguan, for instance, means an officer (guan) surnamed Jin. Many hamlets in the township were originally the bases of the garrison subdivisions, and many of them were named after the officers of the troops stationed there. Among the eleven villages, I chose two to conduct my household survey. They are called Cuihu and Wengpeng. Cuihu village is 3 kilometers east of Jinguan township (see Plate 2). It consists of four hamlets (or "natural villages"): Guguan (after a M i n g officer's name), Ruiguan (after a M i n g officer's name), Bashang ("on the dam") and Longtan ("Dragon Lake"). They were included in Cuihu administrative village in the 1950s. The name Cuihu meaning "green lake" is derived from the numerous lotus ponds in the village. The village has 881 households and 3,589 residents. Most of them are the descendants of M i n g troops. They cannot remember how many generations they have been there, but all know very well that their forefathers came to this place from Hunan province during the reign of the Hongwu Emperor (1368-1398). Wengpeng is 5 kilometers northwest of the township seat. It is a semi-mountainous village (see Plate 4). The village name comes from the surnames of M i n g officers' called respectively Weng and Peng. There used to be 5 hamlets all of which are named after the garrison officers. The hamlets were finally integrated into Wengpeng village in the 1950s. Wengpeng has 505 households and 2,200 residents. The physical layout of the village consists of a dozen homestead clusters spread along the foot of the mountains. Because the residences  are so dispersed, the village was divided into eleven production teams under the collective system (while Cuihu with nearly nine hundred household had only five production teams). This particular settlement is a result of a mixture of immigrating households from different periods of time. Some came with the earliest garrison settlement from Jiangxi and Hunan provinces, and others from Sichuan province and other places such as Ninglang county in more recent history. The most recent immigration started around 1940. Forty per cent of the households under my household survey migrated to Wengpeng after 1940. The main reasons for the recent immigration are adoption and flight from famine, bandits, military conscription or debt.  Ninglang County  The county seat of Ninglang is 116 kilometers north of the county seat of Yongsheng and 216 kilometers east of the prefectural capital, Lijiang (see M a p 1). It is at an elevation of 2,200 meters, and is considerably more modern and in better shape than the 600 year old town of Yongsheng. The name Ninglang is an abbreviation of two place names: Yongning in the north and Langqu in the south (where the present county seat is located). These two places used to be territories of two Mosuo Tusi ("aboriginal headmen"). In 1936, Langqu and Yongning merged into one administrative unit under the Yunnan provincial government. Ninglang has a total population of 204,403 (1991), 59.2 per cent of which is Y i . The second largest population in the county is Han which makes up 22.2 per cent of the total population. The third largest population segment is Mosuo which is 7.4 per cent of the total. The Pumi population is the next largest to the Mosuo, though it only consists of four per cent of the total population in Ninglang. The rest of smaller ethnic groups are Lisu, N a x i , Zhuang, B a i , Tibetan, M i a o and H u i . Because the majority population is Y i , Ninglang was designated as the Y i Autonomous County in 1956.  Ninglang has an area of 6,025 square kilometers. The highest point is 4,510 meters above sea level, and the lowest is 1,340 meters. Fifty-four per cent of the area is at an elevation between 2,800 and 4,500 meters, where the annual temperature is 8.5°C on average with a frost season of 240 days and annual rain fall of 996 millimeters. The area between 2,100 and 2,800 meters above the sea level is 26.5 per cent of the total area, where the average annual temperature is 10-12.7°C with a frost season of 160-200 days and annual rain fall of 1,012 millimeters. The area at the elevation between 1,500-2,100 meters is 18.3 per cent of the total area, where the average annual temperature is higher than 16°C with a frost season of less than 160 days and annual rain fall of less than 800 millimeters. The area of the lowest elevation between 1,340-1,500 meters is only 0.55 per cent of the total area, where the average annual temperature is 20°C with a frost season of no more than 110 days and annual rain fall of less than 700 millimeters. There is a general pattern of association between elevation and the residence of ethnic population. The Y i mainly live in the mountains at elevation of 2,500 meters and above, the Mosuo, Pumi and Han reside in the highland basin at elevation of 2,000 to 2,600 meters, and the areas of the lowest elevation is the Jinsha River valley occupied by mostly by the Zhuang. Communications in Ninglang are difficult compared to other counties in Lijiang prefecture. Though there are bus services daily to the county seat of Ninglang from Yongsheng, Huaping (south of Ninglang) and Lijiang (passing through Yongsheng), the narrow winding mountainous roads make travel very difficult. The road condition is poor. Soon after crossing the border with Yongsheng, the asphalt surface ends and the road becomes narrower and narrower as it enters the mountains deeper and deeper. The mountains are mainly inhabited by the Y i . Their timber cabins sit on the hill slopes around which buckwheat is sparsely planted. In the distance, some forests are visible. Due to the particular topographical situation, the Ninglang economy has a different emphasis from Yongsheng in terms of exploiting natural resources and economic development.  While Yongsheng is one of the most important grain production bases in Yunnan, Ninglang is one of the most important forestry and animal husbandry bases. Ninglang has some three million mu (200,000 hectares) of pasture-land which is about forty per cent of the total area in the county. Livestock farming in Ninglang is divided into three zones. A t an elevation of 3,000 meters in the south, sheep, goats and oxen are grazed. The sheep stock in Ninglang is the second largest in Yunnan province. In the basin area at an elevation of 2,000 meters and above, the main livestock are pigs (making up 37 per cent of the total number in the county), buffalo (making up 98 percent of the total number in the county) and other draft animals such as horses and mules (making up 44 per cent of the total number in the county). The area in the lowland along the Jinsha River is less significant as far as animal husbandry is concerned. The animals raised here include pigs and goats which contribute some twenty per cent of the total output of animal husbandry in the county. (Ninglang Yizu Zizhixian Zhi 1993: 342-3) Forest resources in Ninglang are abundant. The forest area is 460,000 hectare (6.9 million mu) which is 76 per cent of the total area of the county. The forestry produce consists of two categories: agricultural and industrial. Agricultural produce includes saplings, firewood, pine seeds and mushrooms. Industrial produce includes logging and transportation, timber processing, and so on. The revenue from both agricultural and industrial forestry production contributes over 70 per cent to the county's total revenue income (Ninglang Yizu Zizhixian Zhi 1993: 320). The 446,000 mu of arable land only amounts to five per cent of the county's total area. Agriculture in Ninglang can be divided into four zones by elevation and crops. A t an elevation of 3,500-4,500 meters (mostly inhabited by the Y i ) , the main crops are buckwheat, oats and tubers (potato and wuqing).  4  Farming in this area used to be slash-and-burn. Forests were  ^Wuqing is locally called yuangen, i.e. "round root." Its Latin name is Brassica rapa. It is grown for food and fodder.  cleared between August and October every year, crops were rotated (buckwheat-oats-potato) for two or three years, then the land was abandoned and people moved to another place. After 2-3 decades of government interference and administrative control, most of the Y i were forced to settle down and give up the slash-and-burn agriculture. A t an elevation between 2,600 and 3,000 meters, the population is mixed with Y i , Pumi and Mosuo. The main crops are corn, buckwheat, potato and beans. The basin area is at an elevation between 2,000 and 2,600 meters. Ninglang has ten big and small basins, the total area of which is about 145,500 mu which is no more than 1.6 per cent of the total area of the county. The economy in the highland basins is the most flourishing in Ninglang. The majority population of the basin area is Mosuo, Pumi and Han. The main crops are rice, barnyard grass, cereals and vegetables. A t the lowest elevation of 1,400 to 2,000 meters, the main inhabitants are L i s u , Zhuang and Mosuo. This is the warm temperature area. It is the only place in Ninglang where two harvests a year are possible.  5  Yongning and Its Peoples  Yongning is located at the north end of the Ninglang county border, 96 kilometers away from the county seat (see M a p 3). There is a daily bus from the county seat to the Yongning township seat under normal conditions. But the service is often interrupted in the summer and autumn whenever there is a heavy rain and in winter whenever there is snow. Yongning is known to the outside world for two attractions. One is the natural wonder of the Lugu Lake and the other is the cultural myth of the Mosuo matrilineal family. Lugu lake  T h e descriptions about the natural and economic environments are mainly based on the sources from Ninglang Yizu Zizhixian Zhi 1993, and Ninglang Yizu Zizhixian Gaikuang 1985. 5  is 50 kilometers in circumference and 73 meters in depth. The surface of the lake is 2,700 meters above sea level, which is one hundred meters higher than the basin. The border line between Ninglang county and Yanyuan county in Sichuan province passes through the middle of the lake. The lake abounds with a variety of fishes. Fishing is the major sideline production of the households along the shore. The management of fishing in the lake is under the supervision of a joint committee between Ninglang and Yanyuan counties. Tourism has slowly increased in the lake area. There are now a dozen restaurants plus lodges erected on the lake shore. The delicatessen offered to tourists are fresh fish from the lake and home fed chicken. Canoeing is a popular tourist program. The canoe is hand-made out of the trunk of a huge tree, and paddled by a pair of people, mostly Mosuo women from the two hamlets ashore. A n hour trip is to Bird Island in the middle of the lake; and a day trip takes one into Sichuan province. B i r d Island is covered in woods. O n the top of the hill, there two sightseeing sites: a newly built Lamaist temple and a white pagoda which is a tomb of a Yongning T u s i .  6  On the north end of Lugu lake, there lies a mountain known as L i o n Mountain. The name is gained from the look of the mountain which resembles a lion lying prone with head raised facing south and tail stretching to the north. It, in the Mosuo language, called Hidi Gamu, a goddess worshipped by the people in Yongning. A t the foot of the L i o n Mountain, lies the Yongning basin which is 27 kilometers north of the Lugu lake. The basin is the largest basin of the county which is 2,600 meter at the sea level (a hundred meter lower than the lake). It has distinct dry and wet seasons. The dry season is between November and M a y and the rainy season is between June and October. The frost season is from the end of October through m i d April. Yongning is one of Ninglang's sixteen townships, and has a population of 16,875 (1991). Its largest population is Mosuo which amounts to 37.7 per cent of the total. The second  6  T u s i A h Yunshan died in the 1930s.  36 largest population is Han amounting to 26.3 per cent of the total. The third largest is Y i amounting to 16.9 per cent of the total. The fourth in the rank is Pumi which is 10.2 per cent of the total. The remaining 9 per cent of the population is composed of the ethnic groups such as Zhuang, N a x i and Tibetans. Ninety-six per cent of the Yongning population lives on agriculture. The main livelihood of the Mosuo in Yongning is a combination of crop faming and animal husbandry. The per capita income of the Yongning township in 1991 was 418 Yuan (which was 125 Yuan lower than in Jinguan). W i t h a population less than 20,000, Yongning is classified as a xiang comprising six villages which administer sixty-six hamlets and 3,006 households. Less than half of the hamlets have access to electricity and public roads. Kaiji village has the largest 7  M o s u o population of all. It consists of 18 hamlets, 954 households and 5,827 people (which is more than 90 per cent of the Mosuo population in Yongning). The four hamlets that I selected for household survey are all included in Kaiji administrative village.  8  Compared to the concentrated residences in Jinguan, the hamlets in Yongning are much smaller in size and more sparse in space. Gesawa is located in the north, half way up the mountain; Dapo is situated in the southeast, at the foot of the mountain. Both these hamlets are 9  inhabited by Mosuo. Gesawa has 29 households and 225 people; and Dapo has 32 households and 270 people. Neiba is situated right in the middle of the basin. Residences are concentrated, and the entire population is Han. This hamlet did not even exist fifty years ago. According to one of the oldest Han settlers, Neiba hamlet only had 13 households i n 1953, but now it has 103 households and 570 people. Pijiang hamlet is situated on the same street where the township government and the market is located. It has 106 households and 536 people. The  O f the total 66 hamlets in Yongning township, only 26 hamlets have access to electricity, and a number of them have already lost lamp poles and wires. 7  8  T h e name of the administrative village is changed.  9  T h e names of the two Mosuo hamlets are changed.  residents of Pijiang hamlet are mixed. Many of them are Naxi, and the rest are Han, Mosuo and Tibetan. The majority of the residents were originally artisans and traders. This was how the hamlet gained its name Pijiang, meaning "tanners." The name was once changed into Pingjing, meaning "peace," after the hamlet was burned in a blaze by the rebels in the spring of 1 9 5 6 .  10  Accidentally, pingjing ("peace") in the N a x i dialect sounds almost identical to pijiang ("tanner").  The Mosuo China has fifty-five officially classified ethnic minorities, and there are twenty-four of them living in Yunnan. However, Mosuo is not a designated nationality (zu). Mosuo might have been once included among some 260 names of ethnic minorities in Yunnan (and some 400 in all over China) submitted to the central government for identification in the early 1950s, but its insignificant size (in addition to all other historical circumstances) has not won an official recognition of its ethnicity. Instead, it has been defined as a sub-division of a larger ethnic group: N a x i .  1 1  Ethnic identity in the Tibeto-Burmese Highland in southwestern China is very complex as the ethnic groups are numerous but the population of each group is small; and those peoples have lived in this area for so long and yet there is so little written history about them. In addition, one often finds that the existing history and local legends, the other- and selfidentifications do not necessarily coincide with each other.  12  The total Mosuo population in China is estimated at around 40,000 of which 15,000 live in Ninglang, 20,000 in Yanyuan and M u l i and the rest in Yanbian, Xichang and other places (Ninglang Yizu Zizhixian Zhi 1993: 178). There are different names used to refer to the same  10  S e e Liangshan rebellion in Chapter 3.  1  ^ o r an official review of China's ethnic identification, see Hsiao-Tung F e i (1981).  S e e , for example, Stevan Harrell (1990): "Ethnicity, Local Interests, and the State: Y i Communities i n Southwest China." 12  group of people which I call, in this dissertation, Mosuo. Mosuo is a phonetic transliteration of a Chinese term which is an other-referent as well as self-referent used by the peoples (Mosuo and non-Mosuo) living in Yongning today. In the Chinese historical records and contemporary ethnographies, Mosuo and "Moxie" are the commonly used names. In the English literature, there are various terms and forms of transliteration: Hli-khin meaning "the basin people" (Rock 1947: 355); "Moso" (e.g. Jackson 1979; Shih 1993); and "Naze" (e.g.Weng 1993; Harrell 1995). A s far as the provincial government is concerned, the Mosuo in Yongning are N a x i , or a branch of the N a x i . Since the 1950s, the term N a x i has been widely used in the academic research and public media which cause considerable confusion as to which group the name actually refers. A s a matter of fact, the Mosuo were recognized as a zu or nationality in Ninglang county and in Yunnan province until 1959 (Ninglang Yizu Zizhixian Zhi 1993). Following the 13  official ethnic identifications embarked by the central government in the 1950s throughout China, the Mosuo in Yunnan was recognized as the subdivision of N a x i and henceforth incorporated into N a x i zu; and at the same time, the Mosuo in Sichuan were included in Menggu zu or Mongolian (Ninglang Yizu Zizhixian Zhi 1993: 176). Since 1978, the Ninglang county government-on behalf of the Mosuo community-has lodged several petitions to the provincial and National People's Congresses for rehabilitation of the Mosuo ethnic identity. In 1990, the Mosuo in Ninglang were officially granted a title of Mosuo ren ("people"), but still remain as a sub-group of the N a x i zu. However ambiguous the re-definition is, the partial recognition of the Mosuo identity is a political compromise that the Mosuo and the government has so far reached. The Mosuo cadres are not content with this result because the title ren does  T h e term zu is officially translated as "nationality." In China, the officially recognized zu are 56 (including the Han). 