1. See Guowei, Wang 王國維， Guantang jilin 觀堂集林 (Taipei: Shijie, 1975), 2.15a–24b Google Scholar; Cezong, Zhou 周策縱, “Gu wu dui yuewu ji shige fazhan di gongxian” 古巫對樂舞及詩歌發展的貢獻, Qinghua xuebao 清華學報 n.s. 13.1-2 (1981), 1–25 Google Scholar; Wang, C.H., From Ritual to Allegory: Seven Essays in Early Chinese Poetry (Hong Kong: The Chinese University Press, 1988), 1–51 Google Scholar. See also, Ch'en, Shih-hsiang, “The Shih-ching: Its Generic Significance in Chinese Literary History and Poetics,” Bulletin of the Institute of History and Philology, Academia Sinica 39 (1969), 371–413 Google Scholar (repr. in Studies in Chinese Literary Genres, ed. Birch, Cyril [Berkeley: University of California Press, 1974], 8–41)CrossRefGoogle Scholar. In addition to the religious hymns and banquet songs included in the ya and song sections, the pieces of the guo feng 國風 (Airs of the states) part have also been discussed as performance songs. Granet, Marcel, in his Fêtes et chansons anciennes de la Chine (Paris: Bibliothèque de l'école des hautes études, 1919)Google Scholar, set out to reconstruct the folk rites embodied in the guo feng from an anthropological and sociological perspective; and Wang, C.H., in The Bell and the Drum: Shih Ching as Formulaic Poetry in an Oral Tradition (Berkeley: University of California Press, 1974)CrossRefGoogle Scholar, tried to apply the Lord-Parry hypothesis of oral composition to the Shijing in general. Both studies, each of them written strictly within a particular theoretical framework, are children of their times, and they have been aging together with their underlying methodological assumptions.

2. Guowei, Wang, Guantangjilin, 2.15b–17b Google Scholar; Zuoyun, Sun 孫作雲, Shi jing Zhou dai shehui yanjiu 詩經與周代社會硏究 (Beijing: Zhonghua, 1966), 239–72Google Scholar; Shaughnessy, Edward L., “From Liturgy to Literature: The Ritual Contexts of the Earliest Poems in the Book of Poetry ,” Hanxue yanjiu 漢學研究 13.1 (1994), 133–64Google Scholar. For a more skeptical assessment, see Shizuka, Shirakawa 白川靜， Shikyō kenkyū 詩經硏究 (Kyoto: Hōyū, 1981), 339–48Google Scholar.

3. Maspero, , China in Antiquity, trans. Kierman, Frank A. Jr. (Amherst: University of Massachusetts Press, 1978), 272 Google Scholar.

4. Maspero, , China in Antiquity, 274–76Google Scholar.

5. The term guci, very common in later commentarial writing, appears already twice in the Li ji 禮記, “Li yun” 禮運; see Li ji zhengyi 禮言己正義 (Shisan jing zhushu fu jiaokan ji 十三經注疏附校勘記 ed., 1815; repr. Beijing: Zhonghua, 1987), 21.190a, 22.197c. The best study on the guci is still Xu Zhongshu's徐中舒 seminal “Jinwen guci shili” 金文網辭釋例, Bulletin of the Institute of History and Philology, Academia Sinica 6.1 (1936), 144 Google Scholar; see also, von Falkenhausen, Lothar, “Issues in Western Zhou Studies: A Review Article,” Early China 18 (1993), 151-55, 166 Google Scholar.

6. Mao shi zhengyi 毛詩正義 (Shisan jing zhushu ed.), 1/1.4c. Ruan Yuan 阮元i (1764–1849), in his Yan jing shi ji 挈經室集 (in Huang Qing 皇清經解 [1829]; repr. Shanghai: Shanghai, 1988), 1068.249b-250a7 has proposed the interchangeability of song / *zljongs 頌 and rong / *(l)jong 容 (phonetic reconstructions of Zhou Chinese follow Baxter, William H., A Handbook of Old Chinese Phonology [Berlin and New York: Mouton de Gruyter, 1992])CrossRefGoogle Scholar. In light of the assumed first century composition date for the “Great preface,” Ruan Yuan's widely accepted hypothesis is sound; see Karlgren, Bernhard, Loan Characters in Pre-Han Texts (Göteborg: Elanders Boktryckeri Aktiebolag, 1968), nos. 542 and 140Google Scholar. An early example of the interchangeability of song and rong can be found among the texts from the late fourth century B.C. tomb at Guodian 郭店 (Hubei): in slips 21 and 66 of the “Xing zi ming chu” 性自命出 manuscript, song is probably to be read as rong; see bowuguan, Jingmen shi, Guodian Chu mu zhujian 郭店楚墓竹簡 (Beijing: Wenwu, 1998), 179, 181 Google Scholar. For other explanations of song, see Liu Yuqing 劉毓慶,” ‘Song’ shi xinshuo: ‘song’ wei yuanshi zongjiao son gei kao” 頌詩新說：頌爲原始宗教誦辭考 Jinyang xuekan 晉陽學刊 1987.6, 76-83; and Shizuka, Shirakawa, Shikyō kenkyū, 421–31Google Scholar.

7. von Falkenhausen, Lothar, “Ritual Music in Bronze Age China: An Archaeological Perspective” (Ph.D. diss., Harvard University, 1988), 693 Google Scholar.

8. Baxter, William H., “Zhou and Han Phonology in the Shijing,” in Studies in the Historical Phonology of Asian Languages, ed. Boltz, William G. and Shapiro, Michael C. (Amsterdam and Philadelphia: J. Benjamins, 1991), 30 Google Scholar.

9. Lun yu zhushu 論語法疏 (Shisan jing zhushu ed.), 17.69b (Lun yu 17.10).

10. The same has been noted for the Homeric epics; see Assmann, Jan, Das kulturelle Gedāchtnis: Schrift, Erinnerung und politische Identitāt in frūhen Hochkulturen (Munich: C.H. Beck, 1997), 276 Google Scholar.

11. For the structure of the “Minor prefaces” and the related early exegetical tradition, see Zoeren, Steven Van, Poetry and Personality: Reading, Exegesis, and Hermeneutics in Traditional China (Stanford: Stanford University Press, 1991), 80–115 Google Scholar.

12. On the antagonism between eternal text and ephemeral performance, see Shieffelin, Edward L., “Problematizing Performance,” in Ritual, Performance, Media, ed. Hughes-Freeland, Felicia (London and New York: Routledge, 1998), 198–99Google Scholar. As will become evident in the following discussion, this approach to the songs differs from the assumption of a historical development from liturgical to literary texts among the Shijing ritual hymns; for this historical approach see Shaughnessy, “From Liturgy to Literature.”

13. Mao shi zhengyi, 13/2.199b-202b (Mao 209).

14. Mao shi zhengyi, 13/2.199b.

15. Mao shi zhengyi, 13/2.202b-14/1.210c. On the whole group, see Shizuka, Shirakawa, Shikyö kenkyū, 310-19, 617–20Google Scholar; and Masaaki, Matsumoto 松本雅明， Shikyō shohen no seiritsu ni kansuru kenkyū 詩經諸篇の成立に關する研究 (Tokyo: Kaimei, 1981–1982), vol. 2, 792–805 Google Scholar.

16. For Western Zhou dates I follow Shaughnessy, Edward L., Sources of Western Zhou History: Inscribed Bronze Vessels (Berkeley: University of California Press, 1991), xix Google Scholar.

17. Mao shi zhengyi, 13/2.199b-c (including early commentaries).

18. Mao shi zhengyi, 13/2.202b.

19. In English language scholarship, this is particularly true for broader or comparative accounts of early Chinese culture. See, for example, Schwartz, Benjamin I., The World of Thought in Ancient China (Cambridge: Harvard University Press, 1985), 49 CrossRefGoogle Scholar; or Paper, Jordan, The Spirits are Drunk: Comparative Approaches to Chinese Religion (Albany: State University of New York Press, 1995)Google Scholar, the main title of which is a (simplified) translation of stanza five, line five of the hymn.

