A Study of the zhao-mu System from Its Possible Origins to the Han, with an Extended Discussion on the Posthumous Names of the Shang Kings (2017 draft) Matthew M. Anderson The zhao-mu 昭穆 system has been frequently written about for over two thousand years, but much remains unclear about it. The traditional discussions of this system primarily (though not exclusively) concern themselves with the arrangement of burial sites and ancestral temples.1 In the traditional understanding of the system, in the Western Zhou, at the most basic level, fathers and east (that is, left, from the point of view of the south-facing ancestor) were zhao, and sons and west (right) were mu. During the Han, the left–right orientation is said to have been reversed, but the description otherwise should still apply. In the case of ancestral temples, the system is said to have comprised a fixed number of temples, consisting, in most cases, of one permanent temple for the founding ancestor (taizu 太祖) and a set number of temples dedicated to other ancestors. To oversimplify, each time a lineage head died, one of these temples, that of the most distant ancestor who was not the founding ancestor, would be dismantled (hui 毀) and a new temple would be set up for the recently deceased lineage head, leaving the overall number of temples the same. The ancestral tablets (zhu 主) from the dismantled temples were either buried or displayed in the temple of the founding ancestor. This paper represents an attempt to better understand how this system worked during the early imperial period, especially the Western Han and the following Xin 新 dynasty of Wang Mang 王莽, and to trace this system 1 Throughout this paper, I translate miao 廟 as “temple,” though a number of other scholars, including, e.g., Kenneth Brashier and Michael Loewe, render miao in this context as “shrine.” I reserve the English word “shrine” for the word ci 祠. 1 back as far as is possible, with extended discussions of certain aspects of the ritual systems of the Shang and Zhou periods. The natural place to begin is to consult the three ritual compendia which are now included in the Thirteen Classics—the Zhou li 周禮 (the Rites of Zhou), the Yi li 儀禮 (Ceremonial Rites), and the Li ji (Records of Rites).2 Though the dates of particular passages in these texts are by no means clear, the Li ji is perhaps overall the latest of the three (it probably dates to the Han, though it likely does include some earlier material), but it offers the simplest description of the zhao-mu system. Two passages, among a number of other discussions of the system, are especially useful; one describes the purpose of zhao and mu: 夫祭有昭 穆,昭穆者,所以別父子、遠近、長幼、親疏之序而無亂也。是 故,有事於大廟,則群昭群穆咸在而不失其倫。此之謂親疏之殺也。 Sacrificial rites have zhao and mu. Through zhao and mu, the relationships between fathers and sons, distant and near, old and young, and closely related and distantly related are distinguished and are not confused. Because of this, when services are held at the great ancestral temple, the many zhao and the many mu are all present and none lose their proper place. This is what is called the succession of the distance in relationships (between relatives).3 and the other claims to give the numbers of temples used by different social groups during the Western Zhou: 天子七廟,三昭三穆,與大祖之廟而七。諸侯五廟,二昭二穆,與大祖之廟 而五。大夫三廟,一昭一穆,與大祖之廟而三。士一廟。庶人祭於寢。 2 As Michael Loewe points out, the earliest appearance of the term might be its appearance in the “Lu yu” 魯語 section of the Guo yu 國語, but the passages in these ritual classics more effectively describe the system. See Michael Loewe, Problems of Han Administration: Ancestral Rites, Weights and Measures, and the Means of Protest (Leiden: Brill, 2016), 6. He provides a translation of the passage in question on pp. 40–41. 3 Li ji xunzuan 禮記訓纂, ed. Zhu Bin 朱彬 (Beijing: Zhonghua, 1996), 25.729 (“Ji tong” 祭統). 2 As for the seven ancestral temples of the Son of Heaven: there are three zhao and three mu, which together with the ancestral temple of the founding ancestor make seven. As for the five ancestral temples of the many lords: there are two zhao and two mu, which together with the ancestral temple of the founding ancestor make five. As for the three ancestral temples of the Grand Masters: there are one zhao and one mu, which together with the ancestral temple to the founding ancestor make three. Men of Service have one ancestral temple. The common people offer sacrifice in their living quarters.4 There are only a few relevant passages in the Zhou li. Wang Mang, in modeling his government after the Duke of Zhou 周公, heavily relied upon the Zhou li, as it was at that time thought to have been compiled by the Duke of Zhou, though this is not explicitly stated (in any extant text) until significantly later.5 It has often been claimed that the Zhou li was forged by Liu Xin 劉歆 (46 B.C.E.–23 C.E.), but it seems most likely that it dates to the late Warring States period.6 The most interesting example from this text describes the official duties of the Grave Makers7 (zhongren 冢人), who were responsible for preparing and maintaining the official burial grounds: 冢人掌公墓之地,辨其兆域而為之圖。先王之葬居中,以昭穆爲左右。凡諸 侯居左右以前,卿、大夫、士居後,各以其族。凡死於兵者,不入兆域。凡 有功者,居前。以爵等爲丘封之度與其樹數。大喪既有日,請度甫竁,遂爲 之尸。及竁,以度爲丘隧,共喪之窆器。及葬,言鸞車象人。及窆,執斧以 4 Li ji xunzuan, 5.183–85 (“Wang zhi” 王制). William G. Boltz, “Chou li 周禮,” in Early Chinese Texts: A Bibliographical Guide, ed. Michael Loewe (Berkeley: The Society for the Study of Early China and the Institute of East Asian Studies, University of California, Berkeley, 1993), 26–27. 6 Boltz, “Chou li 周禮,” 26–29. 7 I follow Charles Hucker’s translation of the official title zhongren 冢人 here. See Charles O. Hucker, A Dictionary of Official Titles in Imperial China (Taipei: Southern Materials Center, 1985), 191. 5 3 涖。遂入,藏凶器。正墓位,蹕墓域,守墓禁。凡祭墓,爲尸。凡諸侯及諸 臣葬於墓者,授之兆,爲之蹕,均其禁。 Grave Makers are in charge of the cemetery of the ruler. They select the burial ground and draw diagrams of it. The grave of the first ruler8 is in the center, and (the tombs of the later rulers) are (located) to the left and right according to zhao and mu. All the many lords (are buried) to the left and right in front (of the first ruler), and the Ministers, Grand Masters, and Men of Service (are buried) to the back (of the first ruler), each according to their lineage. All those who die in battle are not interred in the burial ground. All those who have merit are buried in front (of the king’s tomb). The dimensions of mounds and their number of trees are determined according to emolument and rank. When the date of grand burials has already been decided, (the Grave Makers) request to make the measurements and begin excavation and thereafter serve as the impersonators of the dead. Upon excavation, according to the dimensions of the mound and the subterranean passageway, they provide the tomb’s burial vessels. Upon (the beginning of) the funeral, they report to the figurines on the ceremonial chariot. Upon burial, they grasp hatchets and attend (to the service) and thereupon inter the burial goods. They rectify the position of the grave, prohibit (unauthorized persons from entering) the burial ground, and enforce the prohibitions of the grave. In all cases of tomb sacrifices, (the Grave Makers) are the impersonators of the dead. When all the many lords and the many ministers are buried in the burial ground, (the 8 “First ruler” (xianwang 先王) here could also be understood as “former kings”; in this case, “former kings” would include both King Wen 文王 and King Wu 武王 of Zhou, as, according to some understandings of the zhao-mu system as it was applied during the Western Zhou, as will be discussed further below, both of these kings would have had permanent ancestral shrines (that is, together with their ultimate ancestor Hou Ji 后稷, who could not possibly have been interred in the royal burial ground, they would have jointly been treated as the founding ancestor for the purposes of the zhao-mu system). If this principle were further extended to tombs, they would both have been buried in the center, neither zhao nor mu, with the following kings only, starting from King Cheng 成王, arranged in zhao-mu order. As this passage seems to be giving a generic description of the proper rites, however, I take xianwang here simply to mean “first ruler.” 4 Grave Makers) provide (the plans) for their graves, prohibit (unauthorized persons from entering) them, and enforce all their prohibitions.9 While the Li ji’s discussion concentrates on the zhao-mu system as it was supposedly applied to ancestral temples, this passage extends the system to cover tomb placement. Some scholars have argued that the zhao-mu system did not control ancestral temple placement or burial placement at all, but that zhao and mu instead simply denote the senior (zhao) and junior (mu) members of a lineage.10 Assuming the above cited passages are accurate reflections of Western Zhou practice (which is certainly not necessarily the case), this theory is difficult, but not impossible to reconcile with the second passage quoted from the Li ji and the passage drawn from the Zhou li. However, it is compatible with the first passage cited above from the Li ji, and it also fits with the only discussion of zhao and mu that I have been able to find in the Yi li, which is perhaps the oldest of the ritual classics, with some of its material possibly dating back to as early as the late Springs and Autumns.11 Both of these passages, however, by no means rule out the possibility that the zhao-mu system was primarily used for arranging ancestral temples or burials, or any of the other theories to be discussed below. The 9 Zhou li zhushu 周禮注疏, ed. Zheng Xuan 鄭玄, Jia Gongyan 賈公彥, and Peng Lin 彭林 (Shanghai: Shanghai Guji, 2010), 24.817–23 (“Chun guan zong bo: Zhongren” 春官宗伯: 冢人). 