_full_journalsubtitle: International Journal of Chinese Studies/Revue Internationale de Sinologie _full_abbrevjournaltitle: TPAO _full_ppubnumber: ISSN 0082-5433 (print version) _full_epubnumber: ISSN 1568-5322 (online version) _full_issue: 1-2 _full_issuetitle: 0 _full_alt_author_running_head (neem stramien J2 voor dit article en vul alleen 0 in hierna): Meir Shahar _full_alt_articletitle_deel (kopregel rechts, hier invullen): Newly-Discovered Manuscripts _full_is_advance_article: 0 _full_article_language: en indien anders: engelse articletitle: 0

T’OUNG PAO Newly-Discovered Manuscripts T’oung Pao 105 (2019) 183-228 www.brill.com/tpao 183

Newly-Discovered Manuscripts of a Northern-Chinese Horse King Temple Association

Meir Shahar Tel Aviv University

In 2001, my friend Ye Derong 葉德榮 (A-de 阿德) purchased at the Bei- jing Panjiayuan 潘家園 antique market a set of manuscripts that hailed from province. Labeled “Horse King Association” (Mawang she 馬王社) (figure 1), the five hand-written notebooks originated in a vil- lage shrine dedicated to the tutelary deity of horses, donkeys, and mules. The manuscripts record the workings of the Horse King temple – its ac- tivities, revenues, and expenditures – from 1852 to 1956, after which the writing ceases in mid-page. Written materials from rural north are rare. To the best of my knowledge the records of temple associations such as those discovered by Ye Derong have never been studied. In this essay I rely upon them to reconstruct the workings of a late imperial Shanxi village. The temple association was “the bedrock of rural social structure in north China,”1 and the newly-discovered Horse King temple manuscripts chronicle its activities in extraordinary detail. Compiled by the villagers themselves, the Horse King notebooks provide us with unique access into the values and methods by which the rural community operated. In other words, the newly-discovered manuscripts help us understand how the villagers managed their public affairs and how late imperial Chinese villages gov- erned themselves. The Horse King manuscripts reveal the inseparability of religion, so- cial organization, and cultural production in a Chinese village. The Horse King association was a rural social organization that was based in

1) As David Holm observes in his “The Death of Tiaoxi (the ‘Leaping Play’): Ritual Theatre in the Northwest of China,” Modern Asian Studies 37 (2003): 868-69.

©T’oung Koninklijke Pao 105 Brill (2019) NV, Leiden, 183-228 2019 DOI: 10.1163/15685322-10512P05

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 184 Meir Shahar

Figure 1: Cover of one of the newly-discovered notebooks. The bottom left reads “Records of the Horse King Association.” The top right reads: “Rotating shifts account book.” The date corresponds to 10 November 1872. a religious establishment. The association managed the local cult of the Horse King (Mawang 馬王), mounting plays in honor of the equine god. As such the Horse King association sponsored the theatrical perform- ances that were the principal manifestation of village culture. Thus, the manuscripts shed light on the society, religion, and culture of rural north China. Combining textual research with fieldwork, this essay mines the manuscripts for those three interrelated aspects of the tradi- tional village. The manuscripts Ye Derong purchased bore no place name (knowing where they lived, the villagers had no reason to mention it). We will fol- low below the detective work that led to the discovery of the southern Shanxi village where they originated. Arriving at the dilapidated ­remains

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 185 of the Horse King temple we were able to contextualize, supplement, and enrich the information gleaned from the dry association notebooks. We interviewed villagers who reminisced about the cult of the equine god; we were introduced (in an adjacent market town) to a local scholar who authored a detailed modern history of the region;2 and, most im- portantly, we unearthed at the temple a nineteenth-century stele that, corroborating the association notebooks, supplemented them with a wealth of information on other organs of rural government (such as the village council, xiangyue 鄉約). In this respect, this essay is a methodo- logical exercise in the writing of microhistory that draws upon two texts: the Horse King manuscripts (which were purchased in Beijing) and the stele inscription (which was excavated in the village of their source). The newly-discovered texts bring to the fore a forgotten religious cult and its ecological background. Tutelary deity of horses, donkeys, and mules, the Horse King was among the most widely worshiped divinities of rural north China. Following the gradual disappearance of his equine protégés, his present-day cult is not nearly as popular as it used to be. The Horse King manuscripts anchor the god’s traditional veneration in the agrarian economy. They reveal that the villagers were expected to make contributions to the temple in accordance with the number of equines in their possession. We will see below that the Horse King as- sociation was entitled to survey the community and punish those who sought to evade taxation: villagers who hid draft animals, misreporting their numbers, were duly fined for their transgression. The manuscripts outline the association’s income and expenses in minute detail – down to every cup of tea and noodle plate. The fasti­ diousness of the financial accounts opens a window into the material culture of the village. It enables us to see what the villagers ate and drank, which modest furnishings adorned their humble shrine, and how they celebrated festive events. Furthermore, the exhaustive accounts ­attest to the paramount value of transparency in the management of village affairs: it was imperative to show community members exactly how much money was levied from them, and precisely how it was spent. ­Financial accountability was joined by another communal norm:

2) See Gucheng zhen zhi 古城鎮志, ed. Liu Wujing 劉武經 et al. (Zhengzhou: Huanghe shuili chubanshe, 2011).

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 186 Meir Shahar collective decision-making. Whether or not they were unanimously agreed upon, the policies of the Horse King association are invariably presented as such. “Everyone was happy with this decision,” is the re- frain that runs through the minutes of the association meetings. Evi- dently, it was tremendously important to reach – or at least give the impression of having reached – a village-wide consensus. In the early 1930s, Sydney Gamble and his Chinese associates con- ducted extensive surveys of the economy and social organization of vil- lages in four northern Chinese provinces: Hebei, Henan, Shandong, and Shanxi. Gamble concluded that, within the very same region, “no two villages were alike,” differing in everything from size and finances to pat- terns of leadership and social organization. At the same time, he also noted similarities that applied to the overwhelming majority of villages across regional boundaries, most notably their autonomy. By and large, early twentieth-century (and presumably in earlier times as well) north- ern Chinese villages managed their own affairs independently of gov- ernment intervention.3 The self-governing ability of village organizations might be partially attributed to the principles of transparency and col- lective decision-making that are apparent in the Horse King manu- scripts. The example of the Shanxi village identified and discussed below might therefore be instructive across provincial boundaries, even though other aspects of its social, religious, and cultural life might have been quintessential or local. For example, the great frequency of theatri- cal performances in it was typical of Shanxi province, even though plays in honor of the gods were regularly performed across north China. The structure of the essay follows the process of my investigation. I begin with a description of the Horse King manuscripts that Ye Derong generously shared with me.4 I tell the story of our joint search for their origins and the discovery of the stele inscription that corroborated and supplemented them. Then follows a topical analysis of the village ad- ministrative organs, the centrality of theater in rural life, and the Horse King’s flourishing late imperial cult. I address the question of village lit- eracy and tell the sad story of the temple’s twentieth-century demise.

3) Sidney D. Gamble, North China Villages: Social, Political, and Economic Activities Before 1933 (Berkeley: Univ. of California Press, 1963), 1. 4) Himself busy with other research projects, Ye Derong allowed me to take photos of the manuscripts in order that I transcribe, study, and publish them.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 187

The conclusion attempts to wrap up the insights that might be gained from this case study of rural temple documents.

The Village The members of the Horse King association knew where they lived. There was no reason for them to name their native place, which they referred to as “our village” (ben cun 本村). They write, for example, “there are so many horses, donkeys, and mules in our village.” From which of the thousands of northern Chinese villages did our manuscripts hail? I will briefly sketch the trials and errors that accompanied our search since they may be methodologically instructive. Ye Derong purchased the five notebooks of the Horse King associa- tion in one batch together with forty others, which all measured 8.85 by 10.23 inches and were wrapped with an identical blue cover. The latter were land-registration notebooks, and they bore a place name: Beizhang li 北張里 (Beizhang parish), Jiangzhou prefecture 絳州府 (currently known as Xinjiang 新絳 county), in southwestern Shanxi province (see map 1). The li was the smallest administrative unit of the Qing empire, usually encompassing some five villages. It was therefore to Beizhang li that we headed in search of our manuscripts’ source. We arrived there on 8 July 2013, accompanied by three faculty members – Hou Huiming 侯慧明, Yang Xueyong 楊學勇, and Xing Weidong 行衛東 – from Shanxi Normal University, which is located nearby in city. We toured Beizhang parish, looking for the remains of a Horse King temple. We visited several hamlets that did not fit the bill, before arriv- ing at the large and affluent Jianxi 澗西 village that seemed promising: not only did it feature an active Horse King temple; it also boasted a gor- geous Qing-period theater stage. The elders we interviewed recalled the festivities that accompanied the god’s birthday, on the 23rd of the 6th lunar month. Perhaps because they sought to humor us, some even seemed to recall the members of the association mentioned in the note- books. We felt gratified. That evening we returned triumphantly to Shanxi Normal University, congratulating ourselves on our discovery. It was on the train ride back to Beijing that I began to be pestered by doubts. True, we did find a Horse King temple. However, during the Qing period, the equine deity was immensely popular, being worshipped in

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 188 Meir Shahar

Map 1: Location of Beizhang and Gucheng in southwestern Shanxi Province. numerous rural shrines. How could we be certain that the temple we visited was the correct one? Beyond the vague recollections of a few elders, we possessed no concrete evidence tying Jianxi village to the manuscripts. For example, we obtained two genealogies of Jianxi fami- lies, but the names therein did not correspond to any of the dozens of names that are listed in the manuscripts. Hence, on 31 July 2014, we un- dertook yet another trip to southwestern Shanxi. Again, we toured the entire Beizhang parish (including some of the villages surrounding it), achieving the same ambivalent results: assuming that the notebooks hailed from the parish, the Horse King temple at Jianxi village must have been their source. Yet there was still no proof of a connection. Further- more, on this second visit, we met an octogenarian who was adamant that Jianxi never had a Horse King association. “We did have a temple,” he readily conceded, “but there was never an association (she).” In the summer of 2016, I shared my lingering doubts with Prof. Hou Chong 侯沖 of Shanghai Normal University. We went through the note- books, discovering something I should have noticed earlier: they do