20. Falkenhausen, , “Reflections on the Political Role of Spirit Mediums in Early China: The Wu Officials in the Zhou Li ,” Early China 20 (1995), 297 Google Scholar; see also his “Issues in Western Zhou Studies,” 148-50, and Suspended Music: Chime-Bells in the Culture of Bronze Age China (Berkeley: University of California Press, 1993), 25–32 Google Scholar. Maspero, Henri, China in Antiquity, 150–54Google Scholar, also bases his reconstruction of a typical ancestral sacrifice on this hymn, “fleshed out with the aid of various chapters” (428–29n.46) from the three ritual canons. He was evidently not aware that he was essentially repeating what the compilers of the ritual canons appear to have done: “fleshing out” the contents of “Chu ci” with additional data from other sources.

21. Jiheng, Yao, Shijing tonglun 詩經通論 (Beijing: Zhonghua, 1958), 11.231 Google Scholar; Yurun, Fang, Shijing yuanshi 詩經原始 (Beijing: Zhonghua, 1986), 11.431 Google Scholar.

22. This observation applies to the English translations by James Legge (The Chinese Classics IV: The She King [repr. Taipei: Southern Materials Center, 1985], 368-73), Waley, Arthur (The Book of Songs [New York: Grove Weidenfeld, 1988], 209–11)Google Scholar, and Karlgren, Bernhard (”The Book of Odes, Kuo feng and Siao ya,” Bulletin of the Museum of Far Eastern Antiquities 16 [1944], 246–47)Google Scholar, as well as to the widely read modern Japanese and Chinese versions.

23. The performer of ritual as its synchronical observer is implied in the widely accepted “social solidarity thesis” of ritual theory; on this, see Bell, Catherine, Ritual Theory, Ritual Practice (New York and Oxford: Oxford University Press, 1992), 171–72Google Scholar, passim. Burkert, Walter, Structure and History in Greek Mythology and Ritual (Berkeley: University of California Press, 1982), 49 Google Scholar, holds that “those systems which attained relative stability over a period of time and are called definite ‘civilizations’ on account of this owe their success, the maintenance of their identity, to group solidarity re-inforced by ritual.” Leach, Edmund R., Culture and Communication (Cambridge: Cambridge University Press, 1976), 45 CrossRefGoogle Scholar, has coined the poignant formula that “we engage in rituals in order to transmit collective messages to ourselves.”

24. The most comprehensive account of the early musical discourse may be found in Keisuke, Kurihara 栗原圭介, Chūgoku kodai gakuron no kenkyū 中國古代樂論の研究 (Tokyo: Daitō bunka daigaku tōyōkenkyūsho, 1978)Google Scholar. In addition, see Falkenhausen, , Suspended Music, 23-66, 310–24Google Scholar; DeWoskin, Kenneth, A Song for One or Two: Music and the Concept of Art in Early China (Ann Arbor: University of Michigan Center for Chinese Studies, 1982), 29–98 CrossRefGoogle Scholar; DeWoskin, , “Early Chinese Music and Origins of Aesthetic Termi-nology,” in Theories of the Arts in China, ed. Bush, Susan and Murck, Christian (Princeton: Princeton University Press, 1983), 187–214 Google Scholar; and my Die Hymnen der chinesischen Staatsopfer: Literatur und Ritual in der politischen Repräesentation von der Han-Zeit bis zu den Sechs Dynastien (Stuttgart: Franz Steiner Verlag, 1997), 23–50 Google Scholar.

25. A brilliant account of the early use of the Shi may be found in Lewis, Mark Edward, Writing and Authority in Early China (Albany: State University of New York Press, 1999), 155–76Google Scholar; see also, van Zoeren, , Poetry and Personality, 38–44 Google Scholar. The most com-prehensive study of the use of the Shi in the Zuo zhuan 左傳 is Qinliang, Zeng 曾勤良, Zuo zhuan yinshifushi zhi shijiao yanjiu 左傳引詩賦詩之詩教研究 (Taipei: Wenjin, 1993)Google Scholar.

26. Langer, Suzanne, here quoted after Tambiah, Stanley J., “A Performative Approach to Ritual,” Proceedings of the British Academy 65 (1979/1981), 164 Google Scholar.

27. See Zoeren, Van, Poetry and Personality, 28–51 Google Scholar.

28. See my “Ritual, Text, and the Formation of the Canon: Historical Transitions of Wen in Early China,” T'oung Pao 87 (2001), 43–91 CrossRefGoogle Scholar.

29. Gombrich, , The Sense of Order: A Study in the Psychology of Decorative Art (Ithaca: Cornell University Press, 1984)Google Scholar.

30. See Rawson, Jessica, Chinese Ornament: The Lotus and the Dragon (London: British Museum Publications, 1984)Google Scholar; Powers, Martin, “The Figure in the Carpet: Reflections on the Discourse of Ornament in Zhou China,” Monumenta Serica 43 (1995), 211–33CrossRefGoogle Scholar; Koerner, Joseph Leo, “The Fate of the Thing: Ornament and Vessel in Chou Bronze Interlacery,” Res 10 (1985), 28–46 Google Scholar. In addition, the conference volume The Problem of Meaning in Early Chinese Ritual Bronzes, ed. Whitfield, Roderick (London: University of London, 1993)Google Scholar, is dedicated to the issue, with Sarah Allan taking the opposite view that ornament, in particular on Shang bronzes, could, and should, be interpreted as primarily referential; see Allan, “Art and Meaning,” in The Problem of Meaning in Early Chinese Ritual Bronzes, 9—33, and “Epilogue,” ibid., 161-76. See also my discussion in Die Hymnen der chinesischen Staatsopfer, 11-22.

31. I am using the term “indexical” in the usual semiotic sense (as defined by Charles Sanders Peirce): the relation between the signifier and the signified is established not by convention or resemblance but through causal association on the basis of experience.

32. Rawson, Jessica, “Late Shang Bronze Design: Meaning and Purpose,” in The Problem of Meaning in Early Chinese Ritual Bronzes, 92 Google Scholar.

33. Bagley, Robert W., “Meaning and Explanation,” in The Problem of Meaning in Early Chinese Ritual Bronzes, 44–45 Google Scholar.

34. According to Rawson, a fundamental reform or even revolution in ritual practice (and possibly in the underlying belief system) in the early ninth century B.C. can be deduced from several observations: while wine vessels were mostly abandoned, food vessels grew larger not only in form but also in numbers, becoming arranged in extended sets; the ornament on these vessels was carried out no longer in minute detail but in larger patterns that could be fully recognized from a distance; and bronze bells were now introduced to the ensemble of ritual artifacts, adding the element of music to the ceremonies. These facts seem to indicate a development from a more private to a more public form of ancestral ritual, with larger numbers of participants perhaps standing at some distance. See Rawson, Jessica, “Statesmen or Barbarians: The Western Zhou as Seen Through their Bronzes,” Proceedings of the British Academy 75 (1989), 87–93 Google Scholar; and Rawson, , Western Zhou Ritual Bronzes in the Arthur M. Sackler Collections (Cambridge: Harvard University Press, 1990), vol. 1, 102–10Google Scholar, passim; see also, Tai, Luo 羅泰 (von Falkenhausen, Lothar), “Youguan Xi Zhou wanqi Iizhi gaige ji Zhuangbai Wei shi qingtongqi niandai de xin jiashe: cong shixi mingwen shuoqi” 有關西周晚期禮制改革及莊白微氏青銅器年代的新假設：從世系銘文說起, in Zhongguo kaoguxue yu lishixuezhi zhengheyanjiu 中國考古學與歷史學之整合硏究, ed. Chenghwa, Tsang 威振華 (Taipei: Academia Sinica, institute of History and Philology, 1997), 651–76Google Scholar. Therefore, if the bronze artifacts of the ancestral temple were now produced to be displayed for a larger ritual—probably also diplomatic—public, their significance as manifestations of status and wealth may have substantially increased.

35. Tambiah, , “A Performative Approach to Ritual,” 134 Google Scholar.

36. Burkert, Walter, Homo Necans: The Anthropology of Ancient Greek Sacrificial Ritual and Myth, trans. Bing, Peter (Berkeley: University of California Press, 1983), 29 Google Scholar.

37. For some of the prominent statements of this analogy, as well as for its critique, see Bell, Catherine, Ritual Theory, Ritual Practice, 43-46, 110–14Google Scholar.

38. Leach, Edmund R., “Ritualization in Man in Relation to Conceptual and Social Development,” Philosophical Transactions of the Royal Society, B, 251 (1966), 404, 407 Google Scholar. Burkert, , Homo Necans, 29 Google Scholar, also seems to imply this reverse statement by pointing to the social function of speech: “In many cases, that which is said seems less important in everyday life than that something is in fact said. Being together in silence is almost unbearable.”