13  not entitle the community and the individuals of the community to certain rights and privileges in both political and economic terms.  14  The N a x i and Mosuo are believed to be the descendants of the same people, that is the the legendary Qiang (or Chi'ang) from northwest China. Both historical records and my observations in the field suggest that the Mosuo might have been related to the Naxi, but lost contact at some point in their history for uncertain reasons. A t present, the N a x i and M o s u o appear to be two different groups of peoples living on two sides of the Jinsha River. The N a x i are concentrated in Lijiang on the west bank of the river; and the majority of Mosuo are located in Yongning ( M u l i and Yanyuan as well) on the east bank of the river. The main differences between the N a x i and Mosuo in political system, religious practice, language and social organization are roughly as follows.  • The N a x i used to be ruled by the Tusi ("aboriginal governor") surnamed M u ; and the Mosuo were subjects of the Tusi surnamed A h . The N a x i commoners share a single surname He (there are more varieties of surnames in modern times), while the Mosuo commoners carry individual family names with double or multiple syllables similar to those of Tibetans.  • The N a x i aboriginal religion is Dongba which has pictographic scripts, and the Mosuo one is Daba which has no written script. Under the influence of the Tibetan Buddhism, the N a x i were converted to the Red Sect (of Lamaism) in the 1600s; and the Mosuo were converted to the Yellow Sect (of Lamaism) in the 1700s (see Mosuo Religion below).  • The N a x i and Mosuo have different languages of which about thirty per cent of the vocabulary is mutually intelligible.  O n e immediate concern of being recognized as Mosuo zu for the local leadership is the legitimacy of establishing the ethnic autonomous township (minzu xiang). A t present, although the majority population in Yongning is Mosuo amounting to nearly 40 per cent of the total, Yongning township is not entitled to be an ethnic autonomous township because the M o s u o are not recognized as a separate "nationality" (zu). 14  • The N a x i are patrilineal and the Mosuo (in Yongning) are matrilineal. • The N a x i bury their dead while the Mosuo cremate them. • The N a x i women wear trousers and a cap, the Mosuo women wear a skirt and a turban.  None of the evidence provides a solution to the puzzle of the relation between the N a x i and Mosuo, and probably w i l l "ever remain a moot question," as Rock suggested a half century ago (Rock 1947: 4). Some Mosuo claim to be Mongolian (Weng 1993; L i 1994). Consistent with what has been elaborated by Shaoming L i (1986) and Xingxing L i (1994) about the relationship between Mosuo (or Naze) and Mongolian, the local people in Yongning gave me an account of the Mosuo identifying themselves with the Mongolians. It is said that the Mosuo do not celebrate the M o o n (mid-autumn) festival like the Han and N a x i do. This is because the M o o n festival originated from the local uprisings against the Mongolian rulers, and the Mosuo apparently sided with the Mongolians for one reason or another (either ethnic or political). Some Mosuo youth claim that their boots and knives are of the Mongolian tradition. Some historical records and ethnographies make the Mosuo Mongolian connection plausible. The Mongolian identity of the Mosuo is apparently associated with the Mongolian presence in the Yunnan-Sichuan border area: once on their way to conquer Dali in the 13th century led by Kublai Khan, and the other time in the 14th century when the Mongolians retreated to the Yunnan-Sichuan border area under the suppression of the M i n g garrison troops.  15  Lincan L i and Joseph Rock, during their ethnographic research conducted separately a  half century ago, noted that the Yongning Tusi family claimed to be the descendants of a Mongolian officer left behind by Kublai Khan (Lincan L i 1984; Rock 1947). But by studying  L a n c a n g wei of the M i n g garrison troops stationed in today's Yongsheng was apparently related to this event. 15  their family nomenclature and ethnic history, both L i and Rock concluded that the Yongning Tusi were not Mongolian but Mosuo. The same conclusion is also reached by other scholars (e.g. Shaoming L i 1986; Xingxing L i 1994). In spite of the lack of support for the Mongolian origin of the Mosuo Tusi, the historical fact is that the establishment of the Tusi system is certainly a legacy of the Mongolian invasion to Yunnan.  The Tusi System Tusi is a Chinese transliteration which literally means "aboriginal governor." A s the Ninglang county history records, Kublai Khan and his troops passed Yongning in 1253 on their journey southward to Dali. The emperor named the place Yongning for it is a good omen as yong means forever and ning means peace. In the meantime, the aboriginal governor submitted himself and his subjects to the rule of the emperor; and in 1279, Yongning became a prefecture of the Chinese empire (Ninglang Yizu Zizhixian Zhi 1993: 46). The establishment of the Tusi system in Yongning (as well as in Langqu) took place in 1381. It was the central policy to let the aborigines rule the aborigines. Once the Tusi system 16  was established, the succession of the Tusi position was required to have the consent from the central court. The Tusi succession followed the patrilineal line specified by the central court. This is probably the reason why the Tusi family in Yongning was patrilineal while the ordinary Mosuo families are matrilineal. Along with the establishment of Tusi system, the system of tribute (from the local Tusi to the central government) and the system of reward and punishment (from the central government to the local Tusi) were enforced. In the M i n g dynasty, the central court reformed the Tusi system for the purposes of restraining the power of the aboriginal rulers in the southwest. This reform is known as gaitu  T h e Tusi system can be compared with the present ethnic minority autonomous administration, but the Tusi government in the past was far more powerful than the ethnic leadership today. 16  guiliu ("submitting aboriginal governors to the control of the imperial commissioners"). A t this time, Yongning and Langqu were brought under the jurisdiction of Beisheng prefecture (i.e. today's Yongsheng county). In some places of the southwest China, the aboriginal reigns were gradually abolished in the Qing dynasty, but the Tusi in Yongning and Langqu persisted until the 1950s. The Yongning Tusi was the ultimate owner of the land and mountains in the area. Under him, there were two powerful administrative positions called Zongguan ("governor") and Kenbu (head of the Lamasery). These two positions were held by the Tusi's most trusted relatives in the patrilineal line. Under these two top ruling positions, there were several subdivision chiefs. They were all related to the Tusi family and shared the same surname A h . A l l the relatives of the Tusi were called Sipi which in the Mosuo language means "landlords" who received enfeoffment from the Tusi. The Tusi and Sipi families received a Chinese education and spoke mandarin Chinese. In the Tusi administration, some Han people were employed for bookkeeping. Under the Tusi system, the Mosuo society was differentiated by a hierarchy. The commoners were called tongren who farmed the Tusi land and were subject to paying certain tributes each year. The lowest class was piazi or wo. They either worked as corvee labor for a period of time each year to offset tribute payments, or as long-term farmhands of the Tusi family (or household slaves). Children born into the piazi family became Tusi household slaves after they had their puberty ceremony at the age of 13. The last Yongning Tusi governor was A h Shaoyi. According to his living relative in Dapo (the grandson of one of A h Shaoyi's wives' sister), A h Shaoyi had three wives, two locals and one from Lijiang. The only known offspring of the Tusi was a daughter born by his N a x i wife from Lijiang, but the child died of illness at the age of seven or so. A h Shaoyi himself survived Liberation and was appointed county deputy chief from 1950 to 1959. H e died in 1967.  Although the Tusi system has long been abolished, the prestige and spiritual power of Tusi have, to some extent, remained. The old Tusi residence is located in Zhongshi, 3 kilometers south of the township seat, and well maintained. It is managed year round by a relative of the old Tusi family. The master of the residence is the younger brother of the late Tusi. H e has two roles, and both symbolic. O n the one hand, he holds a position with the United Front Organization in the county government; and on the other, he is the head of the Yongning Lamasery. He is worshipped by the Mosuo as the "Living Buddha." When he returns to his home from the county seat on the Lunar New Year, the Mosuo visit the Tusi residence to show respect and receive blessing from the "Living Buddha." The traditional Mosuo society is said to be a society where political and religious power merged. Today, the political power of the Tusi family has become totally symbolic, but the religious function is maintained. Different from the people of other ethnic groups living in the Yongning basin, religion plays an important role in the Mosuo life.  The Mosuo  Religions  The Mosuo religious system consists of Jabba, or Lamaism; and Daba, an aboriginal shamanic practice. The Mosuo in general believe in both. The priests of both 17  religions-Jabba  and Daba-are regarded as the most knowledgeable and highly respected people in Mosuo society. The Mosuo Jabba religion derives from the native Tibetan B o n religion in combination with Buddhist religion. According to David C . Graham (1961), Buddhism entered Tibet in the  Lama and Lama jiao (Lama religion) used in the texts are the Chinese translation of the Mosuo term Dabba. The reader should bear in mind that the term Lamaism used here to refer to the Mosuo religion may be controversial in terms of the strictest definition of the Tibetan-related religions. Without a discussion of the series of reforms of Buddhist religions in the TibetanBurmese highland, the complicated religious systems cannot be fully understood. A s this study is not about religion itself and the purpose of introducing the Mosuo religion is for understanding the Mosuo culture and society, I choose to use Jabba (Mosuo Lamaism) for the convenience of the reader. 1 7  44 7th century during the reign of Srongtsan Gampo following the arrival of his two wives, one from India and the other from China. In the first few centuries, Buddhism encountered resistance from the native B o n religion until Padma Sambhava from India brought with him Tantric Buddhism and founded the Red Lamaism (or Nyingmapa). The Red Lamaism succeeded owing to the emphasis of Tantric Buddhism on magic and the exorcism of demons, similar to the Tibetan B o n religion. Between the 10th and 14th centuries, the conflicts between the B o n religion and Buddhism continued and led to further religious reforms in Tibet. In the 14th century, Tibetan Lamaism was reformed by Tsong-ka-pa who established the "victorious sect" (i.e. the Y e l l o w sect of Lamaism, also known as Gelugpa) which introduced ceremonial garments (yellow hats and robes), elaborated rituals and enforced celibacy. F r o m then on, three main sects of Lamaism co-existed in Tibet: Red Lamaism, Yellow Lamaism and Black Lamaism (i.e. the original B o n r e l i g i o n ) .  18  Rock noted that the residents of Yongning were "adherents of the Black B o n sect" before Lamaism penetrated the south (Rock 1947: 356). Anthony Jackson believes that the N a x i K i n g of Lijiang was converted to the Red Sect of Buddhism around 1600 A . D . and that the Fflikhin [Mosuo] were converted to the Yellow Sect of Buddhism around 1700 A . D . (Jackson 1979: 35). N o matter how and when this Tibetan religion entered Yongning, the Y e l l o w sect of Lamaism is the dominant religion among the Mosuo living in Yongning today. It has been estimated that one third of adult men were Lamas in Yongning before the 1950s (Zhan et al. 1980; Y a n and Song 1983; Zhao 1987). This estimation is not very far from Graham's estimation that one out of five persons in Tibet is a Lama and that "every family contributes at least one son to the priesthood" (Graham 1961: 99). The Yongning Lamasery is located in the vicinity of the township seat (see Plate 9). The highest leader of the Lamasery is the younger brother of the late Tusi. The old Lamasery has  T h e above account of Tibetan religions is largely based on the description by David C . Graham (1961: 97-9); and the two Tibetan terms for the Red and Yellow sects of Lamaism are suggested by Stevan Harrell (1995). 18  been under reconstruction since the 1980s. Every year, it holds regular meetings and attracts thousands of worshippers. There are currently some 100 Mosuo Lamas or Jabba in Yongning. Most of them live at home in the hamlet, and provide religious services to the local residents. There are four Jabba in each of the Mosuo hamlets that I studied. A l l of them have received religious education in Tibet and possess scriptures written in Tibetan. Jabba are invited to provide services mainly at funerals to comfort the soul of the dead, and are sometimes consulted for naming, healing wounds and curing illness assisted by herbal medicines. The Lamaist influence is pervasive in Mosuo society. Different from the Jabba religion, the master of Daba does not possess any written scriptures and the religious knowledge is his living memory. According to the local people, in ancient times, there were four Mosuo tribes migrating from the north. Each of the them had a Daba who possessed the knowledge of the history of Mosuo society. Today, Yongning has only one Daba who is 74 years old, living in Dapo hamlet (Plate 10). He has had several disciples, but none of them is, at the moment, considered suitable to succeed him. The Daba's successor is chosen through ritual performances. It is said that the magic and power of a Daba are transmitted through mind and spirit. When he foresees his death, the Daba w i l l hold a dance ritual at which his disciples dance in front of him; and judging from the individual performance, the Daba selects his successor. Both Jabba and Daba are influential in Mosuo life. In some circumstances, people send for a Daba when they are i l l ; and in others they prefer to consult a Jabba. The Daba is believed to possess supernatural forces, prophesy ominous events (e.g. death), and communicate with the spirits of the dead. The Jabba's function at funerals is to release the soul from purgatory, while the power of the Daba is believed to put the soul at rest especially when dealing with accidental death in which the soul of the dead is believed to be harmful to the living. The Pumi living in Yongning claim that the Mosuo Daba religion is very similar to their native religion. The Pumi used to be known as Xifan. In the 1960s, the official name for the Xifan living in Yunnan became Pumi, while the Xifan living in Sichuan province and Tibet  became Tibetan.  19  Culturally, the Mosuo are more similar to the Pumi than to the N a x i . The  Pumi living in Yongning all speak Mosuo. The Pumi language-which contains more Tibetan vocabulary-is only spoken among the Pumi. Their residences are mingled with the Mosuo's in the hamlet. While the Mosuo have two poles (respectively standing for female and male) in the middle of the main house, the Pumi have only one pole in the middle of the house, which is known as "male pole." This architectural difference is consistent with the difference in social organization: the Mosuo are matrilineal, the Pumi are patrilineal with preference for the marriage between mother's brother's daughter and father's sister's s o n .  20  In spite of the differences, the  Pumi, like the Mosuo, are legendarily believed to be the descendants of the Qiang. The ethnic entanglement between Pumi and Mosuo is as complex as it is with N a x i .  2 1  Some believe that the Mosuo ancestors first came to settle down in Yongning after defeating the Tubo at the end of the Han dynasty (206 B . C . - 220 A . D.) (Lincan L i 1984: 2 5 1 ) .  22  Other  historical sources suggest that between 7th and 9th centuries, the Xifan were sent by the Tubo government down to the south to occupy the areas along the Jinsha River (Ninglang  Yizu  Zizhixian Zhi 1993: 220). Rock was inclined to believe that the aborigines of Yongning were Xifan (or Pumi) before the Mosuo migrated from the north (Rock 1947). This is consistent with Gong's study of the Tusi system, in which he claims the ethnic identity of the Yongning Tusi to be Pumi (Gong 1992). If the Mosuo do have some historical connection with the N a x i , it is 23  T h e total population of Pumi in China is around 30,000, concentrated in Lijiang, Nujiang and Diqing prefectures in Yunnan. 1 9  D u e to the interactions with the Mosuo, the practice of "walking marriage" characteristic of Mosuo matriliny, is nowadays not uncommon among the Pumi in Yongning. 2 0  2 1  See Stevan Harrell (996): "The Nationalities Question and the Prmi Prblem."  T u b o is a Chinese term originally for Tibetans. In Yuan and M i n g history, Tubo refers to the aborigines (including Xifan) living in the Tibeto-Qinghai highland. Shaoming L i suggests a connection between some ancient Qiang tribes and Tubo, and the subsequential influence of Lamaism among those Qiang tribes ( L i 1981: 69-70). 2 2  T h i s can be another explanation, in addition to the Mongolian origin, for why the Yongning Tusi was patrilineal while the rest of Mosuo society is matrilineal. 