10 A representative example is Chen Xiaofang 陳筱芳, “Zhao-mu zhidu yiyi” 昭穆制度異議, Shixue yuekan 史學月刊 1 (2010): 17–26; see especially pp. 21 and 24. I find the argument unconvincing, but it is difficult to categorically refute it. 11 The extant Yi li seems to be only the remnants of a larger pre-Han collection, mostly made up of Warring States material, but possibly including some material dating back to the time of Confucius. See William G. Boltz, “I li 儀禮,” in Early Chinese Texts: A Bibliographical Guide, ed. Michael Loewe (Berkeley: The Society for the Study of Early China and the Institute of East Asian Studies, University of California, Berkeley, 1993), 237. 5 passage from the Yi li is taken from one of the many notes (ji 記)12 appended to the “Pin li” 聘禮 (Rites of State Visits) section of the text: 賜饔,唯羹飪。筮一尸,若昭若穆。僕爲祝。祝曰:“孝孫某,孝子某,薦 嘉禮于皇祖某甫,皇考某子。”如饋食之禮。假器於大夫。朌肉及廋、車。 When sacrificial meat is offered, use only the stew of the cooked meat. Divine about the one who will be impersonated, (selecting an ancestor from) either a zhao or mu (generation) (that is, selecting the host’s deceased father, or if the father is living, selecting the host’s deceased grandfather, etc.). A servant is the invocator. He invocates, saying, “filial grandson X or filial son X offers sacrifice (in accordance with) the excellent rites to his august deceased grandfather Father X or his august deceased father Master X.” This resembles the rites of offering the Lesser Sacrifice.13 Borrow the sacrificial vessels from a Grand Master. Award the meat (to attendants, including) Horsemen and Chariot Masters.14 This passage indeed does not seem to be using the terms zhao and mu to refer to the positions of the deceased ancestors’ ancestral tablets, but instead simply to identify zhao and mu generations. I will reiterate here, though, that this does not present evidence against the existence of a specific zhao-mu ritual system, but could instead simply be an example of an extended (or original) usage of the terms zhao and mu; this kind of phenomenon is common with ritual terminology from the Classical and Pre-Classical periods. Before moving on to discuss the ways in which the zhao-mu system may or may not have been employed during the Han, it will be helpful to attempt to come to terms with the concept’s 12 There is no reason to believe that the “Notes” appended to many of the sections of the Yi li date to a later period than the compilation of the sections they accompany. See Boltz, “I li 儀禮,” 236. 13 I follow the commentators here in understanding the phrase “kui shi zhi li” 饋食之禮 as referring to the shaolao 少牢 (Lesser Sacrifice). 14 Yi li zhushu 儀禮注疏, ed. Zheng Xuan 鄭玄, Jia Gongyan 賈公彥, and Wang Hui 王輝 (Shanghai: Shanghai Guji, 2008), 24.743–44 (“Pin li” 聘禮). 6 origins during the Zhou period, as well as to examine the possibility that it might descend from earlier Shang ritual systems, making use, when possible, of the scattered archaeological evidence. The most famous attempt to establish that the origins of the zhao-mu system stretch back to the Shang period probably remains K.C. Chang’s. There are many problems with this attempt, but it remains influential,15 so it is worth discussing at length. First, some background information may be useful. The posthumous names of the Shang kings (as well as the names of the six predynastic ancestors who immediately preceded the first king, Da Yi 大乙) all contain a heavenly stem (tiangan 天干)16 as their last element. Since heavenly stems are used to name days in the Shang calendar, two previous theories as to the origins of these kings’ names are that the kings were named after the days on which they were born or on which they died. Chang conclusively demonstrates that these names could not have been randomly assigned by making two points. One is that the names of Shang kings are distributed in a way that is very unlikely to be random.17 Counting predynastic ancestors, the 15 Hwang Ming-chorng, for example, has produced a significant amount of work which ultimately derives from (though does not agree in detail with) K.C. Chang’s theory that the ten tiangan 天干, when used in the shrine names of Shang kings, refer to ten separate but closely-related lineages; this concept is the foundation for Chang’s ideas about the origins of the zhao-mu system and is tightly bound up with those ideas. See, among others, Hwang Ming-chorng (Huang Mingchong) 黃銘崇, “Shangren jisi yong de qinshu chengwei tixi jiqi yiyi” 商人祭祀用的親屬稱謂體系及其意義, in Guwenzi yu gudai shi, di yi ji 古文字與古代史·第一輯, ed. Chen Zhaorong 陳昭容 (Taipei: Zhongyang Yanjiuyuan Lishi Yuyan Yanjiusuo, 2007), 139–79, and Hwang Ming-chorng (Huang Mingchong) 黃銘崇, “Shangren rigan wei shengcheng yiji tong gan bu hun de yiyi” 商人日干為生稱以及同干不婚的意義, Zhongyang Yanjiuyuan Lishi Yuyan Yanjiusuo jikan 中央研究院歷史語言研究所集刊 78.4 (2007): 705–54. See also Li Hengmei’s work, including Li Hengmei 李衡眉, Zhao-mu zhidu yanjiu 昭穆制度研究 (Jinan: Qilu, 1996), 216–18, for one example, to be discussed more fully below. 16 The ten tiangan are, in order, jia 甲, yi 乙, bing 丙, ding 丁, wu 戊, ji 己, geng 庚, xin 辛, ren 壬, and gui 癸. 17 This discussion can be found in Chang Kwang-chih (Zhang Guangzhi) 張光直, Zhongguo qingtong shidai 中國青銅時代 (Taipei: Lianjing, 1983), 159ff. A fuller discussion of these issues can be found in Adam Smith, “The Chinese Sexagenary Cycle and the Ritual Foundations of the Calendar,” in Calendars and Years II: Astronomy and Time in the Ancient and Medieval World, ed. John M. Steele (Oxford: Oxbow, 2011), 5, 12–13, and throughout. 7 names of 21 Shang rulers contain only three tiangan (jia 甲, yi 乙, and ding 丁), while the names of the remaining 16 rulers are written with the remaining seven tiangan.18 Chang also points out that, in a random selection of 1,295 (mostly or exclusively non-royal) bronze inscriptions which feature heavenly stems as parts of names, only five heavenly stems appear on 1,134 of them (86%).19 This could not possibly be random. Note, too, that the distribution of royal tiangan names is quite different from the distribution on these bronzes, in a way that is unlikely to be random. Chang argues instead that the Shang ruling family was probably subdivided into ten separate lineages, each known (at least in death) by one of the heavenly stems. In this argument, the yi and ding lineages had the most power; the jia, wu 戊, and ji 己 lineages allied with yi (which we can call Group A), and the bing, ren 壬, and gui 癸 lineages allied with ding. Chang calls the geng 庚 and xin 辛 lineages “neutralists.”20 If one follows his argument, yi and ding kings could not take consorts from yi or ding lineages, as this would represent too great a concentration of power; instead kings took consorts of lower status.21 Royal sons, then, had mothers who were from neither yi nor ding lineages, so they did not themselves become king. Instead, a son of the king’s brother by a higher status consort would take the throne. That is, kingship passed from maternal uncles to sororial nephews in two generations.22 Chang allows that this seems an unusual form of royal succession, but he mentions some possible comparisons 18 If predynastic ancestors are not counted, the names of 18 kings are written with the three tiangan mentioned above, and the remaining 13 are written with the other seven tiangan. 19 These are the five “even” tiangan: yi, ding, ji, xin, and gui. See Kwang-chih Chang, Shang Civilization (New Haven: Yale University Press, 1980), 169. 20 Kwang-chih Chang, “Some Dualistic Phenomena in Shang Society,” Journal of Asian Studies, 24.1 (Nov. 1964), 49. 21 Chang, “Some Dualistic Phenomena in Shang Society,” 50. 22 Chang, “Some Dualistic Phenomena in Shang Society,” 50. 8 from other cultures to demonstrate that this sort of system is at least possible.23 The result then, is that kingship alternated between Group A and Group B, with so-called “neutralists” able to take the position of either A or B. Additionally, if Group A was in power, the king’s most important minister would be from Group B, and vice versa. Some illustrations will be helpful in demonstrating Chang’s view of this system. The Shang king list was preserved in the “Yin benji” 殷本紀 section of the Shi ji 史記, and what could be constructed from Shang inscriptions has, for the most part, confirmed the account given in the Shi ji.24 Chang produces the following table, representing the Shi ji’s version of the Shang king list25: 23 Chang, “Some Dualistic Phenomena in Shang Society,” 50–51. See Chang Kwang-chih, Zhongguo qingtong shidai, 155–59 for an overview. This chapter was originally published separately in 1963. 