39. Durand, Jean-Louis, “Ritual as Instrumentality” in The Cuisine of Sacrifice among the Greeks, ed. Detienne, Marcel and Vernant, Jean-Pierre, trans. Wissing, Paula (Chicago: The University of Chicago Press, 1989), 120 Google Scholar.

40. Mao shi zhengyi, 13/2.199c.

41. Mao shi zhengyi, 17/1.263b (Mao 245); this phrase opens the penultimate stanza seven of the hymn.

42. Xiao jing zhushu 孝經注疏 (Shisan jing zhushu ed.), 6.18b; the phrase is a common-place in Han literature. See also my The Stele Inscriptions of Ch'in Shih-huang: Text and Ritual in Early Chinese Imperial Representation (New Haven: American Oriental Society, 2000), 140–47Google Scholar.

43. David Keightley, “Late Shang Divination: The Magico-Religious Legacy,” in Explorations in Early Chinese Cosmology, ed. Rosemont, Henry Jr., Journal of the American Academy of Religion Studies 50.2 (1984), 13–14 Google Scholar, has noted the rigorous control of how the spirits could possibly react as a salient feature already of the Shang divinations. For Zhou and early imperial times, see my Die Hymnen der chinesischen Staatsopfer, 17—22, and my The Stele Inscriptions of Ch'in Shih-huang, 140–47. In bronze inscriptions, the self-referential gesture is most explicit in what Falkenhausen has called the “statement of purpose”; see Falkenhausen, , “Issues in Western Zhou Studies,” 154 Google Scholar, and below, pp. 94-95.

44. See Wheelock, Wade T., “The Problem of Ritual Language: From Information to Situation,” Journal of the American Academy of Religion 50.1 (1982), 50 Google Scholar: “Ritual language is not just an instrument for conveying ideas, but is directly used in accomplishing the ends of the ritual operation. This basic fact gives ritual language a set of charac-teristics that distinguishes it from the discourses of mythology and theology.”

45. Tambiah, , “A Performative Approach to Ritual”, 164 Google Scholar.

46. Tambiah, , “A Performative Approach to Ritual,” 119 Google Scholar.

47. Tambiah, , “A Performative Approach to Ritual,” 140 Google Scholar.

48. Tambiah, , “A Performative Approach to Ritual,” 141 Google Scholar.

49. For the hymns, see my Die Hymnen der chinesischen Staatsopfer, 174-303.

50. I have analyzed all the omens mentioned for the Wudi period in “Religious Anxiety and Political Interest in Western Han Omen Interpretation: The Case of the Han Wudi Period (141-87 B.C.) ；׳ Chūgoku shigaku 中國史學10 (2000), 1–31. In this paper I argue that the Hanshu contains different strata of omen records and interpre-tations for Wudi: while the auspicious records were immediately submitted by the fangshi, the calamitous ones are mostly much later retrospective interpretations. Significantly, a number of omens that were initially celebrated as auspicious were later— decades after Wudi's reign—redefined as highly unlucky and related to major political disasters; these critical retrospective interpretations are mainly preserved in the Hanshu “Monograph on the Five Phases” (“Wu xing zhi” 五行志；ch. 27).

51. Clear evidence for this comes from the initiative of Kuang Heng 匡衡 (chancellor 36-30 B.C.), who in 32 B.C. proposed to change the wording in two of the hymns; see Hanshu (Beijing: Zhonghua, 1987), 22.1057–58Google Scholar. The hymns had been maintained together with the ritual structures, especially cosmic altars, of the Wudi reign; see my “Ritual, Text, and the Formation of the Canon.”

52. Wheelock, , “The Problem of Ritual Language: From Information to Situation,” 58 Google Scholar.

53. Austin, John L., How To Do Things With Words (Oxford: Oxford University Press, 1962)Google Scholar.

54. Searle, John R., Speech Acts (Cambridge: Cambridge University Press, 1969)CrossRefGoogle Scholar; and Searle, , “A Taxonomy of Illocutionary Acts,” in Language, Mind, and Knowledge, ed. Gunderson, Keith (Minneapolis: University of Minnesota Press, 1975), 344–69Google Scholar.

55. A substantial monograph on ritual speech acts is Werlen, Iwar, Ritual und Sprache: Zum Verhältnis von Sprechen und Handeln in Ritualen (Tübingen: Gunter Narr, 1984)Google Scholar.

56. Wallace, Anthony F.C., Religion: An Anthropological View (New York: Random House, 1966), 233 Google Scholar.

57. Tambiah, Stanley J., “The Magical Power of Words,” MAN n.s. 3.2 (1968), 179 CrossRefGoogle Scholar. See also his “A Performative Approach to Ritual,” 132-33:

The passage of new information as such from one person to another is only one aspect of social communication, and in ritual, which we have seen to be formalized and predictable, this aspect may be subordinate and of little relevance.… Social communication, of which ritual is a special kind, portrays many features that have little to do with the transmission of new information and everything to do with interpersonal orchestration and with social integratio n and continuity.

58. Shiji 史記 (Beijing: Zhonghua, 1982), 12.475, 28.1398 Google Scholar; Hanshu, 25A.1235 Google Scholar.

59. Bloch, Maurice, “Symbols, Song, Dance and Features of Articulation: Is Religion an Extreme Form of Traditional Authority?” European Journal of Sociology 15.1 (1974), 55–81 CrossRefGoogle Scholar; and “Introduction,” in Political Language and Oratory in Traditional Society, ed. Bloch, Maurice (London: Academic Press, 1975), 1–28 Google Scholar.

60. Bloch, , “Symbols, Song, Dance and Features of Articulation,” 60–61 Google Scholar.

61. Bloch, , “Symbols, Song, Dance and Features of Articulation,” 71 Google Scholar.

62. Wheelock, , “From Information to Situation,” 56 Google Scholar.

63. While the Yi li reflects the centrality of role and status throughout its meticulous (and tedious) accounts of particular ritual acts, the Li ji, much in the tradition of the Xunzi, presents the underlying ideology of ritual as social stratification.

64. Tambiah, Compare, “A Performative Approach to Ritual,” 153 Google Scholar:

[T]here is on the one hand an ontological and experiential constraint that leads to formalization and archaism through the performance of cosmological archetypes, and on the other hand, a social constraint that allocates to persons in ranked positions and relations of “power of solidarity” a differential access to and participation in a society's major rites, and a differential enjoyment of their benefits.

65. Hymns that explicitly praise a specific ancestor are exceptions. Yet even here, the ancestor is addressed not by his personal but by his temple name; that is, as a paradigmatic embodiment of those virtues that are expressed in this posthumous designation. See the “Shiia jie” 認法解 chapter of the Yi Zhou shu 逸周書 (ch. 6 in the principal editions, e.g., Lu Wenchao's 盧文诏[1717-1796] collated Yi Zhou shu in the Baojing tang congshu 抱經堂叢書 of 1786).

66. Wheelock, , “From Information to Situation,” 61 Google Scholar.

67. Wheelock, , “From Information to Situation,” 63 Google Scholar.

68. The term “ritual coherence,” borrowed from the Egyptologist Jan Assmann (Das kulturelle Gedächtnis, 87—89, passim), refers to ritual as the key factor to bring about cultural continuity.

69. This is even true for times of ritual reforms, for example, under Han Wudi when the fangshi also claimed to design the new rituals according to old writings and charts inherited from the Yellow Emperor. See my “Ritual, Text, and the Formation of the Canon”.

70. Assmann, , Das kulturelle Gedächtnis, 22–23 Google Scholar: “Das kulturelle Gedächtnis deckt sich weitestgehend mit dem, was innerhalb der Gruppe an Sinn zirkuliert.” The following discussion is intended to introduce the main aspects of Assmann's theory as they may be applied to the study of early Chinese ritual culture. This theory must not be confused with notions of the “collective unconscious” of the Jungian type; the cuiturai memory is neither biologically hereditary nor unconscious. Assmann develops his ideas from the work of Maurice Halbwachs, a French sociologist who in the first half of the twentieth century formulated a theory of “collective memory” (mémoire collective), that is, of memory as a social phenomenon; see his On Collective Memory, ed. and trans. Coser, Lewis A. (Chicago: University of Chicago Press, 1992)Google Scholar. Halbwachs, a professor at the Collège de France since 1944, was murdered in the concentration camp of Buchenwald near Weimar on March 16, 1945, one day before Henri Maspero died there. One cannot help pausing for a moment to ponder how these two men, while together in Buchenwald for months, may have enriched each other's existence, or what they were able to maintain of it.

