NOTES

INTRODUCTION

Chapter epigraph from the 1982 Communist Party Document 19, cited in MacInnis 1989, 25–26.

1“Communist Party Document 19” is formally titled Guanyu woguo shehuizhuyi shiqu zongjiao wenti de jiben guandian he jiben zhengce (Basic viewpoints and policies toward the question of religion during the Socialist Period in our country) and contains important guidelines for Party members, including recommendations for facilitating the reconstruction of religious buildings and institutions and for educating clergy (see, e.g., Potter 2003, 319).

2In this book, I will try to maintain the distinction between “religion” and “(religious) beliefs,” with the former referring to social practices and institutions.

3In some areas of Muli and Ninglang, the pronunciation was nanjhi. I will use anji to designate the religious experts operating in my field site in Muli and hangui to designate their colleagues in Ninglang. The latter form has found its way into the Chinese ethnography of the Premi.

4“Bustling Township” and the other names on map 2 are fictitious names. I explain the use of fictitious names further on in the introduction.

5The term “Bön” has different connotations. Besides being the designation of the pre-Buddhist religion of Tibet, it is also the name of a religion that developed in Tibet in the tenth and eleventh centuries and has existed uninterruptedly until today alongside Buddhism with its own textual tradition and monastic system (see, e.g., Kværne 1996, 9–10; Samuel 1993, 10–12).

6The Lòlop’o is one of several ethnic groups or linguistic communities that have been classified by the Chinese state as Yi; this category also includes the Nuosu, who are close neighbors of the Premi (for further discussion, see chapter 2).

7This is of course not unique to China. See, for instance, Paul Dresch on the problems of conducting anthropological fieldwork in the Arab world. In many countries of the Middle East, the prying of the foreign anthropologist is associated with spying, the more so when he or she speaks Arabic well, yet if one does not speak Arabic well, one is not considered worthy of being told anything of any significance (Dresch 2000).

8In 2004, I interviewed the tulku, or incarnate lama, of Muli, Guzyo Pema Rinchin, and lived with a family in Bustling Township’s North Village; in 2006 and 2007, I twice visited Chicken Foot Village, met with central informants such as Nima Anji, interviewed monks at the monastery of Kulu in Muli, and made shorter trips to other Premi areas such as Jiulong County in Sichuan and Lanping Bai and Pumi Autonomous County in Yunnan.

9Between 1987 and 1995, I visited Premi villages in what was then Lijiang County (now Lijiang City and Yulong County), Songziyuan in Taian Township, Zhongluo and Zhonglu in Mingyin Township, Tonghailuo in Jiuhe Township, and Diantou in Ludian Township. During the same time period, I also visited the following Premi villages in Yunnan, all in Ninglang County: in Yongning Township, Tuoqi, Biqi, Bajia, Wadu, Gala (all in the vicinity of Yongning Town), Mudiqing, and Luoshui, and in Labai Township, Guludian, Tuokua, Banawa, and He’erdian. In addition, I visited Zuosuo Township, in neighboring Sichuan.

10Chinese government regulations require that foreign researchers be accompanied in the field by an assistant, usually a researcher or student from the Chinese institution responsible for the project. This assistant is often required to write a report on the activities of the foreign researcher.

11Muge Samten, a famous Tibetan scholar, argued in an article published in 1981 and translated by Janet Upton, that it would be absurd to consider the Premi of Muli anything else but Tibetans. He does so using an argument that is apparently self-explanatory, that is, by not calling them Premi or Pumi but Bod-mi, which means “Tibetan people.” Furthermore, there are even those “who are diligently planning to make Muli [County] a non-Tibetan [County]” (Upton 2000: 8).

1MULI

1The current official appellation “Muli” is probably an early Chinese rendering of “Mi-li,” one of the appellations of the area used in Premi and many Tibetan sources (rMi-li) before the Communist takeover (Rock 1947: 356; Kessler 1982–: 13). Another Tibetan appellation—used, for example, in the autobiography of the fifth Dalai Lama of 1677—is “rMu-le” (Ahmad 1970: 60). Today the official Chinese name is Muli, and official publications in the Tibetan language have adopted the name Mu-li.

2See especially Rock 1925 and Rock 1930. Besides Davies and Rock, only a few other Westerners had passed through Muli around the turn of the twentieth century: the Swedish missionary Edvard Amundsen passed through Muli in 1899 (Amundsen 1900), and the Austrian botanist Heinrich Handel-Mazetti—stuck in Southwest China during the First World War—made a short trip through Muli to collect plant samples (Handel-Mazzetti 1927).

3Chinese maps depicted Muli as officially part of Yanyuan County until the early 1950s.

4Muli was the first new territory added to the Chinese empire since the conquest of the Southwest by Kublai Khan in the thirteenth century, and in the Ming dynasty, it was under the nominal administration of the Jianchang “military station” (wei), with its center located at the present-day town of Xichang, capital of Liangshan Prefecture. Before the Mongol conquest, it was part of the territories of the Dali Kingdom, and before that, it was ruled by the Nanzhao Kingdom, which had taken over the area from the Tibetans (Muli Zangzu zizhixian zhi 1995: 113).

5The Official Chronicles of the Mu Clan (Mu shi huan pu) states that the eleventh Mu king, Mu Qin, conquered several of the villages in the Shuiluo River valley in the western part of Muli in the second half of the fifteenth century. His grandson, Mu Ding, added a considerable number of villages, also in Shuiluo (Rock 1947: 110, 114).

6The core of this area corresponds roughly to the present-day town of Lijiang and the counties of Yulong, Weixi, and Xianggelila (Shangri-la, called Zhongdian until 2001), all in northwestern Yunnan.

7Although this Bönpo monastery is newly built, it is likely that a monastery existed on the site before, since it is inconceivable that the Bönpo would have been allowed to establish themselves in the Mu territory at that time. Today there are no Bönpo in Muli, but there is still a small monastery with a few monks near the town of Zuosuo in Yunnan one or two stages (one stage is one day’s travel by mule or walking) from Liewa, and one with ten monks in nearby Ga’er in Jiulong County.