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 189 mention place names, albeit not of the village itself but of surrounding ones. Under the date of 6 July 1891 we found the following entry: “The Horse King association possesses an embroidered tent (fengbu zhangzi 縫布帳子). Today all members of the association gathered to discuss its lease. It was decided that if rented to residents of our village, the price would be 200 copper coins (wen 文). If rented to residents of other vil- lages, the price would be 300 copper coins.” The records of following years name several villages that – in accordance with this policy – rent- ed the marquee (most likely on the occasion of public gatherings that required feasting on a large scale – funerals, religious ceremonies, or weddings). On 4 September 1913, for example, Guan village 關村 paid the required 300 coins for it, followed (on 17 December) by Beihou 北侯 village (figure 2). All in all six villages are mentioned in the Horse King manuscripts: Guancun, Beihou, Donghou 東侯, Nanhou 南侯, Changcun 常村, and Beiliang 北梁. All six are located within walking distance from each oth- er in Gucheng township 古城鎮, Xiangfen 襄汾 county, southern Shanxi (some 20 miles north of Beizhang) (see maps 1 and 2). While not one of them, the Horse King village must have been situated within their im- mediate vicinity. A glance at the map suggested the most likely candi- date to be Dengcun 鄧村 village. Therefore, on 21 September 2016, Ye Derong and I headed to Dengcun village. Along the way we were joined by the social historian Yao Chunmin 姚春敏 of Shanxi Normal Univer- sity, who proved immensely helpful in our fieldwork and analysis of the documents. Our growing delegation further included a Ph.D. candidate named Zhao Danrong 趙丹榮 and a faculty member named Xu Guang­ dao 許廣道 (both from Shanxi Normal University). The Dengcun Village chief (cunzhang 村張), Mr. Wang Furong 王福 榮 (born 1954), awaited our arrival at the small town of Gucheng. Emerg- ing from our vehicle, we showed him the manuscript pages covering the 1940s and 1950s, whereupon he pointed out one of the names men- tioned: “This is my father,” he exclaimed in delighted astonishment (fig- ure 3). Thus, within minutes of our arrival, our hypothesis proved correct: Dengcun village was the source of the Horse King manuscripts. It turned out that Wang Furong’s father, the late Wang Kegong 王克恭 (1913–2007), had been a member of the Horse King association, and his name appeared several times in its records. During the association’s last

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 190 Meir Shahar

Figure 2: The rented marquee as key for identifying the temple’s location. The text reads: Second year of the Republican period (1913)… The 24th of the 4th lunar month [29 May], income: Wang Qisheng rented the marquee: 200 coins. The 4th of the 8th lunar month [4 September], income: Guan Village rented the marquee: 300 coins. The 20th of the 11th lunar month [17 December], income: Beihou village rented the marquee, 300 coins. year of operation (1956), some of its possessions and documents were deposited at his place for keeping. These might have included the very five notebooks that, at some point sold to a dealer, are the subject of this essay. An entry for 1956 says that “under [Wang Kegong’s] name are kept one box (xiazi 匣子); one small case (xiaoxiang 小箱); five notebooks (ben 本); eleven teacups (chawan 茶碗); five bowls (wan 碗); one red shoulder-pole (hong caidan 紅採擔); one altar-cloth (zhuoqun 桌裙); [and] two loan receipts (jieyue 借約) (figure 4).

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 191

Map 2: Gucheng town and surrounding villages. The manuscripts originated in Dengcun Village. They allude to the villages that are marked by a triangle.

Figure 3: Mr. Wang Furong identifying his father’s name in the Horse King manuscripts. On his right is the author, and on his left is Yao Chunmin. (Photo by Ye Derong.)

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 192 Meir Shahar

Figure 4: Entry dated 1956 of the Horse King manuscripts. It indicates that the manuscripts (totaling 5 notebooks) were at the time in Wang Kegong’s possession.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 193

In subsequent conversations, Mr. Wang pointed out other relatives mentioned in the notebooks. A comparison of the Dengcun population with the Horse King manuscripts revealed an identical composition of surnames. In 2001, the village was home to 209 nuclear families, totaling 885 men, women, and children.5 (It might be noted that many of these – especially young couples – do not live in the village, being employed as migrant workers elsewhere). Eighty percent of the village families be- long to the Wang Clan, whereas the remaining twenty percent are most- ly surnamed Cui 崔, Liang 梁, or Guan 關. This same prevalence of the Wang surname, accompanied by small numbers of Cui, Liang, and Guan, is reflected in the Horse King manuscripts. Surprisingly, at present the surname Deng does not appear in the village so named. It is men- tioned, however, in the records of its early, fourteenth-century, history.6 The Dengcun provenance of the Horse King manuscripts having been established, we headed with Mr. Wang to his village. Along the way, he shared with us a cautionary tale of blasphemy and its dire consequenc- es. The Horse King temple, he said, had featured an elegant stele com- memorating its founding. Someone stole it, to be used as building material for his privy. Having completed the desired structure, he met a sudden and premature death. His descendants got rid of the baleful stone, only to have someone else get hold of it for a similar purpose. That person too suffered an untimely death, whereupon his relatives threw away the stele. The current whereabouts of the inscribed stone are no longer known. The pious sentiments underlying the story notwithstanding, no ef- forts have been made to restore the Horse King temple. Arriving at Dengcun village, we discovered the long-deserted shrine in ruins. Origi- nally, the Horse King temple was a three-room structure, measuring ap- proximately forty by fifteen feet. Exposed to the elements, its roof finally caved in in the 2000s. Nowadays, only one room is partially visible, the debris of the remaining two being overgrown with thick vegetation. This sad fate was shared by other Dengcun shrines. Our informants men- tioned no fewer than seven temples that once graced their village. Focal

5) According to the Gucheng zhen zhi, 445. 6) A person named Deng Zhen 鄧貞 (fl. 1302) is mentioned as a native of the village in the (Chenghua) Shanxi tongzhi 成化山西通志, ed. Li Kan 李侃 et al. (1475 edition; reprinted 1933), 3.33a; see also Gucheng zhen zhi, 22-23.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 194 Meir Shahar points of the traditional community, they no longer exist. The largest one, dedicated to the Emperor of the Eastern Peak (Dongyue dadi miao 東嶽大帝廟) was torn down by Japanese soldiers (who used it as firewood),7 and the remaining six were destroyed following the commu- nist takeover and the Cultural Revolution. Only one shrine – dedicated to the Jade Emperor – was renovated in recent years. Whereas the Horse King temple lies in ruins, the theater stage facing it survives. The elegant structure was saved from destruction, perhaps because at some point it was converted into a tool shed. The pillars sup- porting its roof are engraved with the inscriptions: “Strike the Drums” (gu yue 鼓吹) and “Pluck the strings” (qin xiao 琴嘯). We interviewed an eighty-year-old gentleman, Mr. Wang Yue 王躍, who recalled watching plays at the theater on the occasion of the Horse King’s birthday. Another extant monument is the village gate. Like many northern Chinese villages, Dengcun was protected from marauding bandits by a massive rammed-earth wall.8 Located at the eastern and western ends of the village, two imposing stone gates led in and out of the fourteen- foot tall fortification. The Horse King temple was situated by the west- ern gate, which still towers above its surroundings (figure 5). It is embellished above the entrance with the inscription “Congealed Emer- ald” (ning cui 凝翠). We will see below that the Horse King association contributed to the gate’s nineteenth-century restoration. Rural shrines rarely made it into official publications. Neither the Horse King temple nor any other Dengcun shrine is mentioned in the gazetteers of the relevant Taiping 太平 county, Pingyang 平陽 prefec- ture, or Shanxi province. However, the overwhelming majority of village shrines feature stele inscriptions, which shed invaluable light on local society. This proved to be the case with the Horse King temple. From underneath the foliage-covered debris of the shrine we pulled out an inscribed stone slab measuring two feet by one and a half feet. Dated 1861, it is transcribed in the Appendix. The inscription verified the Deng- cun provenance of our manuscripts. No fewer than twenty of the

7) Gucheng zhen zhi, 372. 8) Travelling through Shanxi Province in the mid-1860s, the Scottish missionary Alexander Williamson was greatly impressed with the villages’ high walls and gateways; see his Journey in North China, Manchuria and Eastern Mongolia (London: Smith, Elder& Co., 1870), 160-61, 325.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 195

Figure 5: The western gate of Dengcun village. In 1861, the Horse King Association contrib- uted to its restoration. (Photo by the author.) individual names mentioned correspond with the notebooks’ entries for that same year. Furthermore, the stele supplements the manuscripts with information on other Dengcun village organizations. In the follow- ing section we will analyze the association’s manuscripts together with this temple inscription.

The Association The late imperial temple associations of northern Chinese villages were known either as she 社 or hui 會 (both terms appear in the manuscripts of the Horse King association). They varied significantly in the scope of their activities. Some were concerned with temple management only. Since most villages featured quite a few shrines (the average number ranging from four to seven),9 several associations were often simultane-

9) In his 1940s surveys of village temples around the towns of Xuanhua and Wanquan

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 196 Meir Shahar ously active in the same community. Sidney Gamble alludes to Shanxi and Hebei villages with two, tree, four, or five temple associations each, based on fieldwork done in the early 1930s.10 Prasenjit Duara mentions a Hebei village that featured no fewer than seventeen associations: “Among these associations were the Association to Offer Fruit (the members of which offered fruit, among other things, to Guandi 關帝, the god of loyalty); associations devoted to the insect god, the god of medicine, and the rain god, all of whom were worshipped on particular dates; and the Drama association and the Lantern association, which organized temple festivities.”11 Whereas some associations limited their activities to the manage- ment of the god’s yearly festival, others were heavily involved in village affairs. In some cases, the temple association functioned as village gov- ernment. Willem Grootaers who, in 1948, conducted an exhaustive study of 115 Hebei villages, concluded that the “real masters of the village” were often members of the temple committee. “The attributions of the [temple] association are rather broad,” he writes. “It takes care of every- thing pertaining to village government, waterways and irrigation, crops, charity works, legal suits, schools, temple feasts and temple revenue…. No important thing is done in the village except through this association.”12 Yao Chunmin, who has studied hundreds of Qing-period temple inscrip- tions from Zezhou 澤州 prefecture (southern Shanxi), similarly ob- serves that “the temple association (she) was the most common self-governing body in rural society. It might even be said that it was the only self-governing body that held power in Zezhou villages. Temple

(northern Hebei), Willem Grootaers discovered an average of 4.2 and 6.5 temples per village respectively. Sidney Gamble notes that, in 1882, there were 435 temples in the 62 villages of Ding County (southern Henan), an average of 7 temples per village. David Johnson counted no fewer than 17 temples in one southern Shanxi village alone; see respectively: Willem A. Grootaers, Li Shih-Yü 李世瑜 and Wang Fu-Shih 王輔世, “Rural Temples around Hsüan-Hua (South Chahar), Their Iconography and Their History,” Folklore Studies 10 (1951): 9; Sidney D. Gamble, Ting Hsien: A North China Rural Community (New York: Institute of Pacific Rela- tions, 1954), 405; and David Johnson, Spectacle and Sacrifice: The Ritual Foundations of Village Life in North China (Cambridge, Mass.: Harvard Univ. Asia Center, 2009), 27. 10) Gamble, North China Villages, 36, 38, 163. 11) Prasenjit Duara, Culture, Power, and the State: Rural North China, 1900-1942 (Stanford: Stanford Univ. Press, 1988), 120. Duara draws upon the surveys of the Japanese South Man- churian Railway Company (Mantetsu), which were conducted in the late 1930s and early 1940s. 12) Grootaers, “Rural Temples around Hsüan-Hua,” 40.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 197 associations were involved in all aspects of village administration: tax collection, public safety, environmental protection, facility mainte- nance, lawsuit mediation, preservation of local customs, relief work, rain rituals, and more.”13 What might have been the reach of the Horse King temple associa- tion? Did it exercise authority over the people of Dengcun village? We begin with the association’s statement of purpose, which appears at the head of the first notebook (dated 13 October 1852) (figure 6):