2 3  47 reasonable to assume that the Mosuo should be even closer to the Pumi. A s estimated, the Mosuo had lived in Yongning for some 700 years before Kublai Khan passed this basin i n the 13th century (Ninglang Yizu Zizhixian Zhi 1993: 178). B y this calculation, the history of the Mosuo inhabiting Yongning is about one and a half millennia. Therefore, it is more than reasonable to say that both the Mosuo and Pumi-whether they are of the same people or not-are de facto natives of Yongning. The significant immigration of other ethnic populations to Yongning began only two and half centuries ago.  Immigrants Immigrants in Yongning include the Y i , Han, Naxi, Tibetan and other minor ethnic groups. The massive Y i immigration started two and half centuries ago, when several Y i clans moved into the Yongning-Langqu area from the north, an area known as Greater C o l d Mountains (Da Liangshan) in Sichuan province. While some of them migrated further south into Huaping and Yongsheng, five clans stayed in Ninglang. The Y i tribal leaders first rented lands and forests from the Mosuo Tusi. A s the Y i population expanded and their power grew (mainly through opium cultivation in the mountains), some parts of the Mosuo Tusi mountains eventually fell into the hands of the Y i . B y the end of the Qing dynasty (1644-1911), the Y i population in Ninglang had reached twenty thousand. In 1956, the first census showed that the Y i population in Ninglang was nearly fifty thousand (Ninglang Yizu Zizhixian Zhi 1993). Today, most of the Y i inhabit the mountains in the middle and southern parts of Ninglang. Because the Y i originally came from the Greater Cold Mountains (Da Liangshan), the mountainous area that the Y i live in today in Ninglang is called the Lesser Cold Mountains (Xiao Liangshan). The first Han immigration to Ninglang started in 1838. Most of them were attracted by the silver mine in Langqu. A majority of miners came from Sichuan province; some came from the south of Yunnan. In 1908, the silver mine was closed down, and the miners became farmers and settled in the basin area near today's county seat. But the H a n - o w i n g to the  Yongning Tusi's strict rules-were essentially kept out of the Yongning basin until the 1940s. The massive Han immigration into Yongning was, in a way, encouraged initially by the Republican government. In 1933, the government launched a rural economic reform to develop China's southwest. Land survey was conducted as the government believed that the actual amount of arable land was more than what was shown in the statistics. A s recorded, on the basis of the land survey in the first ten counties (out of the total 98 at the time) in Yunnan, the government concluded that the actual amount of arable land was up to three times higher than previously assumed (Yunnan Sheng Nongcun Diaocha 1935). This land survey was motivated and accompanied by taxation enforcement in the border regions. In Ninglang, the land survey was conducted around 1937; and in the following year, the provincial government issued assignments of land tax to Ninglang. Due to the resistance from the local Tusi, the tax assignment was not fulfilled which eventually led to the removal of two Ninglang government officials from their posts on charges of duty negligence (Ninglang Yizu Zizhixian Zhi 1993). The first Han household settled in Yongning in 1940. O l d Huang (now in his 60s) whose family came to Yongning among the first few told his story. His great grandfather (whose forefathers were originally from Nanjing and had been drafted in the garrison troops to Yunnan) came to Ninglang from the Kunming area to work in the silver mine. After the mine was closed down, the great grandfather assumed farming in a basin area around Langqu. In the generation of O l d Huang's father, life became more and more difficult: land became scarce due to the population expansion in the basin and the constant looting by the Y i in the mountains which contributed more to the subsistence crisis. A s O l d Huang's father was looking for a way out, he met a government official who came to Ninglang after conducting the land survey from Lijiang through Yongning. Sympathetic with the situation associated with O l d Huang's father, the official suggested that he move to Yongning where a large area of land had not been cultivated. But at the time, everybody knew that the Yongning Tusi did not allow the Han to enter the Yongning basin. The official assured O l d Huang's father that the government had notified the Tusi that i f he continued to refuse letting the Han migrate into the basin, the state  would confiscate the waste land. This is how O l d Huang and his family came to Yongning. A t that time, he was 9 years old and had six siblings (four brothers and two sisters). H i s father rented land from one of the Mosuo landlords, and began to grow rice. Since then, O l d Huang's family has lived in Neiba hamlet for four generations, and grown into 14 households of 95 people today. The Han population in Neiba hamlet is divided into a Yunnanese group and a Sichuanese group. The people of the Yunnanese group are those whose ancestors were originally the M i n g troops from Jiangxi, Jiangsu and Hubei, and who have long settled down in other parts of Yunnan before coming to Yongning. The people of the Sichuanese group are the peasants who were driven by poverty and migrated directly to Yongning from the southern part of Sichuan. Before the Han immigration, the Naxi were already in Yongning. The first N a x i family from Lijiang settled down in Yongning about one hundred years ago. Unlike the Han immigrants who were farmers, most of the N a x i immigrants were artisans. This is how the name of the hamlet Pijiang ("Tanners") came into being. In the early 1950s, Pijiang hamlet had some 50 households. Half of them were professional tanners, silversmiths, copper smiths and blacksmiths. The other half of the 50 households comprised traders, a mixture of N a x i , Han and Tibetans. The tanners made Tibetan boots for sale outside Yongning, tanned skins, sewed ordinary leather shoes and fur coats, and made leather harness cords for the locals. In the 1950s, the tanners were organized and worked in the cooperative, which was dissolved at the same time as the the household production responsibility system was implemented. In recent years, the growth of the local market and the availability of consumer goods have put the tanners and silversmiths out of business. Today, there are only three tanners left in the hamlet, and many silversmiths have retired without successors. The N a x i in Yongning have the highest percentage of employment outside agriculture. Some of them work in the government institutions, and others are engaged in food processing and trades. A l l residents of Pijiang hamlet including the Han and Mosuo speak the N a x i language. The N a x i can understand the Mosuo language but rarely speak it. The N a x i i n  50 Yongning do not have their own religion, and it would appear as i f the early N a x i immigrants had left their religion behind when they came to Yongning from Lijiang. For some religious purposes, the N a x i in Yongning consult the Mosuo Jabba. Occasionally, older N a x i women are seen going to the Yongning Lamasery to worship and make offerings. The Han and N a x i immigrations have had much impact on the Yongning economy i n terms of the development of both rice cultivation and commercial trade. In day-to-day life, the Han and N a x i have had much interaction with the Mosuo. In the early years of the settlement, the Han exchanged rice and vegetables with the Mosuo for salt and tea, and exchanged bean curd for home-brew. Today, this exchange relationship is still maintained. Every year, the Mosuo go to the Han hamlet for rice seeds, and the Han go to the Mosuo for firewood. M a n y Han in Yongning understand the Mosuo language. The younger generation can understand some 50 per cent of the Mosuo language and the old generation understand more. This suggests that either the Han had more and closer contact with the Mosuo in the past, or more Mosuo today have learned to speak the Han language. B y comparison, the relationship between the N a x i and Mosuo is less on the subsistence level, due to the nature of the N a x i crafts which are somewhat commercialized. The N a x i crafts have provided the Mosuo as well as the Han with daily necessities from shoes to utensils. Today, much of the N a x i produce is still much appreciated by the locals such as corn spirit, processed food (e.g. doufu, and pea jelly), and copper and silverwares. Despite the contacts, intermarriage between Mosuo and Han is rare (though a couple of Han households under survey have Mosuo wives). This social distance, according to the Mosuo, is not a matter of prejudice, rather a matter of custom. Apart from religion, social organization and the related customs are the main factors that maintain the cultural diversities among the peoples in Yongning, which have direct and indirect influence on their economic activities.  51  CHAPTER 2 L O C A L ADMINISTRATION A N D ECONOMIC A U T O N O M Y  The institutional changes brought about by the economic reform have had significant impact on China's economic development; one of the most important institutional changes has been the decentralization which has altered the political relations between the central and local governments and had great implications for the management of local resources. In the past decade, the reform policies have allowed local governments more and more latitude to manipulate political and economic means to accelerate economic development. The transition to a market economy and increased productivity have helped strengthen the power of local governments. The decentralization-characterized by revenue sharing and a shift i n power from the central government to the local governments-has been noted as the direct outcome of the national fiscal reform in the 1980s (Oksenberg and Tong 1991). The economic incentives for the local administrations and the political tolerance of the central government have created a situation where local officials take great initiative in the reform. The economic authority assumed by local officials during the process has been referred to as "local state corporatism" (Oi 1992), "neolocalism" (Nee 1992), and "federalism" (Montinola et al. 1995), and these terms all suggest a greater or lesser degree of local autonomy. However, it is important to note the extent to which the economic basis of this local autonomy differs in different parts of China. In most coastal and economically more advanced  regions, local power has been built and strengthened through the accumulation of revenues generated by a rapidly developing local industry. In many other places in the hinterland where the economy remains less developed, the economic autonomy of the local administrations has been benefitted largely from the state preferential treatment including financial subsidies, relief funds and the like. The governments of the two counties included in the present study both fall within the latter category. Their power is built upon-apart from a monopoly of local economic resources-a broad range of bargaining strategies for dealing with the higher authorities who for a variety of reasons do not want to alienate the local governments. Not surprisingly, one government has more to bargain with than the other, owing to its "locality uniqueness in policy implementation," a common political phenomenon in China under the economic reform (Zweig 1992: 341). The locality uniqueness in this study (despite the two counties being under the jurisdiction of one prefecture) is their ethnic and poverty status. In both counties, the bargaining relationship with the state has fostered a local autonomy characterized by a greater or lesser degrees of freewheeling local leadership, and it has had highly significant economic consequences. The decisive role that the local government plays in the local economy is the focus of this chapter. It centers on a range of economic incentives that guide local cadres in the implementation of economic policy, and the extent to which the policy and its implementation interferes with peasant economic life. The chapter begins with an introduction to the government organizations as a framework to understand the bureaucratic relations. The position of each level of government in the bureaucratic system and the hierarchical interrelationships between them illustrate how bargaining relationships are formed. The county government in this study is placed as the first level of local administration. Its economic autonomy is reflected in the management of local finances and natural resources. The difference in ethnic and economic status in the two counties highlights the particularity of local autonomy.  The township government is at the interface between the state and the peasants. O n the one hand, it represents the state; and on the other, the local residents. The political and economic power of the township government is manifested in its administrative and redistributive functions. A s an agent of the township government, the village administration uniquely assumes a semi-governmental and semi-collective role. The change in the village administrative functions in the aftermath of decollectivization has impacted the peasants' attitude towards public welfare. The conflict between "public" and "individual" has complicated the peasant-state relations; and much of the situation is, to a great extent, related to the behavior of local cadres. The discussion of all these issues associated with local administration in this chapter is intended to provide a perspective which will facilitate a better understanding of the relation between the roles of the local government and economic change in rural society in a remote province like Yunnan.  Bureaucratic Structure of Government Administration  In the contemporary Chinese bureaucratic system, there are four levels of government (siji zhengfu), namely the center represented by the State Council, province (including municipalities directly under the central jurisdiction or zhixia shi), county (and its urban equivalent, city or shi) and township (and its urban equivalent, district or qu). Through this bureaucratic hierarchy, a high degree of central control is achieved. The bureaucratic system of government is illustrated in Figure 4 on the following page.  54  Figure 4. Levels of government administration  County  Township  © Xiaolin Guo, 1996  55 A s shown in the figure, each level of government-except for the center- is subordinate to the upper level and at the same time superior to the lower level. The two rhomboid signs in the chart indicate two special organs: the prefectural (diqu) and village administrations. Neither is defined as a level of government, but both are attached to, and included in the government administrative system. The prefectural administration is an intermediate unit between the provincial government and county governments. In the bureaucratic jargon, it is a "dispatched organ" (paichu jigou)  of  the provincial government. U p till 1975, the official title of the prefectural governor was "commissioner" (zhuanyuan), and the prefectural administration was called the "commissioner's district" (zhuanyuan qu or zhuanqu). The role of the commissioner (or prefectural governor) is that of an agent of the provincial government, and the function of this dispatched organ is to supervise the work of county governments on behalf of the provincial government. Yongsheng and Ninglang in this study are two of the four counties under the jurisdiction of the Lijiang prefectural government. Another dispatched organ shown in Figure 4 is the village administration. It is metaphorically described as a "leg" of the township government. The intermediate position of the village administration is between the township government and peasant households. It is officially called "administrative village" (xingzheng cun ) because its organization is not based on the naturally formed settlement. The administrative village is a collective organization since it functions as the headquarter of several subordinate hamlets. A t the same time, because it is a "leg" of the township government, the village administration may well be defined as a government agency. The village administration in contemporary rural China is, therefore, best described as a semi-governmental and semi-collective organization. The local administrations in this study are mainly at two levels: county and township. From the administration of the county government, we shall see how local political and economic autonomy are gained and maintained; on what basis the local government bargains with the state; and what impact the local autonomous administration has on the local economy.  56 Because the township government is positioned between the county government and village administrations, subordinate to the former while superior over the latter, it ends up playing a dual role: to the higher level of the government it represents the local peasants; and to the local peasants it represents the higher level of government (or the state). The dual role that the . township government plays inevitably involves conflicts and tradeoffs between state and local interests. In these two levels of government, we shall see many important aspects of statepeasant relations unfold.  Political and Economic Autonomy of County Government  In late imperial China, the county magistracy was the lowest level of state bureaucratic administration. The average population under each county jurisdiction in the 1800s was around 300,000. Because of the "law of avoidance" which prevented Chinese officials from serving in their own home province, the county magistrate was always an outsider; this rule of government contributed to the situation where a "thinly spread and weakly rooted state apparatus had a limited ability to penetrate local society and much of the governance fell to local elites operating outside the formal bureaucracy" (Esherick and Rankin 1990: 3). This situation began to change at the turn of the twentieth century when commercialization and industrialization gradually became salient along with the introduction of Western style education and the subsequent abolition of imperial examination systems (ibid). This change was followed, after the fall of the Qing dynasty in 1911, by a strengthening of state power by the Republican government especially in the 1930s referred to as "state making," and characterized by a deeper bureaucratic penetration initiated and rationalized by revenue extraction (Duara 1988). This indeed coincided with what was happening to the government administration in Yongsheng and Ninglang during the same time.  