8It appears that the monastery already existed but had been a Bön center (Ganze zhou zhi 1997: 316).

9In Tibetan, bari (‘ba’ ri) (Muli chöchung 1993: 18, line 11). Rock has Bar-ri (1947: 357).

10Other sources on Muli argue that the tusi system started in 1648 with the rule of the first head lama (see, for example, Muli Zangzu zizhixian zhi 1995: 543). Nevertheless, History of the Dharma mentions that the father of the boy who would become the first head lama already held the golden seal of hereditary native chieftainship, a title locally called (w)onmatsang, translated in Chinese as guan shijia, “aristocratic family of officials” (Muli chöchung 1993: 4, line 17 [Chinese part]; 18, line 15 [Tibetan part]). The area controlled by the Bar clan at that time was situated around their manorial estate in the southeast of Muli and was considerably smaller (in that it did not, for example, include the Shuiluo River valley).

11Although the Qing designated only men and their paternal line as tusi, during the Ming there were quite a number of women tusi among the Yi in the Southwest (Herman 1997: 51).

12The presence of several Naxi villages in Muli today may be traced to this strategy, as could the ruins of watchtowers in the Shuiluo River valley and a few other places.

13In paintings, Mu Zeng is depicted as a monk, suggesting he was at least initiated as a Buddhist monk (Rock 1947: 161). He is also credited with establishing or rebuilding several of the monasteries and temples around Lijiang (ibid.: 206, 210).

14Although the Mu kings were strong supporters of the Karma Kagyu School from at least the end of the fifteenth century, the autobiography of the third Dalai Lama mentions that the king of Sa-tham (Lijiang) provided laborers and artisans for the construction of the Chokhorling monastery at Lithang and even invited the third Dalai Lama to come to Lijiang (Ahmad 1970: 59). At that time, large parts of Lithang were under the control of the Mu kings (Xu 1993: 265).

15Referred to in Shakabpa’s Tibet, which remarks that Tibetans at the Simla conference in 1913–14 used the records of this census to refute Chinese claims to these areas in eastern Kham (1967: 113).

16Data on the administration of Muli before 1949 are drawn from the History of the Dharma (Muli chöchung 1993: e.g., 56, 57) and the Annals of Muli Tibetan Autonomous County (Muli Zangzu zizhixian zhi 1995: 544–48).

17The Baiwu bazong was of Naxi origin, and his territory was situated in what is now Yanyuan County. See Harrell 2001: 134 and Muli Zangzu zizhixian zhi 1995: 948.

18See chapter 3 for a discussion of whether the position of baise corresponds to that of besé in Walnut Grove.

19The Wu Sangui Rebellion, also called the Rebellion of the Three Feudatories (San Fan Zhi Luan), took place from 1673 to 1681 in Southwest China. Wu Sangui, a Han general who had helped the Manchu defeat the Ming, rebelled in 1673 and joined forces with the remnants of the Ming in the south. The rebellion ended in 1681 with the fall of Yunnan Fu (Kunming) and the suicide of Wu Shifan, Wu Sangui’s grandson (see, e.g., Wakeman 1985).

20An extensive presentation of the available Tibetan and Chinese sources on communications between the fifth Dalai Lama and the Qing court regarding Wu Sangui is in Ahmad 1970: 205–29. These documents, including the Dalai Lama’s autobiography and the dynastic history of the Qing, seem to leave little doubt about the Dalai Lama’s hesitant attitude.

21One of the consequences of this change of policy was the annexation of Dartsendo (today Kangding), to the northeast of Muli, after closer examination of the border region revealed that this area had been administered by the Ming (ibid.: 225–29).

22Within the Qing military native chieftain system, anfushi is a mid-level rank, at 5b (in the eighteen-level system ranging from 1a to 9b) (Herman 1993: 26).

23Rock mistakenly transcribes the character as Hang (see, e.g., 1947: 357). From this point onward, I refer to Muli rulers only by the Chinese name form, although it is clear from the sound of the characters Chinese sources give for the names of the head lamas that the part of the name following the family name Xiang is undoubtedly Tibetan, as, for example, Xiang Songlang Zhashen.

24The tusi was required to send tribute to Beijing on regular occasions but was not required to participate personally in the mission.

25This is how his name is cited in the Annals of Yanyuan County (Gu 1894: 13). The Annals of Muli Tibetan Autonomous County gives the name as Xiang Zhashen (Muli Zangzu zizhixian zhi 1995: 543).

26In the Qing system of administrative grades, level 3a—the rank of xuanweishi—was the sixth-highest position out of a total of eighteen (Hucker 1985: 95; Herman 1993: 26).

27The large monasteries of Bathang and Chatring were completely destroyed in 1905 and 1906, respectively, after they rose in rebellion, and all surviving monks were summarily executed (Teichman 1922: 21–22).

28Local hereditary rulers in Kham were called either depa or gyelpo. Depa, or governors, were lay administrators and were often less independent from Lhasa than were gyelpo, or kings. The Qing had invested both depa and gyelpo with tusi titles (see, e.g., Coleman 2002: 33; Samuel 1993: 74–75).

29Banditry was one of the major scourges of Western travelers to the region, and their travelogues abound with descriptions; see, for example, Bacot 1912: 148–49 and Rock 1931: 14. Spengen 2002 gives an informative overview of these descriptions.

30At the end of the Qing, there were still 150 tusi in the whole of Sichuan, but according to Gong Yin, only 50 had any real authority, and those were mostly tusi of smaller areas (1992: 175).

31Eric Teichman, who did not visit Muli, noted the area’s special administrative status, but he mistakenly designated it as “a Tibetan native state under the suzerainty of the Yunnanese Government” (1922: 204). Although the Muli head lama did enter into an alliance with the Yunnan warlord Long Yun at one point, Muli was never officially administered by Yunnan.

32At that time, Chen Xialing was stationed at Dartsendo as frontier commissioner, but since he acted rather independently to preserve his own power base (as did many of the other high-ranking Chinese officials in the region), the term “warlord” would be more appropriate. Chen was a longtime survivor of the theater of the Sino-Tibetan border region and had held various positions in the region since the days of Zhao Erfeng (see ibid.: 47, 50, 57).