Since ancient times, the Horse King temple has held three-day theatricals [each year]. In our village, many raise draft animals, and they all depend upon the god for their livelihood. For many years the theatricals were held without there being a ritual head (shenshou 神首). People who were randomly chosen for the task often declined. Therefore, we invited everyone for public discussion at the temple. We came to the conclusion that it would be better to select by rotation the teams [of ritual heads] in charge of the [temple] accounts. No one was unhappy with this decision. When the time of festivities arrives, the members of the community share the expenses of “pipes and gongs.” Each one brings his required contribution to the Horse King temple. If anyone fails to make a contribution, he is publicly penalized by the ritual heads. During the three-day theatricals the ritual heads are required to arrive early at the temple. If someone does not show up, he is fined the price of oil that will last the entire festival. This is how the rotating shifts of ritual heads [operate]. 古有,馬王廟演戲三天,村中养牲極多,全賴神聖扶持。歷年演戲,無有神 首,散股邀人,數次推拖。余等仝請眾人廟中公義。情愿排班落賬,無不欣 然。演戏之日,鳴鑼摊钱,送至神廟。倘有一家不到,神首公罰。演戲三天神 首人早到,有一人不到,罰油足用戲完之日,輪流交班。

13) Yao Chunmin 姚春敏, Qingdai Huabei xiangcun miaoyu yu shehui zuzhi 清代華北鄉村 廟宇與社會組織 (Beijing: Renmin chubanshe, 2013), 16. Yao argues that, in Shanxi, temple associations titled she were involved in village management, whereas most of those named hui had a more limited, religious, function. On the she and the hui temple associations see also Chen Baoliang 陳寶良, Zhongguo de she yu hui 中國的社與會 (Taibei: Nantian chuban- she, 1998); Qin Jianming 秦建明 and Lü Min 呂敏 (Marianne Bujard), Yaoshan shengmu miao yu shenshe 堯山聖母廟與神社 (Beijing: Zhonghua shuju, 2002); Marianne Bujard, “Construction, organisation et histoire du territoire liturgique de la Dame du Yaoshan,” Bul- letin de l’École française d’Extrême-Orient 99 (2012-2013): 228-93; Ekaterina Zavidovskaya, “Temple Committees and Religious Life in Rural North China: The Case of Shaanbei and Western Shanxi,” Zhongzheng daxue zhongwen xueshu niankan 中正大學中文學術年刊 2011.1: 291-336; and Johnson, Spectacle and Sacrifice, 184-86.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 198 Meir Shahar

Figure 6: Opening of the first notebook, dated 13 October 1852.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 199

The charter of the Horse King association is narrow indeed. It indi- cates that the association was not involved in local government. Rather, it was established for the sole purpose of administering the Horse King’s yearly festival, the principal feature of which was the mounting of plays. Like the overwhelming majority of northern Chinese village shrines, the Horse King temple housed no resident cleric.14 Instead the villagers themselves rotated as “ritual heads” (shenshou 神首), taking care of fes- tivities and finance.15 We will see below that they were divided into shifts (ban 班) that took turns managing the temple. Each shift num- bered ten members, and it was required to record its activities, income, and expenses. The material possessions of the temple – furniture, cos- tumes, and ritual implements – were likewise accounted for, prior to the handing over of keys to the following shift. Even though it did not concern itself with village governance, the Horse King association did exercise authority over the people of Deng- cun. The association expected the villagers to make the required contri- butions for its yearly festival, apparently possessing means of coercion. Those who did not pay were duly punished. The minutes of a meeting dated 6 August 1884, clarify that the Horse King’s beneficiaries were re- quired to fund his festival. The owners of the god’s protégés – horses, donkeys, and mules – were expected to contribute to his temple’s up- keep:

The previous shift and the coming one held a meeting at the temple. It was agreed that all households in the village that rear draft animals should make a contribu- tion towards the mounting of plays on the occasion of the god’s yearly festival.

14) In the early 1930s, Sidney Gamble found that only two of his eleven sample villages housed clerics (the rate of clerics per temple was even lower, since there were several tem- ples per village). In his earlier study of Ding County (Hebei province), he found an average of 1 cleric per 10,000 people (i.e. less than 1 cleric per 10 villages). In the early 1940s, Willem Grootaers conducted an exhaustive survey of village temples southeast of City (Shanxi Province), noting that “the overwhelming majority of the 401 temples I visited were perfectly empty [of any resident cleric];” see respectively Gamble, North China Villages, 120; Gamble, Ting Hsien, 21, 401; and Willem A. Grootaers, “Les temples villageois de la région au Sudest de Tat’ong (Chansi Nord), leurs inscriptions et leur histoire,” Folklore Studies 4 (1945): 171. 15) The term shenshou is reminiscent of shenjia 神家, recorded by David Johnson some thirty miles southwest of Dengcun, at Renzhuang village. Johnson renders shenjia as “Godly Ones;” see his Spectacle and Sacrifice, 33-35, 51-55. The terminological resonance does not necessarily imply similar functions.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 200 Meir Shahar

Their draft animals should be counted and contributions be made accordingly. If the heads of the association learn that someone has cheated, hiding his draft ani- mals, he will be heavily penalized. The heads of the association discussed it, and agreed that, without exception, the fine for one unreported donkey would be three-hundred [copper coins] worth of oil. The regulations of this association must be respected by all draft-animal rearing families. 前後班當廟公義所有閤村喂16牲口之家,每年敬神演戲槽頭點數,按數出錢。倘 有瞞昧藏躲,將會首查出加賠重罰。衆會首公議,查出驢一頭,罰油三百,不 得異說,所喂牲口之家,各尊會規。

Thus, the Horse King association possessed village-wide authority in regard to draft animals. It was entitled to survey the entire community, levying contributions from all owners of horses, donkeys, and mules. It might be noted that the owners in question constituted the majority of households, even though some likely possessed only a share of the given draft animal. Studies conducted in the late 1930s and early 1940s by the researchers of the Japanese South Manchurian Railway Company (Man- tetsu) have taught us that, in Hebei and Shandong, it was common for two, three, or even four households to share one horse, one donkey, or one mule, between them.17 Even though the manuscripts of the Horse King association are mute on the issue, we might assume that several families that split the expense of feeding the same draft animal likewise shared the burden of offerings to its tutelary deity. Whereas some Deng- cun residents contributed the required sum to the Horse King temple in full, others divided it between them in accordance with the percentage of the draft animal they owned. The villagers were expected to contribute to the Horse King associa- tion out of respect for its regulations (huigui 會規). They were required to abide by the time-honored practice of yearly contributions to the temple. In this respect, Dengcun village illustrates what Huaiyin Li has described as the principal force shaping the moral behavior of Chinese peasants, namely the commitment to community norms. Li argues that the decisions of the villager were determined, on the one hand, by the

16) Amending 喟 to 喂. 17) In Lujiazhai 盧家寨 village (northeastern Hebei) two or three families commonly shared one donkey; see Minami Manshū Tetsudō Kabushiki Kaisha, Tenshin jimusho chōsaka 南滿洲鐵道株式會社, 天津事務所調查課, Junka ken Rokasai nōson jittai chōsa hōkoku 遵化縣盧家寨農村實態調查報告 (Tianjin, 1936), 116-20, 123-24.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 201 pursuit of self-interest and, on the other hand, by adherence to conven- tions variously referred to as “village regulations” (cungui 村規), “local regulations” (xianggui 鄉規), or “old regulations” (jiugui 舊規). What is unique about the Horse King association is that its regulations have come down to us in written form whereas, according to Li, “although acknowledged by all community members, these regulations usually lacked codified texts.”18 Since the Horse King association was not directly involved in ­Dengcun administration, how was the village managed? The answer is suggested by the 1861 stele that we discovered at the temple. Commemorating the renovation of the village wall and the adjacent Horse King temple, it unravels the relations between the association and other village organs, of which the most important one was the village council (xiangyue). The initiative to restore the crumbling wall was taken by the village council, which enlisted for the purpose several other organizations in- cluding the Horse King association. The inscription indicates that the village council functioned as a village government (consult the original Chinese text in the Appendix):

Stele Commemorating the Renovation of the Village Wall and the Restoration of the Temple We have it in the Poetry Classic: “Let not the walls collapse,” and from Mencius: “Build high your walls.”19 Whenever the city wall is damaged, the families are in danger. Under such conditions, it is impossible to guarantee the peace and safety of the multitude. Now the southwestern corner of our village wall has been dam- aged by the torrential rains. Therefore the village council gathered and in a coop- erative spirit (xiehe 協合) decided that its members, together with the other villagers, would each make a donation of the grain harvest. The income gathered was used to repair the wall, making it five feet taller. Restorations were also carried out to the Three Kings [i.e. Horse King]20 temple, the various deities’ halls,21 the

18) Huaiyin Li, Village Governance in North China, 1875-1936 (Stanford: Stanford Univ. Press, 2005), 12; compare Gamble (North China Villages, 33): “The village association was… based on local custom rather than law and with unwritten tradition rather than written constitu- tion.” 19) Shijing, Mao 254 (“Ban”); and Mengzi 1B.13. 20) In Shanxi province, the Horse King is often worshiped together with the Ox King (Niu- wang 牛王) and the Medicine King (Yaowang 藥王), hence the name of “Three Kings” by which his shrines are commonly known. 21) These might have included the “Goddess temple” (Niangniang miao 娘娘廟) that ac- cording to our informants was situated next to the Horse King temple.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 202 Meir Shahar

theater stage, the gate-tower, the interior of the temple, and the wall’s immediate vicinity. On the day the work was declared completed, the donors’ names were in- scribed on the left side of this stele. The Zhangzheng canal donated 10,000 wen [copper coins]. The Horse King association donated 6,000 wen. The Dengcun canal donated 1,300 wen.

[The individual names of villagers who each donated seven sheng (approximately 1.8 gallons) or three sheng (approximately 0.8 gallons) of grain follows. At some point (perhaps during the Cultural Revolution) the stele was intentionally dam- aged, and hence the names of quite a few donors cannot be deciphered. Nonethe- less, at least twenty of the legible names feature in the Horse King manuscripts.]

The above donations amounted to 17,500 wen. The grain donated equaled 17,077 wen. The stone [slabs] cost 10,000 wen; lime: 4031 wen; nails: 155 wen; bricks and tiles for temple drains: 2020 wen; the artisans were paid 13,380 wen; their food amounted to 734 wen; the stele and its engraving cost 2600 wen. The seven items listed above totaled 32,920 wen, the remaining [1,657 copper coins] were used for the mounting of the stele and offerings to the gods on the occasion. Heads of the village council [fourteen names follow, at least five of which are listed in the man- uscripts as members of the Horse King association]. First day of the First lunar month, the Eleventh Year of the Xianfeng Reign [10 February 1861].