71. Assmann, , Das kulturelle Gedächtnis, 52 Google Scholar; see also 75–78.

72. Assmann, , Das kulturelle Gedächtnis, 53 Google Scholar.

73. For a summary of this thesis in connection with early China, incorporating the relevant scholarship, see Falkenhausen, , “Issues in Western Zhou Studies,” 161–67Google Scholar.

74. See my “Ritual, Text, and the Formation of the Canon.”

75. There is a vast amount of scholarship on the highly complex issues of “orality” and “literacy” in ancient Greece. A perspicacious study is Thomas, Rosalind, Literacy and Orality in Ancient Greece (Cambridge: Cambridge University Press, 1995)Google Scholar. For Egypt, see Assmann, Jan, Das kulturelle Gedächtnis, 167–95Google Scholar, passim. In emphasizing the status of the written text in comparison to ritual performance, I wish to make it clear that I do not suggest that the songs of the Shi are essentially oral compositions, as claimed by CH. Wang, The Bell and the Drum. It is impossible to transfer the Lord-Parry hypothesis of oral composition from the case of the Yugoslav bards to any early Chinese text. In a literate world like that of Eastern Zhou China (or of ancient Egypt and Greece), the question of oral composition must be distinguished from oral performance and oral transmission. On this distinction—which helps to guard against simplistic demar-cations between “orality” and “literacy”—see Finnegan, Ruth, Oral Poetry: Its Nature, Significance, and Social Context (Bloomington and Indianapolis: Indiana University Press, 1992), 16–24 Google Scholar.

76. Chunqiu Zuo zhuan zhengyi 春秋左傳正義 (Shisan jing zhushu ed.), 27.209b (Cheng成13). I quote here the translation given by Shaughnessy, Edward L., “Military Histories of Early China: A Review Article,” Early China 21 (1996), 159 CrossRefGoogle Scholar. Shaughnessy's perspicacious translation of this sentence (and correction of the usual “… reside in sacrifice and war”) is of utmost importance.

77. See shi bowuguan, Jingmen, Guodian Chu mu zhujian, 194 Google Scholar.

78. Qiu Xigui 裘錫圭 suggests that the graph might be read as either zhi 志 or shi 詩； see shi bowuguan, Jingmen, Guodian Chu mu zhujian, 200n.6.1Google Scholar doubt that shi 詩 is a meaningful option here; my choice of zhi 志 and its translation as “aspirations” is, of course, related to the early notion of shi y an zhi 詩言志 (“the songs express aspirations”) that first appears in the “Yao dian” 堯典 chapter of the Shang shu; see Shang shu zhengyi 尙書正義 (Shisan jing zhushu ed.), 3.19c (here, in the “old text” Shang shu, in the “Shun dian” 舜典 chapter).

79. Jeffrey Riegel has characterized the early understanding of the Shi as “history told in verse”; see his “Eros, Introversion, and the Beginnings of Shijing Commentary,” Harvard Journal of Asiatic Studies 57 (1997), 171 Google Scholar.

80. shi bowuguan, Jingmen, Guodian Chu mu zhujian, 195 Google Scholar.

81. shi bowuguan, Jingmen, Guodian Chu mu zhujian, 194 Google Scholar.

82. shi bowuguan, Jingmen, Guodian Chu mu zhujian, 188 and 179 Google Scholar.

83. shi bowuguan, Jingmen, Guodian Chu mu zhujian, 129-31, 167-68, and 149–51Google Scholar.

84. For example, of the twenty-three Shi quotations in the “Zi yi” manuscript, nine come from the xiao ya, nine from the da ya, two from the “Cao feng” 曹風song “Shi jiu” 鳴鳩 (Mao 152), two from the “Zhou nan” 周南 section, and one from a song not included in the received text. This quotation pattern matches those in other Li ji chapters like “Zhong yong” 中庸 and “Da xue” 大學： the “Zhong yong” quotes twice from the xiao ya, eight times from the da ya, three times from the Zhou song 周頌, once from the Shang song 商頌, once from the guo feng, and once from a text that cannot be identified with certainty. The “Da xue” quotes four times from the xiao ya, three times from the da ya, once from the Zhou song, once from the Shang song, and three times from the guo feng.

85. For a study of the Shi quotations in Warring States and early Han manuscripts, see my “The Shi in Excavated Manuscripts,” a paper delivered at the conference “Text and Ritual in Early China,” Princeton University, October 2000, and forthcoming in the conference volume of the same title.

86. For duan zhang qu yi, referring to the practice of taking isolated Shi lines out of their original context and quoting them according to one's own needs, see Chunqiu Zuo zhuan zhengyi, 38.298a (Xiang 襄 28)Google Scholar.

87. Assmann, As, Das kulturelle Gedächtnis, 75 Google Scholar, has noted, the “fiction” (myth) / “reality” (history) antagonism “has been up for discharge for quite some time.” White, Hayden, Tropics of Discourse: Essays in Cultural Criticism (Baltimore and London: Johns Hopkins University Press, 1985), 126 Google Scholar, has made the same point: “We are no longer compelled … to believe—as historians in the post-Romantic period had to believe — that fiction is the antithesis of fact (in the way that superstition or magic is the antithesis of science) or that we can relate facts to one another without the aid of some enabling and generically fictional matrix.” For the relevant discussion that substantiates this claim, see White, Tropics of Discourse, esp. 51-134; White, , The Content of the Form: Narrative Discourse and Historical Representation (Baltimore and London: Johns Hopkins University Press, 1990), esp, 1-57, 142–84CrossRefGoogle Scholar; and White, , Metahistory: The Historical Imagination in Nineteenth-Century Europe (Baltimore and London: Johns Hopkins University Press, 1975)Google Scholar.

88. Assmann, , Das kulturelle Gedächtnis, 56–57 Google Scholar; see also 143–44. Another excellent study that develops the same point is Connerton, Paul, How Societies Remember (Cambridge: Cambridge University Press, 1989)CrossRefGoogle Scholar.

89. Wang, , From Ritual to Allegory, 75 Google Scholar. These pieces are numbered 236 (“Da ming” 大明), 237 (“Mian” 緜), 241 (“Huang yi” 皇矣), 245 (“Sheng min” 生民), and 250 (“Gong Liu” 公劉) in the order of the Mao shi. Wang, aiming to reconstruct what he calls the Weniad (the epic of King Wen文 of Zhou [r. 1099/56-1050 B.C.]), proposes that their logical order should be 245, 250, 237, 241, and 236 (see From Ritual to Allegory, 74). A similar logic is applied by those who have, as mentioned at the beginning of the present study, attempted to reconstruct the da wu suite. These re-orderings of texts which are not placed together in the received Shijing text reveal an interesting assumption on the side of the modern scholars: if their well-reasoned rearrangement of the hymns should indeed restore “the original order” it would mean that the compiler(s) of the Mao version had missed the original coherence and significance of these hymns.

90. Assmann, , Das kulturelle Gedächtnis, 71 Google Scholar.

91. Shaughnessy, , Sources of Western Zhou History, 176–77Google Scholar.

92. This, again, is typical. Assmann, , Das kulturelle Gedächtnis, 71—72, 83–86 Google Scholar, does not forget to mention the “alliance of sovereignty and oblivion” and the terror of dictatorship trying to eliminate memory; under the latter circumstances, memory turns into an instrument no longer of control but of resistance.

93. Based on the above discussion on material ornament and the pragmatic aspects of ritual language, I understand the inscriptions also as ritual performance texts. Even if we may be unable to prove that the contents of the inscriptions were recited during the ceremony, they still show all the paradigmatic characteristics outlined above.

94. On this aspect of the cultural memory, see Assmann, , Das kulturelle Gedächtnis, 79–80 Google Scholar.

95. Mengzi zhushu 孟子注疏 (Shisan jing zhushu ed.), 8A.63c. Shi 詩 here may be taken as a generic category, referring to the royal ritual hymns, rather than specifically to those hymns that are preserved in the Shi jing.