33The warlord of Yunnan, Tang Jiyao, was forced to relinquish power by his three commanders, Long Yun, Hu Kuyu, and Zhang Ruji, in 1930. Immediately afterward, the commanders turned against one another in the struggle for control of Yunnan. Long won out over Hu and chased him and the remnants of his troops in the direction of Muli (Muli Zangzu zizhixian zhu 1995: 78–79). This was the start of Long Yun’s rule as warlord of Yunnan. He was an ethnic Yi and Liu Wenhui’s enemy (Peng 2002: 65).

34“Nomihan” or “Nominhan” is a Manchu title that means “Dharma King.”

35Although both Liu Wenhui and Jiang Jieshi were members of the Nationalist Party, Liu was part of its reformist wing, which hoped to oust Jiang from power (Peng 2002: 65). This explains why Xiang Cicheng Zhaba sought support directly from Nanjing.

36Ethnic categories are explained more specifically in the glossary.

37This traditional form of corvée existed all over Tibet and Kham and obligated local people to provide transportation services to those whom the local ruler designated as entitled to such services. In Muli, this meant one family must provide one mule or horse, or two persons, for covering a distance of half a stage (a half day of traveling, usually between ten and fifteen kilometers or six and nine miles). Providing food, water, and encampment or lodging was also part of corvée duties (Liu 1939: 67).

38One social division in the highly stratified Nuosu society in Liangshan was between the aristocratic Black Nuosu and the majority White Nuosu, which was in itself stratified (see, e.g., Sichuan sheng Liangshan Yizu shehui diaocha ziliao xuanji 1987).

39This was an initiative by Hu Zongnan, deputy commander of the Nationalist Southwest Military and Administrative Headquarters (Xinan Zhangguan Gongshu), and He Guoguang, the governor of the province of Xikang (Muli Zangzu zizhixian zhi 1995: 84).

40This paragraph draws from the Annals of Muli Tibetan Autonomous County, an officially sanctioned version of Muli history and contemporary conditions. I have corroborated some of the facts through talks with government officials but have not been able to interview surviving witnesses of these events outside the Bustling Township area. When stripped of ritual remarks such as “the PLA troops received a warm welcome from all people of all ethnic and religious groups,” the inner logic of the narrative of the Communist takeover of Muli is nevertheless consistent and corresponds to better-researched cases on the United Front policy (see, e.g., Goldstein 1998: 23).

41In view of the fact that the label “Tibetan” is also used in Liangshan to designate other Tibeto-Burman speakers such as Premi or Namuyi, Mu’s real ethnic affiliation is unclear. Mu later held several high-ranking positions within the Prefectural People’s Congress of Liangshan and its Chinese People’s Political Consultative Conference (CPPCC). He was still around for the fiftieth anniversary of the Liangshan CPPCC in Xichang in March 2006. http://zx.lsz.gov.cn/showdetailokok.aspx?infoid=26259 (accessed 14 May 2006).

42In 1993, the provincial government of Sichuan posthumously recognized Wang Peichu Qudian as a revolutionary martyr (Muli Zangzu zizhixian zhi 1995: 940).

43One jin equals five hundred grams or about one pound.

44See, for example, the work by Lin Yaohua, one of the progenitors of ethnic classification work in China (1987) as well as the recent reappraisals of this process of state-controlled labeling by Nicholas Tapp (2002) and the authors whose work is collected in a separate issue of China Information, edited by Thomas Mullaney (Gros 2004; Mackerras 2004; Mullaney 2004a, 2004b; Caffrey 2004).

45One article on the Naxi of Eya in the southwestern corner of Muli, by the Naxi scholar Guo Dalie, was published in the volume on the Naxi (Guo 1986).

46These are mainly Zhongguo Shehui Kexue Yuan Minzu Yanjiu suo and Guojia Minzu Shiwu Weiyuanhui Wenhua Xuanchuansi 1994: 813; Li 1986; and Sun 1983.

47The Annals of Muli Tibetan Autonomous County acknowledges that the Tibetans of Muli call themselves by different names and lists the townships inhabited by these different categories of Tibetans (Muli Zangzu zizhixian zhi 1995: 833).

48Altogether, the proportional figures exceed the total figure of 35,000 because some townships are inhabited by more than one group and, except for Shuiluo, the Annals of Muli Tibetan Autonomous County does not distinguish between these groups at the township level.

49The text contains a strong defense of the concept of Zhonghua minzu (the Chinese nation), an attempt by some Chinese intellectuals in the 1990s to defend the indivisibility of the peoples living within the current political borders of China and Taiwan on the basis of “natural” or even racial grounds. In the text, Long also launches a vitriolic attack on Western writings on China.

50For a discussion of the terms “elite” and “minority elite,” see chapter 5.

51Xiang Zhaba Songdian became a member of Sichuan’s CPPCC in 1956, and Lin Jiayong participated in the 1957 National People’s Congress in Beijing, where he had a personal meeting with Chairman Mao Zedong’s secretary, who conveyed to him Mao’s support of his progressive work in Muli. In that same year, Lin participated in an important meeting on ethnic minority work in Qingdao, where he would meet Zhou Enlai, Mao, and Ulanfu (Muli Zangzu zizhixian zhi 1995: 946–47).

521 mu equals 0.0667 hectare or one-sixth of an acre.

53See, for example, Smith 1996: 399–412.

54The last rebel surrendered in 1974 and was given amnesty (Muli Zangzu zizhixian zhi 1995: 94).

55Xiang Zhaba Songdian financed the building of Muli’s first hydropower station in 1957, among other things (ibid.: 951).

56Xiang Zhaba Songdian died in 1964 and Xiang Songdian Chunpin in 1961. Lin Jiayong and former gatekeeper Han Jiayang had died the year before, in 1960. Former bazong Shu Yuanyuan, who played a central role in establishing the first two “state-run livestock farms” (guoying muchang) in Muli, died of cancer in 1963 at the young age of forty-seven (ibid.: 948).