The stele unravels the elaborate social and political structure of a small Shanxi village in the 1860s. Even though Dengcun was probably home to no more than eighty or ninety families,22 the inscription alludes to no fewer than four administrative organs within it: the village council (which initiated and supervised the wall’s restoration); the Horse King association (which partially funded the project); the Zhangzheng irriga- tion canal; and the Dengcun irrigation canal (both acting as sponsors). Considering that there were six other temples in Dengcun, the number of associations in the village was probably larger still. We may assume that at least the biggest Dengcun temple, dedicated to the Emperor of

22) Extrapolating population figures backwards is hazardous. Nonetheless, considering that in 2001 there were 209 families in the village, it is likely that in the 1860s there were no more than eighty or ninety. We know that between 1949 and 2001 the population of Gucheng township (including Dengcun village) almost tripled (from 13,467 to 31,578); see Gucheng zhen zhi, 441-42, 445.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 203 the Eastern Peak, had its own religious association which, like the Horse King temple, levied contributions for its yearly festival. The centrality of the village council in Dengcun administration illus- trates the merging of organizations originally intended for distinct pur- poses. Ming and Qing period officials established xiangyue councils for the purpose of political indoctrination, just as they envisioned a lijia 里甲 organization for tax collection and a baojia 保甲 system for rural self-defense. However, by the mid-Qing period these three organs of lo- cal administration often merged, the xiangyue assuming the taxation and policing functions of the lijia and baojia respectively. The combined body was sometimes referred to by such hybrid terms as xiangbao 鄉保. The Dengcun example illustrates the Shanxi tendency to name its joint body of rural government xiangyue.23 Significantly for our purpose, be- tween five to seven members of the Dengcun village council were simul- taneously fellows of the Horse King association.24 The contribution of two irrigation authorities – the Zhangzheng ca- nal and the Dengcun canal – to the wall-restoration project is notewor- thy. The fields of Dengcun and neighboring villages were watered by an elaborate system of canals and sluice gates that drew upon the nearby Jianhe 澗河 stream. The Zhangzheng and Dengcun canals were proba- bly two sections of this irrigation system, each one having its own Canal head (quzhang 渠長) appointed by the villagers. Studies of nearby Hongtong county have taught us that the Canal head would levy money from the villagers in accordance with the size of their estates.25 The pro- portional system of taxation enabled the maintenance of the canals. The Dengcun case illustrates that the public funds thus gathered were

23) See Chang Jianhua 常建華, “Ming-Qing Shanxi beike li de xiangyue” 明清山西碑刻裡的 鄉約, Zhongguoshi yanjiu 2010.3: 117-38; Kung-Chuan Hsiao, Rural China: Imperial Control in the Nineteenth Century (Seattle: Univ. of Washington Press, 1960), 201-5; and Philip C.C. Huang, The Peasant Economy and Social Change in North China (Stanford: Stanford Univ. Press, 1985), 224. 24) Five council members listed in the stele appear in the Horse King manuscripts. Two ­additional names are partially damaged, making the identification uncertain; see below note 69. 25) Li Sanmou 李三謀 and Li Zhen 李震, “Qingchao Hongtong xian de hequ guangai yu guanli” 清朝洪洞縣的河渠灌溉與管理, Nongye kaogu 2003.3: 172. The elaborate irrigation system of Dengcun and neighboring villages is mentioned in Qing period gazetteers such as (Yongzheng) Pingyang fu zhi 雍正平陽府志, ed. Fan Anzhi 范安治, Zhang Tinggui 章廷珪 et al. (1736 edition), 13.14a.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 204 Meir Shahar

Figure 7: Association members listed in the manuscripts. Wang Jindong was replaced by his son Kentang, and Liang Wangxing by his son Dingmen. sometimes used for other purposes such as the renovation of the village wall and its adjacent Horse King temple. Returning to the Horse King association, its manuscripts name fifty members (who, in the 1860s, likely represented approximately sixty per- cent of the village families).26 They were divided into five shifts of ten

26) The 1861 stele lists at least twenty-four donors who were not members of the Horse King association. This would bring the number of families to seventy-four. However, since the stele is partially damaged there might have been others still.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 205

Figure 8: Association members listed in the manuscripts. Note the items listed under Wang Sigong’s and Guan Maoquan’s names; Liang Yongfeng’s expulsion (dated 13th year of the Guangxu reign [1887]); and the nickname Puppy Cui (Cui Xiaogou). members each. When someone passed away, his name was marked by a triangle, and the name of his descendant who replaced him was added underneath. In some cases it is specified that the descendant is a son or even a grandson (see figure 7). The dry listing betrays occasional bursts of internal feuds. In 1887, a person named Liang Yongfeng left the asso- ciation, and the notebooks tell us that “he would never be admitted back” (see figure 8). Why he was excommunicated remains a mystery.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 206 Meir Shahar

Keeping track of the temple’s belongings was an important function of the rotating shifts. The members of the Horse King association jeal- ously guarded their possessions, which were apparently stored in their individual homes for safekeeping. The notebooks record which items were whose responsibility, beginning with the temple keys, and includ- ing furniture (tables, chairs, benches, stools, cupboards, and boxes); ta- bleware (bowls and teacups); lights (iron lamps and gauze lanterns); ritual implements (incense burners and ceremonial robes); decorations (inscribed wooden plaques, door curtains, and altar cloths, sometimes woven of silk); various utensils (such as shoulder poles), and finally the notebooks themselves (and other financial documents such as loan re- ceipts). The example reproduced in figure 8 reads:

Wang Sigong: under his name are kept: two boxes (xiazi 匣子); four account books (zhangben 賬本); one door-curtain (men lian 門簾); one curtain of red satin (hong duanzhang 紅緞帳27). In addition, under his name are kept: one robe (paomei 袍袂 or pao’ao 袍袄); one red silk altar-cloth (hongling zhuoqun 紅綾棹裙); one temple gate-key (miaomen yaoshi 廟門鑰匙); one red soft (?) curtain (hong chuo lian 紅綽 簾); one bamboo pole (zhugan 竹竿28); and two lengths of rope (sheng 繩). Guan Maoquan: under his name are kept: one pair of gauze lanterns (sha deng­ long 紗燈籠) with covers; one low cupboard (guizhuozi 櫃桌子); three rectangular tables (tiaozhuozi 條桌子); fourteen chairs (yizi 椅子); two benches (bandeng 板 凳); two iron lanterns (tiexi denglong 鐵系) with iron handles; one wooden in- cense-burner (mu xianglu 木香爐); and one pair of [inscribed] wooden plaques (mu duizi 木對子) with covers.29

People who lost items for which they were responsible were penal- ized, as the minutes of an 1875 meeting attest. That case concerned an association member named Wang Zhigong under whose name were kept two door curtains, two gauze lanterns and an altar cloth. Wang failed to show up for the transition of shifts (when he was supposed to

27) Amending 賬 to 帳. 28) Amending 干 to 竿. 29) The same meticulous concern with temple belongings is evident in the extant notebooks – dating from the late imperial and Republican periods – of the Goddess of Mt. Yao temple in Pucheng County, Shaanxi province. See the photographic reprints in Qin Jianming and Lü Min (Marianne Bujard), Yaoshan shengmu miao yu shenshe, 130-50. I am grateful to Mari- anne Bujard for sharing with me high-quality photos of these notebooks.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 207 hand over these items). His colleagues determined that he should be fined fifteen-hundred copper coins:

On the first year of the Guangxu reign period (1875), the second shift handed over temple responsibility to the third one. Wang Zhigong [of the third shift] was given several items to keep under his name. However, when his [third] shift was over, he did not show up. Thus, when time came for the yearly festival, items were missing. There were mutual recriminations [between the various shifts] but no conclusion could be reached. Wang Zhigong agreed to compensate for the missing items: a pair of door curtains, an altar cloth, and a pair of gauze lanterns. A general discus- sion was held and, considering that he had no previous offenses, it was unani- mously agreed that he should be fined the large sum of fifteen-hundred copper coins. In the future, if items listed under a given person’s name are lost, that person – whoever he might be – must pay for them. Leniency will not be tolerated. We record [this decision]. 光緒元年因二班交三班。王執中存并一切物件,交班之時,執中而未到場。待 週30年演戏,而此物不全。你推於我,我推於你,终無可考。執中情願包賠門幔 一對,桌裙一條,紗燈一對,公同商議,念其非罪,與一着中賠出壹千五百 文。如再侯何人名下失物何人一人保賠,不許容情,為叙。

The members of the Horse King association are referred to in the manuscripts by their adult names. However, two (in 1887 and 1891 re- spectively) are alluded to by their childhood nicknames. One is called “Cui the Mule” (Cui luozi 崔騾子), and the other is “Puppy Cui” (Cui ­xiaogou 崔小狗) (see figure 8). Distanced from them by over a century, we are unable to tell why, but they might have been orphans (no one there to bestow upon them a proper name) or poor. Alternatively, they might have sought to keep their good luck into adulthood – such names as a “Dog Does not Care [for him]” (Goubuli 狗不理) were sometimes given to toddlers to ward off the evil eye of jealousy. If even a mangy mongrel ignored him, the lurking demons would leave the infant in peace.

Revenue The Horse King was the tutelary deity of equines: horses, donkeys, and mules. Appropriately, his cult was funded by his beneficiaries, the owners of draft animals. Each year, the temple association counted the

30) Amending 舟 to 週.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 208 Meir Shahar number of equines in the village and, in accordance with expected ex- penses, determined the sum to be levied per head. For example, the en- try for the year 1910 reads: “All in all the number of mules, horses, and donkeys [in the village] is 72.5 [The “half” might refer to a foal]. For each head a sum of 600 should be paid. Therefore, the total sum of contribu- tions comes to 43,500 copper coins (wen)” 共騾馬和驢七十二頭半。每 頭應出六百。共該收四十三千五百文。 Thus, the temple’s income depended upon a system of proportional taxation: the more animals a villager owned, the greater his contribu- tion to the Horse King association. Other northern Chinese villages sim- ilarly levied contributions in proportion to their residents’ individual property. In their 1930s surveys, Sydney Gamble and his Chinese col- leagues found that proportional taxation was the principal source of in- come of most rural temples, which varied not in regard to the principal of a fixed tax rate but in regard to the property used for its calculation. If the god was responsible for livestock, their numbers were the yardstick of contribution. By contrast, if the deity safeguarded agricultural crops, the villagers’ landed estates were relied upon to calculate their contribu- tions. Gamble mentions a Henan temple dedicated to the Ox King (Niu- wang 牛王) the yearly festival of which was financed by the “assessment of twenty cents for a cow, seventeen cents for a horse, fifteen cents for a mule, ten cents for a donkey and seven cents for a sheep.”31 He also cites the example of the Dragon King who was responsible for the weather, and hence for agricultural yields. Contributions to the ophidian god were calculated in relation to the size of the individual peasant’s farm. In some northern Chinese villages, temple contributions were deter- mined by the number of animal and human members of the household. A curious combination of a poll tax and an animal toll, the system as- sumed that beasts of burden and their human masters were equally valuable in economic terms. Gamble writes of a Shanxi village in which “the plays given to the gods in the second month were to ask for health for both men and animals. In figuring the amounts to be charged to the village families, a man, a horse, or a mule was counted as one unit, a donkey or a cow as two thirds of a unit, and a child as a half unit.”32