Throughout the M i n g and Qing dynasties, Yongsheng was a prefecture (zhou, orfu)? A t the beginning of the Republican era, Yongsheng became a county (xian) which had jurisdiction over Ninglang. In the mid 1930s, Yongsheng and Ninglang were separated and became two county administrations: Yongsheng county (xian) and Ninglang Administration Bureau (shezhiju), respectively under the jurisdiction of the Yunnan provincial government (Yongsheng Xianzhi 1989; Ninglang Yizu Zizhixian Zhi 1993). Following the establishment of the Ninglang Administration Bureau, the provincial government enforced control over taxation. According to the official history of Ninglang, two years after the Ninglang Administrative Bureau was established, in 1938, the provincial government started to issue annual tax assignments to Ninglang. Since 1950, these two counties have been administered by the Lijiang prefectural government. Today, the Yongsheng county government administers eighteen townships and a population of 335,000 (1991), the majority of which is Han. Ninglang county has a population of 204,403 (1991) and its government administrates sixteen townships. Because the majority of its population is Y i , Ninglang is a Y i autonomous county. In addition, it has remained a "state designated poverty-stricken county" (guojia ji pinkun xian) since 1986. In the past decade of economic reforms, the county governments in both Yongsheng and Ninglang have gained much juridical power. Due to differences in the ethnic composition of the population and economic state of affairs, the two county governments employ different administrative strategies in their political and economic leadership. The local autonomy and distinct functions of the county government are manifest in three crucial aspects of administration, namely local leadership, local management of finance and monopoly of local resources.  ' D u e to the changes in jurisdiction, the organization of zhou and/u switched back and forth. In the M i n g and the early Qing dynasties, it was Beisheng zhou. Around the end of 17th century, it became Y o n g b e i / u ; and later became Yongbei zhiliting (Yongsheng Xianzhi 1989).  58 Ethnic Leadership and Local Autonomy The Y i population comprises about sixty per cent of the total population in Ninglang. In the county capital, the Y i language is commonly spoken, or at least understood, by many peoples of different ethnicities. Mandarin Chinese, locally called Hanhua, is only the second commonly spoken language, but it is the official language, i.e. the language used in government documents and 'school textbooks. Most of the county leaders are locals. A s the county officials estimate, sixty per cent of the middle- and high-ranking cadres of the county government are Y i . The county magistrate is Y i , and many government bureau chiefs are Y i . The ethnic composition is the core of the legitimacy of the local administration. This is a lesson learned by the central government from a period and an event in contemporary Ninglang history known as the "Cold Mountain rebellion" (Liangshan panluari). The rebellion was mainly led by the Y i . It began in the Greater Cold Mountains in neighboring Sichuan province, and soon spread to the Lesser Cold Mountain areas in Ninglang county, and further to the Y i and L i s u areas on the Yongsheng border. The rebellion in Ninglang broke out in A p r i l of 1956 when the land reform was just under way. It was an almost simultaneous uprising in several townships. The political message of the rebellion was to let the locals rule the locals. The rebellion was put down by thousands of troops sent from the Kunming Military Region, Yunnan Military District and Lijiang Military Sub-District plus hundreds of local militia from neighboring counties (Yongsheng Xianzhi 1989). In the C o l d Mountain areas, many rebels took refuge in the mountains. In Yongning, some rebelling Lamas tried to escape to Tibet, and a few made it while the rest were arrested. According to the locals in Yongning, the current head of Yongning Lamasery (the younger brother of the late Tusi) was shot in the leg on horse back when fleeing from Yongning. It took about two years before the situation in Ninglang finally calmed down in 1958. It is noteworthy that it was not a coincidence that Ninglang was granted the title of Y i autonomous county by the State Council in September 1956, five months after the rebellion broke out. It was also not coincidental that the magistrate of the newly established Y i autonomous county  was a Y i , while his three predecessors from 1950 to 1956 had been N a x i and Han. This situation changed again during the Cultural Revolution when turmoil swept the whole nation. From 1967 when military control was imposed to 1980 when the Revolutionary Committee was dissolved and the county government was reinstated, the county was headed by Han officials. Since 1980, the ethnic status of the county magistrates as well as the first county party secretaries has remained Y i . (Ninglang Yizu Zizhixian Zhi 1993) The importance of the ethnic identity of the local leadership lies in enabling the officials to act as the legitimate representatives of the locals with whom they share culture, language and religion. However, all levels of local governments are, in theory, agencies of the state. In this respect, the ethnic status of the local leadership creates a special relationship between the local government and the state. This special relationship involves many sensitive political issues which often have economic bearings. It is, in short, a bargaining relationship between the local government and state, and between the local government and the local population. This bargaining relationship also exists in non-ethnic autonomous counties such as Yongsheng where the economic situation, though better than the average in the prefecture, is still dependent on state support. But the mechanisms of the bargaining relationship work differently in the two counties. One important aspect of the differences in the two government administrations is manifested in the management of local finance.  Financial Management: A Crucial Aspect of Local Autonomy Management of local finance is an important aspect of the government administration. Not only does it mirror the economic situation of the region, it also reflects the power of the local government in political bargaining and the latitude of the local government in manipulating economic resources. The power of local governments in finance management may be attributed to the fiscal reforms of the 1980s. The fiscal policy of the central government in 1980 was formulated as "dividing revenue and expenditure" (huafen shouzhi) and "target assignment by level" (fenji  baogan). This policy is metaphorically described as "eating from separate pots" (fenzao chifari). In this system, the revenues and expenditures of the central government and all levels of local government were fixed on the basis of the 1979 budget. The target assignments defined the amount of revenue to be turned over to the central government, the portion to be retained by the local and the figure to be subsidized. In 1985, the fiscal system was redefined as "dividing tax types" (huafen shuizhong), "settling baselines for revenue and expenditure" (heding  shouzhi),  and "target assignment by level" (fenji baogan). This system is a continuity of the 1980 policy and further specifies the fixed amount of the central and local revenues and the fixed amount of the revenue shared between the central and local governments (Caizhengbu Difang Yusuansi 1992). Both of the reforms are intended to regulate expenditure and stabilize revenues at all levels o f governments. The mechanisms of the revenue and expenditure regulations designated by the fiscal reforms were intended to work to the advantage of both central and local governments. The revenue and expenditure responsibility system termed "target assignment by level" has eased the financial burden of the central government; and by allowing the local governments to retain the excess-quota income, it has created incentives for the local governments. In the revenue sharing agreement between the central and local governments, the share of the overfulfilled revenue quota is retained by the local government as an extrabudgetary fund, and the use of the extrabudgetary fund is within the power of the local government. The 2  overfulfilled revenue quota is divided between the higher and lower levels of government by ratios of 30 to 70, 40 to 60 or 50 to 50. In Yongsheng, for instance, the percentage of revenue shared between the county and township governments is 50 to 50. This system is intended by the central government to stimulate the local government to generate revenues.  Extrabudgetary fund (yusuanwai zijin) is, by definition, a fund that exists outside the state budgetary plan, which is managed and disposed by the local finance. It is composed of various tax surcharges, revenue shares, special appropriation funds, and other administrative incomes. See also Jean O i (19992: 105). 2  61 In the poor areas that suffer from chronic budget deficits, the financial arrangements between the central and local governments are different. In these areas, both budgetary and extrabudgetary revenues often depend on subsidies from the central government. Yunnan is one of the "eight minority provinces/regions" (including five ethnic minority autonomous regions and three provinces with a large ethnic minority population) that enjoy favorable treatment from the central government in economic development. In the fiscal reform 3  of 1980, the policy of "dividing revenue and expenditure" and "target assignment by level" granted fixed amounts of subsidy to these areas with an annual increase of 10 per cent. A s 4  estimated, during the 12 years between 1980 and 1991, the financial subsidy that the central government issued to the eight minority provinces/regions amounted to 120 billion Yuan, or on average 10 billion Yuan a year (Caizhengbu Difang Yusuansi 1992: 8). Most of these areas are categorized as "old-minor-peripheral-poor" (lao shao bian qiong) referring to old revolutionary bases, ethnic minority areas, remote regions and poverty-stricken areas. The central financial subsidies in these areas embrace moral intention. To the so-called old revolutionary base areas, the central government is indebted for their contribution to the revolution in the past; to the ethnic minority regions, the central government is compelled to provide support for their symbolic role in the nation's unity; and to the poverty-stricken areas, the central government is obliged to provide help for they are lagging behind the national economy for a variety of reasons. The aim of the financial subsidies is twofold: a long term interest in economic recovery and a short term guarantee of regional stability. Yunnan has a reputation for "subsistence finance" (chifan caizheng or "bread and butter finance"). This refers to a situation where local revenue can barely sustain the administrative body (paying salaries and expenses for the maintenance of bureaucratic organs), but lacks the.  T h e five ethnic minority autonomous regions are Inner Mongolia, Xinjiang (Uighur), Guangxi (Zhuang), Ningxia (Hui) and Tibet; the three provinces are Yunnan, Guizhou and Qinghai.  3  T h e 10 per cent annual increase was discontinued in 1987 due to the financial strain on the central government. Instead, the amount of subsidy for each area was fixed for five years.  4  strength to support economic construction. In fact, 102 out of 127 county-level administrations (eighty per cent of the total) in Yunnan are in financial deficit. This means that even the "bread 5  and butter" needs to be subsidized by the central government. Both Yongsheng and Ninglang are in the category of the 102 finance deficit counties. Among the four counties in Lijiang prefecture, Yongsheng county has the highest output value and per capita income, and Ninglang is the poorest. B y the special financial arrangement for the eight minority provinces/regions, each of the two counties receives a fixed subsidy of over 10 million Yuan a year since 1984-85. Because Ninglang is an ethnic minority autonomous county and at the same time a state designated poverty-stricken county, Ninglang is i n a more favorable position to receive more financial subsidy. According to the Ninglang county history, from 1953 to 1989, the state subsidy amounted to 71.39 per cent of the total revenue income of the county (Ninglang Yizu Zizhixian Zhi 1993: 426). Given the population difference, the discrepancy in local revenue per capita, state subsidy per capita and government expenditure per capita between the two counties is remarkable. O n the basis of the sources available, a comparison of county finances is shown in Figure 5. W h i l e the local revenue does not show much discrepancy, which would suggest that the economic situation in both counties in 1985 was almost equally poor, there is a large discrepancy in state subsidy per capita and consequently in expenditure per capita. A s recorded, the state subsidy received by Ninglang county in 1989 exceeded 20 million (Ninglang Yizu Zizhixian Zhi 1993: 427).  It should be noted that the actual financial situation in the 102 counties may not be consistent with the official definition, as the county governments have considerable leeways to manipulate statistics for the purpose of receiving state financial support. 5  63  Figure 5. Comparison of finance situation between Yongsheng and Ninglang Yongsheng county  Ninglang county  Population  331,445  180,135  Local revenue Per capita (Yuan)  6,597,000 19.90  3,258,000 18.08  State subsidy Per capita (Yuan)  11,438,000 34.50  12,031,000 66.78  Total expenditure Per capita (Yuan)  15,805,000 47.68  14,069,000 78.10  S O U R C E S : Yongsheng Xian Zhi 1989; Ninglang Yizu Zizhixian Zhi 1993.  The ethnic status of the population earned Ninglang its title of autonomous county. The ethnic autonomy of the Ninglang county government today exists not only in name as many outsiders would assume. In terms of financial management, the county government of Ninglang certainly enjoys tremendous autonomy compared to Yongsheng. B y bureaucratic arrangement, the county bureau of finance is subject to the dual leadership of the county government and the prefectural/provincial bureau/department of finance. To the former, its relation is administrative; and to the latter it is operational. The administrative leadership is basically personnel control. Though the county bureau of finance is subject to the administrative orders of the county government, the operational discipline controls the actual financial management such as drawing up budgets and appropriating funds. In the operational system, the county bureau of finance is supervised by the prefectural bureau of finance on behalf of the provincial department of finance. Although the county bureau of finance in Ninglang is subject to the dual leadership of the county government and prefectural/provincial bureau/department of finance, it operates in a different way. The difference is reflected in a weak operational link which means that the higher-level financial authorities (prefecture and province) have less control over its actual  64 operation. In other words, the bureau of finance in Ninglang is more responsible to the county magistrate and his leadership, than to the prefectural and provincial superiors in the financial operational system. This particular situation makes the bureau of finance in Ninglang less powerful than its counterpart in Yongsheng in terms of appropriation of funds and expenditure control. In fact, its role is no more than the county government's cashier, which means that the bureau of finance appropriates funds according to the orders of the county leaders, regardless of the actual financial situation. The weak financial system in Ninglang is also attributable to the county's economic situation. That is, there is little revenue to be extracted, and therefore, there is not much for the bureau of finance to control. According to the county history and my interviews with the county finance officials, between 1979 and 1984, a series of deductions in agricultural taxes and state grain procurement quota was granted to Ninglang. Since 1985, all townships of Ninglang have been exempted from agricultural taxes on grain and animal slaughtering by the provincial Department of Finance. This situation is likely to remain for as long as Ninglang retains its 6  status of poverty-stricken county. The combined status of an ethnic autonomous county and state poverty-stricken county have provided financial advantages to Ninglang. The gap between its revenue income and expenditure is covered by various state subsidies and earmarked funds which are issued through the provincial Department of Finance. In addition, the status of poverty-stricken county also entitles Ninglang to a series of state loans on favorable terms (i.e. central-finance subsidized low-interest loans) which are issued by the Bank of Agriculture, a special agency through which much of the state poverty alleviation funds are channelled. This special loan is called "central interest subsidized loan" (zhongyang tiexi daikuari). The monthly interest of the  I n technical terms, agricultural taxes consist of three major items: grain tax, animal slaughtering tax and property transaction (e.g. house) tax. In the tax exemption, property transaction tax was not specified. According to the township finance office, the property transaction tax has never been enforced in Yongning. 