57The Four Olds (Si Jiu) are old thoughts, old culture, old customs, and old habits.

2BUSTLING TOWNSHIP

1In 1998, production was 3.3 tons per hectare, while the national average was 3.7 tons. For the national figure on wheat yield, see, for example, the Web site of the US Department of Agriculture, http://www.pecad.fas.usda.gov/high-lights/2006/01/china_27jan2006/ChinaWheatArea.htm (accessed 20 March 2006). This was a year of exceptionally bad yields in China. The local government compiled but did not publish the township figures. For the entire county of Muli, only figures up to 1990 were available, and these showed a wheat yield that was markedly lower than the national average. This is probably because the northern townships are much higher above sea level and consequently have lower yields, reducing the total average yield for all of Muli.

2In Shuiluo, a township in the west of Muli, the rate was between 0.7 and 1.1 mu per person in a family in 1982 (see Naef 1998: 56). This means that the quality of the land in Bustling Township was rather low, even by Muli standards.

3Muli itself was classified as a “key poverty county” (zhongdian fupin xian) in Sichuan and received special allocations from the province (Muli Zangzu zizhixian zhi 1995: 928). These funds were distributed to families officially registered as poor and were earmarked for special purposes such as the purchase of seedlings or breeding of livestock and help with medical expenses. In 1998, 42 percent of the households in Bustling Township received such support; the allocations were discontinued in 1999 (interview with the Party secretary in Bustling Township).

4Its Latin name is cordyceps sinensis. In Chinese, the fungus is called chongcao, a shortened form of the Chinese translation of the Tibetan name, dong-chong-xia-cao (winter insect, summer plant), which refers to the life cycle of the fungus: after living underground as host to the larva of a white ghost moth, the fungus breaks through the head of the caterpillar and appears above ground to spread its spores. This shoot, still attached to the empty body of the caterpillar, is used as medicine.

5In 2004, prices were considerably higher in the shops of the major tourist city of Lijiang: chongcao sold for ¥40–80 per gram, and a 250-gram package of songrong went for ¥120. Collection of medicinal plants has become an important source of income for the whole region of Southwest China, and in Muli, it has reached the limits of sustainability for several species, according to ethnobotanist Karoline Weckerle (pers. comm.).

6China enacted this radical prohibition on logging in the natural forests of seventeen Chinese provinces in 1998, after the disastrous flooding of the Yangzi.

7A large company in Yunnan organized this enterprise. Each tree could provide up to several kilos of resin, which the workers tap into plastic bags that are then emptied into large containers. These containers are picked up at regular intervals by trucks making the difficult journey to Bustling Township and, for that matter, to many other places in the pine-covered regions of Southwest China.

8See also chapter 1 for a general discussion of ethnic classification in Muli and the glossary for the different ethnic categories used in this book.

9An ethnic group in Sichuan became classified as Mongol (Mengguzu) even though it had no apparent cultural similarities with the people living on the grasslands of Inner Mongolia because the Na chieftains legitimated their rule by claiming descent from Mongol officers of Kublai Khan who were “left behind” to administer the area (McKhann 1998: 33; Harrell 2001: 218).

10On the Na of Ninglang County in Yunnan, see chapter 5.

11Joseph Rock has written an article on the Rek’ua and their ritual practice, calling them “Zher-khin” (1938). Rock calls the Na “Hli-khin”; the Shuhiⁿ or Xumi from Shuiluo Township in Muli he calls “Shu-khin” and considers them to be a sub-branch of the “Zher-khin.” Charles McKhann—who uses the spelling “Rerkua”—has classified the Rek’ua as a separate ethnic category but one that is closely related to the Naxi (1998).

12The spelling dtô-mbà was introduced by Joseph Rock. This religious practice has also been rendered in Western literature as dongba (Chinese pinyin transcription) or dobbaq (using the so-called Naxi pinyin, a romanized spelling designed for use in literacy campaigns among the Naxi). It is also the name of the religious specialists at the center of this ritual practice. Dtô-mbà is the indigenous religion of the Naxi people of the Lijiang and Zhongdian areas of northern Yunnan and the adjacent areas in Sichuan. A comprehensive corpus of ritual texts written in pictographs has brought the Naxi a certain fame in China and the West, to such a degree that it is now possible to talk about the field of Naxi Studies (Pan 1998; Jackson and Pan 1998; Mathieu 2003).

13I am concerned here mainly with the immediate periphery around my area of research in the south of Muli. This includes, besides southern Muli, the northern half of Ninglang County in Yunnan and the adjacent areas of Yanyuan County in Sichuan. In the larger periphery of southwestern Sichuan and northwestern Yunnan, there are large areas where non-Han languages dominate in small market towns and administrative centers; these are mainly areas with a concentrated Nuosu (Yi) or Tibetan (Khampa) population.

14Even in regions with a large non-Han majority, such as the Tibetan Autonomous Region, local people lose out at all levels of the job market to in-migrating Han who are better skilled (Benjor 2001: 181).

15In Party or “government administrative organs” (xingzheng jiguan), the term bianzhi translated as “authorized personnel” or “the establishment of posts,” refers to the number of personnel authorized by the higher authorities (see Brødsgaard 2002: 363–64).

16It was only coincidence that the township head for Bustling Township was Premi. Since the majority of the population in Bustling Township were officially designated as Tibetan, they belonged to the same minzu for which the autonomous county had been created and were therefore only a “normal township” (putong xiang). In order to show official consideration for the special needs of other minority minzu, five so-called minzu townships (minzu xiang) had been established in Muli in areas where these other minzu made up a large proportion of the township population (although not necessarily the majority): two Na (Menggu), two Miao, and one Naxi. Except for receiving extra funding from the Commission of Ethnic Affairs of Liangshan Yi Autonomous Prefecture (Liangshan Yizu Zizhizhou Minzu Weiyuanhui) (in 1998, this amounted to ¥130,000, divided among the thirteen minzu townships of the whole prefecture), the only obvious difference between a “normal” township and a minzu township was that in the latter the township head had to be of the same minzu as the one for which the autonomous township had been created.