31) Gamble, North China Villages, 123. 32) Gamble, North China Villages, 123.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 209

In Dengcun village, donkeys were the most widely used draft animals. Successive years of the manuscripts allude to no other beast of burden. The significance of the donkey to the village economy accords with the findings of Republican-period agricultural experts. Surveys conducted in the 1930s indicate that, in rural north China, donkeys were the most prevalent equines, outnumbering horses by an average ratio of ten to one.33 The horse is a high-maintenance animal requiring relatively large quantities of high-quality fodder. The harsh environment of the arid north China plains, coupled with growing population pressure, made it impractical to raise the noble equine. For this reason, the members of the Horse King association – like their counterparts in other northern Chinese villages – preferred to employ its inexpensive and hardy rela- tion. Even when alluding to other equines, the manuscripts calculate their worth in donkey units. The key term in the account books is zhe lü 折驢 (“conversion into donkeys”) by which the dues to be paid for a horse and a mule (equaling two donkeys each) were calculated. Dating from 1917, the example reproduced in figure 9 reads: “The total number of [don- keys and] mules converted into donkeys in the village is 62.5 [the half probably refers to a foal]. For each head a sum of 1050 copper coins should be donated. The total income is 65,625 copper coins.” The entry tells us that, in 1917, donkeys and mules alike were part of the village economy. However, the accurate number of each species escapes us. For example, the figure sixty might mean forty donkeys and ten mules, or it might designate fifty donkeys and five mules. The practice of converting animal labor into donkey units is known to us from other northern Chinese villages. Whereas in England “horse- power” was the measure of productivity, the salience of the donkey in the Chinese landscape was such that it was commonly chosen as the yardstick for animal labor. The surveys of the Japanese South Manchu- rian Railway Company reveal that in 1930s Hebei and Shandong, labor was calculated by means of donkeys. According to the peasants’ rule of thumb, a horse equaled two (small) donkeys, a mule also equaled two (small) donkeys, an ox equaled one and a half (small) donkeys, and a

33) See, Meir Shahar, “The Donkey in Late-Imperial and Modern North China,” Asia Major 32.2 (2017): 71-100.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 210 Meir Shahar

Figure 9: Conversion into donkey units. Dated 1917, the entry reads: “Income: The total number of [donkeys and] mules converted into donkeys is 62.5. For each head a sum of 1050 should be donated. Thus, the total income is 62,625 copper coins. [In addition] Guan Wei- zhou donated a hundred and twenty copper coins, and Cui Fengyi donated a hundred cop- per coins.” large donkey was likewise the equivalent of one and a half (small) don- keys.34 The ubiquity of the donkey was such that it served as the meas- ure for all draft animals. The yearly taxes levied from equine owners were the principal in- come source of the Horse King association. The sums to be levied from the villagers were determined in relation to the expected expenses of

34) See Minami Manshū Tetsudō Kabushiki Kaisha 南滿洲鐵道株式會社, Dainiji Kitō Nōson Jittai Chōsa Hōkokusho, Tōkeihen, Dai ichiban: Heikoku ken 第二次翼東農村實態調查 報告書, 統計篇, 第一班, 平谷縣 (Dalian: 1937), 48-51 (Table 9); and Huang, The Peasant Economy, 144 (Table 8.4).

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 211 the yearly festival, adding a small margin for unforeseen expenditures such as temple repairs and replenishments. Following are the total sums levied from the villagers in sample years: 1863 (27,600 copper coins); 1872 (28,420); 1890 (30,450); 1905 (29,500), and 1909 (40,015). The infla- tion of the early Republican period is evident in the sums levied in 1919 (78,750) and in 1926 (100,650). Across rural north China, the increased devaluation of copper coins led to a gradual shift to silver dollars (dayang 大洋). In 1930, the amount gathered from Dengcun villagers was ninety- three silver dollars and sixty cents. From its founding in 1852 through 1890, the Horse King association relied for its operations solely upon the yearly amount levied from equine owners. However, beginning in 1891, new sources of modest rev- enue made an appearance in its account books. First was the embroi- dered marquee which, as noted above, provided the clue for our identification of the association’s Dengcun location. It was rented out to people who presumably needed to feast guests on such occasions as fu- nerals or weddings. In 1898, for example, the rental of the marquee brought in 700 copper coins, supplementing the much larger equine-tax revenues of 32,560 copper coins. Another source of modest income was individual donations. Mem- bers of the Horse King association sometimes contributed small sums of money above and beyond the required levy on their draft animals. The example reproduced in figure 9 shows that the 1917 equine-tax revenues of 65,625 copper coins were supplemented by Guan Weizhou and Cui Fengyi, who contributed 120 and 100 copper coins respectively. Appar- ently, such modest donations were meant to cover short-term budgetary deficits. Relying upon the exhaustive 1930s surveys of the Japanese South Manchurian Railway Company, Prasenjit Duara noted that some temple associations functioned as credit institutions.35 The manuscripts con- firm his observation. In its difficult final years, the Horse King associa- tion resorted to lending. On several occasions dated 1948 through 1951 the association lent its members grain, charging a yearly interest of twenty percent. That the loan was given in kind, rather than money, might indicate the currency shortage that followed the devastating Jap-

35) Duara, Culture, Power, and the State, 121.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 212 Meir Shahar anese occupation and civil war. It is noteworthy that the 1948 contribu- tions to the temple festival were similarly calculated in grain rather than cash.

Expenditures The Horse King association was established for the purpose of adminis- tering the equine god’s yearly festival. Dated 1852, its charter tells us that it was founded to manage the three-day theatricals in honor of the Horse God. Accordingly, the festival was the association’s single most important item of expenditure. Indeed, the yearly amounts levied from the villagers were calculated so as to cover its estimated cost. The manuscripts list in minute detail the expenses of the festival, which might be broadly divided into three categories: (a) Expenditures directly related to the mounting of plays (including payments to the theater company and its agent; stage-construction costs, and mats for the spectators); (b) Ritual expenditures (including offerings of food, in- cense, and firecrackers); and (c) The cost of the feast (including food, alcohol, tea, and tobacco for the guests and fodder for their mounts). Following are the expenditures of the 1888 festival, listed in the same order as in the manuscript. Note that it is not always clear whether given items of food were meant for sacrifice or for consumption (or possibly for both: first to be offered to the equine god, and then to be distributed among his devotees). Weights are in catties (jin 斤; approximately 1.1 pound), and the prices are quoted in copper coins (wen 文):

Performance cost (xijia 戲賈) 20,000 Food (fan 飯) for 156 guests, each 50 wen36 7800 Wheat bran (fu 麩) [as fodder], 60 catties 570 Hay (cao 草) 230 Charcoal (tan 炭) 300 Fried edibles [?] (youshi 油食), 16 catties 930 Oil (you 油), 30 catties 1920 Lamb-fat Candles (yang zhu 羊烛), 4.5 catties 472

36) The 156 guests might have included either the male members of the village, or the male members of families that owned draft animals (and paid their appropriate dues). Since most villagers owned at least a share of a draft animal, the two figures – of all males and all equine- owning males – were likely similar.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 213

Tea (cha 茶) and firecrackers (huoyao 火藥) 400 Theater agent (lantou 攬頭) 200 Rented mats (lin xi 賃席) 50 Cotton (mianhua 棉花) 130 Bowls (wan 碗) 28 Stage workers (guan tai 管臺) 800 Alcohol (jiu 酒), 5 catties 280 Pork (zhu rou 猪肉), 3.5 catties 245 Hookah (煤心水烟) 163 Bangers, incense, and pastries (paobiao xiang dianxin 炮表香点心) 1200 Ganshui cabbage (ganshuicai 干水菜) 530 Return of previous-year advance payment37 to Mr. Wu Jun 181 Return of previous-year advance payment to Mr. Wang Yuying 170 Vinegar and salt seasoning (cu yan tiaohe 醋鹽調和) 150

Total festival expenses: 36,749

The centrality of theatrical performances in the association’s expen- ditures is not exceptional. A growing body of scholarship has revealed to us the primacy of drama in Chinese rural culture. Plays were the center- piece of all religious festivals. Performed in temple courtyards they in- troduced their spectators to the heroes and villains of Chinese history as well as the gods and demons of its religious lore. Permanent theater stages were established in Shanxi temples as early as the Song period (960-1279). By late imperial times, they spread across the province, em- bellishing humble village shrines such as the Dengcun Horse King tem- ple. Stone inscriptions tell us of the pride Shanxi villagers took in their elegant theater stages. For example, a stele dated 1802 from county records the communal efforts of the villagers who renovated their local temple stage. They contributed labor to the project in accord- ance with their family size, and money in proportion to their landed estate.38

37) Amending dian 佃 to dian 墊. 38) The stele is transcribed in Cao Fei 曹飛, Shanxi Qingdai shenmiao xi bei ji kao 山西清 代神廟戲碑輯考 (: Sanjin chubanshe, 2012), 237. On theater and religion in rural China (with a special emphasis upon Shanxi province) see Feng Junjie 馮俊杰, Gu juchang yu shenxi shenmiao yanjiu 古劇場與神系神廟研究 (Xian: Xian jiaotong daxue chubanshe, 2013); idem., Shanxi shenmiao juchang kao 山西神廟劇場考 (Beijing: Zhonghua shuju, 2006); and Johnson, Spectacle and Sacrifice.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 214 Meir Shahar

Prior to 1949, Dengcun was home to six temples in addition to the Horse King’s. It is likely that at least some of these held their own yearly festivals. Thus, Dengcun villagers were treated to theatrical perform- ances by diverse temples on different occasions. In their Republican- period surveys, Gamble and his colleagues noted the frequency of Shanxi village plays (compared to their relative scarcity in some other provinces): “In Hebei only… a few villages had plays every year. The more usual practice was to perform them on special occasions, such as the dedication of a rebuilt temple.… In Shanxi, on the other hand, practi- cally every village gave plays once a year at least, and some gave them each month. It was estimated by some village elders that before 1911 a fairly large village would have from eight to ten plays per year [each last- ing several days].”39 Its detailed expenditure lists are a measure of the association’s ac- countability. Transparency is the primary concern that runs through the Horse King’s manuscripts. Recall that the members of the association were divided into shifts, which took turns managing the temple’s finan- cial affairs. It was the responsibility of each shift to provide the following one with an accurate report of all sources of income and items of ex- penditure. Other Shanxi temple associations were similarly concerned with financial accountability. Yao Chunmin has investigated a series of nineteenth-century temple inscriptions from White Mulberry village (Baisang xiang 白桑鄉) in the province’s southeast. As suggested by its name, the local villagers reared silkworms. They took turns running the local temple, holding office for three years at a time. Like their Dengcun counterparts, they reported their earnings and expenditures to their successors. In their case, the financial records were inscribed in stone.40 Other than their cost, we know nothing about the plays that were performed at the Horse King temple. The manuscripts allude neither to the genre of theatrical performance nor to the topics of individual plays. There is one exception: the 1882 entry specifies that these were shadow