96. Ban Gu 班固 (32-92), in the preface to his “Liang du fu” 兩都賦 (Rhapsody on the two capitals); see (Liu chen zhu 六臣注) Wen xuan 文選 (Sibu congkan 四部叢刊 ed.), 1.1b. Kings Cheng and Kang are eulogized together in “Zhi jing” 執競 (Mao 274); see Mao shi zhengyi, 19/2.321c.

97. See Maspero, Henri, “Le Ming-t'ang et la crise religieuse avant les Han,” Mélanges chinois et bouddhiques 9 (1951), 1–171 Google Scholar.

98. In general, I follow the rhymes as identified in Baxter, , A Handbook of Old Chinese Phonology, Appendix B, 584–743 Google Scholar (for “Chu ci” see 679–81), and note the few but significant cases where I deviate from his order.

99. In addition to the Shisan jing zhushu edition, I use the following major works of Qing philological scholarship: Jihengr, Yao Shijing tonglun, 11.229–31Google Scholar; Ma Ruichen 馬瑞辰 (1782-1835), Mao shi zhuan jian tongshi 毛詩傳箋通釋 (Beijing: Zhonghua, 1989), 21.698–708 Google Scholar; Huan, Chen 陳奥 (1786–1863), Shi Mao shi zhuan shu 詩毛氏傳疏 (Beijing: Zhongguo shudian, 1984), 20.26a–32b Google Scholar; Yurun, Fang, Shijing yuanshi, 11.430–433 Google Scholar; and Xianqian, Wang 王先謙 (1842–1917), Shi sanjia yi jishu 詩三家義集疏 (Beijing: Zhong-hua, 1987), 18.749–55Google Scholar. In addition to the Qing sources, I have consulted the most important Song commentary, Zhu Xi's 朱熹 (1130-1200) Shi ji zhuan 詩集傳 (Sibu congkan ed.), 13.6b-11a. Useful commentaries may also be conveniently found in Wanli, Qu 屈萬里, Shi jing quanshi 詩經證釋 (Taipei: Lianjing, 1984 [Qu Wanli xiansheng quanji 屈萬里先生全集 5 ]), 403–6Google Scholar; or in Heng, Gao 高亨， Shi jing jin zhu 詩經今注 (Shanghai: Shanghai guji, 1987), 321–25Google Scholar. For Karlgren, see his “Glosses on the Siao Ya Odes,” Bulletin of the Museum of Far Eastern Antiquities 16 (1944), 132–38Google Scholar.

100. I am most grateful to Donald Harper for alerting me to this issue as a crucial problem in reading “Chu ci.”

101. Beginning in Stanza 1, the recurrent first-person pronoun wo refers to the “host” (zhuren 主人) or principal descendant in charge of the offerings, and, by extension, to the ritual community he is representing. Zheng Xuan comments on the present stanza that “the invocator (zhu 祝)uses the host's phrases to urge him/them (i.e., the impersonator[s] of the spiritfs]), to assist the offering son (xiao zi 孝子) to receive great blessings.” (On xiao as “offering” see below.) Kong Yingda accepts the reading, noting that “all the words of the host and the impersona tor (s) are transmitted by the invocator” (Mao shi zhengyi, 13/2.200a). This commentary is supported by the functional descriptions of the Great Invocator (da zhu 大祝)and the Minor Invocator (xiao zhu 小祝)in the Zhou li, according to which the impersonator (shi 尸) is ushered into the temple and seated at his position by these officials, and not by the host himself; see Sun Yirang 孫詒讓 (1848–1908), Zhou li zhengyi 周禮正義 (Beijing: Zhonghua, 1987), 49.2021–23Google Scholar (“Da zhu” 大祝) 50.2034–35 (“Xiao zhu” 小祝).See also, Yi li zhushu 儀禮注疏 (Shisan jing zhushu ed.), 45.239c-240a (“Te sheng kui shi li” 特牲饋食禮), 48.257b (“Shao lao kui shi li”少牢饋食禮); and Li ji zhengyi, 26.229c (“Jiao te sheng” 郊特牲). Huan, Chen, Shi Mao shi zhuan shu, 20.26b Google Scholar, believes that the final lines of the stanza are the words from the spirits, as presented to the host by the invocator; I do not follow this interpretation but read the passage as a prayer directed towards the spirits, which is then matched by their answer at the end of the following stanza. It is important to recognize the tally-like match between prayer and answer as a structurai feature of guci wording: it precludes any doubts about the efficacy of the descendant's efforts (see Xu Zhongshu, “Jinwen guci shili,” 9).

102. Ci 茨 “caltrop” is Tribulus terrestris, L.; see Bernard E. Read, Chinese Medicinal Plants from the Pen Ts'ao Kang Mu A.D. 1596: 3rd Edition of a Botanical, Chemical and Pharmacological Reference List (repr. Taipei: Southern Materials Center, 1982), no. 364. Chu chu 楚楚 “thorny, thorny” is problematic; my translation follows the Mao commentary. Karlgren, , “Glosses on the Kuo Feng Odes,” Bulletin of the Museum of Far Eastern Antiquities 14 (1942), 228 (no. 360)Google Scholar, has argued against this reading on the basis of the only parallel case of chu chu in the Shi in “Fu you” 特辦 (Mao shi zhengyi, 7/3.116b [Mao 150]): yi shang chu chu 衣裳楚楚. There Mao glosses chu chu as “the appearance of freshness and brightness” (xian ming mao 鮮明貌) ， The issue of Chinese reduplicatives , and also the practice of translating them into repetitions of English words, requires some basic considerations. Reduplicatives (chongdie 重暴)—a linguistic device very common to both Eastern Zhou inscriptions and Shijing songs—are a peculiar phenomenon primarily of poetically structured texts. For the Shijing, see Fagao, Zhou 周法高, Zhongguo gudai yufa: gouci bian 中國古代語法：構詞編 (Taipei: Institute of History and Philology, Academia Sinica, 1962), esp. 114-28, 154–60Google Scholar; and Zhou, 34 (1962/63), 661-98. For the Chu ci 楚辭, see Gallagher, Martha Wangliwen, “A Study of Reduplicatives in the ‘Chu Ci’” (Ph.D. diss., Yale University, 1993)Google Scholar. In bronze bell inscriptions, reduplicatives are usually employed as onomatopoeia to convey the sound of the bells on the linguistic level; the same function as words expressing sounds can be identified in a number of Shijing songs. George A. Kennedy has argued that reduplicatives are not constructed by duplicating a common word but that they are primary forms themselves, usually based on a rather rare or particular word; see his “A Note on Ode 220,” in Studia Serica Bernhard Karlgren Dedicata. Sinological Studies Dedicated to Bernhard Karlgren on his Seventieth Birthday, ed. Egerod, Søren and Glahn, Else (Copenhagen: E. Munksgaard, 1959), 190–98Google Scholar. Kennedy's observation explains some of the trouble in glossing or translating these terms. Like rhyming or alliterative binômes in early Chinese literature, reduplicatives do not have a fixed meaning that they bring to the text; instead, they assume their meaning according to the specific context (see also the discussions in Knechtges, David R., Wen xuan, or Selections of Refined Literature, vol. 2 [Princeton: Princeton University Press, 1987] 3–12 CrossRefGoogle Scholar; and my Die Hymnen der chinesischen Staatsopfer, 194). From this perspective, Karlgren's argument (which he tries to corroborate by references to the single use of chu elsewhere in the Shi) seems to lose much of its strength. Even if some reduplicatives appear to convey the meaning of the word that is written by the single character, I remain inclined to see the vast majority of them operating primarily on the aesthetical level of the text, especially as sounds. Knechtges (quoted above) has shown that graphic interpretations of reduplicatives can be very problematic. Sima Xiangru's 司馬相如 (179-117 B.C.) fu 賦—for some of which we are fortunate to have different versions— provide compelling evidence that (a) the writing of such words was not yet fixed in Western Han times and (b) they were conceived essentially as sounds in texts that were recited (i.e., performed); see Kamatani Takeshi签谷武志, “Fu ni nankai na ji ga ōi no wa naze ka: Zen-Kan ni okeru fu no yomarekata” 陚に難解な字が多いのはなぜか：前漢における賦の讀まれかた, Nihon Chūgoku gakkai hō 日本中國學會報 48 (1996), 16–30 Google Scholar. This performative dimension leads me to translate the reduplicates into repetitions of English words although the representation of sound operates along different principles in the two languages. As such, my translations are hybrid: trying to catch both the meaning derived from the context and the very pattern of repeated sound (Knechtges, in his Wen xuan translation, has chosen a similar approach to rhym-ing and alliterative binomes). Sacrificing the latter dimension would be a high price to pay for a more “natural’, reading.