17The campaign is also called Tuigeng Huanlin Huancao (Return Farmland to Forest and Grass), and its major aim is to stop or reverse the deterioration of the natural environment and protect water resources (see, e.g., Ye, Chen, and Fan 2003). According to a Tibet Information Network Testimonies report of 17 August 2004, since 1999–2000, farmers all across the Tibetan Plateau have been urged and—according to some of the testimonies—sometimes forced by the government to plant trees and scrub on farmland. http://www.tew.org/development/tin.testimonies.08.17.04.html (accessed 20 March 2006).

18The campaign was also launched in other places in Muli. According to the ethnobotanist Karoline Weckerle, people in the township of Shuiluo in the western part of Muli were also planting chestnut trees on their contracted lands (pers. comm.).

19In a restricted publication on the work of the civil administration in Muli, propagation of Chinese marriage laws is put forward as an area in which a lot of work has been done but also one in which a lot remains to be done, promoting monogamy being one of the major tasks. According to this book, another challenge linked to the issue of polygamy is marriage registration (not registering a marriage is one way of keeping the state at bay): in a 1998 survey, of 9,626 marriages in Muli County, 34.5 percent were “not registered” (wei dengji), only 10 percent were “concluded with the mutual consent of both partners” (zizhu hunyin), and 2.22 percent were so-called “arranged mercenary marriages” (baoban maimai hunyin), in which the bride had been bought by the groom’s family (see Muli Zangzu Zizhixian Minzhengju 1998: 111–13).

20Unlike many other ethnic minority regions that had fully complied with national standards on organizing local elections (see, e.g., Benjor 2001: 88), Bustling Township had to wait until 2005 before residents could propose their own candidates.

21This is not necessarily the standard situation. In a comparative study of ten villages dispersed throughout rural China, Stephan Feuchtwang points out that according to regulations in Yunnan, officials at the administrative village level should not originate from the village in which they are posted (1998). In the Tibetan Autonomous Region, in contrast, candidates for the village committees were proposed and chosen by the villagers, at least in some counties in the 1996–98 period (see Benjor 2001: 88).

22National statistics showed that in 2002, 90 percent of Chinese had completed the nine years of compulsory education. Even leaving aside for a moment the thought that this figure is undoubtedly the result of overreporting, there are still 450 counties in China that had not fully implemented the law of 1986 on compulsory education (see, e.g., Renmin ribao, 3 November 2003, 14 March 2003). Ethnic minority areas, which make up a large part of these areas, were given more time to reach the required level (see Hansen 1999: 22, describing the situation in Yunnan, where all minority areas were supposed to establish six years of compulsory education by the year 2000). In a speech at the Chinese Women’s Ninth National Congress on 2 September 2003, Premier Wen Jiabao vowed to make nine years of compulsory education available to all children in China by 2008 (Renmin ribao, 3 September 2003).

23According to Regulation 51 of the Autonomy Regulations adopted at the First Meeting of the Seventh Session of the National People’s Congress of Muli County in 1990 (ratified by the National People’s Congress of Sichuan in 1992), “Tibetan-language courses will be established, alongside the centralized curriculum, in secondary schools with a majority of Tibetan students, in minority classes [shaoshu minzu zhongdianban], and in primary schools in townships with a concentrated Tibetan population, aiming to enable the student to master both spoken and written Tibetan and Chinese” (Dai Zuomin 1992: 24). This regulation stipulates further that where other ethnic minorities constitute a majority, classes or courses in their spoken and written languages are to be started. This can apply only to the Nuosu/Yi in Muli, but I have no data on the existence of any formalized courses in their language. The six Nuosu (out of 201 pupils) at the district junior middle school of the district in which Bustling Township is situated had no other option than to participate in Tibetan classes.

24This seems to be the case in other places in Muli as well and is apparently related to the fact that Premi is quite different from standard Tibetan. A comparative investigation reveals that among the ninety (Kham) Tibetan speakers of Xiangyang Village in Donglang Township, thirteen people could read Tibetan, while among two communities with a total of 176 Premi speakers in the townships of Kala and Taoba, only one person could be found who was able to read Tibetan (Zhongguo Shehui Kexue Yuan Minzu Yanjiu Suo 1994: 423).

25Despite its name, this school was established to train teachers for the whole prefecture of Garze, not just Kangding County. Its full name is Prefectural Normal School at Kangding (Zhou Kangding Shifan Xuexiao). Although Muli County is not in Garze Prefecture, the school accepts students from Muli because there are no schools for educating Tibetan-language teachers in Liangshan Prefecture.

3THE PREMI HOUSE

1In order to distance themselves from traditional kinship theories, several anthropologists have argued for using the term “relatedness” (see, e.g., Bodenhorn 2000; Carsten 2000; Edwards and Strathern 2000; Hutchinson 2000). Carsten holds that the use of this new term signals openness to indigenous idioms, and it facilitates comparisons “without relying on an arbitrary distinction between biology and culture” (2000: 3).

2The pronunciation of dzèⁿ differed slightly throughout Bustling Township.

3When discussing the organization of landownership in certain western Polynesian societies, Stevan Harrell defines “household” in the sense not of a co-resident group but of a group that “eats together,” that is, the “processing group” (1997: 242).

4Levine keeps to the concept of the corporate household in analyzing Nyinba society and positions herself against, among others, Carter (1984) when she holds that it is a mistake to interpret the fact that because common residence necessarily precedes the establishment of the domestic group, such co-residence is therefore generative of social relationships “where no such ties existed before” (1988: 128).

5For Naxi kinship, see, for example, McKhann 1992; for the Na, see Wang and Zhan 1988, Zhou and Zhan 1988, Yan and Liu 1986, Yan 1986, and Yan and Song 1983.