39) Gamble, North China Villages, 123. 40) The inscriptions are transcribed in Yao Chunmin, Qingdai Huabei xiangcun miaoyu, 340-52. Financial records were sometimes written on wooden tablets, which were suspend- ed from temple eaves. Grootaers visited one such temple in Hebei province. Dating from 1908, the price of theatrical performance (16,950 wen) was not very different from the one Dengcun villagers paid that same year (14,000 wen); see Grootaers, “Rural Temples around Hsüan-Hua,” 15-17.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 215

Figure 10: Shadow puppet of the Horse King (late Qing, ca. 1850-1911, Gansu province). Note the equine god’s multiple limbs and third eye (in the middle of his forehead). (The Richard Hardiman Collection, Wilfrid Israel Museum, Kibutz Hazorea, Israel). plays (yingxi 影戲). The rich repertoire of Shanxi shadow theater in- cludes hundreds of dramas on topics ranging from the romantic and the heroic to the supernatural. It is tempting to speculate that at least one of the plays offered to the Horse King celebrated the god himself as the principal protagonist. The equine god figured in shadow theater (see fig- ure 10). Scripts of shadow plays celebrating his martial adventures have been preserved in Gansu and Sichuan provinces alike.41 It is therefore possible that the Horse King association commissioned a play on its own tutelary deity for the occasion of his birthday.

41) Titled “Mawang juan” 馬王卷 and dated 1843, the script of a Gansu shadow play on the Horse King is transcribed in Zhao Jianxin 趙建新, Longying jilüe 隴影紀略 (Beijing: Zhong- guo shehui kexue chubanshe, 2007): 296-332. A Qing-period manuscript of the same play has been preserved in Sichuan; see Li Long 李龍, “Mashen xingxiang de jingdian suzao yu xiju chonggou” 馬神形象的經典塑造與戲劇重构, Zongjiaoxue yanjiu 2015.1: 260. Li writes that the Horse King shadow play was extremely popular throughout Shaanxi, Gansu, and Si- chuan provinces. Admittedly, we possess no concrete evidence of its performance in Shanxi.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 216 Meir Shahar

The Equine God The newly-discovered manuscripts unravel the ecological background of a late imperial religious cult. Dengcun villagers worshiped the Horse King because their livelihood depended upon his equine protégés. The Horse King’s flourishing cult mirrored the ubiquity of horses, donkeys, and mules in the economy.42 The geographic contours of the cult betray a correlation between animal husbandry and religious practice. In north China, where equines were the preferred draft animals, the Horse King was widely worshiped. By contrast, in the rice fields of south China, where the water buffalo reigned supreme, his presence was modest. A preliminary survey of local histories reveals a large concentration of Horse King temples in such northern provinces as Shanxi, Shaanxi, He- bei, and Shandong, whereas in southern ones such as Zhejiang, Fujian, and Taiwan, his temples were fewer and far-between.43 The Horse King’s ancestry dates back to medieval times. His multiple limbs and eyes (see figure 10) are one indication of his Tantric Buddhist origins. The equine god is a descendent of the Horse-headed Hayagrīva Avalokiteśvara, whose cult was introduced to China from India during the Tang period.44 The Buddhist Horse-headed divinity was adopted by popular religion and Daoist clergy alike. His lay devotees venerated him as the Horse King or Horse God (Mashen 馬神), whereas Daoist

42) The Horse King has received only scant scholarly attention. The few exceptions include: Deng Qingping 鄧慶平, “Ming Qing Beijing de Mashen chongbai ji qi gongneng, yiyi de zhuanbian” 明清北京的馬神崇拜及其功能意義的轉變, Beijing shehui kexue 2006.2: 71- 77; and Li Qiao 李喬, Zhongguo hangye shen chongbai 中國行業神崇拜 (Beijing: Zhongguo Huaqiao chubanshe, 1990), 295-303. On the Horse King’s origins see Meir Shahar, “The Tan- tric Origins of the Horse King: Hayagrīva and the Chinese Horse Cult,” in Chinese and Tibetan Esoteric Buddhism, ed. Yael Bentor and Meir Shahar (Leiden: Brill, 2017), 147-89. 43) My rough estimate of the relative prevalence of Horse King temples in diverse prov- inces is based upon the Erudition (Airusheng 愛如生) gazetteer database (chuji 初集). I have counted the number of references to Horse King temples (miao 廟 and ci 祠), as well as (us- ing his other name) Horse God (Mashen 馬神) temples. I have arrived at a total, per prov- ince, of: Shanxi (202 references); Shaanxi (218); Hebei (353); Shandong (366); Zhejiang (29); Fujian (54); Guangdong (84); Guangxi (26); Taiwan (5). The figures should not be taken as indications of the actual number of temples in each province (because many temples did not make it into gazetteers, whereas conversely successive gazetteers count the same temples over and over again). Nonetheless, they might provide a tentative indication of the relative prevalence of Horse King temples in different provinces. 44) See Shahar, “The Tantric Origins of the Horse King.”

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 217 scriptures baptized him as the Horse Marshal (Ma Yuanshuai 馬元帥).45 By late imperial times, the divine protector of equines was among the most widely worshiped in north China. The detailed 1940s surveys con- ducted by Willem Grootaers, Li Shi-Yü, and Wang Fu-Shih, attested his prominence in rural religion. In their survey of 115 villages in northwest- ern Hebei, they discovered that the Horse King ranked fourth in the number of cult units dedicated to him. The equine deity was preceded only by such nationally (or locally) famed divinities as the Dragon King, the Wudao 五道, and the Bodhisattva Guanyin.46 The Horse King was worshiped in cities and villages alike. Merchants were as dependent upon his equine beneficiaries for the transport of goods as peasants were for ploughing their fields. Hence, the tutelary deity of draft animals was dedicated shrines in merchant guild halls.47 Government officials vied with horse traders in their support of the equine god. Horses were widely used in the diverse government agen- cies engaged with transportation and defense. Civil officials in charge of the “horse administration” (mazheng 馬政) worshiped the Horse King, as did army officers whose success in battle depended upon his divine

45) Daoist scriptures such as those included in the fifteenth-century Daofa huiyuan 道法 會元 (Daoist Methods United in Principle) (DZ 1220) allude to the equine god by the addi- tional title of “divine official” (lingguan 靈官) (hence: Ma Lingguan), and the added name Huaguang 華光. The latter features in works of fiction and drama on the equine god, such as the sixteenth-century novel The Heavenly King Huaguang’s Journey to the South (Huaguang Tianwang Nanyou zhizhuan 華光天王南遊志傳); see Shahar, “Tantric Origins of the Horse King,” 171-85. 46) Conducted in the midst of the raging 1948 civil war, the survey covered the villages sur- rounding the town of Xuanhua 宣化 (approximately a hundred miles northwest of Beijing), see Grootaers, “Rural Temples around Hsüan-Hua,” 54-57. In a slightly earlier survey of 93 villages in Wanquan 萬全 County (some 110 miles northwest of Beijing), the equine deity was discovered to rank sixth in the number of his cult units; see Willem Grootaers, Li Shi-Yü 李世瑜, and Chang Chi-Wen 張冀文, “Temples and History of Wanch'üan (Chahar): The Geographical Method Applied to Folklore,” Monumenta Serica 13 (1948): 210. By “cult unit” Grootaers and his colleagues refer either to an independent temple or to a clearly delineated shrine within another god’s temple. 47) Extant to this day, the Horse King hall (Mawang dian 馬王殿) within the Beijing temple of the Eastern Peak (Dongyue miao 東嶽廟) was patronized by horse merchants, muleteers, and donkey-drivers, who hailed from the nearby Horse Market thoroughfare (Mashi dajie 馬市大街). Stele inscriptions dated 1720 and 1749 commemorate its renovation by the “Horse King’s devout disciples of the Horse Market” and “his disciples of the donkey guild,” respectively. The steles are transcribed in Niida Noboru Hakushi shū Pekin kōshō girudo shiryōshū 仁井田陞博士輯北京工商ギルド資料集, ed. Saeki Yūichi 佐伯有一 and Tanaka Issei 田仲一成 (Tokyo: Tōkyō Daigaku Tōyō Bunka Kenkyūjo, 1975), 697-98, 700-1.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 218 Meir Shahar support of the cavalry. Horse King temples were established at all levels of government, from the imperial palace to local yamens and courier stations.48 Furthermore, the divine protector of the cavalry was dedi- cated shrines in each and every military base.49 State patronage played a major role in the proliferation of the Horse King’s cult. The long history, diverse clientele, and rich lore of the Horse King go beyond this article’s scope. Here I will merely comment upon one con- figuration of his cult that informs his Dengcun shrine. In Shanxi prov- ince, the equine god is commonly worshiped together with two other tutelary deities: the Ox King and the Medicine King (Yaowang 藥王). The “Three Kings” (Sanwang 三王), as they are jointly known, take care respectively of equine, bovine, and human well-being. In some temples, the Horse King takes pride of place at the center, and is flanked by the icons of his bovine and human counterparts, whereas in other temples the Ox King occupies center stage, being accompanied on his left and right by the Horse and Medicine Kings. The former are usually named after their central divinity Horse King temples, whereas the latter are known as Ox King shrines.50 In his Dengcun temple, the Horse King was worshiped side by side with the Ox King and the Medicine King. The 1861 stele unearthed at his dilapidated shrine alludes to the restoration of the “Three Kings tem- ple,” a term which also appears in the association’s notebooks. The joint worship of the two animal tutelary deities might explain the occurrence of oxen in the manuscripts. The entries of several successive years al- lude to the total number of “horses, mules, donkeys, and oxen” in the

48) On the Horse King shrine in the imperial palace see Wang Mingzhen 王銘珍, “Zijin­ cheng de Chenghuang miao he Mashen miao” 紫禁城的城隍廟和馬神廟, Beijing dang’an 北京檔案 2004.12: 45. 49) A fine example is the Ming-Period Horse King temple that is situated to this day within the Juyong guan 居庸關 Fortress, some forty miles northwest of Beijing. A stele inscription dated 1792 records its renovation by a Qing military commander of the Mongol Blue Banner (information gathered during my visit to the site on 24 August 2014). 50) An early example of a Shanxi temple that features the “Three Kings” is the renowned Ox King temple at Weicun 魏村 village, approximately twenty-five miles north of Dengcun. Dating back to the thirteenth century, the temple has the Ox King flanked by his equine and human counterparts. A more recent example is the Horse King temple at Xizuang 西莊 ­village (some thirty miles northwest of Dengcun), in which the equine god occupies the center. On the Weicun Ox King temple see Yan Baoquan 延保全, “Weicun difang shehui yu Niuwang miao saishe yanju” 魏村地方社會與牛王廟賽社演劇, Minsu quyi 民俗曲藝 138 (2002): 7-65.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 219 village. Because the Ox King was enshrined in its temple, the Horse King association was entitled to levy contributions from the owners of bovines. The Dengcun cult of the Horse King, Ox King, and Medicine King might indicate a similar theological standing for their respective equine, bovine, and human protégés. The joint worship of human and animal tutelary deities implies that people and their beasts of burden are equal- ly vulnerable and similarly in need of divine protection. In this respect, Chinese religion differs from the Western monotheistic faiths, which posit an existential abyss between draft animals and their human mas- ters. The Bible has it that man was fashioned in god’s image, whereas all other animals were created to serve and be eaten by him.51 (Admittedly, however, the practice of Christianity, as distinct from its theology, per- mitted the veneration of animal tutelary deities.)52