6  65 loans set aside by the Ministry of Finance is 0.74 per cent in which 0.5 per cent is subsidized by the central finance. According to the county government's poverty-alleviation office, between 1986 (when the state poverty alleviation program started) and 1992, Ninglang county received a total of 18.18 million Yuan (on average 2.6 million Yuan a year) in interestsubsidized loans.  7  Because the finances of Ninglang are under the autonomous administration of the county government, the management of state subsidiary funds and loans is totally in the hands of the county government administration. This means, once the funds are available, it is up to the county leaders to dispose of them in whatever area and in whatever way. In the past few years, much criticism has been directed at the state support to the poor areas and many questions have been raised about the effect of poverty alleviation ( e.g. Zhou 1992; Park and Rozelle 1994; Larsen 1992; Jiushi Niandai de Fupin Zhanlue 1993). However, the problems involved are many-fold and the explanations are complex. Many say that the poorer the county, the richer the county government, which may not be an over-statemertt. It is common knowledge in Lijiang prefecture that Ninglang (the only state poverty-stricken county in the prefecture) has more and fancier official motor vehicles than any of the three counties in the prefecture. The means at the disposal of the local government in poor areas deserve attention in studies of local economy. A s the county government of Ninglang enjoys large amounts of state subsidies in various forms, one may expect a lack of administrative momentum in generating revenue since  O n e of the major projects of poverty-alleviation in Ninglang is the much advocated "Project 3358." It was launched in 1987 and is part of the county's 7th Five Year Plan (1986-1990). It aims at utilizing the county's unexploited mountain areas. The four digits of the project designation respectively stand for the development of 30,000 mu of plum trees, 30,000 mu of pepper trees, 50,000 mu of apple trees and 80,000 raw of alfalfa. The centrally subsidized loans have been used primarily to subsidize orchard construction, nurseries (saplings are provided to peasants for free), and training of technicians. It is still too early to tell how successful the project will be since it will take a few years before the trees bear fruit. Also, because the 190,000 mu of orchards are contracted to individual peasant households, it may be difficult to estimate the actual return on the government investment. 7  deficits can be set off by subsidies and even loans demand no rush repayment. Yongsheng 8  county, on the other hand, enjoys less favorable treatment because it is neither an ethnic minority autonomous county nor a poverty-stricken county. The county leaders of Yongsheng have mixed feelings about their economic status. O n the one hand, they are proud of being the number one in the prefecture with its highest output value and per capita income; and on the other, they quite envy the favorable treatment accorded their neighbor Ninglang by the state. They think i f Yongsheng were a poverty-stricken county, they would receive more state subsidies and their economic construction would be much easier. Though Yongsheng finances have also been substantially subsidized by the provincial government, the financial arrangement between county and province (the policy of setting the baseline for revenue and expenditure, and target assignment) does not permit Yongsheng to rely completely on the state support. Rather, it is being forced to solve, at least partially, its financial problems by generating more revenue on its own. Since the latitude that Yongsheng has to bargain with the state for subsidies is limited, the county government resorts to the promotion of a cash economy. The urge to generate higher revenue has made the county administration so desperate that it had to exert coercive power. The development of cash economy in order to raise government revenue took the form of tobacco production, which created conflicts between the local government and peasants.  Monopoly of Resources Due to the differences in government administration and finance management in the two counties, each government administration employs different economic strategies to its own  A t the time of my interview with the official of the county government office of poverty alleviation, the repayment of the first loans issued in 1987 was not yet due. The official made it quite clear that his duty is to approve projects and distribute funds, while collecting payments is the responsibility of the Bank of Agriculture. 8  67 advantage. The different strategies involve different measures of resource monopoly which have had different economic consequences for the local communities. Forests are among the richest natural resources in Ninglang. Seventy-six per cent of the total area of the county is forest which contrasts sharply to the arable land area amounting to only five per cent. The industrial exploitation of forest resources began in the mid 1960s. A t the time, two Forest Industry Bureaus were set up in Ninglang, both subject to the leadership of the provincial government. In succession, seven logging centers were established under the management of the two bureaus (Ninglang Yizu Zizhixian Zhi 1993: 330). For nearly twenty years, logging was under the monopoly of the provincial government. In 1982, the provincial government transferred the management of the logging industry to the Ninglang county government, and at the same time the state monopoly on timber sales was abolished. In 1986, the two Forest Industry Bureaus were dissolved. The forest resources and industry today are in control of the county government. Logging is the most profitable industry in Ninglang. The annual volume of logging in the 1980s was around 180, 000 to 200,000 cubic meters, and most of the timber-about 50 to 70 per cent-was sold outside the county, mainly to Sichuan, Hubei, Henan, Shandong, Zhejiang and Shaanxi provinces. Part of the timber is sold for cash and part in exchange for materials like steel, cement, plastic film, diesel and grain. The revenue from the logging industry and other forest products contributes seventy per cent of the county's total revenue income (ibid.: 320). The autonomous administration of Ninglang county gives the local government considerable leeway to monopolize natural resources and manipulate revenue income. The county government has a special forest fund which exists outside the budget, called the "petty cash bank" (xiao jinku) though the amount of the fund is by no means "petty." The existence of this fund is public knowledge but its exact size is not known. This special forest fund operates under the direct control of the county government. The fund is accumulated on the profits made  from timber sales. Apart from the state owned forests which are under the direct control of the 9  county government, there are also collectively owned forest areas. Each year, the county forest fund committee issues commercial logging quotas to townships and v i l l a g e s . The purchase of 10  timber is monopolized by the county government, i.e. the forest fund committee. A s for how this forest fund is used, it is strictly the business of the county administration-partly appropriated for subsidizing township finances, partly for subsidizing residence construction (e.g. helping the Y i people settle down), partly for the construction of the county capital (including building and recreational establishments), and partly for subsidizing various administrative expenses, most notably feasts. The ethnic and poverty status of Ninglang also makes the economic development in the area assume priority. The year after the former provincial Forest Industry Bureaus were dissolved, the Ninglang county government proposed to set up a pulp mill to utilize its abundant forest resources and to create employment for some 900 persons. The proposal was approved by the provincial government in 1991. The total capital investment in the construction amounted to 150 million Yuan, some 60 per cent of which came from the state special allocation and poverty-alleviation fund, and the rest from loans of different channels.  11  Ever since Ninglang  It mainly consists of the division of revenue on timber sale. U p till 1994, 10 per cent of the profit was retained by the county's special forest fund. After that, the percentage retained by the county government was raised to 15. 9  Y o n g n i n g is not a logging center, and logging quotas are not as high as in the mountainous areas. In 1991, the village where I conducted fieldwork received a logging quota of 192 cubic meter from the Yongning township. The village administration office redistributed the quota to the 18 hamlets; then the hamlet leaders divided the quota and assigned it to the households. Each household with the quota assignment needed to obtain a logging permit stamped by the township Forest Control Station from the village administration office. According to the village book keeper, as long as the peasants have the logging permit, the actual felling volume is not strictly controlled. In 1991, the actual volume of timber cut was between 500-600 cubic metertwice higher than the quota allowed; and in 1992, the actual volume of timber cut was three times higher than the previous year. According to the township cadre who is in charge of logging quotas in the township, the government purchases timber from peasants at a price of 100 Yuan per cubic meter, and sells outside the county for 360 Yuan. 10  1  ' T h e construction of the pulp mill was completed in 1995.  made the decision to set up the pulp mill, the neighboring counties Yongsheng and Huaping have complained for fear of the water contamination in the rivers flowing down through their paddy fields. But the provincial government has chosen not to interfere mainly for two reasons: (1) Ninglang is an ethnic minority autonomous county; and (2) Ninglang is a state povertystricken county. Fro these two reasons, i f the provincial government were to interfere, it would be under pressure to cough up a large amount of fund to compensate the loss of opportunity for Ninglang to "get rich." In Yongsheng, forest resources do not constitute an economic advantage, but the county has fertile arable land. Grain crops in Yongsheng have the highest yield in Lijiang prefecture and a yield higher than in most places in Yunnan. But the price for grain products is relatively low. In order to increase its revenue, the county government cast its eyes on a theoretically high profit product: tobacco. Tobacco is a major cash crop in Yunnan. The climate in many places is suitable for tobacco cultivation. In 1991, the tobacco sown area in the province amounted to 46 per cent of the total cash crop area (Yunnan Tongji Nianjian 1992). Tobacco production makes a significant contribution to the provincial economy, as there is no other agricultural product that generates more profits. Cigarette manufacture is the most profitable industry in Yunnan. The output value of the tobacco manufacturing sector in 1991 was over 10 billion Yuan which made up thirty per cent of the Gross Product Value of Industry in the province (Yunnan Tongji Nianjian 1992). Beginning in the 1980s, the provincial government of Yunnan offered preferential conditions for new tobacco-growing counties by suspending provincial or central taxes during the first five years of production. For the county government, tobacco production is profitable since a 38 per cent sales tax can be extracted (which is much higher than agricultural taxes and any other taxes). Tobacco sale is monopolized by the county government through its purchasing agencies in the townships. There is no tobacco market outside the government channels. This monopoly guarantees the control of profits by the county government and eliminates commercial competition.  70 Eleven out of eighteen townships in Yongsheng were assigned tobacco growing areas in 1991 and 1992. Only the townships in the mountainous areas were spared for a lack of the necessary conditions such as warm temperature and a secure grain base. In order to encourage township governments to implement the tobacco production policy, the county government provided economic incentives by giving target revenue assignments, and allowing for the overfulfilled revenue to be shared between the county and township governments. In this revenue-sharing system, the county government receives 30 per cent while the township government retains 70 per cent. The county government's decision to promote tobacco production, at first, did not appear irrational considering the history of tobacco cultivation in Yunnan and the presence of the necessary conditions in Yongsheng: climate, irrigation, trained technicians and so on. However, it became problematic in practice. To the individual cultivators, tobacco cultivation runs counter to their economic interests (see Chapter 5). Encountering reluctant response from the peasants, the county government decided to make tobacco production compulsory. It started in 1991. Many peasants resisted the coercive order passively, and some expressed their defiance explicitly. A t the end of the year, several peasants in Jinguan delivered a petition to the provincial government protesting the coercive order of tobacco production imposed by the county and township governments and claiming their rights to choose what to grow on the land contracted to their households. The provincial government did not want to get involved in the local conflict, and transferred the petition to the county government, demanding that a local settlement be reached. Since the peasants were under the magistracy who promoted tobacco production in the first place, the local settlement could not possibly have ended in their favor. However, the county government had no intention to magnify the matter, being well aware of the unpopularity of the tobacco policy. Eventually the case was settled quietly "out of court," and the incident did not have much impact on government policy. In 1992, when I was in Jinguan, the conflict between township cadres and peasants became more acute as the assignment for tobacco growing area was increased, and the township  government took formal coercion to enforce the tobacco policy (see Economic  Intervention  below). A year later, tobacco production ended in disaster after the autumn harvest. The disaster was initially caused by the poor harvest. Then, the harsh conditions set by the government agencies for purchasing cured tobacco added burden to the peasants and fueled their pent up discontent. Finally the peasants took violent action towards local cadres. A l l this was happening in the vicinity of the county capital, and the government was much shaken by the consequences of its policy. In 1994, the compulsory tobacco production was abolished. The development of tobacco production in Yongsheng is a product of the power expansion of local autonomy. The power is both economic and political. O n the economic level, the circumstances provide the county government with incentives to generate revenue. O n the political level, the local autonomy of the county government enables the administration to employ coercive means to exploit local resources to achieve its economic goals. In the past few years, there have been many spectacular incidents of peasant uprisings against excessive levies, delayed grain purchase payment, embezzled compensations and all kinds of bullying by officials reported in the media (see, for example, Wedeman 1996). M u c h of the blame has fallen on the central government. While the central government is responsible for many undesirable situations and outcomes of the economic reforms, some other explanations are certainly worth seeking at lower levels in the same bureaucratic system. W h y did peasant uprisings take place in some counties but not in others? W h y are peasants more exploited in one place but less in another? If we acknowledge the trend of decentralization in Chinese politics, we should recognize the considerable autonomy of the local governments that has been significantly expanded and strengthened in the past decade of economic reforms, and has had a tremendous impact on the development of rural economy. When we come to scrutinize these problems, we shall discover that, at least, part of the responsibility may very well lie with the local administrations.  Local Autonomy and the State: A Bargaining  Relationship  What factors have strengthened local autonomy? On what basis has the local autonomy of Yongsheng and Ninglang county governments been sustained? These issues concern diversities of politics at all levels of government, which create latitude for political bargaining. Recently, there have been notable studies on the spatial aspects of power between the central and provincial governments though mainly concentrated on the eastern coastal regions.  12  However, there have been comparatively few studies dealing with bargaining relationships between county governments and the state. Both Yongsheng and Ninglang are financial deficit counties. Once the local government has deficits, it relies the state for support, otherwise even the administrative body cannot be sustained. A s stated earlier, both Yongsheng and Ninglang receive about the same amount of fixed financial subsidy (around 10 million a year each) according to the special financial arrangement for the eight minority provinces/regions, despite the population in Ninglang being no more than 60 per cent of the population in Yongsheng. But this subsidy is far from sufficient to make ends meet in either county. So each county has to apply for additional subsidiary funds from the state. The amount of the additional funds is, however, not fixed. A s for how much each county can get depends on how well the local administration can bargain with the higher authorities. A s for how well each local administration can bargain, it depends on what chips it has to bring to the bargaining table. Ninglang's combined ethnic minority and state poverty-stricken status provides the county government with solid bargaining power and entitlement to state support. This is regarded by outsiders (people in Yongsheng, cadres in Lijiang and officials in the central institutions of poverty alleviation, whom I have interviewed) as the "advantages of being poor." Though it may sound blunt and unsympathetic, such a remark does nonetheless tackle the core  12  J a e H o Chung (1995) offers a useful summary and critique about studies in this field.  