103. Here, yan 言， which I translate as “so” in this line, functions like the conjunction nai 乃 “then”; see Xuehai, Pei 裴學海, Gu shi xuci jishi 古書虛字集釋 (Beijing: Zhonghua, 1982), vol. 1, 434–35Google Scholar.

104. In accordance with the “Minor preface” to this hymn, Zheng Xuan under-stands this line as referring to the ancient kings; the recurrent first-person pronoun wo 我, beginning with line four, does not seem to support this reading.

105. For ji 稷 “panicled millet” (Panicum miliaceum, L.) and shu 泰 “glutinous millet” (P. miliaceum, L. var. glutinosa, Bretsch.), see Read, Chinese Medicinal Plants, nos. 751, 752. Complementing the sacrificial meat, the two species of millet are used for the principal offerings (including the alcoholic beverages) in Zhou ancestral rites; see, e.g., the sacrificial hymns “Xin nan shan,” “Fu tian,” and “Da tian” (Mao shi zhengyi, 13/2.2033; 14/1.205b, 206b, 207c, 209b), as well as the sacrificial chapters of the Yi li, “Shao lao kui shi li” and “Te sheng kui shi li” (see Yi li zhushu, 45.239c-240c; 48.256c-257a, 257c, 259c).

106. Mao glosses yi 億 as wan wan 萬萬 “one hundred million”; Zheng Xuan, whom I follow here, gives shi wan 十萬 “hundred-thousand”; either number represents fullness in an innumerable measure; see Xianqian, Wang, Shi san jia yi jishu, 18.750 Google Scholar, Ma, Ruichen, Mao shi zhuan jian tongshi, 21.700–701 Google Scholar.

107. Mao glosses tuo 妥 “to assuage” as an zuo 安坐 “to seat at ease,” and you 侑 “to provision” as quan 勸 “to urge.” Elaborating on these very concrete glosses of other-wise more general terms, Zheng Xuan and Kong Yingda see the addressees of these actitvities as the ancestors and their impersonators who are led “to take places at their spirit positions” (shi chu shen zuo 使處神坐)—a reading that is supported by descriptions in several of the San li (see the passages mentioned in the introductory note to this stanza, n. 101).

108. Jie 介, often glossed as da 大 “great,” is used as a verb here and could be under-stood as “to increase”; Zheng Xuan glosses the term here as zhu 助 “to assist.” However, as Xu Zhongshu, “Jinwen guci shili,” 6, has shownf jie is interchangeable (and almost homophonous) with gai “to pray”; both words are common in texts directed to the ancestral spirits. A parallel for jie in the meaning “to pray” (and again glossed as zhu by Zheng Xuan) may be found in the sacrificial hymn “Fu tian” (Mao shi zhengyi, 14/1.206b [Mao 211]): 琴瑟擊鼓/以御田祖/以祈甘雨/以介我稷黍/以穀我士女 (“[We] play the zithers, strike the drums, I to invoke the Father of husbandry, / to beg for sweet rain; / to pray for our glutinous millet and panicled millet, / to request a bun-dance for our men and women”). Zheng Xuan glosses jing 景 as da 大 but I am inclined to maintain the light metaphor implied in the primary meaning of the term. In Zhou ritual language, metaphors of light are most commonly associated with the ancestors and their blessings; for a discussion of epigraphic examples, see Constance Cook, A., “Auspicious Metals and Southern Spirits: An Analysis of the Chu Bronze Inscriptions” (Ph.D. diss., University of California, Berkeley, 1990), 218–37Google Scholar. According to the traditional commentaries, the whole stanza comprises three distinctive parts: an introduction ending with the rhetorical question of line three, an idealized account of the agricultural preparations for the sacrifice (11. 4-8), and the ultimate purpose of these preparations (11. 9–12). This semantic structure is enhanced by formal characteristics: all the lines of the middle part begin with the first-person pronoun wo 我, while all the lines of the final part begin with yi 以. In the middle part, lines five and six are syntactically identical, as are lines seven and eight. In the final part, lines ten and eleven are syntactically identical, and lines nine and twelve share a very similar structure with one another. The final portion is almost entirely composed of gu ci “auspi-cious words” formulae that also occur in other Shi jing hymns of the ya and song sections: line nine, yi wei jiu shi 以爲酒食 (“With them, [we] make ale and food”) is again used in “Xin nan shan” 信南山 (Mao shi zhengyi, 13/2.203a [Mao 210]). The final line, yi jie jing fu 以介景福 (“to pray for radiant blessings”), appears in four other hymns: “Da tian” 大田 (Mao shi zhengyi, 14/1.209b [Mao 212]), “Han lu” 旱麓 (Mao shi zhengyi, 16/3.248a [Mao 239]), “Hang wei” 行葦 (Mao shi zhengyi, 17/2.267c [Mao 246]), and “Qian” 潛 (Mao shi zhengyi 19/3.327c [Mao 281]). In “Han lu”, the formula appears at the end of stanza four, the only stanza (of altogether six stanzas) that deals directly with the sacrifice; in the three other hymns, “to pray for radiant blessings” is the very last line of the whole text. Moreover, in both “Han lu,” and the Zhou song “Qian,” the formula is preceded by yi xiang yi si 以享以祀 (“to offer, to sacrifice”), which is the formula of line ten in Stanza 1 of “Chu ci.”

109. The perspective of speech in this stanza hinges mainly on the understanding of line two and lines ten to twelve. I follow Kong Yingda, who takes jie 絮 (= 潔) as a full verb, er 爾 as the personal pronoun “you,” and “oxen and sheep” as the direct object of jie. See Mao shi zhengyi, 13/2.200b: 乃鮮絜爾王者所祀之牛羊 (“[you] then purify the oxen and sheep that your king will sacrifice”) where “you” refers to the dignitaries who are described in the preceding line. The translation of er as “you” necessarily transforms the text—probably the entire stanza—into a direct speech instead of a narrative description. Legge, , The She King, 369 Google Scholar, reads er as a modal auxiliary, rendering the preceding jie as “pure-like are,” since with er “as the pronoun, the line is to me altogether out of connection.” The speaker of the stanza, if it is a direct speech recapitulating the preparations for the sacrifice, seems to be a ritual official; since by the end of the stanza he utters “auspicious words” towards the descendant, he is most probably one of the invocators. (See also n. 101 above, quoting Kong Yingda that the words between the host and the impersonator[s] are always transmitted by the invocator.) I deviate from Kong Yingda's reading in assuming that “you” refers not to the dignitaries but to the host, i.e., the offering descendant mentioned in the congratulatory formula of line ten. Note that this line, “The offering descendant shall have benison,” is parallel to a line in the hymn “Bi gong” 悶宮 (Mao shi zhengyi, 20/2.347c [Mao 300]) where it is clearly addressed to the offering descendant (again according to the use of the pronoun er in the following lines).

110. Zheng “winter sacrifice” and chang 嘗 “autumn sacrifice” are standard technical terms of ancestral sacrifices, as identified by the commentators; see also, Zhou li zhengyi, 33.1330 Google Scholar (“Da zong bo” ״大宗伯), and Hequan, Zhang 張鶴泉, Zhou dai jisi yanjiu 周代祭祀研究 (Taipei: Wenjin, 1993), 148–53Google Scholar. The chronological inversion in the present line may be due to the need to fit the line into the sequence of yang 陽(-ang) category rhymes.

111. The lines describing the preparations of the sacrificial meat represent the notion of order and division of labor that governs early Chinese ritualism, with every task being assigned to a particular specialist. According to the Zhou li, the nei yong 內賽 “palace chefs” are in charge of flaying and boiling at the ancestral sacrifices ( Zhou li zhengyi, 8.274 Google Scholar); the peng ren 亨人” stove attendants” are in charge of cooking under the supervision of the net yong (Zhou li zhengyi, 8.282).

112. According to Zheng Xuan, the spirits were thought to move around inside the gate of the ancestral temple; the offering descendant ordered the invocator to attract them by sacrificing to them near the gate.

113. Both this and the next line also occur in the hymn “Xin nan shan” (Mao shi zhengyi, 13/2.203c [Mao 210]). In “Chu ci,” Zheng Xuan glosses ming 明 “shining” as bei 備 “complete”; as in line twelve of Stanza 1,1 prefer to maintain the light metaphor.