Literacy Historians and anthropologists have noted the significant role of literate specialists in Chinese rural culture. The overwhelming majority of vil- lage writings were produced by a small portion of their population. Most peasants were at best semiliterate, and their literary needs were supplied by “the very limited group of people who were equipped with the rare ability of making texts.”53 The literate specialists were of varied backgrounds: some were the scions of eminent local families, others

51) Genesis, 1, 26; on the Christian attitude towards animals see Andrew Linzey and Dan Cohn-Sherbok, After Noah: Animals and the Liberation of Theology (London: Mowbray, 1997), 1-16; and Benjamin Arbel, “The Renaissance Transformation of Animal Meaning: From Petrarch to Montaigne,” in Making Animal Meaning, ed. Linda Kalof and Georgina M. Mont- gomery (East Lansing: Michigan State Univ. Press, 2011), 62-64. 52) On the veneration of equine tutelary deities in Brittany (France) and Russia respec- tively see Glaoda Millour, Les saints vétérinaires en Bretagne (Morlaix: L’Ecole Bretonne, 1990); and Ann M. Kleimola, “Visions of Horses: The Evolution of the Russian Cult of Florus and Laurus,” in The Place of Russia in Europe and Asia, ed. Gyula Szvák (Boulder: Social Stud- ies Monographs, 2010), 79-92. 53) Ren-yuan Li, “Making Texts in Villages: Textual Production in Rural China during the Ming-Qing Period” (Ph.D. diss., Harvard Univ., 2014), 435; compare James Hayes, “Specialists and Written Materials in the Village World,” in Popular Culture in Late Imperial China, ed. David Johnson, Andrew Nathan, and Evelyn Rawski (Berkeley: Univ. of California Press, 1985), 75-111.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 220 Meir Shahar were hired school teachers, and others still were religious professionals (such as Daoist priests). They functioned as village scribes, composing a range of texts from contracts and family genealogies to exorcistic charms and the handsome new-year couplets that embellished people’s homes. The above-quoted stele that was unearthed at the Horse King temple was likely authored by such a specialist. Albeit far from a dazzling com- position, it features two references to the classics that only a person with a modicum education could have been familiar with. Whereas its stele was likely penned by a literate specialist, the asso- ciation’s manuscripts (with the possible exception of their elegant opening page) were written by rank and file association members. Re- call that the account books were compiled by generations of villagers who in turn took responsiblity for the temple’s possessions and financ- es. As attested by diverse handwritings, many hands were involved in their compilation. It is possible that the most educated (or gifted) mem- ber of each shift was assigned the task of writing. However, as educated as he might have been, he did not belong to anything like the literate elite. In its entire history, Dengcun did not produce as much as one pro- vincial degree holder (not to mention a jinshi graduate). Its lack of suc- cess in the examinations might be contrasted with surrounding villages that boasted successive generations of degree holders (often bearing the same surnames). Whereas no landowning gentry family resided in Dengcun, some elite families lived in the neighboring communities of Donghou and Jing’an 京安.54 Riddled with orthographic errors, the manuscripts themselves attest that their authors received no more than rudimentary education. Ac- quainted with a limited number of characters, the villagers often substi- tute a homophone for the required character. In other cases, characters are erroneously drawn, betraying the writers’ lack of calligraphic prac- tice. The case of numerals is especially interesting. The Horse King’s ac- count books feature multiple standard and improvised methods of number writing, often on the same page. They include the standard modern form, the accountant form (used to prevent forgery), and the so-called Suzhou numerals (Suzhou mazi 蘇州碼子), which were com-

54) See the Gucheng zhen zhi, 497-99, for a list of all successful degree holders from the township and surrounding villages. Dengcun village boasted one Qing-period successful holder of a military degree (wuju 武舉).

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 221 monly used by merchants.55 To give just two examples: the character si 四 (four) is often written in the notebooks ; and the character wu 五 (five) is commonly drawn as . Errors and idiosyncrasies notwithstanding, the notebooks do accom- plish their purpose. They provide an accurate account of the Horse King temple’s activities, revenues, and expenditures. In this respect, the man- uscripts might exemplify the functional literacy that was enjoyed by at least some segments of the northern Chinese rural population. Albeit by no means fully literate, the association members recognized the limited number of characters that were indispensable for daily life. They could write their own names, and they were fully capable of rather complex arithmetic calculations. In this respect, the notebooks support Evelyn Rawski’s claim that “it was possible for a broad cross-section of Qing males to attain some degree of literacy.”56 Our manuscripts indicate that at least some Dengcun villagers possessed rudimentary writing, reading, and arithmetic skills.

Aftermath Beginning in 1852 and continuing through the late 1920s, the Horse King association maintained its activities with hardly any interruption. For over seven decades, the temple levied contributions from Dengcun equine-owners and mounted plays for the villagers’ entertainment. However, by the early 1930s, twentieth-century history caught up with the temple’s fortunes. Shanxi province was torn by warfare during two long decades. The devastating wars, and the communist takeover that followed, had a detrimental impact upon the Dengcun cult of its tute- lary deity. Let us briefly review the military catastrophes that befell Shanxi ­province in rapid succession. In 1929, the local warlord Yan Xishan 閻錫山 (1883-1960) joined forces with Feng Yuxiang 馮玉祥 (1882-1948)

55) The Suzhou numerals are also known as shangma 商碼 (merchant numerals); see Joseph Needham and Wang Ling, Science and Civilization in China, vol. 3: Mathematics and the Sciences of the Heavens and the Earth (Cambridge: Cambridge Univ. Press, 1959), 6-7, Table 22. 56) Evelyn Rawski, Education and Popular Literacy in Ch’ing China (Ann Arbor: Univ. of Michigan Press, 1979), 140. Consult also her discussion of arithmetic skills, 125-28.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 222 Meir Shahar against Chiang Kai-shek’s Kuomintang armies. In the war that followed, the Shanxi warlord was badly defeated, causing enormous loss of life and economic hardship in the province.57 The Shanxi population hardly had time to recuperate when, in February 1936, thirty-four thousand communist troops crossed the Yellow River from neighboring Shaanxi province, occupying the very region of Dengcun village in Xiangfen County.58 The communist troops were repelled by Yan Xishan, who by now had allied himself with the Kuomintang. For his part, Yan Xishan was defeated a year later by the invading Japanese. Capturing the pro- vincial capital of Taiyuan in November 1937, the Japanese troops ad- vanced southward to Xiangfen County. On 2 March 1938 they marched into Gucheng town, less than a mile east of Dengcun village. Gucheng was to serve as a Japanese army post, terrorizing the local population until the Japanese surrender in the Pacific War in August 1945.59 Blank pages in the Horse King manuscript testify to the long years of war and hardship that were endured by its authors. The association held no activities during the two decades of warfare between warlords and nationalists, between nationalists and communists, and between na- tionalists, communists, and Japanese invaders. It held its last festival in 1930, after which it ceased operation. Other Shanxi temples fared no bet- ter. Conducting his surveys in the early 1930s, Sydney Gamble noted the disruption of religious life that followed Yan Xishan’s defeat at Chiang Kai-shek’s hands: “In Shanxi the demands of the military were so heavy that the theatrical performances given annually by the villages had to be discontinued for lack of funds.”60 Heavy population loss made the re- sumption of the yearly festivals impossible. During the seven years of Japanese occupation numerous Shanxi villagers found refuge across the

57) For general background see Lloyd Eastman, “Nationalist China during the Nanking dec- ade, 1927-1937,” in The Cambridge History of China, vol. 13: Republican China 1912-1949, Part 2, ed. John K. Fairbank and Albert Feuerwerker (Cambridge: Cambridge Univ. Press, 1986), 126-28. 58) See Gucheng zhen zhi, 100. 59) On the Japanese occupation of Gucheng Town and its surrounding villages see Gucheng zhen zhi, 86-100. For general background see Edward L. Dreyer, China at War 1901-1949 (New York: Longman, 1995). 60) Gamble, North China Villages, 5.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 223

Yellow River. Fleeing to the Chinese-controlled Shaanxi province, they left behind empty homes, barren fields, and deserted temples.61 Following the Japanese surrender in August 1945, the battle for China between communists and nationalists resumed. Early in the war the southern Shanxi countryside fell into communist hands. On 2 June 1947, almost two years before they would take over the provincial capital, communist troops entered the town of Gucheng.62 The communist vic- tory brought respite to Dengcun villagers from almost two decades of constant warfare. For a short while it seemed as if the ritual traditions of old might be revived.63 The Horse King manuscripts record the reorgan- ization of the rotating shifts in charge of the temple. On 13 July 1948, the Horse King festival was held once more. Funds for the festivities were levied in grain, reflecting the shortage of currency at the time. The 1948 festival was the last ever celebrated at the Dengcun temple. If they had hoped for a resumption of their venerable tradition, the members of the Horse King association were to be disappointed. The establishment of the People’s Republic of China brought major eco- nomic, social, and cultural transformations to the countryside. The pen- etration of the modern state into the villages was such that the social and religious organizations of old could no longer survive. After 1948 the Horse King association no longer levied money from the Dengcun vil- lagers. The yearly festival with its elaborate theatricals was replaced by a modest sacrifice of incense and firecrackers. The last of these frugal af- fairs was held on the 15th of the first lunar month, 1956. After this date, no more entries were recorded in the Horse King manuscripts.