of the matter. B y comparison, Yongsheng has less of an advantage in bargaining as it is neither an ethnic minority nor a state poverty-stricken county. The latitude of bargaining, therefore, is comparatively restricted. The differences in bargaining by the two local governments have resulted the substantially large state subsidy flowing into Ninglang and administrative interferences with peasant production in Yongsheng. In the bargaining relationship, the state and local government have different goals and gains. Generally, the state subsidies are viewed as having more social effect (shehui xiaoyi) than economic effect (jingji xiaoyi). It is clear that the special efforts that the central government has made in recent years by providing billions of Yuan in subsidies to the "old-minorperipheral-poor" areas, is motivated by a desire to safeguard social stability in these areas. To some extent, the poor financial situation and slow economic development which are de facto problems of the local government have become the de jure problems of the state. In the bargaining relationship, the political interests of the state are realized in exchange for economic gains of the local governments. A t the end of the bargaining, the state retains the peace and the local governments gain the wealth. The political and economic behavior of local administrations is much related to the relationship between the state and the local government. The different relationships between the state and local governments determine the different degrees of political and economic autonomy in different localities. The different degrees of local autonomy, in turn, have much influence on economic polities in the area. The impact that monopoly of resources by the county government in both Yongsheng and Ninglang has had on the peasant economy can be observed further from the administration of township government which-situated between the county government and villages-carries out the policies from the higher authorities while managing its day-to-day agenda of administration work in the local community. The study of township government administration shows how the state reaches the grassroots society and how the peasants interact with the state on both practical and symbolic levels.  74  The Grassroots Leadership: Township Government  Townships are called either xiang or zhen in Chinese. B y official definition, the pre-condition for the establishment of zhen is a township with a total population less than 20,000 and an urban population in excess of 2,000; or a township with a population of over 20,000 of which the urban population comprises over 10 per cent (Zhongguo Tongji Nianjian 1993: 9 3 )  1 3  By  this definition, Jinguan is a zhen, and Yongning a xiang. The additional conditions that Jinguan possesses to qualify as a zhen, according to the township head, are being a market center and having access to fairly convenient public highways. Generally speaking, xiang seems more rural than zhen.  Evolution of Township Administration The official history of Yongsheng county indicates that the township, as a government administration, only emerged in the mid 1930s. It was first called qu and later changed into xiang and zhen (Yongsheng Xianzhi 1989: 39). This establishment of township administration as a sub-county government coincided with what has been called the "state making" and "state strengthening" of the Republican era in the wake of the disintegration of state bureaucracy in the late imperial time (Duara 1988). Since 1950, the official name of the township government has been subject to a series of changes, as has that of the village administration. Throughout the government documents and records of modern history, we find various official names used to designate township and village administration.  T h e official definition also indicates that exceptions apply to ethnic minority areas, sparsely populated remote and mountainous regions, small-scale mining-industrial areas, small port and tourist areas, etc.. where the urban population may be less than 2,000. 13  Figure 6. Rural administration in Yongsheng Township  Village*  1988-  xiang/zhen  cungong suo/banshi chu  1984-87  qu  xiang  1968-83  gongshe  dadui  1962-67  qu  xiao gongshe (or subcommune)  1960-61  gongshe  dadui  1950-59  qu  xingzheng cun  SOURCE: Yongsheng Xianzhi, 1989: 47-55. *Generally called xingzheng cun ("administration village") in contrast to ziran cun ("natural village" or hamlet).  Figure 7. Rural administration in Ninglang Township  Village  1987-  xiang/zhen  cungong suo/banshi chu  1984-86  qu  xiang  1969-83  gongshe  dadui  1962-68  qu  xiang  1958-61  gongshe  dadui  1950-57  qu  xingzheng cun  SOURCE: Ninglang Xianzhi, 1993: 49-53.  Figures 6-7 show how these official names have changed repeatedly in the two counties during different periods of time from 1950 to the present. Each time, the change of institutional names involved institutional rearrangement of merger or separation; and all changes have turned out to be related to the general political situation in China at different times:  • 1958 was the year when the People's Communes were established nation-wide, as well as in Yongsheng and Ninglang, though Yongsheng did not officially use the name of Commune until 1960. However, the official history of Yongsheng county does indicate that the three-tier accounting system characteristic of the communes was adopted in 1958, which suggests that there was indeed an institutional change. • 1962 was the year following the severe three-year famine caused by the Great Leap Forward. W i t h the intention to alleviate the burden of the peasantry and come to terms with the drop in productivity, the Party abandoned many of the key policies of the original Commune movement. In the autumn of 1962, the Party Center passed the "Regulations on the W o r k of the Rural People's Communes (Revised Draft)" (Zhongguo Nongye Dashiji, 1982: 1 0 6 ) .  14  Following various "liberalizing" decisions, the Communes were dissolved in some parts of China, especially in the poor and remote areas such as Yongsheng and Ninglang. The earlier qu and xiang administrative division was thus revived. The production teams were replaced by the earlier co-ops, and private plots were restored. In the remote regions, household farming was permitted. • 1968-1969 was the high point of the Cultural Revolution following the imposition of military control over parts of China. The People's Communes all over the country were encouraged by the national media to "Learn from Dazhai," a so-called model brigade in Shanxi province, which practised such radical policies as the abolition of private plots and extreme "self-reliance." B y 1969 at the latest, it would appear as i f all the qu and xiang administrations that had survived since 1962 had again been redesignated People's Communes which persisted throughout the 1970s. • 1982-1984, rural China was subject to administrative reform following the implementation of the household production responsibility system.  14  15  A t the end of 1983, the  I t is also known as the "Sixty Articles on Agriculture."  F r o m this period on, political changes are preceded by economic reform which is completely different from what had happened from the late 1940s to 1970s. 1 5  policy of the "Separation of Government Administration from Economic Management" (zhengshe fenkai) was formally put forward by the central government {Xinshiqi  Nongye,  1992: 220-2). Further institutional change came along in the years that followed. Eventually, Commune as the name for township administration was abandoned, and the qu administration was resumed. • 1986-1987 is an important period during which "Strengthening the Rural Grassroots Government Power" was brought forward on the C C P agenda (Xinshiqi Nongye, 1992: 4119). Following this, townships became xiang or zhen, and this seems to have concluded the institutional change of the current economic reform so far. This is the time when the economic power of the local government began to expand.  Institutional Organizations and Roles of Township Cadres The present township government consists of a core leadership and two or three dozen officials. The core leadership is the Party secretary and his two deputies, the township chief (zhenzhang or xiangzhang) and his two (and sometimes three) deputies, the chair of the People's Congress (renda zhuxi), and the administration clerk (zhengfu mishu). Under this leadership, there are offices responsible for various affairs such as Party school, Party organization, youth league, women's work, civil administration, legal affairs, township enterprise management, local militia, land management, township construction management, accounting, statistics and general service. Each of these offices is usually held by one or two officials. The township government also administers several agencies dispatched by the county government, e.g. family planning, rural economic development, fishing (in Jinguan) and orchard development (in Yongning) agricultural technology and land administration. The number of the township personnel varies according to the size of the population under its jurisdiction. In Jinguan, the total number of township officials is around forty, while i n Yongning it is around thirty.  M a n y township government officials come from the same township, and some from other townships within the county. Township cadres from other counties are rare. A l l the township cadres-except for the employees under special contracts-are on the state pay-roll; have urban household registration (which entitles them to monthly commercial grain ration, but not contracted land); and enjoy state welfare benefits. The appointment of township cadres is based on two prerequisites: education and experience. Generally speaking, the township cadres in Jinguan have a higher level of education than the cadres in Yongning. Most of the young officials, especially those who hold specialized and technical positions (e.g. as agricultural technicians, accountants, statisticians) have their job assigned upon graduation from secondary technical schools or senior middle schools. The senior positions are held by middle-aged and old cadres whose appointments are based on experience and who are likely to be transferred from one township to another, or in the most successful cases, from the township to the county government. Needless to say, personal connections always play a role in the assignments of jobs. The ethnic identity of cadres is not important in Jinguan, but it is in Yongning where the majority of the population is Mosuo (see Ethnic Leadership and Local Autonomy earlier in this chapter). In Yongning, more than one third of the township cadres are Mosuo. A s a matter of fact, about 80 per cent of the township cadres in Yongning are a combination of Mosuo, Pumi, N a x i , Y i and Tibetan. The top leadership, i.e. the township head and the Party secretary are both Mosuo. The three township deputy heads are respectively Mosuo, Y i and Han; and the two Party secretary deputies are both P u m i .  1 6  The township cadres in Jinguan and in Yongning hold different attitudes towards the changes brought by economic reforms in the last decade. The basic ideological difference can be illustrated by the two heroes respected by the majority of the cadres in the two places: M a o  T h e personnel of the township leadership, mainly the party secretary, township chief and their deputies, are frequently changed. 1 6  79 Zedong in Jinguan and Deng Xiaoping in Yongning, the former representing tradition and the latter representing change. From my daily contacts with them in the fields, I observed that Jinguan cadres have mixed feelings about economic reforms-they praise the household responsibility system, but at the same time they are rather negative and ambiguous about other concurrent changes which pose threat to social and personal security. B y comparison, the township cadres in Yongning seem to enjoy every freedom and benefit that the economic reform has brought along to them; and are much more relaxed and pragmatic about what is happening to them and to life in general in their territory. In Jinguan, the administrative work of the township government is both year round and seasonal. B y year round, I mean that the cadres in general work seven days a w e e k .  17  M a n y of  them take regular days off once a month to see their families. B y seasonal, I mean that much of the administrative work is concentrated in the busy farming seasons. In Yongning, the administrative work proceeds in a different pattern. The three-floor office building is vacant most of the time with only three or four cadres (sometimes one or two who happen to have families living in the office compound) holding the fort. The majority of cadres are invisible on ordinary days and many of them only appear once a month briefly on the payday. The real busy working season for the township government in Yongning is in the winter around the end of year when all offices are required to submit statistics and annual reports.  18  Despite the significant institutional changes in the countryside brought by economic reforms, the township government remains functional. The basic functions of today's township government, in theory, can be generalized as follows: implementing policies, maintaining social  1 7  F r o m 1995, they began to have weekends off.  O n e of the economic advantages for the township cadres in Yongning in the reform era is the opportunity to be engaged in some kind of moonlighting business including pool houses, video-rentals, mushroom and herbal sales and other trades. Some of these businesses are run by their families (e.g. wives and brothers), but often the cadres themselves are directly involved in the businesses. Doubtlessly, the official positions held by the cadres permit their family business to gain access to important social and economic resources. 18  order, protecting public property, enforcing contracts, managing irrigation systems, administering state subsidies and relief funds, and supervising economic activities in general. The most important functions which have direct effect on peasant life are redistributive administration, irrigation management and economic intervention. Generally speaking, the cadres in Jinguan are more conservative and politically conscious of their roles as government cadres and more active in performing their duties. The cadres in Yongning, on the other hand, seem to have less administrative pressure imposed on them by the higher authorities and are therefore less inclined to interfere with the local life.  Redistributive  Administration  The township redistributive administration mainly involves social relief (shehui jiuji), disaster relief (zaiqing jiuji) and poverty alleviation (fupiri). The government institutions involved in these affairs are the civil administration (minzheng), and finance (caizheng) offices. The redistributive mechanisms work differently in Jinguan and Yongning. B y comparison, the poorer the township is, the more redistributive funds are available to be manipulated, and consequently the more powerful the local government becomes. Social relief consists of state-resold grain (fanxiao Hang), cash subsidies, and interestsubsidized loans (tiexi daikuan). These relief materials are aimed at supporting the individuals or individual households that live below the "subsistence level" (wenbao xian). In Jinguan, the township chief estimates that about ten per cent of the households in the township live below the subsistence level. In Yongning, according to the township Party secretary, about 6,000 people (or about thirty-five per cent of the total population) live below the subsistence level, and most of them are the mountainous or semi-mountainous dwellers. The reasons for living conditions being below the subsistence level are varied, but mainly the result of a shortage of labor force in the household, chronic illness of family members, or cyclical economic burdens. State-resold grain is part of the grain procured by the state from the peasants in the previous year which is resold to peasants whose annual output is not sufficient to meet the basic  food needs. The state-resold grain is usually a combination of corn, wheat and rice. The price is lower than both the state grain purchase price and the market price (on average about 40 per cent lower). For the households that do not have money to buy the state-resold grain, there is a small amount of cash subsidies, on average 30 Yuan per household. The distribution of state-resold grain is controlled by the township government and village administration. The list of households in need of relief is first drawn up by the village administration office, then submitted to the township government. Every year, the township government, based on the list provided by the village offices, applies to the county government for state-resold grain and corresponding subsidiary materials. Once the subsidies are granted, the township government redistributes them to the village administration offices which further allocate to the households. According to the township social relief officer in Jinguan, the subsidy materials arrive twice a year: in winter (or early spring) and in mid summer. The difference related to the situation of state-resold grain between Jinguan and Yongning is shown below in Figure 8.  