25 The names of kings differ on the charts below in various superficial ways; these should not have an affect on my discussion below, but I will list the variants here in the interest of clarity. Wei 微 is another name for Shang Jia 上甲. Bao Yi 報乙, Bao Bing 報丙, and Bao Ding 報丁 are alternate forms of 匚乙, 匚丙, and 匚丁, respectively. Zhu Ren 主壬 and Zhu Gui 主癸 are different names for Shi Ren 示壬 and Shi Gui 示癸. Tian Yi 天乙 is interchangeable with Da Yi 大乙. Otherwise, the graphs Da 大 and Tai 太 are interchangeable in names. Yong Ji 雍己 and Lü Ji 呂己 are interchangeable. Zhong Ding 仲丁 and Zhong Ding 中丁 are interchangeable. Wai 外 and Bu 卜 are interchangeable in names. He Dan Jia 河亶甲 and Jian Jia 戔甲 are different names for the same person. Wo Jia 沃甲 and Qiang Jia 羌甲 refer to the same person. Hu Jia 虎甲, Yang Jia 陽甲, and Xiang Jia 象甲 are all the same person. Pan Geng 盤庚 can also be written 般庚. Tai Ding 太丁 (number 34 in the series, not number 8) is also known as Wen Wu Ding 文武丁. I also provide king numbers (prefixed with K-). These do not conflict with Chang’s numbers but simply provide additional information—that is, they begin counting from the first king (Da Yi/Tian Yi; known in many received texts as Cheng Tang 成湯) instead of from the first predynastic ancestor (Shang Jia Wei). 24 9 Figure 1. Table from Chang Kwang-chih, Zhongguo qingtong shidai, 156 A revised version of the above table, taking into account information from Shang inscriptions: 10 Figure 2. Table from Chang Kwang-chih, Zhongguo qingtong shidai, 171 Chang then divides this chart up into three columns, representing Group A, Group B, and, sandwiched in between them, the “neutralists.” Each row represents one generation, by Chang’s reckoning: 11 Generation 世代 (each row represents one generation) (每格代表一世) Group A A組 Irregular and temporarily ungrouped temple names 不和規律及暫不分組之廟號 Group B B組 Shang Jia 上甲 Bao Yi 匚乙 Bao Bing 匚丙 Bao Ding 匚丁 Shi Ren 示壬 Shi Gui 示癸 Da Yi 大乙 Da Ding 大丁, Wai Bing 外丙, [Zhong Ren 中壬] Da Jia 大甲 Da Geng 大庚 [Wo Ding 沃丁] Xiao Jia 小甲, Da Wu 大戊, Yong Ji 雍己 Jian Jia 戔甲 Zhong Ding 中丁, Wai Ren 外壬 Zu Yi 祖乙 Hu Jia 虎甲, Xiao Yi 小乙 Qiang Jia 羌甲 Zu Ding 祖丁, Nan Geng 南庚 Pan Geng 盤庚, Xiao Xin 小辛 Zu Ji 祖己, Zu Jia 祖甲 Zu Geng 祖庚 Zu Xin 祖辛 Wu Ding 武丁 Lin Xin 廩辛, Kang Ding 康丁 Wu Yi 武乙 Wen Wu Ding 文武丁 Di Yi 帝乙 Di Xin 帝辛 Table 1. After Chang Kwang-chih, Zhongguo qingtong shidai, 172–73 Note that this diagram does not strictly follow Chang’s description. Several of the predynastic ancestors (Bao Yi, Bao Bing, and Shi Ren) are forced into the neutralist category when they would more naturally belong to A or B.26 Additionally, three kings, Jian Jia 戔甲, Qiang Jia 羌甲, and Zu Ding 祖丁, are put into the neutral category when they would seem to belong to A or B. A different view of essentially the same chart may help show the movement between groups A and B that Chang has in mind: 26 Again, Chang’s A consists of jia, yi, wu, and ji, and his B consists of bing, ding, ren, and gui; only geng and xin are neutral. 12 Figure 3. From Chang Kwang-chih, Zhongguo qingtong shidai, 21727 This idea is interesting, and there is certainly something to it, but there seem to me to be too many exceptions necessary for the entire system to work. I took a slightly revised 27 The chart is labeled with the tiangan, written in Shang script, in an unusual order across the top: it reads [from group A:] ji, wu, yi, jia; [from the “neutralists”:] geng, xin; and [from group B:] bing, ding, ren, and gui. 13 chronology28 and prepared the following chart, presented without making any exceptions to my understanding of Chang’s system29: Group A Unaligned Group B 甲乙戊己 庚辛 丙丁壬癸 1. Shang Jia 上甲† 2. Bao Yi 匚乙† 3. Bao Bing 匚丙† 4. Bao Ding 匚丁† 5. Shi Ren 示壬† 6. Shi Gui 示癸† 7. (K1) Da Yi 大乙 8. (K2) Da Ding 大丁 9. (K3) Da Jia 大甲 10. (K4) Bu Bing 卜丙‡ 12. (K6) Xiao Jia 小甲‡ 11. (K5) Da Geng 大庚 13. (K7) Da Wu 大戊 14. (K8) Lü Ji 呂己‡ 18. (K12) Zu Yi 祖乙 17. (K11) Jian Jia 戔甲‡ 15. (K9) Zhong Ding 中丁 16. (K10) Bu Ren 卜壬‡ 20. (K14) Qiang Jia 羌甲‡ 19. (K13) Zu Xin 祖辛 22. (K16) Nan Geng 南庚‡ 26. (K20) Xiao Yi 小乙 23. (K17) Xiang Jia 象甲‡ 21. (K15) Zu Ding 祖丁 24. (K18) Pan Geng 般庚‡ 25. (K19) Xiao Xin 小辛‡ 27. (K21) Wu Ding 武丁 30. (K23) Zu Jia 祖甲 28. (K0) Zu Ji 祖己* 29. (K22) Zu Geng 祖庚‡ 31. (K23a) Lin Xin 廩辛‡ 32. (K24) Kang Ding 康丁 33. (K25) Wu Yi 武乙 34. (K26) Wen Wu Ding 文武丁 35. (K27) Di Yi 帝乙 36. (K28) Di Xin 帝辛 Table 2. 28 Following the divisions between generations given in David N. Keightley, Working for His Majesty: Research Notes on Labor Mobilization in Late Shang China (ca. 1200–1045 B.C.), as Seen in the Oracle-Bone Inscriptions, with Particular Attention to Handicraft Industries, Agriculture, Warfare, Hunting, Construction, and the Shang’s Legacies (Berkeley: Institute of East Asian Studies, University of California, Berkeley, Center for Chinese Studies, 2012) 244–45. Also see David N. Keightley, Sources of Shang History: The Oracle-Bone Inscriptions of Bronze Age China (Berkeley: University of California Press, 1978), 185–86. 29 Each row represents one generation. 14 This table does not, at first glance, seem to very neatly divide up into Chang’s groups. I have appended symbols to the ends of some of the names on the chart for ease of understanding. The names followed by a single dagger (†) (the ancestors numbered 1–6) are the predynastic ancestors. There seems to me to be no need to force some of these ancestors into the neutral category, as it is clear that, whatever system the kings are named by, the predynastic ancestors are being named by a different system. If one converts the tiangan into numbers, these ancestors are named in a simple sequence: 1, 2, 3, 4, 9, 10. I have no explanation for the jump from 4 to 9, but there is no dualism evident in this section of the ancestral line. The king whose name is followed by an asterisk (*) (ancestor number 28) is likely never to have taken the throne and can be skipped over.30 Other names are followed by a double dagger (‡) (the ancestors numbered 10, 12, 14, 16, 17, 20, 22, 23, 24, 25, and 29). Note that in every generation in which one or more double dagger appears, exactly one name appears without a following symbol. The unmarked names are in the main line of descent, or the trunk lineage (da zong 大宗 or da shi 大示, as it is referred to in oracle-bone inscriptions). The names followed by double daggers are in the lesser lineage. The traditional understanding of the main line of descent is that it refers to the group of deceased kings who are both sons of and fathers of kings (so, there can naturally only be one per generation). Chang dismisses this possibility, as it is incompatible with his theory of ten separate descent groups within the royal family, and, from what I have been able to find, does not seem to consider the issue much further. 30 For Zu Ji, a son of Wu Ding, see Keightley, Sources of Shang History, 208, note ad. It is also possible that Lin Xin did not reign, as he does not appear in the later Shang ritual cycle (see Keightley, Sources of Shang History, 187, note h), but see the discussion of Qiang Jia below, which suggests that there are other possible reasons for a king’s absence in the ritual cycle. 15 If, however, one simply eliminates all kings who are not part of the main line of descent, the result is the following table: Group A Unaligned Group B 甲乙戊己 庚辛 丙丁壬癸 7. (K1) Da Yi 大乙 8. (K2) Da Ding 大丁 9. (K3) Da Jia 大甲 11. (K5) Da Geng 大庚 13. (K7) Da Wu 大戊 15. (K9) Zhong Ding 中丁 18. (K12) Zu Yi 祖乙 19. (K13) Zu Xin 祖辛 21. (K15) Zu Ding 祖丁 26. (K20) Xiao Yi 小乙 27. (K21) Wu Ding 武丁 30. (K23) Zu Jia 祖甲 32. (K24) Kang Ding 康丁 33. (K25) Wu Yi 武乙 34. (K26) Wen Wu Ding 文武丁 35. (K27) Di Yi 帝乙 36. (K28) Di Xin 帝辛 Table 3. The result is a nearly perfect presentation of (a modified version of) Chang’s theory. The kingship switches between A and B, not with every king, but with every generation. If this presentation is correct, it is unimportant which group kings who are not of the main line of descent belong to, but kings from the trunk lineage must alternate between A or B, with the neutral category still serving for either, as in Chang’s original version. A few problems do still remain, however. The transition between 18, 19, and 21 is especially problematic, as, though it 16 never happens that A or B repeats exactly, if 19 is serving for either A or B, then one of them effectively repeats. It turns out, though, that there is a simple solution to this problem. Chang himself has argued that Qiang Jia might be the son (not brother) of Zu Xin and the father (not uncle) of Zu Ding (though it is possible he is speaking generationally here, not about literal sons and fathers).31 As David Keightley goes on to point out, this would put Qiang Jia into the main line of descent; in fact Qiang Jia was treated as a member of the trunk lineage during the middle part of the late Shang period, though he was treated as a member of the lesser lineage during the early and late parts of the late Shang. There was perhaps some sort of struggle between certain of the Shang kings and Qiang Jia’s descendants.32 If one treats Qiang Jia as a member of the trunk lineage (with the understanding that certain later kings were only trying to remove him, and thus his descendants, from this important position), then this problem is resolved. We can additionally do away with the “Unaligned” column, by simply assigning the geng and xin groups to Group B: 31 32 Chang Kwang-chih, Zhongguo qingtong shidai, 163. Keightley, Sources of Shang History, 187, note f. 17 Group A 甲乙 戊(己) Group B (壬?) (丙)丁 庚辛 (癸?) 7. (K1) Da Yi 大乙 8. (K2) Da Ding 大丁 9. (K3) Da Jia 大甲 11. (K5) Da Geng 大庚 13. (K7) Da Wu 大戊 15. (K9) Zhong Ding 中丁 18. (K12) Zu Yi 祖乙 19. (K13) Zu Xin 祖辛 20. (K14) Qiang Jia 羌甲 21. (K15) Zu Ding 祖丁 