114. To bring out the dynamic nature of this line (see also the following line, as well as line ten of the Stanza 3), I translate shi 是 as strictly grammatical, i.e., as the resuming pronoun which refers back to the preposed object; see Pulleyblank, Edwin G., Outline of Classical Chinese Grammar (Vancouver: University of Bristish Columbia Press, 1996), 70 Google Scholar. I reconstruct the subject as “you” (i.e., the offering descendant), who seems to be also the object of the congratulatory formula in line ten. The precise meaning of xianzu 先祖 “former ancestor(s)” is difficult to determine. It is not honorific in the sense of huangzu 皇祖 “august ancestor(s)״ but can still be used both in narratives and also to address the ancestors directly; for the latter, see the hymn “Yun han״雲漢 ( Mao shi zhengyi, 18/2.294b [Mao 258]Google Scholar). Xianzu seems to be a more generic term and interchangeable with zukao 袓考； neither term appears in the 爾雅 nomen-dature of kin designations (see Er ya zhushu, 4.26b–27c Google Scholar). Xianzu may refer not to one specific ancestor but to the ancestors in general and could be taken both in the singu-lar or in the plural. In the Liji, zukao is in one instance defined as the first and highest of the seven ancestors to whom the Son of Heaven presents his sacrifices ( Liji zhengyi, 46.361a Google Scholar [“Ji fa” 祭法]), but the terminology used there is at odds with that of another Li ji passage ( Liji zhengyi, 12.107a Google Scholar [“Wang zhi” 王制]); see also Songdao, Guo 郭嵩煮 (1818–1891), Li ji zkiyi ！記質疑 (Changsha: Yuelu, 1992), 563–65Google Scholar. Huang 皇, glossed as da 大 “great” by Mao, has been read as wang BS by Zheng Xuan; wang needs to be read as a loan character for wang 往 “to return” (like gui wang 歸往), see Ruichen, Ma, Mao shi zhuan jian tongshi, 21.702–3Google Scholar. Although Zheng's reading has been widely accepted, it has been rejected by Mou Ting 牟庭 (1759–1832), Shi qie 詩切 (Jinan: Qi Lu, 1983), vol. 3, 1567, and by Karlgren, , “Glosses on the Kuo Feng Odes,” 134 (no. 661)Google Scholar.

115. The term shenbao 神保 “divine protector” has been a puzzling issue. Within the whole Confucian canon, the term appears only in the present hymn. While the earliest commentators, Mao and Zheng, take bao 保as a verb meaning an 安 “peaceful,” Xi, Zhu, Shi ji zhuan, 13.8a Google Scholar, as well as more recently Guowei, Wang 王國維, “Song Yuan xiqu kao” 宋元戲曲考 (in Wang Guowei yishu 王國維遺書 [Shanghai: Shang-hai guji, 1983], vol. 15), 2a-bGoogle Scholar, and Zhongshu, Qian 錢鍾書, Guan zhui bian 管錐編 (Beijing: Zhonghua, 1986), vol. 1 156–58Google Scholar, take shenbao as an honorific designation for the impersonator(s), shi 尸. Heng, Gao, Shi jing jin zhu, 323 Google Scholar, and Wanli, Qu, Shi jing quan shi, 404 Google Scholar, take shenbao as referring to the spirits in general. (The most speculative interpretation has been offered by Ting, Mou, Shi qie, vol. 3, 1567 Google Scholar, who reads bao 保 as the feathered bao 傑cover on the spirits' chariot.) All these assumptions lack concrete support from any early parallel passages (though Wang and Qian try to read shenbao as identical with lingbao 靈保,a term that appears in the Chuci 楚辭 anthology), and some of them (Zhu, Wang, Qian) clearly contradict the text as a whole. The most fruitful approach to the problem appears to be that by Yingda, Kong (Mao shi zhengyi, 13/2.200c)Google Scholar, who is followed by Jiheng, Yao, Shijing tonglun, 230 Google Scholar, and Ruichen, Ma, Mao shi zhuan jian tongshi, 21.703 Google Scholar. According to these commentators, shenbao refers to the spirit(s) but in a particular way: while terms like xianzu 先祖or zukao 祖考 denote the ancestors as the deceased persons, shenbao denotes their essence as ancestral spirit(s) in heaven. Following this interpretation (for which Ma furnishes textual support), I understand shenbao as the ancestral spirit(s) called down into the impersonator(s): the “divine protector(s),” as he or they appear in the text, would be the ancestral spirit(s) vis-à-vis the descendants. As for xianzu in the preceding line, I have no con-elusive evidence whether shenbao should be taken in the singular or plural.

116. As noted above, this line occurs verbatim in the hymn “Bi gong” 悶宮. In pre-Confucian texts, there is no reason to translate xiao sun 孝孫 as “filial descendant”; instead, we have ample evidence that xiao 孝 means “to give offerings to one's anees-tors” in early Zhou texts, and that the later transformation of the word to mean “filial piety” towards one's living parents represents a step from ritual practice to social ideology. For one recent assessment of the question, see Knapp, Keith N., “The Ru Reinterpretation of Xiao ” Early China 20 (1995), 195–222 CrossRefGoogle Scholar. Knapp considers the possibility that xiao in early times also could refer to some concrete material support towards one's living parents, in contrast to the later notion of general filial piety. See also n. 131 to Stanza 4, line five, below.

117. Another possible reading would be “[The spirits] will requite [your] praying for blessings.” However, in this line one may accept the common interpretation of jie as da “great,” as it is given in the traditional commentaries.

118. Very similar to Stanzas 1 and 3, Stanza 2 closes with a typical guci formula. In Stanza 2, Xianqian, Wang, Shi sanjia yi jishu, 18.752 Google Scholar, notes that the final three lines are “the words by which the invocator on behalf of the impersonator(s) conveys the bless-ings to the host” ( Huan, Chen, Shi Mao shi zhuan shu, 20.28a, is similarGoogle Scholar). I share this reading though I think the whole stanza is actually uttered by the invocator. As in Stanza 1, only the latter part and in particular the final couplet is highly formulaic in the sense that most of these lines can be found elsewhere in the Shijing. The concluding couplet is identical with the final lines of “Xin nan shan” (Mao shi zhengyi, 13/2.203c [Mao 210])Google Scholar and “Fu tian” (Mao shi zhengyi, 14/1.207c [Mao 211]Google Scholar). The final line of Stanza 2 appears again in “Qi yue” 七月 (Mao shi zhengyi, 8/1.124b [Mao 154]), “Tian bao” 天保 ( Mao shi zhengyi, 9/3.144b [Mao 166]Google Scholar), and “Nan shan you tai” 南山有臺 (Mao shi zhengyi, 10/1.151c [Mao 172]). Note that with the exception of “Qi yue,” all verbatim parallels in the Shi jing to full lines of “Chu ci” are in the ya and song sections.

119. There is no strong evidence to determine the speaker of Stanza 3. Since the concluding guci are almost identical with those of Stanza 2 and since Stanza 4 begins with a statement by the offering descendant, I assume that the speech of the present stanza—if it is a direct speech at all—is probably directed to the descendant as in Stanza 2. A key to the relation between Stanzas 2 and 3 may be found in the comparison of the ritual personnel mentioned. In both stanzas, the divine protector is necessary as the subject of the final “auspicious words,” but with respect to all the otheritual partici-pants, the two stanzas seem to represent distinctively differergroups not of persons but of ritual roles: in Stanza 2, we meet the invocator, the ancestor, and the offering descendant; Stanza 3, by contrast, mentions the furnace managers, the noble wives, and the guests and visitors. Registering the communal efforts to perform the ancestral sacrifice, the two stanzas do not duplicate but complement each other, with Stanza 2 clearly representing the leading stratum of those involved in the spiritual exchange proper (including the spirits themselves; only the impersonator is missing) while Stanza 3 is confined to the lower auxiliary roles. As such, it is not unlikely that Stanza 3 may indeed continue the speech of Stanza 2; the redundancy in the almost identical words that close both stanzas can be seen as a typical element of rhythmic ritual litany.

120. The nominalizing reading of zhi cuan 執爨 as “furnace managers” follows Kong Yingda's commentary.

121. The line is exactly parallel to line five below. In both cases, wei 爲should be understood in the meaning “to prepare”; kong 孑匕,which I translate as “grand,” referring to the offerings in the same way as shuo 碩 “magnificent” and shu 庶 “numerous”, may also be taken as the modifying intensifier “very.”