Figure 8. State-resold grain in Jinguan and Yongning Jinguan township  Yongning township  Total state-resold grain received (kg)  17,835  190,385  Percentage of the total annual procurement  0.7  55  Total number of households in the township  8,165  3,006  Number of households benefiting from the stateresold grain  620  1,034  Total population of the township  34,402  16,875  Number of people benefiting from the state-resold grain  2,579  5,996  Average amount of stateresold grain per person (kg)  6.9  31.8  SOURCES: Jinguan Township Statistics 1991; Yongning Township Statistics 1991.  In the two townships, the ways of distributing the state-resold grain are different. In Jinguan, it is distributed to the households with special difficulties (tekun hu); in Yongning, in one Mosuo hamlet the state-resold grain is distributed evenly to every household and in another it is used to subsidize the households that have had funerals in the y e a r .  19  While the state-resold grain is intended as an immediate relief of subsistence difficulties, the state interest-subsidized loans are aimed to support rural production and development. In the mid 1980s, the township government in Jinguan started to issue centrally interest-subsidized loans to peasant households. The loans come from two channels: the civil administration office and finance office. The amount of loans controlled by the civil administration office is smaller than that controlled by the finance office. U p to 1991, the total amount of loans issued by the  19  T h e major Mosuo household expenditure is on feasts and gifts involved in funerals.  township civil administration office in Jinguan was some 3,000 Yuan. The amount given to each household was between 300 and 500 Yuan. The finance office of Jinguan township offered loans to the peasant households for the first time in 1985. The amount was 20,000 Yuan. The maximum amount of loan each household received was 200 Yuan and the minimum amount was 20 Yuan. For the loan issued by the township civil administration office, a 3 per cent administration fee is charged, and the repayment period is one to two years. According to the civil administration officer, most of the households have paid back the loans. N o interest is charged on loans issued by the township finance office, i f they are paid back within a period of three years. U p to 1992, eighty per cent of the loans had been paid back, and the rest was "hung up" waiting to be written off by the higher authorities. In Yongning, the township finance office started to issue loans in 1980. In the first four years, the total amount o f loans was 320,000 Yuan and no interest was charged. These loans were aimed at supporting peasant households developing sideline production, i.e. raising pigs and chickens. The loans were issued through the village administration offices. The average amount of loans each household received was between 30 and 500 Yuan. From 1987, the township finance office started to collect the loans issued in the early 1980s, and the result was meager. For this reason, a new regulation was imposed by the county government that the amount of the loans that the township finance office receives each year is linked to the repayment of the previous year. That is, the amount of the loan payment collected in the previous year is the amount of the loans to be issued in the next. Between 1987 and 1992, the amount of loans that the township Finance Office received each year from the county government and in turn, issued to the village administrations was 20,000 Yuan on average. This time, 0.24 per cent monthly interest was imposed. The loans were primarily issued to the households to develop small enterprises such as apple orchards. The average size of the loan received by each household rose from 500 Yuan to 3,000 Yuan. According to the cadres in both townships, the interest-subsidized loans issued in these years have, to some extent, served social purposes, but have so far had not much economic  effect. This is because the amount of the loans is too small to be used for any substantial investment i n production. Instead, many peasants just used the loans for temporary relief, e.g. buying some rice and other food stuff from the market, or other commodities. For this reason, the township finance office in Jinguan no longer issues such loans to peasant households. The finance office in Yongning complained that it is difficult to pursue loan payments for the reason that the money was spent but not invested. But, different from the finance office in Jinguan, it continues to issue such loans. Disaster relief is different from social relief. In social relief, the relief materials are provided on a regular basis (annually or monthly) and the beneficiaries are selected according to certain criteria. In the case of disaster relief, the relief funds are issued after natural disasters occur, and the beneficiaries are the households in the disaster-stricken area. The disaster relief consists of "tax exemption" and "subsidies for farmland restoration," both handled by the finance office. In 1989, the farmland of several hamlets in Wengpeng village was flooded. The township finance office in Jinguan issued relief funds to the peasant households to restore the land destroyed by pebbles and sand. For every mu of land, the compensation was 25 Yuan. In 1991, Wengpeng was struck by a hail storm which damaged the rice harvest. A t the end of the year, the government subsidized the peasants by paying for the grain extracted for agricultural tax: for every jin of grain, the peasants received a compensation of 0.25 Yuan. The amount of disaster relief that the government provides is not significant-if compared to social reliefbecause natural disasters are comparatively rare in the basin areas. Poverty alleviation is especially relevant to Yongning because it is under the jurisdiction of a state designated poverty-stricken county, Ninglang. Owing to this poverty-stricken county status, peasants in Yongning (as well as in other townships all over Ninglang county) have been exempted from the payment of agricultural tax, and the state grain procurement is no longer compulsory. Besides, every year Yongning receives various forms of agricultural aid 20  T h e exemption means that the entire output of the household is retained by the peasants. Before the exemption, the agricultural tax rate in Yongning was already much lower than in 2 0  from the government: subsidized grain seeds, nursery stock, plastic sheets and chemical fertilizer. In addition to the chemical fertilizer ration at subsidized price rewarded by the state for the fulfillment of grain procurement (see Chapters 5-6), the peasants in Yongning receive an extra 2 kg chemical fertilizer subsidy for every mu of contracted land plus 7.5 kg more for every mu of corn cultivated. Both of these additional rations are not available to the peasants 21  in Jinguan. The residence subsidy is also special in Ninglang. Since the m i d 1980s, Ninglang county government has provided subsidies for residence building in the rural areas. This residence-building subsidiary fund (minfang buzhu kuan) is issued under the county's policy of "Three Settlements" (san ding): "settlement of mind" (ding xin), "settlement of roots" (ding gen) and "settlement of residence" (ding ju). This policy aims to help, particularly the Y i , abandon their slash-burn agricultural tradition. Because most of the peasant houses in Ninglang are built out of timber and in order to save forests, the residence-building funds are especially used to subsidize the cost of bricks and other building materials like cement. The total amount of residence-building subsidiary funds issued through the Yongning township government from 1986 to 1990 was 252,931 Yuan which benefitted 565 households. O n average, each household received 300-500 Yuan (the per capita income of 1991 was 418 Yuan). However, the number of households that benefitted from the residence-building funds was not even in the four hamlets under my household survey. In Pijiang hamlet (where the N a x i are concentrated), about 1.1 per cent of households received the subsidy; in Neiba hamlet  Jinguan. According to the Mosuo hamlet leaders, on average for every mu of land, the tax was 2-3 k g of grain, while in Jinguan, the average tax rate is 40-50 kg per mu . Originally, the exemption of agricultural tax in Ninglang was partly intended to reduce the state-resold grain, but the amount of the state-resold grain has, in fact, remained about more or less the same with a slight increase in the past years in Yongning (the township has the highest grain output in the county). Although the state grain procurement is no longer compulsory, the peasants in Yongning still choose to sell their grain surplus to the state granary for the price is guaranteed. and the grain sale is attached to rewarding rations of some production materials. b e c a u s e corn is a high yield crop, the chemical fertilizer subsidy is intended to encourage peasants to grow corn. 2  (Han hamlet), 1.4 per cent; in Gesawa (Mosuo hamlet), 27.5 per cent; and in Dapo (Mosuo hamlet), 40 per cent. The markedly higher percentage of beneficiaries in Dapo is not accidental. Neither is it because the residents there are more in need for help than others. Rather, the favorite treatment is a result of that one of the residents in the hamlet happens to be the most powerful person in the local administration. The redistributive administration has provided the township government much power to manipulate resources. The power of the township government plays a crucial role i n deciding who enjoys the benefits, and who gets more and who gets less, while negotiating with the higher authorities for relief materials. This redistributive function of the township government and the associated power is likely to persist as long as the area remains poor and as long as state subsidies are available to the administration. The socio-political and economic differences in these two areas have resulted in differences in the categories of state subsidy and differences in redistributive administration. They also have influence on the government behavior in the two townships in control of local resources, namely, irrigation, land and rural enterprises.  Irrigation Management The government control over irrigation has been described as characteristic of the "Oriental Despotism" (Wittfogel 1959). Despite the criticism of the "universal applicability of the irrigation hypothesis" by Julian Steward and others (Ulmen 1978), the importance of irrigation management to the government is obvious, as irrigation systems are referred to as "heavy water works" which "feed the ultimate agrarian producer one crucial auxiliary material: water;" and "fulfill important protective functions for the country at large" (Wittfogel 1959: 28). Water conservancy has long been regarded as the lifeline of agriculture in China. Irrigation construction after 1950 has been largely managed by the government by providing substantial funds, while labor force (known as "voluntary labor") and small funds (known as "collective funds") have been contributed by local communities.  A large scale construction of irrigation systems was launched during the 1950s and 1960s throughout China. A t that time, rural labor was organized by the Communes which made endeavors that required much concentration in manpower easy and possible. In Jinguan, there are three major irrigation facilities. One is the Dragon Lake dam built in the 1950s; the second is the Grand Northern Canal built in the 1960s; and the third one is the Grand United Canal built in the 1970s. When these projects were underway, up to 60 per cent of the able-bodied labor force was mobilized and engaged in the intensive construction work. The agriculture in the Three-River basin has since benefitted from them. The household production responsibility system has, in fact, made government control of irrigation in terms of operation and maintenance even more necessary. B y comparison, irrigation management work in Jinguan is more demanding than in Yongning. This is perhaps explained by the fact that the more intensive the agriculture is in an area, the more organized work is needed (in addition to the difference in general administrative performance of the township cadres). In Jinguan, one of the most important seasonal tasks of the township government is irrigation management. Before the transplanting season, the township cadres are mobilized to inspect the condition of major canals and to draw up irrigation schedules for each village. When the transplanting season begins (at the end of A p r i l and beginning of May), the cadres are organized in small teams and dispatched to the villages to supervise the irrigation process. The villages along the main water routes in particular require strong supervising teams. Each year, as reported, there are dozens of disputes over water between villages, hamlets and households, and the role played by the township cadres in policing the irrigation system is crucial. The coordinative roles that the township cadres and the village leaders assume in irrigation control may be described as "over-all organizers" and "on-spot disciplinarians" (ibid.: 26). The township government issue orders to the villages. When water passes through the village, the village director (cunzhang) takes charge of the main canals. Because the work is demanding, he usually employs a "water officer" (shuiguan) to assist. The water officer gives orders to the  team leaders who further guide the peasant households to water their land in due time. The cadres work on day and night shifts throughout the season. Conflicts that cannot be settled between households are mediated by village cadres; and problems that cannot be solved between villages are handled by the township officials. Interventions from the township-village cadres are generally effective. A s far as irrigation is concerned, there is only administrative management involved, but there is no control by any kin-based organizations, nor individual rights that may exert influence on irrigation management  22  The water station in Jinguan is a cooperative unit that manages the routine technical work and collects water fee. The water fee is 1.50 Yuan per 100 cubic meter, and the cost for watering one mu of land is 4-5 Yuan. In Yongning, there is no fee charged for irrigation. Each year, the state provides funds for maintenance of the irrigation system. The annual amount that Jinguan township receives for irrigation construction is between 300,000 to 400,000 Yuan. The huge cost is owing to the extensive irrigation system involving the frequent redirecting of courses and flood control in the major rivers flowing across the Three-River basin, in addition to the maintenance of reservoirs and a power station. The heavy investment of the state in the irrigation work in Jinguan is to protect the important grain production base. B y comparison, the scale of irrigation work in Yongning is far smaller, and there is no major construction for flood control on a regular basis. The annual fund from the state ear-marked for irrigation maintenance is around 10,000 to 15,000 Yuan which is mainly used for subsidizing hamlets in purchasing building materials and tools for maintaining ditches. The maintenance of major canals and other large construction, i f needed, are supported by other special appropriated funds. The actual repair work of the water ways is organized by the village and hamlet leaders. Since land was contracted to the households, organization of water way repairs has become more difficult than under the collective system. The problem is twofold. O n the one hand, the  T h e individual right over land and water as discussed by E . Leach (1968) is not relevant here because in China only land is contracted to the individual peasant household while the whole irrigation system remains under public control. 2 2  village collective has become less effective in manipulating the labor force since production is no longer organized by the collective; and on the other, since land was contracted to the peasant households, peasants have showed less interest in collective work and public activities (see Village Administration: the Semi-Governmental Organization). While the administrative intervention in irrigation management by the township government seems to be necessary under the circumstances and has a positive effect on the organization of production in the village, some other forms of resource management in the two townships show a rather more controversial aspect of township administration.  