26. (K20) Xiao Yi 小乙 27. (K21) Wu Ding 武丁 30. (K23) Zu Jia 祖甲 32. (K24) Kang Ding 康丁 33. (K25) Wu Yi 武乙 34. (K26) Wen Wu Ding 文武丁 35. (K27) Di Yi 帝乙 36. (K28) Di Xin 帝辛 Table 4. Of the kings on this revised list, none has bing, ji, ren, or gui in his name. It would be possible to simply eliminate ren and gui from the list33; in that case, Group A would be made up 33 The reason for eliminating ren and gui (9 and 10) but not bing or ji (3 or 6) is that the grouping of bing and ji with groups B and A, respectively, is clear from the overall paired order of the stems, but ren and gui are difficult to conclusively assign to either group. It is interesting that the stems ren and gui appear in the names of the last two pre-dynastic ancestors, Shi Ren and Shi Gui. This is only idle speculation at this point, but could it be possible that all the main trunk (da shi, written with the same graph as the first graph in these two ancestors’ names) ancestors in Group A are somehow connected with Shi Ren and all those in Group B with Shi Gui? The heading for Group A would then read 甲乙 戊己 壬 and for Group B 丙丁 庚辛 癸. Shi Ren and Shi Gui are certainly significantly more frequently sacrificed to than the three ancestors whose names begin with Bao, suggesting a special importance (see Chang 18 of jia, yi, wu, and ji, and Group B would consist of bing, ding, geng, and xin. This would divide up very neatly—if numbers were substituted for tiangan, Group A would be 1, 2, and then 5, 6 and Group B would be 3, 4, and then 7, 8. That is, it would be a perfect alternation between pairs of numbers. It is also interesting to note that the names of their consorts are known from oraclebone inscriptions for all the kings on this revised list, except for the last two (and the last king, of course, was never sacrificed to by the Shang, as the period ended with his death). The only other Shang king for whom a named consort has been identified is Wai Bing 外丙,34 the fourth king overall (of 28), who was the very first king not from the main line of descent to reign, eight or nine generations before the reign of Wu Ding, the first king to leave written records. If Wai Bing’s spouse has been correctly identified, this suggests that either the primacy of consorts to kings from the royal line of descent, or perhaps the full ritual system, was not quite settled during the earliest period. It is most likely, though, that this spouse is not correctly identified,35 in which case this aspect of the ritual record also supports the case that these kings were uniquely important, which may be why they are the only ones who seem to matter for this dualistic system. The above supports the idea that there is a strong dualistic nature to the selection of the posthumous names of Shang kings, but it does not address the question of how these names were Yuzhi 常玉芝, Shang dai shi—juan ba: Shang dai zongjiao jisi 商代宗教祭祀:商代史·卷八, ed. Song Zhenhao 宋鎮豪 (Beijing: Zhongguo Shehui kexue, 2010), 224–5 for a brief discussion of the frequencies with which these ancestors were sacrificed to). 34 A convenient list of all known royal consorts can be found in Wang Yuxin 王宇信 and Yang Shengnan 楊升南, Jiaguxue yibai nian 甲骨學一百年 (Beijing: Shehui kexue wenxian, 1999), 444. 35 The chart in Wang and Yang, Jiaguxue yibai nian lists the findings of seven scholars. Of all the kings listed, Wai Bing is the only one for whom only one of these seven scholars has identified a spouse. In almost every other case (excepting the latest two kings for which spouses are known, of which five and four scholars, respectively, agree on the names of their spouses), all seven scholars have identified at least one consort for each king. 19 selected. A variety of solutions to this problem have been suggested.36 As recently proposed by Adam Smith, the theory that perhaps best fits the data is that these names were selected by divination, as were other aspects of funerary preparation.37 Part of Smith’s evidence is an oraclebone inscription which seems to describe the selection of the day (ri 日) of a royal consort, one who other inscriptions suggest had recently deceased.38 Since the publication of this article, a new collection of oracle bone inscriptions has appeared,39 and one of the inscriptions contained within it also looks to be a record of divination about the selection of a day for a deceased member of the royal family: 己亥貞:王日隹丁,若。 On jihai (day 36), divined: The king’s day will be ding, (which will be) approved.40 This inscription, then, deals with the assignment of the day name ding to the Shang king Wu Ding.41 And it forms another piece of evidence for the idea that Shang posthumous names were chosen (at least in part) by divination. 36 A useful and concise summary of some of these theories can be found in Smith, “The Chinese Sexagenary Cycle and the Ritual Foundations of the Calendar,” 14–15. 37 See Smith, “The Chinese Sexagenary Cycle and the Ritual Foundations of the Calendar,” 15–17. 38 Smith transcribes this inscription, which is written on three separate scapulae (Heji 23712, 23713, and 23714), as “作小后 X 日,惠癸” and translates this as “In making a day for Lesser Queen X, it should be day 10/10.” Smith, “The Chinese Sexagenary Cycle and the Ritual Foundations of the Calendar,” 16. 39 Zhongguo Shehui Kexueyuan Kaogu Yanjiusuo 中國社會科學院考古研究所, eds. Yinxu Xiaotun cun zhong cun nan jiagu 殷墟小屯村中村南甲骨 (Kunming: Yunnan Renmin, 2012) (abbreviated CZN). 40 CZN 158. 41 CZN is a Li 歷 group inscription. This group of inscriptions is best dated to the reigns of kings Wu Ding and Zu Jia; in this interpretation it must date to the beginning of Zu Jia’s reign, immediately after the death of Wu Ding (or perhaps to the very end of Wu Ding’s reign). Another school of thought holds that Li group inscriptions date to the reigns of Wu Yi and Wen Wu Ding; if this were the case, then the inscription in question would be either Kang Ding (Wu Yi’s predecessor) or Wen Wu Ding himself. 20 Li Hengmei 李衡眉 is probably the current scholar who has spent the largest part of his career working on the issue of zhao and mu.42 He suggests that K.C. Chang’s theory is quite possibly correct, and that, if it is correct, a reflection of the zhao-mu system can be seen in the late Shang tombs at Xibeigang 西北崗,43 again following an earlier theory of Chang’s.44 First, here is a plan of the Shang royal cemetery: Figure 4. Plan of Xibeigang from Yang Hsi-chang, “The Shang Dynasty Cemetery System,” 51. In the Anyang period, if Lin Xin but not Zu Ji is counted,45 there were 12 kings, starting from Pan Geng; since the last king, Di Xin (also known as Zhòu 紂) was deposed by the founders of the Western Zhou, there should thus be 11 royal tombs. Eleven large cross-shaped 42 A selection of his work includes “Yinren zhao-mu zhidu shitan” 殷人昭穆制度試探, Lishi yanjiu, 1991.9, 2–6, “Lun zhao-mu zhidu” 論昭穆制度, Shehui kexue jikan 社會科學輯刊 4 (1992), 110, and “Zhao-mu zhidu yu zongfa zhidu guanxi lunlüe” 昭穆制度與宗法制度關係論略, Lishi yanjiu, 1996.2, 26–36, though most of the material contained in the above items can be found in Li Hengmei, Zhao-mu zhidu yanjiu, which is an expanded revision of his dissertation, Lun zhao-mu zhidu 論昭穆制度 (Taipei: Wenjin, 1992), originally 1991 from Jilin University. 43 Li Hengmei, Zhao-mu zhidu yanjiu, 216–17. 44 Chang, “Some Dualistic Phenomena in Shang Society,” 46ff. 45 See n.30 above for a brief discussion of Lin Xin and Zu Ji. 21 tombs are in fact found at Xibeigang, though, as Yang Hsi-chang points out, they are not necessarily all royal tombs. He suggests that only eight are large enough to qualify as royal tombs.46 Nonetheless, K.C. Chang argues that all are royal tombs. Chang goes on to identify his A group with zhao and his B group with mu. According to Chang’s system (and this is not directly contradicted by my reinterpretation of his system), four of the Anyang rulers who are potentially buried in these tombs belong to Group A (Wu Ding, Lin Xin, Kang Ding, and Wen Wu Ding) and seven belong to Group B (Pan Geng, Xiao Xin, Xiao Yi, Zu Geng, Zu Jia, Wu Yi, and Di Yi). As can be seen from the figure above, seven of the cross-shaped tombs are on one side and four are on the other. Chang associates the four tombs to the east with zhao and the seven to the west with mu,47 as does Li Hengmei.48 Li does not stop with the late Shang, however. He additionally proposes an ultimate, and much earlier, origin for the zhao-mu system, arguing that the system originated in a Neolithic system of combining two clans in marriage. As this society transitioned from boys identifying with the clan of their maternal uncles to identifying with the clans of their fathers, he argues that it became necessary to make clear the distinctions between the generations.49 Li goes on to explain the name of the zhao-mu system. It should actually be written, he argues, zhao-mu 昭㣎.50 This means, he argues, something like “a clear and fine crack,” which represents the 46 Yang Hsi-chang, “The Shang Dynasty Cemetery System,” trans. Laurence Scott Davis, in Studies of Shang Archaeology: Selected Papers from the International Conference on Shang Civilization, ed. K. C. Chang (New Haven: Yale University Press, 1986), 51. Note that K.C. Chang was the editor of the collection in which this piece appears. 47 Chang, “Some Dualistic Phenomena in Shang Society,” 52–53. 48 Li Hengmei, Zhao-mu zhidu yanjiu, 217. 