61) See Gucheng zhen zhi, 86. 62) See Gucheng zhen zhi, 101. For general background see Suzanne Pepper, “The KMT-CCP conflict 1945-1949,” in The Cambridge History of China, vol. 13: Republican China 1912-1949, Part 2, 723-88. 63) The return to normalcy that followed the communist victory brought a similar short- term resumption of ritual activities in rural Hebei. Stephen Jones notes that following the communist takeover of Gaoluo 高洛 village (south of Beijing) “ritual associations through- out the area were enjoying a certain revival,” see his Plucking the Winds: Lives of Village Musi- cians in Old and New China (Leiden: Chime Foundation, 2004), 112.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 224 Meir Shahar

Conclusion The Horse King manuscripts prove the inseparability of social organiza- tion, religious practice, and cultural production. The Dengcun villagers were treated by a cooperative of their members to theatrical perfor- mances mounted in a religious establishment. A social organization based in a temple took care of the villagers’ entertainment. A traditional village might be likened to a tripod in which social organization, reli- gious observance, and artistic creativity were mutually reinforcing. When one leg was severed, the vessel collapsed. This is what happened when the communist state entered the village, replacing its age-old in- stitutions with new socio-political entities. When the Horse King asso- ciation dissolved, its temple fell into ruins. The plays that for centuries had entertained and edified the villagers could no longer be performed. The dilapidated remains of the Horse King temple are a physical mani- festation of social disintegration and cultural loss. The fate of the Dengcun Horse King temple was not universal. In some regions of China, religious activities have continued to the present. In recent years, other localities have experienced a revival of ritual and theatrical traditions.64 Nonetheless, many Shanxi villages evince the same religious and cultural void as Dengcun. Situated some thirty miles from it, Renzhuang village was in 1909 home to no fewer than nineteen temples, none of which survives. David Johnson’s comments are appli- cable to both communities: “Students of popular mentalities in modern China might wish to consider the implications for the collective psy- chology of the people of Renzhuang of the obliteration of all their tem-

64) This appears to be the case in northern Shaanxi (Shaanbei); see Adam Yuet Chau, Mi- raculous Response: Doing Popular Religion in Contemporary China (Stanford: Stanford Univ. Press, 2006). Discontinued since 1957, the cult of the Lady of Yao Mountain (in Pucheng County, central Shaanxi) was revived in the 1990s; see Qin Jianming and Lü Min, Yaoshan shengmu, and Bujard, “Construction, organisation et histoire.” On the uneven revival of tem- ples (and sectarian activities) in Cang County 滄縣, southeastern Hebei see Thomas David DuBois, The Sacred Village: Social Change and Religious Life in Rural North China (Honolulu: Univ. of Hawai’i Press, 2005), 41, 54-56. Stephen Jones similarly records both decline and re- vival of religious activities in Laishui county 淶水縣, central Hebei; see his Plucking the Winds. Consult also Paul Katz, Religion in China and its Modern Fate (Waltham, Mass.: Bran- deis Univ. Press, 2014), and Vincent Goossaert and David A. Palmer, The Religious Question in Modern China (Chicago: Univ. of Chicago Press, 2011).

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 225 ples in the space of a single generation, and ponder as well the total absence of any efforts to restore them.”65 Historians and sociologists marveled at the degree of autonomy that was enjoyed by late imperial Chinese villages. Naitō Konan 內藤湖南 (1866-1934) argued that the villagers took care of each and every aspect of their lives, from education and defense to relief and tax collection. The state did not penetrate the local community, which managed its own affairs. Village associations (xiangtuan 鄉團), referred to by Naitō as “self-governing bodies” (zizhi tuanti 自治團體), were the fundamen- tal building blocks of Chinese society.66 Writing at approximately the same time, Max Weber (1864-1920) contrasted the Chinese city (which was administered by a government appointee) to the Chinese village (which ran its affairs independently of state intervention): “A city was the seat of the mandarin and was not self-governing; a village was a self- governing settlement without a mandarin.”67 The newly-discovered Horse King manuscripts offer insights into the workings of a late imperial village association. Two principles are paramount in the Horse King notebooks: transparency and collective decision-making. It was tremendously important to show community members exactly how much money was levied from them and precisely how it was spent. It was similarly crucial to reach – or at least present the appearance of having reached – village-wide consensus. The em- phasis upon collective decision making runs through the manuscripts, from the very first page that records the establishment of the Horse King association: “Therefore, we invited everyone for public discussion at the temple. We came to the conclusion that it would be better to select by rotation the teams [of ritual heads] in charge of the [temple] accounts. No one was unhappy with this decision (wu bu xinran 無不欣然).”

65) Johnson, Spectacle and Sacrifice, 27. 66) Naitō Konan 內藤湖南, “Shinaron” 支那論 (1914), in Naitō Konan zenshū 內藤湖南全集, ed. Naitō Kenkichi 內藤乾吉 and Kanda Kiichirō 神田喜一郎 (Tokyo: Chikuma shobō, 1969-1976), 5: 133-38; for an analysis see Joshua Fogel, Politics and Sinology: The Case of Naitō Konan (1866-1934) (Cambridge: Council on East Asian Studies, 1984), 185-87. 67) Max Weber, The Religion of China: Confucianism and Taoism, trans. Hans H. Gerth (New York: Free Press, 1968), 91. Huaiyin Li suggests that village self-governance served the inter- ests of the state, which for ideological and practical reasons alike preferred not to intervene in local affairs. Cooperating with self-elected village authorities proved more efficacious for state purposes than coercion from above; see his Village Governance, 256-60.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 226 Meir Shahar

Sydney Gamble conducted his fieldwork in northern Chinese villages in the early 1930s. Publishing his findings some thirty years later, he struck a nostalgic note: “All of our material describes what is now a by- gone period in north China’s history, but as history and as a description of factors in the life of the villages during the period from the late 1800s to the early 1930s, we feel that it depicts the large amount and wide vari- ety of democratic activity that went on at the village level.”68 Semantically overloaded with ancient and contemporary Western connotations, the adjective “democratic” might be misleading. However, the terms “com- munal” or “cooperative” would do nicely to describe the organizations studied by Gamble, and documented by the newly-discovered Horse King manuscripts.

Appendix: The Stele Inscription Unearthed at the Dengcun Horse King Temple (The individual names that are marked in bold type also appear in the Horse King manuscripts) 修城根兼補修廟宇碑記 聞之詩曰: 無俾城壞,孟子云: 築斯城也。凡以城郭不完,則室家錐云,保聚 究不得恃為安堵無恐也□□□見村之西南城隅,因雷水沖壞,協合鄉約,鄉約 會同村人在村收麥數石,聚石以修理之,增高五尺,并修三王廟,諸神殿宇, 戲臺,門樓,廟內,城根。工成告竣之日,因將捐銀姓氏開於碑左。 張正渠施錢拾千文,馬王社施錢六千文,鄧村渠施錢一千三百文。 趙奇昇,王錫瑛,王殿魁,王保清,王成龍,王思溫,關清元,王吉慶,吳 天祐,王長春,梁有福,王思明,王天貴,關興旺,崔茂雲,王肯堂,王思 聰,王金德,崔書臨,王執中,崔書達,關大英,崔長榮,崔林龍,關茂松, 關貴元,王貴興,崔書清,何文杰,王奎□,王思恭,關泰和,梁玉盛,梁玉 鐘,關連陞。以上各施麥七升。 王錫珍,梁義孟,崔林福,梁有康,崔□成 ….. [Several lines of names cannot be deciphered, the stele having been intentionally damaged]…. 以上各施麥三□ 以上施錢拾七五百文,共麥貲錢拾七千零一七七文。石頭使錢拾錢文,石灰 使錢四千零三十一文,釘子使錢壹百五十五文,廟□泥沟磚瓦共施錢貳錢零二 十文,匠工使錢壹拾三千三佰八十文,敬匠人□錢七佰三十四文,買碑刻字寫 碑共使錢貳十六百文。以上七宗共使錢叁拾貳千九百二十文,下餘錢敬 神立碑 使□

68) Gamble, North China Villages, viii (the emphasis is mine).

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access Newly-Discovered Manuscripts 227

鄉約首事人 王思□,王學金,吳三□,王肯堂,王成龍,崔瑞成,王貴清,梁 有文,梁有康,崔□□,梁義成,崔長泰,□□麒,梁有讓。69 咸豐拾壹年正月吉日立

Acknowledgments I am deeply grateful to Ye Derong 葉德榮 for his generosity. I am ­similarly indebted to Marianne Bujard, Hou Chong 侯沖, Hou Huiming 侯慧明, Hannibal Taubes, and Yao Chunmin 姚春敏. Mark Gamsa provided val- uable comments on an early draft, and my son David Shahar drew the maps. My deepest thanks are also directed to the anonymous readers for T’oung Pao. My research benefitted from the support of the Israel Sci- ence Foundation (Grant no. 325/15) and from the hospitality of the Fon- dation maison des sciences de l’homme (Paris).

Abstract Written documents from rural north China are rare. This essay examines the newly- discovered records of a Shanxi village association, which was dedicated to the cult of the Horse King. The manuscripts detail the activities, revenues, and expendi- tures of the Horse King temple association over a hundred-year period (from 1852 until 1956). The essay examines them from social, cultural, and religious perspec- tives. The manuscripts reveal the internal workings and communal values of a late imperial village association. They unravel the social and economic structure of the village and the centrality of theater in rural culture. Furthermore, the manuscripts bring to the fore a forgotten cult and its ecological background: the Horse King was among the most widely worshiped deities of late imperial China, his flourishing cult reflecting the significance of his protégés – horses, donkeys, and mules – in the agrarian economy.

Résumé Les documents écrits anciens sont rares pour la société rurale de la Chine du nord. L’article examine les manuscrits récemment découverts d’une association villa-

69) 王思□ might be either 王思溫 or 王思恭 who are mentioned as association members in the Horse King manuscripts. □□麒 might be 關夢麒 who likewise appears in the manu- scripts.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access 228 Meir Shahar geoise du Shanxi, consacrée au culte du Roi Equin. Ces textes documentent en détail les activités, les revenus et les dépenses de l’association du temple du Roi Equin pendant un siècle (de 1852 à 1956) ; ils sont ici étudiés du point de vue de l’histoire sociale, culturelle et religieuse. Ils révèlent le fonctionnement interne et les valeurs d’une communauté villageoise à la fin de l’ère impériale, et mettent en lumière la structure sociale et économique du village ainsi que le rôle central du théâtre dans la culture rurale. De plus, les manuscrits attirent l’attention sur un culte négligé par la recherche, ainsi que son contexte écologique : le Roi Equin était l’une des divinités les plus vénérées dans le monde chinois à la fin de l’empire, son culte reflétant l’importance de ses protégés (chevaux, ânes, mules et mulets) dans l’économie agraire.

提要 中國北部鄉村的書面文獻罕有存世。本文檢視了一組新近發現的來自某山西村 落組織致力於馬王教團體活動的記錄。這些手寫文獻詳細記錄了一個世紀里 (1852-1956) 馬王廟組織的活動、收入和支出。本文從社會、文化和宗教角度對 它們加以檢視。這些手寫文獻呈現了一個帝國晚期村落組織的內部運作及其所 共有的價值觀念。它們揭示了村落的社會和經濟結構,以及劇場在鄉村文化中 的中心地位。此外,這些手寫文獻呈現出了一個被遺忘的異教組織及其生態背 景:馬王是帝國晚期中國最被廣爲祭拜的神祗之一,其香火的興盛反映了他所 司職的對象——馬、驢、騾——在農耕經濟中的重要性。

Keywords North China villages – Shanxi province – Chinese religion – animals and religion – Horse King – temple associations.

T’oung Pao 105 (2019) 183-228

Downloaded from Brill.com10/02/2021 11:41:16PM via free access