6In his examination of the numaya institution among the Kwakwaka’wakw (Kwakiutl) people in what is now British Columbia, Canada, Lévi-Strauss turned to history and found similarities with European noble houses of medieval times as well as other institutions like those in Japan in the Heian period (794–1192). He defined such a “house” as a personne morale or corporate body “holding an estate made up of both material and immaterial wealth, which perpetuates itself through the transmission of its name, its goods, and its titles down a real or imaginary line, considered legitimate as long as this continuity can express itself in the language of kinship or of affinity and, most often, of both” (1983: 17).

7See, for example, Howell 1995: 151, 169; Carsten and Hugh-Jones 1995: 10; and Sparkes and Howell 2003.

8This constitutes one of the four ideal types of terminologies distinguished by Lowie in the generation of ego’s parents, that is, “bifurcate merging.” The other three are “generational,” “lineal,” and “bifurcate collateral” (Lowie 1928, cited in Barnard 1996: 477).

9Due to some initial confusion on my part regarding Premi relationship terminology, I cannot eliminate the possibility that there were a few more cases of cross- and parallel-cousin marriages among the forty-six registered marriages in Walnut Grove.

10This study is cited in Ma 2001: 83.

11In their separate studies of two neighboring villages in the west of Muli, Naef and Schiesser each mention a few cases of polygamy. In his house survey of the eighty-two marriages in a Menggu village, Naef registered four cases of sororal polygyny and only one case of fraternal polyandry (1998: 52–53). Schiesser registered a slightly higher occurrence among the seventy-six marriages in a Premi village: five polyandrous and four polygynous marriages (2000: 183). Liu Longchu found seven cases of polygyny among 130 Naxi households in Eya Township in Muli in 1982 (1986).

12This does not differ from other regions in China. While regular rounds of complete land redistribution have become rare, minor rounds of readjustment do occur in many places in the countryside, usually every five years, for example, in relation to changes in family size. As in Bustling Township, major redistribution apparently did not take place in Benjor’s field site, making land use rights more stable and predictable.

13The ashes and bone remains of the deceased Ts’uop’i are placed on Shogunaⁿbasaⁿ, a mountain on the border between Yunnan and Sichuan.

14The symbolic association of the number seven with female and nine with male is widespread among Tibeto-Burmans. Populations as diverse as the Gurung in Nepal and the Lisu in northern Thailand, for example, hold the belief that women have seven souls and men nine (Macfarlane 1981: 56; Larsen 1984: 117). The Naxi funeral pyre for women traditionally has seven layers of logs, while there are nine for men (Shih 1998: 120). The same is true for the Na (Mathieu 1998: 220) and the Nuosu (Stevan Harrell, pers. comm.).

15It is possible to identify the places mentioned in the beginning of the ritual, but as the recitation continues, the identification of these places with actual, present-day, place-names becomes more questionable. Educated Premi and Chinese historians and anthropologists have analyzed and compared many different so-called kai lu (opening of the road) ceremonies and proposed many different locations for the Premi’s place of origin, several of which are situated in present-day Qinghai. There is supposed to be a place called Zhewo in Qinghai that is said to be identical to Jewopöjedaⁿ (Hu Jingming, pers. comm.).

16The nungk’e is made with threads of different colors and is also known to other peoples in the region, such as the Naxi and the Tibetans. In Naxi, it is naká (see Fang and He 1981: 346) or Na-k’wai (Rock 1952: 52), and in Tibetan it is nam mkha or mdos (see also Beer 1999: 322). It serves as a temporary dwelling for divinities during rituals, as, for example, in Naxi ritual, in which it symbolizes the hill Ts’a-nyî-gyû-k’ò mbù, the residence of the king of the demons of love suicide and suicide by hanging (see He and He 1998: 146). Moreover, through the symbolism of the ensnaring net, the nam mkha are also often used to keep evil spirits away. After some time, the nam mkha are often burned, thereby destroying the trapped evil spirits (Nebesky-Wojkowitz 1993: 369–97).

17According to the African Nyakusa, the spirits of the dead survive underground in the place of the shades (ubusyuka), whence they emerge to visit surviving relatives in their dreams and affect their lives (Wilson 1957: 17).

18See Aziz 1978: 51–57 as well as Levine 1988: 41–45 for the Tibetan Nyinba in Nepal.

19This is one of the arguments that led Shih to believe that the tusi family of Yongning was originally of Premi origin, a plausible explanation for the special status of the Premi in his territory (2001: 394–95).

20Several sources (Yan 1993: 38; Shih 2001: 390) mention that the Premi practice the patronymic linkage system common among Tibeto-Burmans. Under this system, the last one or two syllables of the father’s name are the first one or two syllables of the son’s name (Lo 1945). This system was not in use in Bustling Township, perhaps because of the prevalence of Tibetan names.

21House names become an integral part of personal names, so this practice might be the reason some ethnographers have mistakenly concluded that the Premi use the patronymic linkage system (see preceding note). On the one hand, if the son in a house that carries the name of his father uses this name as the first part of the name of his neolocal residence, this would make it seem as if he were following the patronymic linkage system. On the other hand, the possibility cannot be excluded that the custom of inserting the name of one’s native house into the name of one’s own house is a vestige of a once prevalent patronymic linkage system.

22Drwè is the word for the three stones on which the cooking pot was placed in former days, before the use of iron (hsin).

23I am grateful to Toni Huber for pointing this out to me.

24Such symbolism is known from other societies that rely heavily on little-worked timber for house building. In Southeast Asia, the life-giving symbolism of posts that follow the direction of the growing tree is supported by the opposite phenomenon of inverting the posts in constructions related, for example, to funerals (Domenig 2003).

4PREMI COSMOLOGY

1Samten Karmay, personal communication, 17 June 2002.

2See Zhongguo Ge Minzu Zongjiao yu Shenhua Da Cidian Bianshen Weiyuanhui 1990: 495. According to Rock, the name of the mythical founder of the Naxi dtô-mbà religion, Dtô-mbà Shilo, is derived from Tibetan. Dtôm-bà is equivalent to tönpa, “teacher,” and Dtô-mbà Shi-lo would refer to the founder of Bön religion, Tönpa Shenrab (1937: 7–8). The Premi pronunciation of shenro lends support to the argument that the religious traditions of the peoples of the southeastern periphery of historic Tibet have been influenced by Bön practices and beliefs.