49 Li Hengmei, Zhao-mu zhidu yanjiu, 67ff. 50 He points out that Duan Yucai’s 段玉裁 commentary to the Shuowen jiezi states that the graph 穆 is typically used for 㣎 in received texts (Li Hengmei, Zhao-mu zhidu yanjiu, 86). 22 border between the generations.51 The extended version of this theory reaches too deeply into the past to be supported or disproven by current evidence, so it remains beyond the scope of this paper. I do not find K.C. Chang’s exact theory satisfactory, especially as the trunk lineage is such an important part of my solution to these problems, and this particular aspect does not fit in well with Chang’s theory. I do think, however, that Chang has made a brilliant observation about the dualism present in the system of Shang king names.52 It does seem quite possible, though, that, whatever this system was, it formed (at least part of) the origin of the later zhao-mu system. Another common argument holds that the zhao-mu system was created in the Western Zhou and takes its name from Zhou kings Zhao 昭 and Mu 穆, the fourth and fifth kings of the Western Zhou.53 As Kenneth Brashier and others have pointed out, however, this does not seem likely to be the case. King Wen is generally said to belong to a mu generation in the Western Zhou system; if this is not to be believed, then neither is the existence of the zhao-mu system at the time of Kings Zhao and Mu. Following from Wen’s position, King Zhao should belong to a 51 The original definition is “mingxi er you xi de liewen” 明晰而又細的裂紋 (Li Hengmei, Zhaomu zhidu yanjiu, 86). I do not find this explanation especially convincing, but it is certainly conceivable. Some other, more far-fetched proposals have also been offered. Li Jin 李瑾, “Huaxia zongzu zhong ‘zhao-mu’ mingcheng suyuan—Xia–Yi ronghe zhi renleixue toushi” 華夏宗族中“昭穆”名稱溯源— —夏夷融合之人類學透視, Xuchang Shizhuan xuebao (shehui kexue ban) 許昌師專學報(社會科學版 ), 16.1 (1997), 76–82, for example, explains the zhao-mu system as being the result of intermarriage between the Shang and various “non-Chinese” (yi 夷) groups, with the words ‘zhao’ and ‘mu’ deriving from the names of two non-Shang groups, the Shaofang 召方 and the Mufang 目方. This is phonologically conceivable, as shao 召 and zhao 昭 are in the same xiesheng 諧聲 series, and thus could be loans for each other, and mu 目 and mu 穆 are phonologically identical (both are *mjuwk in William H. Baxter and Laurent Sagart’s reconstructions), but I nonetheless do not find the explanation to be plausible. 52 I can go no further with this at this point; that is, I do not have a satisfactory explanation for the origins of this system. 53 See, for example, Michael Loewe, “The Imperial Tombs of the Former Han Dynasty and Their Temples,” T’oung Pao, Second Series, 78.4–5 (1992), 314. 23 mu generation as well, and King Mu should be zhao.54 It seems extremely unlikely that the system would derive from these kings and yet be named in the reverse order. This argument does not establish that the system predated these kings (it could just as easily postdate them), but it does seem to render it unlikely that the system (by its given name, anyway) started at this particular point in the mid-Western Zhou. Sometimes the name is simply understood as having the literal meaning of the words normally written with those characters—zhao and mu, then, would mean “shining” and “sombre”, respectively.55 However it is understood, perhaps its origins really do lie in earlier periods (though I am not satisfied by any of the evidence presented that projects the system back into Neolithic times—this is certainly possible; it is just neither supported nor contradicted by the available evidence). No Western Zhou royal tombs have yet been identified,56 so it is impossible to use archaeological evidence to discuss the royal burial system or the kind of role that the zhao-mu system might have played in it. The tombs of Western Zhou lords, however, have been discovered. One such plot is the Tianma-Qucun 天馬曲村 site, the location of the burial ground of the lords of Jin 晉. The layout has been held up as an example of a Western Zhou tomb site which does not follow the zhao-mu system.57 David Nivison and Edward Shaughnessy 54 K. E. Brashier, Ancestral Memory in Early China (Cambridge, Mass.: The Harvard University Asia Center for the Harvard-Yenching Institute, 2011), 366–7n46. For a different perspective, see Robert H. Gassmann, Verwandtschaft und Gesellschaft im alten China: Begriffe, Strukturen und Prozesse (Bern: Peter Lang, 2006), 75. 55 See, for example, Constance A. Cook, “Wealth and the Western Zhou,” Bulletin of the School of Oriental and African Studies, University of London, 60.2 (1997), 264n82. 56 See, among others, Jessica Rawson, “Western Zhou Archaeology,” in The Cambridge History of Ancient China: From the Origins of Civilization to 221 B.C., ed. Michael Loewe and Edward L. Shaughnessy (Cambridge: Cambridge University Press, 1999), 393. 57 Jay Xu, “The Cemetery of the Western Zhou Lords of Jin,” Artibus Asiae 56.3–4 (1996), 199; cf. Lothar von Falkenhausen, Chinese Society in the Age of Confucius (1000–250 BC): The Archaeological Evidence (Los Angeles: Cotsen Institute of Archaeology, University of California, Los Angeles, 2006), 85n22, who writes, “So far, indeed, there seems to be no archaeological evidence of the use of the zhao mu system anywhere in pre-Imperial China.” 24 acknowledge that the layout of the graves appears to break “a cardinal rule” of the zhao-mu system in that one of the rulers was the younger brother of the previous ruler but, when this ruler is removed as is required by the system, a father and grandson are placed on opposite sides of the temple.58 They go on to convincingly argue, though, that this is not in fact the case, as the chronology in the Shi ji must be incorrect (and not for the first time), not only for this particular reason, but also a variety of other evidence.59 This suggests, then, that there may be at least one archaeologically excavated example of the zhao-mu system from the Western Zhou. An interesting passage from the Chunqiu Guliang zhuan 春秋穀梁傳 (The Guliang Tradition of the Springs and Autumns) makes a natural transition between the discussion of zhao-mu in the Zhou (the period the text purports to be explaining) and a discussion of zhao and mu in the Han. The passage is a commentary on one line from the Springs and Autumns: 八月丁卯,大事于大廟,躋僖公。 大事者何?大是事也,著祫、嘗。祫祭者,毀廟之主陳于大祖,未毀廟之主 皆升,合祭于大祖。躋,升也,先親而後祖也,逆祀也。逆祀,則是無昭穆 也。無昭穆,則是無祖也。無祖,則無天也。故曰:文無天。無天者,是無 天而行也。君子不以親親害尊尊,此《春秋》之義也。 In the eighth month, on the day dingmao, there was a great (sacrificial) affair at the grand ancestral temple, and (Duke) Xi’s 僖公 (tablet) was elevated. What is meant by da shi (great (sacrificial) affair)? It makes great this (sacrificial) affair and shows that it is the xia (collective) sacrifice and the chang (autumnal) sacrifice. In the xia sacrifice, the tablets of dismantled ancestral temples are displayed in (the temple of) the founding ancestor. Those tablets of ancestral 58 David Nivison and Edward L. Shaughnessy, “The Jin hou su Bells Inscription and Its Implications for the Chronology of Early China,” Early China 25 (2000), 41. Thanks to an anonymous reviewer for pointing out that Nivison and Shaughnessy’s discussion of this point is directly based on the "Jin shijia" chapter of the Shi ji, not the Tianma-Qucun site itself. 59 Nivison and Shaughnessy, “The Jin hou su Bells Inscription and Its Implications for the Chronology of Early China,” 42, 48. 25 temples which have not yet been dismantled are all elevated and sacrificed to together with (that of) the founding ancestor. Ji means to elevate; (Duke Wen) placed his father first and put back his ancestor,60 which was contrary to sacrificial (rites). (This was) contrary to sacrificial (rites), and it was thus without (= did not comply with) zhao and mu. (This) was without zhao and mu, and it was thus without ancestral (etiquette). This was without ancestral (etiquette), and it was thus without (the way of) heaven. Therefore it is said: “(Duke) Wen did not have (the way of) heaven.” To not have (the way of) heaven is to not practice (the way of) heaven. The superior man does not use that which is close to him to harm that which is venerated; this is the meaning of the Springs and Autumns.61 In the Han, then, the zhao-mu system was apparently viewed as a subject of key importance. But did the Western Han emperors follow the system? During the Western Han, rather than group imperial tombs together in a single burial ground, as described in the Zhou li and other ritual texts, and as was the case during the Shang, there was an individual funerary complex for each emperor, separated by quite some distance from the others.62 The tombs were arranged as follows: 60 This is a rather oblique reference. Duke Wen was the son of Duke Xi and the nephew of Duke Min 閔公. Dukes Xi and Min were half-brothers. After Duke Min died just into his short (two-year) reign, Duke Xi took power. What is being described here is that Duke Wen took his father’s (Duke Xi’s) tablet and placed it in the (higher) position deserved by the tablet of his uncle and his father’s predecessor, Duke Min. 61 Chunqiu Guliang jing zhuan buzhu 春秋榖梁經傳補注, ed. Zhong Wenzheng 鍾文烝 (Beijing: Zhonghua, 1996), 13.366–70 (“Wen Gong er nian” 文公二年). 