Economic  Intervention  The main issues of economic intervention by township government discussed here concern ownership and government jurisdiction. Under the economic reform, the development of two forms of property have made the control of resources, e.g. rural enterprises and land contracts, more complicated. The administrative strategies of the two township governments in Jinguan and Yongning in the management of resources reveal different political motives which are related to ethnic and poverty situation and have had different economic impact on the peasant economy in the area. Unlike other parts of China, neither Jinguan nor Yongning has significantly developed township enterprises (xiangzhen qiye). In the two townships, the governments have different attitude towards, and apply different definitions to township enterprises. The attribution of ownership and the management of township enterprises by the township government illustrate from yet another aspect how the local leadership can manipulate economic resources, and how they serve the political careers of the local cadres. In statistical terms, Jinguan has five "township enterprises." They are the construction company, the Qinghe power station, the machine repair shop, the garment manufacturer and the zinc melting workshop. The construction company is a privately managed business with scores of employees. The power station is a collectively run enterprise with a dozen employees. The  machinery repair shop and the garment manufacturer used to be two cooperative organizations that employed scores of workers who now work on an individual basis. The zinc smelting workshop is jointly organized by seven households. None of these enterprises has anything to do with the township government in business management, but each has somehow become involved with the township government in one way or another. The arrangement between the construction company and the township government in designating the business as "township enterprise" is that the company pays an annual "administration fee" (guanlifei),  1,000 Yuan to the township government. In exchange, the  township government allows the company to use the official title of the administration (Jinguan township government) to sign contracts. This gives the construction company access to the official letter-head, which is believed to have favorable effect on winning contracts. Other than the administration fee to the township government, the construction company does not even pay income tax, according to the township chief, because it is privately run, plus it is difficult to calculate and estimate the company's income because most of its business is contracted outside the township. The Qinghe Power Station has a similar arrangement with the township government. It also pays an annual 1,000 Yuan of administration fees. Because this enterprise is locally based, there is no use of official letter-head involved. The function of the township government involved in this collective enterprise is nothing more than official supervision. Following the implementation of the household production responsibility system, the machine repair shop and the garment manufacturer were dissolved. Though the workers are still engaged in machinery repair and garment manufacture, they are no longer collectively organized. But, the income of these individuals is still included under the category of township enterprises. Aside from this formal arrangement, there is no fee charged by the township government, nor any production supervision involved. The zinc smelting workshop consists of several furnaces built on public propertyabandoned farmland (which used to be a farm of urban school graduates in the early 1970s)  under the control of township government. The land lease was a lump sum payment of 1,500 Yuan which is a security deposit-in case the land cannot be restored when the zinc smelting workshop withdraws from the site. In addition, there is an annual charge of administration fee of 1,000 Yuan by the township government. The appropriation of the working premise which is public property apparently legitimizes the zinc smelting workshop as a "township enterprise." The logic of including these enterprises under different arrangements into "township enterprises" seems to be a pure administrative strategy. When asked what township enterprises they have, many of the the local cadres would respond: "We do not have any township enterprises by definition; and those we have are for the purpose of statistics." The "purpose of statistics" is partly related to the cadres responsibility system which is called "target administration responsibility system" (mubiao guanli zerenzhi). Under this system, the official performance of the township administration is measured by assignments or quotas.  23  A s the  government statistics act as a kind of formal control over the administrative work and the existing category such as "output value of township enterprises" serves as an index demonstrating economic accomplishment, the local administration is compelled to come up with figures to meet the administrative requirements. Different from the Jinguan township government, the Yongning township government does not seem to have the need to include township enterprises in statistics. In statistical terms, there is no "township enterprise" listed. However, there are two enterprises which are more qualified as "township enterprises" than the five in Jinguan. One of the enterprises is the yak farm first established in 1958. The livestock was originally confiscated from the former Tusi and Sipi (landlords). Under the ownership of the township government, the yak farm operated under a collective accounting system, though the  T h e responsibility system contains two categories of administrative achievements: ideological (sixiang) and economic (jingji). The full accomplishment of the tasks in each category is marked by 100 points. Bonuses are rewarded for high point and a salary deduction is forced upon low points. 2 3  yaks were kept and raised by individual households. The yak farm remained as the only township enterprise in statistics until 1980. When the household responsibility system was implemented, the township government contracted the total of 824 head of livestock to sixteen households living in the mountains. Under the contract, the township government still holds the ownership of the 824 head of yak, on which the township government each year collects 4 Yuan of administration fees per head. The contracts allow the contracted households to make and retain the profits from the reproduction of the livestock. A s estimated by the Party secretary of the township government, the size of the herd has already grown to some two thousand head. Another enterprise in Yongning which is not listed as a township enterprise in the government statistics but produces a significant income for the township government is the government guest house. It was built in 1990, partly on a special development fund and partly on the centrally subsidized loans acquired by the township government, for a total amount of 420,000 Yuan. The two-storey building has about two dozen rooms. Four of them are reserved for the township government's guests, and the rest are leased to a migrant from Sichuan who has turned the guest house into a public hotel. The lease specifies a monthly administration fee of 400 Yuan to be charged by the township government. Together with the guest house, the township government also leased its canteen which was built on a separate fund of 60,000 Yuan to the same entrepreneur who has turned it into a public restaurant. This lease specifies a monthly administration fee of 500 Yuan to be charged by the township government. These two leases alone bring the township government an extra income of more that 10,000 Yuan a year (which is about 5 per cent of the total budget of the township government). Comparing the "township enterprises" in Jinguan and Yongning, there is a difference. In Jinguan, the five enterprises claimed to be township enterprises are in fact privately and collectively organized businesses. The establishment of these businesses and the management of the production in these enterprises do not involve the township government, but by some special arrangements, they all became township enterprises. In Yongning, despite the yak farm  and guest house having been contracted to individuals, their assets are owned by the township government and the government is directly involved in the business contract. However, neither of the enterprises is claimed to be a township enterprise. This difference is a result of different government administrations in the two townships. "No industry, no richness" (wugong bufu) is a set phrase used by the local officials in both Jinguan and Yongning when explaining their economic situation. However, this phrase has different meanings when spoken by different government administrations. When the Jinguan cadres say "no industry, no wealth," they refer to developing local enterprises as the only way to get rich. However, the lack of resources, technology and market sets obstacles. Under these circumstances, to the cadres, to include the enterprises that are not really township enterprises as township enterprises is better than nothing-at least, it satisfies the needs of their political career. In a different context in Yongning, this administrative strategy seems not necessary. When asked about their township enterprises, the township leaders would simply reply that they do not have any. When asked what was the reason for not having township enterprises, I was told: "First of all, the quality of the people (ren de suzhi) here is low; and secondly, we do not have capital to invest." The argument that the slow economic development (especially in the ethnic minority areas) is related to the "quality of people" often conveyed in Chinese writings, has been much criticized by Western scholars as an expression of great Han chauvinism. But why do the ethnic minority cadres make exactly the same reference to their own people? O n another occasion, the township leader voiced his opinion of getting rich: "We are waiting for others to get rich first, they they can help us." This attitude suggests that (if not a matter of indifference to the local economic development), the situation of "no industry, no richness" (wugong bufu) is not really a problem for the government administration in Yongning. This is, no doubt, related to the availability of state support in the ethnic and poverty area. The differences of the government's involvement in township enterprises illustrate differences in political motivation and economic incentives for the local administration. These  differences are also observable in the government's interference with the peasant household production in the two townships. In Yongning, the township government generally does not step in the normal economic activities of the peasants. The township government in Jinguan, on the other hand, exerts much influence and has much tougher hand on the peasant livelihood. A n example of this is the implementation of tobacco production. Tobacco production was advocated by the county government in Yongsheng for the first time in 1985. In the first few years, it remained as a kind of experiment which was termed "free development" (ziyoufazhan).  Six years later, tobacco cultivation became the subject of a  coercive order from the county government. In some villages, the tobacco assignment was evenly distributed among the peasant households, and in others, the assignment was subcontracted to a number of households. In whichever arrangement, the village administration encountered a great deal of resistance from the peasants.  24  In order to accomplish the task given  them by the county government, the township cadres had no choice but to get directly involved in enforcing tobacco production. In 1992, soon after the rice transplanting season was over, the township government switched their central work to tobacco production. In order to avoid tension between village leaders and peasants, the township cadres descended in person to the villages to solve the toughest cases that the village administration had difficulties handling. The township cadres first persuaded the peasants themselves to pull out the rice or corn seedlings planted in the designated tobacco areas themselves, and to plant tobacco instead. But i f the peasants did not comply with the orders, the cadres would pull the seedlings out for them. The tobacco work teams were larger than the work teams organized for irrigation management. Each team was usually composed of one dozen or more cadres, sometimes accompanied by one or two policemen. The size of the team was to ensure that there was enough manpower to pull out the seedlings i f necessary; and to provide moral support and  24  R e a s o n s for this w i l l be discussed under peasant economic concepts in Chapter 5.  security for the cadres involved. In most cases, problems were solved without making much of a scene on either side, that is, the peasants gave in after one or two sessions of talk. But, there were a few serious incidents in which confrontations intensified to the point where there was vocal abuse and fights erupted. O n one occasion, a peasant was arrested by the township police accused of "interfering with the government's administrative work." Despite the efforts that the township government has made, of the 3,300 mu tobacco assignment in 1992 (about 10 per cent of the total arable land in Jinguan), only 2,500 mu were actually planted. Later in the summer of the year, the township cadres already found themselves having to cope with the tobacco assignment for 1993, the sown area of which was as twice high as the previous year. The tobacco production in 1993, however, ended in disaster. The bad harvest was caused primarily by the weather (drought in the early growing period and severe water-logging in the later period) and insect pests. On top of this, the government purchasing agents set harsh conditions on the grades of cured tobacco which reduced the peasant income considerably. This added anger to the peasants' frustration already caused by the crop loss in the bad harvest. Some peasants became so enraged that they simply dumped the rejected tobacco leaves on the road in protest. What happened in Jinguan was not as serious as in another township located in the vicinity of the county seat where 12,000 mu of tobacco was cultivated that year but only half was harvested, and the peasants' protest turned into an uprising. They went to the government office to protest; they hit and wounded the purchasingagents; they stopped the cars of the county officials and asked them to take the tobacco leaves to the provincial government; and the township chief there was forced to take refuge in the police station.  25  After these quite unexpected incidents, the Jinguan township government was forced to back off, and the county government leaders had to abandon the coercive order of tobacco cultivation. The change in local policy, however, coincided with the central promulgation of the  2 5  I learned about these incidents through my correspondence with people in Jinguan.  96 Agricultural L a w in 1993, wherein the contractor's decision-making power over production and management was clarified.  26  It has been assumed that the decollectivization in the Chinese countryside has significantly reduced the power of government administration on production. The argument has been made that: " A s long as peasants remain farmers, they do not need to submit themselves to the power of local officials" (Zweig 1992: 358). However, the implementation of tobacco production in Jinguan by the local government may suggest that this general assumption needs to be qualified. This situation, on the one hand, is an outcome of the specific pattern of local administration in areas associated with particular economic problems (e.g. finance deficit, insufficient state support); and on the other, there is a matter of the peasant position in society. Compared to the bargaining relationship existing between government administrations as discussed earlier, while each level of government has something with which it can bargain in its relation with the other, the peasants at the grassroots level have little to bargain with; and the likely choice for them is to obey orders from the local government on which they depend for state redistributive goods. The enforcement of tobacco production by the township government, however, would not have been achieved without the coordination of the village administration that demarcated land area for tobacco cultivation, assigned quotas to the peasant households, and assisted the township work teams to enforce disciplines. Compared to the township government, the village has more interactions with peasant households in day-today life due to its position. The  I n Agricultural L a w of 1993 specifically states: "Unless stated otherwise in the agricultural work contract, a contractor shall enjoy decision-making power over production and management, as well as the right to dispose of products and earn profits; it must also fulfil the contractual obligations" (Agricultural Law 1993: 17-8). In contrast, the earlier Land management L a w published in 1986 only states: "The collective or individuals under the land contract have obligations to protect and utilize the land according to the contractual agreement. The land contract and management right is protected by law" (Xinshiqi Nongye, Article 12, 1992: 533). 2 6  coexistence of governmental and collective functions of the village administration enable the village administration to play a unique role as a bridge between the state and peasants.  Village Administration: the Semi-Governmental Organization  The term "village" used here refers to what is known as xingzheng cun, or "administration village." Because Jinguan is a zhen, the village administration offices in the township are called banshi chu, and because Yongning is a xiang, the village administration offices in the township are called cungong suo(see Figures 6-7). Each village office administers several hamlets (a subvillage unit, also known as "natural village" or ziran cun which is officially called "villagers' committee" or cunmin weiyuanhui)?-  1  brigade,  28  The village administration in the collective era was called  and most of the hamlets formed production teams. More than ten years after the  dissolution of the People's Communes, peasants in the two townships today still call the village administration office "brigade" (dadui) and refer to the hamlet as "team" (xiaodui) or alternatively, "small village" (xiaocun). This appears to be just a matter of habit. In Jinguan, Cuihu village which has 881 households and 3,598 residents consists of four hamlets: Guguan, Ruiguan, Bashang, Longtan. They used to be organized in five production teams (Ruiguan hamlet was divided into two), and are now organized in three villagers' committees (the two smaller hamlets: Bashang and Longtan are merged into one). Wengpeng village consists of five hamlets: Liiguan, Shangtian, Lujia Jie, Wengguan, X i n y i n g .  T h e organization of the villagers' committee is more or less on the basis of the hamlet. Sometimes, smaller hamlets may be emerged into one villagers' committee,while larger hamlets may be divided into two villagers committees. The same with the organization of production team in the collective time.  2 7  28  B e f o r e 1950, under the old baojia system, the village administration was called Bao gongshu.  They used to be organized in eleven production teams, and are now in seven villagers committees. Unlike in Jinguan, the hamlets in Yongning are much more dispersed. Kaiji village consists of eighteen hamlets. The l