3In Walnut Grove today, a very smart girl may be called a zènbuma, in reference to the disguised shep’a king who managed to trick Dingba Shenro into marrying him.

4Although other ethnic groups in the region also told versions of this story, to my knowledge their stories do not end with the wild chicken putting the lofty Dingba Shenro in his place.

5Brö demons are supposed to have originated among the Na. In the area of Wenquan, to the north of Yongning, where the first infected bride came from, the local Premi children wear small bags with medicinal herbs around their necks to protect them from brö demons carried by their Na classmates.

6This is why some of the more educated Premi identify brö demons with gu, or poison spirits, traditionally associated with the Miao people (see Diamond 1988).

7The phenomenon of demons entering a household with a new bride is also known among Mongolian nomads, where it is explicitly related to such uncertainties (Maria Tatár, pers. comm., April 2004).

8For the sake of clarity, I use the term yèma only to refer to the Buddhist lay priest residing in the village, and the term anji only to designate the non-Buddhist ritual specialist.

9The ritual use of the text translated by Lopez is prescribed as “a cure for a wide range of calamities, misfortunes, dangers, and afflictions, including epidemic, possession by demons, sick livestock, loss of wealth and property, dying under a bad star, false accusations, and bad dreams” (Lopez 1997: 511).

10While Muli Gompa demands that all monks reside at the monastery, the monks of the subordinate monastery of Zhamei in Yongning all live at home and go to a monastery only on certain ritual occasions (see chapter 5).

11Harrell notes that the Na and Premi in Yanyuan County to the south of Muli use the term djaba (2001: 199).

12Although Bön distinguishes between deities and famous teachers who have obtained divine status (Kværne 1996: 116, 119–20), such distinctions are not made by anji, who consider all of them to be powerful deities that can be invoked.

13This exposé on ritualists and mediums among Naxi, Na, Tibetans, and other ethnic groups of the region includes a graphic description of a séance carried out by a famous Tibetan sungma in Yongning, organized for the purpose of casting a spell on an invading rebel army.

14The available figures on the number of resident monks at Muli Gompa for the period 1930–50 are not consistent. According to the Annals of Muli Tibetan Autonomous County (Muli Zangzu zizhixian zhi 1995), after 1919 the number never again reached 300; in contrast, Liu Lirong, who stayed one month in Muli in 1938, put the figure at approximately 1,500 at that time (1939); Joseph Rock mentions 700 monks during his visit in 1924 (1925). This last estimate is close to the quota of 770, or one-tenth the number of monks in Drepung. Several of the surviving monks or Buddhist lay priests who were in Muli Gompa at the time it was shut down in 1959 put the number of resident monks at 600–700. On the total monk population of Muli in various periods, see chapter 3. It should also be kept in mind that few statistics differentiate between fully ordained monks and novices.

15On a short revisit in 2004, I was told by some of the monks that the number of monks had risen to around seventy, but I was unable to obtain the precise number of monks officially registered and the quota of monks allowed. Such a rise would be significant for the overall development of monasticism in Muli, but the number of monks from Bustling Township, as far as I could verify, was the same as in 1999.

16Wang Yangzhong, Director of the Religious Affairs Bureau of Muli Tibetan Autonomous County (Muli Zangzu Zizhixian Zongjiao Shiwuju Zhuren), pers. comm. All these monasteries belong to the Gelug sect, except for one in the north of Muli, which belongs to the Sakya sect.

17On two official delegation visits to China in 2002 and 2004, within the framework of a human rights dialogue between Norway and China, the State Administration of Religious Affairs acknowledged the general policy of not allowing organized religious education of children. This is, arguably, in violation of the United Nations Convention on the Rights of the Child, ratified by China, since such a policy may be viewed as limiting children’s right to practice religion (Lothe, Arnesen, and Wellens 2002). At the same time, actual implementation of this policy was not rigid. In response to a question on several child-monks we had seen in Tibetan monasteries, Thubten, director of the State Administration of Religious Affairs for the Tibetan Autonomous Region, answered: “When I visit a monastery and see a child-monk, I close my eyes.”

18The others are Islam, Protestantism, Catholicism, and Daoism.

19In particular, Documents 6, 144, and 145 specify how religious practice must be administered and controlled by the government and the Party in order to promote stability. Furthermore, they state that foreign intervention in religious practice in China, even from Taiwan, must be curbed (Potter 2003: 319–21).

20See, for example, the article by Ye Xiaowen, the longtime leader of the State Administration of Religious Affairs, “Why We Advocate Religious Freedom” (Women wei shenme zhuzhang zongjiao xinyang ziyou) (Ye Xiaowen 1999).

21Interestingly, there is no mention of China having only five religions.

22See Goldstein 1998: 159; and Wei Jing 1989: 61.

23See note 15.

24State Council General Office Document 39 of 1991 contains “Circular on the Relevant Items Concerning the Reincarnation of Living Buddhas in Tibetan Buddhism” (referred to in Muli Zangzu zizhixian zhi 1995: 906).

25Between 1987 and 2004, the institute educated more than four hundred tulku, and the staff grew to thirty-eight. But ambitions are even higher. According to Vice Director Li, the institute plans to augment its teaching capacity so that it will be able to offer a standardized geshe degree (the highest level of education in Buddhist studies within the Gelug School) for monks from all Tibetan and Mongol areas of China. One underlying motivation is to counter the trend of monks from Tibet traveling to India to study for this degree because they believe the level of geshe studies is not sufficiently high in Tibet. Such a move was not uncontroversial, and the institute dispatched a delegation of two tulku to monasteries in Tibet in an attempt to convince local clergy to support these plans (Li Guoqing, pers. comm., June 2004).

26Although the Gelugpa monastic tradition accepts the establishment of monasteries for Buddhist nuns, there are none in Muli, and there were no plans to establish any at the time of my most recent visit in 2004.

27In an article on hail protection ceremonies, Anne C. Klein and Khetsun Sangpo discuss the view that such ceremonies are dirty business because they involve the harming of malevolent spirits, a violation of the Mahayana spirit (1997: 539).