62 Liu Xujie, “The Qin and Han Dynasties,” in Chinese Architecture, ed. Nancy S. Steinhardt (New Haven: Yale University Press; Beijing: New World, 2002), 49. See also Liu Qingzhu 劉慶柱, “Archaeological Discovery and Research into the Layout of the Palaces and Ancestral Temples of Han Dynasty Chang’an—A Comparative Essay on the Capital Cities of Ancient Chinese Kingdoms and Empires,” trans. Anthony Barbieri-Low, Early China 31 (2007), 113–43. 26 1. Changling (Gaodi) 2. Anling (Huidi) 3. Baling (Wendi) 4. Yangling (Jingdi) 5. Maoling (Wudi) 6. Pingling (Zhaodi) 7. Duling (Xuandi) 8. Weiling (Yuandi) 9. Yanling (Chengdi) 10. Yiling (Aidi) 11. Kangling (Pingdi) Figure 5. The Eleven Western Han Imperial Tombs (After Loewe, Divination, Mythology, and Monarchy in Han China, 275, fig. 18) The layout of these tombs does not appear completely systematic, but Michael Loewe has suggested that they may in fact be arranged by (a modified version of) zhao-mu order,63 with some of the tombs, such as those of the founding emperor and Emperor Hui, constructed in accordance with the system, but others, such as that of Emperor Wen, not so constructed.64 K.E. Brashier and others, on the other hand have convincingly argued that the case for the tombs following any form of zhao-mu order requires too many exceptions, both to the understanding of the tomb layout itself and to the versions of the zhao-mu system reported in Han-edited ritual 63 See Michael Loewe, Divination, Mythology, and Monarchy in Han China (Cambridge, Cambridge University Press, 1994), 278ff. 64 Loewe, Problems of Han Administration, 101-2. 27 texts.65 In this case, then, unlike the case of the tombs of the Western Zhou Jin lords, the tombs do not seem to be arrayed in zhao-mu order. This order, though, did become a major point of discussion in the later part of the Western Han, and, by some accounts at least, it was at least partially returned to by Wang Mang’s Xin dynasty. Before that time, though, many discussions were held about the topic. Some that are reported in the Han shu 漢書 show that many at court believed it to be of utmost importance. Some examples should suffice to show the regard in which the system was held. 至元帝時,貢禹奏言:“古者天子七廟,今孝惠、孝景廟皆親盡,宜毀。及郡國 廟不應古禮,宜正定。”天子是其議,未及施行而禹卒。永光四年,乃下詔先議 罷郡國廟,曰:“朕聞明王之御世也,遭時爲法,因事制宜。往者天下初定,遠 方未賓,因嘗所親以立宗廟,蓋建威銷萌,一民之至權也。今賴天地之靈,宗廟 之福,四方同軌,蠻貊貢職,久遵而不定,令疏遠卑賤共承尊祀,殆非皇天祖宗 之意,朕甚懼焉。傳不云乎?‘吾不與祭,如不祭。’其與將軍、列侯、中二千 石、二千石、諸大夫、博士、議郎議。” At the time of Emperor Yuan (the eighth emperor; r. 48-33 B.C.E.), Gong Yu submitted a memorial saying, “In antiquity, the Son of Heaven had seven ancestral temples. Now that the close relatives (who would worship at) the ancestral temples of the deceased Hui (the second emperor) and Jing (the fourth emperor) Emperors are all extinguished, it would be proper to destroy (those temples). It would also be proper to set right the (imperial) ancestral temples of the commanderies and kingdoms, which do not conform to the ancient rites.” The Son of Heaven thought this idea correct, but, before it could be enacted, Yu died. In Yongguang 4 (40 B.C.E.), (Emperor Yuan) issued an edict about the earlier idea of abolishing the ancestral temples of the commanderies and kingdoms, saying, “I have heard that the enlightened monarch’s controlling of the world follows the times in drafting decrees and accords with circumstances in governing properly. In the past, when all-under-heaven was first put in order, and the distant regions had not yet 65 Brashier, Ancestral Memory in Early China, 87. 28 submitted, (the former Emperors) consequently established (imperial) ancestral temples (that is, temples to their own ancestors) wherever they had personally (inspected). This was to build prestige and prevent the spread (of rebellion) and to unify the people under the highest authority. Now we rely on the spirit of heaven and earth and the good fortune of the ancestral temples; the four regions are on the same path, and the barbarians offer tribute. We have long deferred (to the existing rites) and have not put them in order (that is, have not revised them). This causes the distant and the base to offer sacrifice (to the former Emperors), which was probably not august Heaven’s intention for the ancestral temples. I am deeply afraid. Does not the tradition speak of this? ‘If I do not participate in the sacrifice, it is as if I did not sacrifice’ (Analects 3.12). What are the opinions of the Generals, Adjunct Marquis, officials of the Full Two Thousand Bushels, (regular) Two Thousand Bushel officials, and the various Grand Masters, Erudite officials, and Court Gentlemen for Consultation?”66 丞相玄成、御史大夫鄭弘、太子太傅嚴彭祖、少府歐陽地餘、諫大夫尹更始等七 十人皆曰:“臣聞祭,非自外至者也,繇中出,生於心也。故唯聖人爲能饗帝, 孝子爲能饗親。立廟京師之居,躬親承事,四海之內各以其職來助祭,尊親之大 義,五帝三王所共,不易之道也。《詩》云:‘有來雍雍,至止肅肅,相維辟公, 天子穆穆。’《春秋》之義,父不祭於支庶之宅,君不祭於臣僕之家,王不祭於 下土諸侯。臣等愚以爲宗廟在郡國,宜無修,臣請勿復修。”奏可。因罷昭靈后、 武哀王、昭哀后、衛思后、戾太子、戾后園,皆不奉祠,裁置吏卒守焉。 Counselor-in-Chief Xuancheng (Wei Xuancheng 韋玄成), Censer-in-Chief Zheng Hong, Grand Mentor of the Heir Apparent Yan Pengzu, Chamberlain for the Palace Revenues Ouyang Diyu, Grand Master of Remonstrance Yin Gengshi, and others, 70 people in all, said, “We have heard that sacrificial rites do not arise from outside but emerge from inside and are born of the heart. Therefore, only imperial personages are able to sacrifice to emperors, and only filial sons are able to sacrifice to their fathers. If you establish an ancestral temple in the capital where you dwell and personally attend to its affairs, then 66 I follow Hucker, A Dictionary of Official Titles in Imperial China on these official titles, here and for the rest of this passage. 29 all within the four seas will, according to their duty, come to assist in the sacrifice; this will venerate the significance of your ancestors together with the Five Emperors and the Three Sovereigns and not change their Way. It is said in the Odes: “He comes harmoniously, he arrives respectfully, his assistants are princes and lords: the Son of Heaven in his majesty.” The meaning of (the relevant passage in) the Springs and Autumns is that fathers should not be sacrificed to in the dwellings of common people who are not in the main line of descent, lords should not be sacrificed to in the homes of servants, and monarchs should not be sacrificed to by the many lords (who rule over) lower territories. We humbly believe that it is proper not to maintain ancestral temples in the commanderies and kingdoms, and we beg that they no longer be maintained.” The memorial was approved. Consequently, the cemetery parks of Empress Zhao Ling (Liu Bang’s mother), Prince Wu Ai (Liu Bang’s elder brother), Empress Zhao Ai (Liu Bang’s elder sister), Empress Wei Si (Wei Zifu 衛子夫, a wife of Emperor Wu), and Heir Apparent Li (Liu Ju 劉據, a son of Emperor Wu) were all abandoned and sacrifices were no longer offered at them, and the number of officials and soldiers guarding (the tombs) was reduced.67 and 罷郡國廟後月餘,復下詔曰:“蓋聞明王制禮,立親廟四,祖宗之廟,萬世不毀, 所以明尊祖敬宗,著親親也。朕獲承祖宗之重,惟大禮未備,戰栗恐懼,不敢自顓, 其與將軍、列侯、中二千石、二千石、諸大夫、博士議。” Over a month after the ancestral temples of the commanderies and kingdoms were abolished, (Emperor Yuan) again issued an edict, saying, “I have heard that when an enlightened monarch regulates the rites, he establishes four ancestral temples to his forefathers, and an ancestral temple to his founding ancestor, which is never destroyed, in order to make manifest his veneration of and respect for his ancestors and to make clear how close he is to his fathers. I have inherited the greatness of my ancestors, but because the great rites have not been perfected, I tremble in fear; I do not dare to act ignorantly. What are the opinions of the Generals, Adjunct Marquis, officials of the Full Two 67 Han shu 漢書 (Beijing: Zhonghua, 1962 [1964]), 73.3116–17 (“Wei xian zhuan” 韋賢傳). 30 Thousand Bushels, (regular) Two Thousand Bushel officials, and the various Grand Masters and Erudite officials?” 玄成等四十四人奏議曰:“《禮》,王者始受命,諸侯始封之君,皆爲太祖。以下, 五廟而迭毀,毀廟之主臧乎太祖,五年而再殷祭,言壹禘壹祫也。祫祭者,毀廟與 未毀廟之主皆合食於太祖,父爲昭,子爲穆,孫復爲昭,古之正禮也。祭義曰: ‘王者禘其祖自出,以其祖配之,而立四廟。’言始受命而王,祭天以其祖配,而 不爲立廟,親盡也。立親廟四,親親也。親盡而迭毀,親疏之殺,示有終也。周之 所以七廟者,以后稷始封,文王、武王受命而王,是以三廟不毀,與親廟四而七。 非有后稷始封,文、武受命之功者,皆當親盡而毀。成王成二聖之業,制禮作樂, 功德茂盛,廟猶不世,以行爲諡而已。《禮》,廟在大門之內,不敢遠親也。臣愚 以爲高帝受命定天下,宜爲帝者太祖之廟,世世不毀,承後屬盡者宜毀。今宗廟異 處,昭穆不序,宜入就太祖廟而序昭穆如禮。太上皇、孝惠、孝文、孝景廟皆親盡 宜毀,皇考廟親未盡,如故。” Xuancheng and others, 44 people in all, presented a memorial of their opinion, saying, “(It is said in) the Rites that the monarch who first receives the mandate and the lords who are first enfeoffed as rulers become the founding ancestors. After this, once five ancestral temples (have been built), they are dismantled in succession (one each time a new temple is constructed). The tablets of the dismantled ancestral temples are stored in the temple to the founding ancestor, and five years (after this event) a magnificent sacrifice is held, which is called ‘one di sacrifice and one xia sacrifice.’ For the xia sacrifice, the tablets of the dismantled ancestral temples and of the ancestral temples that have not yet been dismantled are all feasted together at (the ancestral temple of) the founding ancestor. The father’s generation is zhao, the son’s generation is mu, and the grandson’s generation is again zhao: this is the orthodox rite of antiquity. It is said in ‘The Meaning of Sacrifice’: ‘The monarch performs the di sacrifice to whence his ancestors came as well as to his ancestors, and therefore establishes four ancestral 31 temples.’68 This says that, when (the monarch) first receives the mandate and (begins to) rule, he sacrifices to Heaven and at the same time to his ancestors, but does not build ancestral temples, because the familial relationship is exhausted (= is too distant). (The monarch) erects four ancestral temples for his close relatives, (in order to express that) the close relatives are close. As the family relationship becomes exhausted, (the ancestral temples) are dismantled one by one, and when the relationship gradually becomes more distant, this represents the start and finish. The reason that the Zhou (built) seven ancestral temples is that Hou Ji was first installed and King Wen and King Wu received the mandate and then ruled, so there were three ancestral temples which could not be dismantled. Together with the four ancestral temples for close relatives, this makes seven. If Hou Ji had not been first installed, and if Wen and Wu had not had the merit of receiving the mandate, (their ancestral temples) would have all properly been dismantled after their close relatives had been exhausted. King Cheng completed the work of the two sage-kings (Wen and Wu), established the rites and created the music, and his merit and virtue were luxuriant and flourishing, but his ancestral temple was not to last for generations; his conduct led to his posthumous name and that is all. (It is said in) the Rites that the ancestral temples are (built) inside the great gate, as (the descendants) must not dare to become distant from their fathers. Your servant humbly thinks that Emperor Gao (-zu) received the mandate to settle all-under-Heaven, so it is proper that Emperor Taizu’s (that is, Gaozu’s) ancestral temple not be dismantled for generations after generations; but it is proper to dismantle (the ancestral temples of) those that carried on after him, as their descendants gradually become more distant. Now, the ancestral temples are in different places, and zhao and mu are not properly arranged. It is proper to put (the tablets) into Taizu’s ancestral temple and arrange them in zhao and mu order, according to the rites. The close relatives of the (the founding) Emperor’s Father and the deceased Emperors Hui, Wen, and Jing (the first three emperors after Gaozu) have all been exhausted, so it is proper to dismantle their temples. The ancestral temple of the 68 This passage does not seem to be in the received version of the Li ji section called “Ji yi” (The Meaning of Sacrifice). In Han references, the one-syllable title Li typically refers to the Yi li, not the Li ji (See Boltz, “I li 儀禮,” 234), but the Li ji does contain a section entitled “Ji yi”; there is no such section in the extant Yi li. 32 (current) Emperor’s deceased father (Emperor Xuan 宣帝), (because his) close relatives have not yet been exhausted, should remain as it is.”69 These discussions continued after the Western Han,70 as can be seen from this passage from the Hou Han shu 後漢書: 二十六年,有詔問張純,禘祫之禮不施行幾年。純奏:“禮,三年一祫,五 年一禘。毀廟之主,陳於太祖;未毀廟之主,皆升合食太祖;五年再殷祭。 舊制,三年一祫,毀廟主合食高廟,存廟主未嘗合。元始五年,始行禘禮。 父爲昭,南嚮;子爲穆,北嚮。父子不並坐,而孫從王父。禘之爲言諦。諦 諟昭穆,尊卑之義。以夏四月陽氣在上,陰氣在下,故正尊卑之義。祫以冬 十月,五穀成熟,故骨肉合飲食。祖宗廟未定,且合祭。今宜以時定。”語 在純傳。 In (Jianwu 建武) 26 (during the reign of the founder of the Eastern Han, Emperor Guangwu 光武帝; 50 C.E.), there was an edict inquiring of Zhang Chun for how many years the di and xia rituals had not been carried out. Chun submitted a memorial: “According to the rites, every three years there should be one xia sacrifice and every five years there should be one di sacrifice. The tablets of dismantled ancestral temples are displayed (in the temple of) the founding ancestor, and the tablets of ancestral temples which have not yet been dismantled are all elevated and feast together with the founding ancestor. There are two great sacrifices every five years. In the old system, when every three years there was one xia sacrifice, the tablets of dismantled ancestral temples were feasted together with (the tablet of) Gaozu’s 高祖 ancestral temple, and the tablets of the existing ancestral temples did not receive the chang sacrifice together (with the tablet of the high ancestor). In Yuanshi 5 (during the reign of Western Han Emperor Ping 平帝; 5 C.E.), they began practicing the di ritual. Fathers were zhao and faced 69 Han shu, 73.3118–19 (“Wei xian zhuan”) See Loewe, Problems of Han Administration, 16ff for more on these discussions during the early Eastern Han. 70 33 south; sons were mu and faced north. Fathers and sons were not situated together, and grandsons followed their grandfathers. What “di” means is “caution.” Zhao and mu are cautiously examined because of the meanings of (=the relationship between) the honored and the base. In the fourth month, in the summer, yang qi is at its highest and yin qi is at its lowest; thus the relationship between the honored and the base can be rectified. The xia sacrifice is held in the tenth month, in the winter, when the five grains have ripened; thus the relatives (both zhao and mu) can drink and feast together. The ancestral temple of the founding ancestor has not yet been determined, so (the zhao and mu relatives) can be sacrificed to together. Now it is proper to fix their schedules.” These words are included in “The Biography of Zhang Chun.” 上難復立廟,遂以合祭高廟爲常。後以三年冬祫五年夏禘之時,但就陳祭毀 廟主而已,謂之殷。太祖東面,惠、文、武、元帝爲昭,景、宣帝爲穆。惠、 景、昭三帝非殷祭時不祭。光武皇帝崩,明帝即位,以光武帝撥亂中興,更 爲起廟,尊號曰世祖廟。以元帝於光武爲穆,故雖非宗,不毀也。後遂爲常。 The Sovereign thought it difficult to restore the ancestral temples, so he made joint sacrifices at Gaozu’s ancestral temple to be the standard practice. Afterwards, at the time of the triennial winter xia sacrifice and quinquennial summer di sacrifice, when they went to display and sacrifice to the tablets of the dismantled ancestral temples, this was called the great (sacrifice). To the east of the founding ancestor, Emperors Hui, Wen, Wu, and Yuan were zhao, and Emperors Jing and Xuan were mu. The three emperors Hui, Jing, and Zhao were not sacrificed to except at the time of the great sacrifice. When Emperor Guangwu died, and Emperor Ming took his position, he thought Emperor Guangwu had brought order out of chaos and (caused the Han) to be resurgent, so he again made an ancestral temple (for Emperor Guangwu), and gave it the honored appellation “the Ancestral Temple of Shizu.” Because of Emperor Yuan (preceding him), Emperor Guangwu was (classified as) mu, so even though (Yuan) was not his ancestor, (his 34 ancestral temple) would not be dismantled. Afterwards this accordingly became the standard practice.71 While the discussions of the zhao-mu system which appear in the three ritual classics were undoubtedly filtered by the Han, discussion of the system becomes quite common starting from the late Western Han, and a number of texts preserve various, sometimes contradictory, explicitly Han understandings of the system. There are a number of contemporary political issues which strongly color these discussions, but they remain very useful for understanding Han views about zhao and mu. Moving on, most likely, to just beyond the Western Han: a previously unknown large ritual complex was discovered to the south of Han Chang’an in the 1950s. The site is a huge square, approximately 1,960,000 square meters (about three quarters of a square mile) in area— each side was approximately 1400 meters long—almost a quarter of the length of the Han city wall.72 The site, which shows signs of having been badly burned, is constructed with building materials which date to roughly the late Western Han, with the inscription on one column base “Shijianguo” 始建國 (this reign period covered the first four years of Wang Mang’s 王莽 reign, that is, the beginning of the Xin 新 dynasty, 9–13 C.E.).73 Within the remains of a hangtu 夯土 wall appear to be the remains of 12 square structures—11 of these are identical in size, with one (numbered 12 in the plan reproduced below) substantially larger. 71 Hou Han shu 後漢書 (Beijing: Zhonghua, 1965 [1973]), 3194–95 (“Zhi di jiu: jisi xia” 志第九: 祭祀下). 72 Michael Loewe, “Wang Mang and His Forbears: The Making of the Myth,” T’oung Pao, Second Series 80.4–5 (1994), 217. 73 Loewe, “Wang Mang and His Forbears,” 217. 35 Figure 6. From Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 262 The largest building sits in the south of the enclosure, centered from east to west; the other 11 structures are arranged in 3 rows—the top and bottom row each contain four structures and the middle row contains three. The foundation of each structure was surrounded by a square wall approx. 260–280 meters per side.74 The foundation of the largest structure (no. 12) had sides of about 100 meters; the sides of the other 11 structures were roughly 50 meters each.75 Each structure had a main chamber (zhu shi 主室) with chambers projecting from each corner (jia shi 夾室), giving each structure something of a ya-character (亞) shape. The structures are numbered 74 Huang Zhanyue 黃展岳, “Guanyu Wang Mang jiu miao de wenti—Han Chang’an cheng nanjiao yi zu jianzhu yizhi de dingming” 關於王莽九廟的問題——漢長安城南郊一組建築遺址的定名, Kaogu 1989.3, 261. 75 Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 264. 36 (by the archaeologists) from right to left and top to bottom; structure 3 was the only one to be fully excavated; structure 6 could not be worked on at all, as a factory is located above its expected location.76 Despite its seemingly incongruous number of structures, the site has generally been identified with Wang Mang’s “Nine Ancestral Temples” 九廟, though not all scholars agree with this identification. Those that identify it as the Nine Temples tend to identify a separate ritual site to the west of these 12 structures as the Ming tang 明堂 and one to the east as the Pi yong 辟雍.77 The Nine Temples are described in the Han shu 漢書 and in accounts by Yang Xiong 揚 雄 (53 B.C.E.–18 C.E.), Huan Tan 桓譚 (43 B.C.E.–28 C.E.), and Wei Ao 隗囂 (died 33 C.E.); following these accounts, it is generally believed that the structures were completed.78 After Wang Mang, who had been a regent during the late Western Han, seized power, he created a new genealogy for himself, presumably as part of an effort to gain legitimacy for his new dynasty. He traced his ancestry back to the legendary Huangdi 黃帝, through Shun 舜, through a previously minor figure named Shao Hao 少昊 and on to his closer ancestors—in doing so, he altered the ancestry of Yao 堯 and Shun from the accounts given in the Shi ji.79 This marginalized the ancestry of the house of Liu 劉 (the Han royal house) and the branch of descent (through Yao) which they claimed. As stated in the Li ji, quoted above, the prescribed number of ancestral temples for rulers was seven. According to the accounts of the Nine Temples, Wang Mang broke with this tradition primarily by constructing five permanent temples, for Huangdi, 76 Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 262. See Nancy Shatzman Steinhardt, “The Han Ritual Hall” in Chinese Traditional Architecture, ed. Nancy Shatzman Steinhardt (New York: China Institute in America, 1984). 78 Loewe, “Wang Mang and His Forbears,” 217. 79 Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 265. 77 37 Shun, Guiman 媯滿 (King Hu of Chen 陳胡王), Jingzhong 敬仲 (King Jing of Qi 齊敬王), and King Min of Jibei 濟北愍王, together with four temples to his immediate ancestors.80 It was therefore expected that the Nine Temples would consist of one central temple (dedicated to Huangdi) and two columns of four temples each, alternating between zhao and mu temples. If this site is in fact that of the Nine Temples, then, the question of to whom the three extra structures are dedicated must be dealt with. Huang Zhanyue’s 黃展岳 proposal is that they were dedicated to three other ancestors who were especially important in passing down the lineage—Di Ku 帝嚳, according to the Shi ji, one of the five emperors who reigned shortly after Huangdi’s time; Tian He 田和, who took possession of the state of Qi 齊; and Tian Jian 田建, who lost Qi to Qin 秦, but protected the lineage.81 He gives a diagram of this theory: 80 81 Loewe, “Wang Mang and His Forbears,” 217–18. Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 266–67. 38 Figure 7. From Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 266 In the above diagram, Tian He is in position 5, Di Ku is in position 6, and Tian Jian is in position 7. A different proposal was suggested by Gu Jiegang 顧頡剛, that these were Xin dynasty temples—the central one would, after his death, be dedicated to Wang Mang himself, and the other two would be dedicated to the second and third emperors of the dynasty (following the Han model, with permanent temples dedicated to the first three emperors).82 A diagram of his theory follows: 82 Loewe, “Wang Mang and His Forbears,” 218. 39 Figure 8. From Huang Zhanyue, “Guanyu Wang Mang jiu miao de wenti,” 267 In the above proposal, numbers 1, 2, and 3 represent Wang Mang and his descendants. Wang Entian 王恩天, however, strongly disagrees that this site is Wang Mang’s Nine Temples. He argues that the number of temples is wrong (12 instead of nine); that the zhao-mu system, which is clearly described in the Han shu, is not clearly followed; that the size of the complex is not exactly right (though Huang argues that it is approximately what would be expected); and that it is not located exactly where the records say it should be (Huang argues that all the geographical descriptions except the one from Shuijing zhu 水經注 match up with the site, and that this one description must be in error). Wang proposes that this complex is instead a group of Han ancestral tombs built under the stewardship of Wang Mang—he goes on to identify this complex (and not the complex to its west) as the Ming tang.83 He reconstructs the expected 83 Wang Entian 王恩田, “‘Wang Mang jiu miao’ zai yi” “王莽九廟”再議, Kaogu yu wenwu 1992.4, 96–106. 40 layout of Wang Mang’s Ancestral Temples (on the left) and his theory of the site under discussion as: Figure 9. From Wang Entian, “‘Wang Mang jiu miao’ zai yi,” 98, 100 The rendition on the above left shows Wang Mang’s ancestral temples arranged in strict, traditional zhao-mu order, and the one on the right shows Wang Entian’s reconstruction of what he identifies as Han ancestral temples in the Ming tang. It is not certain that this structure is Wang Mang’s Nine Temples (though I do believe that is the most likely identification), but I think that Wang Entian is too quick to dismiss the possibility based on his understanding of zhao-mu order. Whatever the specific identification of this site, it shows a return to (what was believed in the Han) to be the Western Zhou tradition, in 41 collecting the ancestral temples into a single complex, instead of spreading them around the city or beyond. As reconstructed by both Huang Zhanyue and Gu Jiegang, the site certainly complies with zhao and mu. It may not be divided into two columns, but the dualism of left and right and east and west can still be strictly maintained. But why was Wang Mang so intent on following the zhao-mu system, and supposed Western Zhou rituals in general? In Michael Puett’s words (he is speaking of Wang Mang’s favor for the Zhou li, but his argument also applies to Wang Mang’s classicizing in general), it was because “it offers a means of organizing the world simply through the ordering acts of the ruler.”84 That is, the strict use of the zhao-mu system and other actions taken by Wang Mang for which he appealed to high antiquity did not reflect his conservatism, but rather gave him cover for his activist form of government. As has been pointed out by many scholars, Wang Mang’s apparently ancient policies were actually very influenced by late Western Han dynasty theory, notably that of yinyang and Five Phases.85 The same also applies to some of the earlier actions aiming to return to antiquity taken during the late Western Han discussed above.86 Whatever the nature of the zhao-mu system before the late Western Han or the Xin dynasty, when its structure becomes more evident, the system unquestioningly continued to become more and more prominent, in various different manifestations. Examples of later usage are practically unlimited, but a few, spread over a long period of time, should suffice.87 The 84 Michael Puett, “Centering the Realm: Wang Mang, the Zhouli, and Early Chinese Statecraft,” in Statecraft and Classical Learning: The Rituals of Zhou in East Asian History, ed. Benjamin A. Elman and Martin Kern (Leiden: Brill, 2010), 153. 85 Mu-chou Poo, In Search of Personal Welfare: A View of Ancient Chinese Religion (Albany: State University of New York Press, 1998), 113. 86 Michael Loewe discusses at length the political consequences of the zhao-mu system, especially during the Eastern Han, which is beyond the scope of this paper, in Loewe, Problems of Han Administration. 87 A fuller discussion of the post-Han zhao-mu system can be found in Loewe, Problems of Han Administration, 74–96. 42 figures depicted in the ca. 150 C.E. Wu Family Shrines 武氏祠, for example, seem to be arranged according to zhao-mu order.88 Lineage heads in the Song dynasty were known to explicitly state that descendants should be buried together in zhao-mu order, so that the graveyard would have the appearance of a family tree (puxi 譜系).89 A Yuan dynasty figure is known to have forced a family of commoners who maintained an ancestral temple containing 21 generations of ancestral tablets arranged in zhao-mu order to bury the tablets instead, as he felt it violated correct zhao-mu principles, in that it contained tablets for family members who had no descendants and, especially, in that he felt only nobles were entitled to such a temple.90 Finally, the users of a contemporary Hakka mausoleum in Taiwan practice secondary burial, in which they disinter skeletons after a set period of time and place them in urns; these urns are then arranged in the mausoleum in zhao-mu order, with sons, great-grandsons, and so on displayed on stage left, and the other generations on stage right.91 Howard Martin additionally argues that the separation between the urns containing the bones of male and female ancestors in this mausoleum (with females grouped on one side and males on the other) represents for the users of this mausoleum an extension of the zhao-mu principle.92 If this is so, it would represent only one more development in a system that, however unclear its origins, seems to have changed radically over time. 88 See Toshio Nagahiro, “A Study on the Central Pavilion Scenes of the Wu Family Temples,” Acta Asiatica 2 (1961), 40–58. 89 Patricia Buckley Ebrey, Women and the Family in Chinese History (London: Routledge, 2003), 115. 90 Ebrey, Women and the Family in Chinese History, 139. 91 Howard J. Martin, “Hakka Mausoleums in North Taiwan,” Ethnology 30.1 (1991), 97. 92 Martin “Hakka Mausoleums in North Taiwan,” 92. 43