28For a study on the concept of syncretism, see the essays and the introduction in Stewart and Shaw 1994. An interesting alternative approach to coexisting religious traditions has been explored in another volume that focuses on Shinto and Buddhism in Japan (Teeuwen and Rambelli 2003).

29A comparable Tibetan version tells of the contest between Buddhist Milarepa and Naro Bonchung, a Bönpo, or adherent of the Bön religion (Dás 1881).

30Such stories are also told among the Na. Like the Premi, they participate in Tibetan Buddhism but also call on the services of non-Buddhist ritualists, the ddaba or Nda-pa (see also chapter 5). Unlike the anji, these religious experts use only orally transmitted “texts.” According to Na legends, the ddaba did have texts written on ox hide, but once, when they had no food, they were forced to eat their texts. As a result, all of their knowledge is in their stomachs (see, e.g., Rock 1938: 174; or Mathieu 1998: 210–11).

31See Samuel 1993. Per Kværne distinguishes three significations for the use of Bön in Western scholarship: (1) the pre-Buddhist religion of Tibet; (2) the religion that appeared in Tibet in the tenth and eleventh centuries, is still prevalent in some parts of Tibet, and has many points of similarity with Buddhism; and (3) a designation for popular beliefs in Tibet, including divination and the worship of local deities (1996: 9–22).

32It is interesting to note that, in accordance with the policy practiced during the time of the “Lama Kingdom,” leading monks limit the universalist claims of Tibetan Buddhism to the assertion that Buddhism should be the religion of the Tibetans, and possibly the Naxi and the Menggu, but not the Yi or the Han. As explained in chapter 1, since the Yi/Nuosu and Han populations in Muli were counted as commoners, they were not forced to send their third sons to the monasteries.

5MODERNITY IN YUNNAN

1These two figures are based on Public Security data from Yunnan for the year 2000 as found in “Report on an Investigation into the Socioeconomic Development of the Pumi in Yunnan” (Yunnan Sheng Minzu Shiwu Weiyuanhui Pumi Jingji he Shehui Fazhan Wenti Diaoyanzu 2002: 178). They tally with the results of the PRC’s fifth population census of the same year.

2This case is a thorn in the eye for Chinese evolutionist theory, because it clearly undermines the deterministic argument that human society passes through fixed stages of development, whereby a matrilineally organized society always precedes a patrilineal one.

3Wang and Yan trace this change to the year 1863, when the tusi, prompted by a particular case, ruled that the Premi practice of killing babies born out of wedlock was henceforth forbidden. Even if a woman refused to marry, the resulting child could not be killed (1990: 194).

4In Na and Premi, the monastery is called “Jamige.” Its full Tibetan name is Thar-lam dgan-ldan Thub-stan bde-bskyid gling (see also Rock 1959: 807).

5In 2006, there were around twenty monks and novices living at the monastery. As mentioned earlier, many informants did not differentiate between novices and fully ordained monks, a difference that easily involves up to seven or eight years of study. My use of the term “monk” as a translation of the word yèma covers both categories as well.

6According to the abbot, the local people made available seven mu of land and also supplied the manpower for the construction of the monastery. The county government provided ¥20,000, the Provincial Committee on Ethnic Affairs (Sheng Minzu Weiyuanhui) ¥20,000, the Provincial Religious Affairs Bureau (Sheng Zongjiao Shiwuju) ¥30,000, and the town government ¥16,000.

7One of the standard Chinese ethnographies of the Na, the 450-page-long The Matrilineal System of the Yongning Naxi (Yongning Naxizu de muxizhi) (Yan and Song 1983) includes only thirteen pages on Na religious practices. Two chapters by Shih and Mathieu in Naxi and Moso Ethnography (Oppitz and Hsu 1998) on the Na and the Naxi are based on more recent ethnography and are concerned with religion among the Na. Mosuo Daba Culture (Mosuo daba wenhua), a volume with Chinese translations of ddaba ritual “texts,” was published in 1999 (Lamu Gatusa 1999).

8This was part of a larger survey on Premi religious practice involving villages in the two counties of Lijiang and Ninglang. Villages surveyed in Ninglang include the following: in Labai Township, Guludian, Tuokua, Banawa, and He’erdian; in Yongning, Gala, Luoshui (Lugu), Tuoqi, Biqi, Bajia, and Wadu; in Mudiqing, Sanjiacun, Xiaocun, Zhongcun, Dacun, and Sanwangcun; and in Hongqiao, Lakua.

9In the Premi areas of Lijiang County, the survey found that people in many Premi villages called on the services of a Naxi dtô-mbà ritualist (Wellens 1998).

10Mentuhui (The Society of Disciples) is a Christian-inspired millenarian sect especially prevalent in poor rural areas. In Southwest China, it was widespread in certain minority areas—especially among the Nuosu (Yi) (see Luobu 1998)—until the government conducted a series of crackdowns in the wake of the anti-Falungong campaigns in 1999 and the early 2000s. In Ning-lang, it was mainly prevalent among the Nuosu; in Tuoqi near Yongning, three Nuosu committed suicide after a campaign in 2003. One of my monk informants in Tuoqi claimed that Premi and Na were not susceptible to such cults because of their adherence to Buddhism. Nevertheless, I collected several tales of Na and Premi who suffered because of their involvement with Mentuhui. Believing that the end is near, people slaughter all their domestic animals for food, stop working their land, and burn material goods. This makes post-crackdown readjustment very hard and places a terrible strain on the extended family, which is forced to pool resources in order to repair the financial situation of affected households.

11Hu Jingming, pers. comm. Story also referred to in Harrell 2001: 211.

12This is part of a slogan used in official campaigns or speeches in Yunnan to promote development: “Establish a strong province with a green economy and establish a province with ethnic culture” (Jianli lüse jingji qiang sheng minzu wenhua sheng).

13Hu had previously told Stevan Harrell that they were planning to finance this project with iron-ore and timber revenues from Muli (2001: 213). The 1999 ban on logging deprived these cadres of an important source of funding.