122. For the same line, see “Hang wei” 行葦 (Mao shi zhengyi, 17/2.266b [Mao 246]).

123. According to several passages in the sacrificial chapters of the Yi li, the zhufu 主婦 “host's wife,” leading the noble wives of other dignitaries as well as a group of female assistants, presents dishes of meat and millet (and later also drinks) during the ancestral sacrifice; see Yi U zhushu, 44.239b Google Scholar (“Te sheng kui shi li”), and 48.256c-257a (“Shao lao kui shi li”). On the role of the noble women to serve and to remove dishes during the sacrifice, see also Zhou li zhengyi, 14.554 (“jiu pin” 九樓), and Li ji zhengyi, 47.365a (“Ji yi”祭義).In the entire Confucian canon, the term junfu 君婦” noble wives” is used only in this hymn (again in Stanza 5).

124. Zheng Xuan glosses shu 庶 as na 月多 “fat meat” and explains that on the occasion of the sacrifice, the most delicious fat meat must be chosen both for the spirits and for the guests. In Zhu Xi's interpretation (Shi ji zhuan, 13.8b), shu is read as duo 多 “many”: “preparing the sacrificial plates in great numbers.” Both Ma Ruichen, Mao shi zhuan jian tongshi, 21.704, and Karlgren, “Glosses on the Siao Ya Odes,” 134 (no. 663), argue explicitly against Zheng Xuan in favor of the common meaning “many, numerous.” The modern commentators Wanli, Qu, Shi jing quanshi, 405 Google Scholar, and Heng, Gao, Shi jing jin zhu, 323 Google Scholar, also accept this reading.

125. The “guests” and “visitors” may be two different categories of participants, yet the commentators do not specify them. According to Mao's laconic note, the host “treats as guests” bin 賓the impersonator(s) (shi 尸 ) and also the guests (ke 客): 賓尸及賓客. This, however, may not mean that Mao understands both to be included in this line of the hymn (that is, including the shi 尸 as bin 賓,as some later commentators have interpreted his note; see Zhuo, Huang 黃悼, Shi shu pingyi 詩疏平議[Shanghai: Shanghai guji, 1985], 379)Google Scholar.

126. For the complex ceremonial exchange of drinks between the host, the invocator, the impersonator(s), the host's wife, and the guests, see Yi li zhushu, 45.240c-247b (“Te sheng kui shi li”), and 48.258b–259b (“Shao lao kui shi li”).

127. The multiple voices of Stanza 4 have already been noted by Zheng Xuan, who identifies the first two lines as the statement by the offering descendant and lines five through twelve as the “auspicious words,’ which the invocator, in response, invokes from the impersonator and delivers to the host. Zheng Xuan identifies the first-person pronoun wo 我 as referring to the offering descendant; as in Stanza 1, I take his voice to be speaking as a representative for the ritual community vis-à-vis the spirits. For the full argument on the structure of Stanzas 4 and 5 see the analysis following the translation.

128. One cannot determine the speaker of this purely narrative segment of two lines. It is entirely possible, however, that even this portion was spoken within the sacrificial act, probably by some minor official; yet note that the two lines are not rhymed (see the discussion in the analysis below).

129. Grammatically, the line may also be read “The officiating invocator delivers the [spirits'] announcement”; for zhi 致 “to cause to arrive” the direction of motion is not specified lexically and must be derived from the context. Since the announcement is “presented” in the following line, I assume it is first “invoked.” In the identical line in Stanza 5 I translate zhi, again according to the context, as “to deliver.” The directional distinction is explicit only in the translation, not in the original. The reading of zhi as “to invoke” is clear in Kong Yingda's commentary on the difference between the two announcements (Mao shi zhengyi, 13/2.202a): “The invocator first [i.e., in Stanza 4] invokes the intention of the impersonator(s) to announce it to the host. Then in turn [i.e., in Stanza 5] he invokes the intention of the host to make an announcement to the impersonator(s)״ (祝先致尸意告主人乃更致主人之意以告尸). See also n. 136 to Stanza 5, line four, below.

130. The following lines are the “auspicious words” through which the spirits confirm their blessings; the perspective of speech is clear from (a) the introduction of lines three and four, and (b) the use of the personal pronoun er 爾 in lines seven and eleven. Kong Yingda, in a long note on this speech, has noted the close parallel between Stanza 4 and a passage in the Yi li; see Yi li zhushu, 48.258c (“Shao lao kui shi li”), and the translation of this passage below. However, Kong—following Mao and Zheng Xuan—believes that “Chu ci” was performed at the royal Zhou court to commemo-rate the ideal order of the past; he therefore concludes that the “auspicious words” are abbreviated in this stanza, since at the royal court, they would have been longer than those for the aristocrat in the Yi li account. Despite the early claim of the “Minor preface,” there is no evidence that “Chu ci” was indeed a royal hymn. It may very well have been performed in the rituals of an Eastern Zhou aristrocrat; one therefore does not need to follow those parts of his commentary where Kong tries to reconstruct the present hymn as representing royal sacrificial rites, as opposed to the rites of an aristocrat.

131. As noted above (n. 116), xiao is not “filial” in the later Confucian sense of the word but a technical term meaning “to offer” or “offering” in the ancestral sacrifice; see also Zhuo, Huang 黃焯 Mao shi Zheng jian pingyi 毛詩鄭著平議 (Shanghai: Shang-hai guji, 1985), 6.253 Google Scholar, and Ruichen, Ma, Mao shi zhuan jian tongshi, 21.706 Google Scholar. The compound xiao si 孝祀 is probably equivalent to xiang si 享祀 (as in line ten of Stanza 1: 以享以祀) and may be taken as “offering sacrifice”; see Huan, Chen, Shi Mao shi zhuan shu, 20.30b Google Scholar. The Er ya (Er ya zhushu, 2.11a) defines: 享,孝也. Bronze inscriptions frequently include the compound xiaoxiang 孝萏(享)； see Jinwen gulin 金文言古林, ed. Zhou Fagao周法髙 et al. (Hong Kong: The Chinese University Press, 1975), vol. 10, 5289-92. Even, Lun yu 8.21 Google Scholar still employs xiao in this sense, juxtaposing “offering” to “drinking and eating”: “With Yu, I do not find fault. He was frugal in drinking and eating but offered (xiao) to the utmost to the ghosts and spirits” (see Lun yu zhushu, 8. 32a—b). For studies on the pre-Confucian meaning and use of xiao, see Masaru, Ikezawa 池澤優, “Chūgoku kodai no ‘kō’ shisō no shisōteki imi: ‘kō’ no shūkyōgaku, sono go,” 中國古代の「孝」思想の思想的意味：「孝」の宗教學，その五, Shakai bunka shigaku 社會文化史學 31 (1993), 12–26 Google Scholar; Nomura Shigeo 野村茂夫, “Jukyōteki ‘kō’ no seiritsu izen: Shōsho O tegakari toshite” 儒教的孝の成立以前：尙書を手がかり として, Aichi kyōiku daigaku kenkyū hōkoku (jinbun, shakai) 愛知教育大學研究報告 (人文,社會) 23.1 (1974), 17–28 Google Scholar; and Changguo, Zha 査昌國, “Xi Zhou ‘xiao’ yi shi tan” 西周孝義試探, Zhongguo shi yanjiu 中國史研究 1993.2,143–51Google Scholar.

132. It is difficult to decide whether line eight refers to the preceding line or to the following ones. I accept Karlgren's arguments (against Mao and others); see “Glosses on the Siao Ya Odes,” 136 (no. 668).

133. Reading, with Karlgren, “Glosses on the Siao Ya Odes,” 136-37 (no. 669), qi 齊 as zi 齊 “sacrificial grain.” In the following line I again follow Karlgren, “Glosses on Siao Ya Odes,” 137 (no. 670).

134. In the first four lines of Stanza 5,1 reconstruct the two different perspectives as parallel to the preceding stanza. There is no other indication regarding the speaker of the first two lines; note, however, that lines three and four are again without rhyme.

135. This line occurs also in a speech delivered at a capping ceremony, as it is pre-served in Yi li zhushu, 3.13c (״Shi guan li״).

136. Zheng Xuan understands the following announcement as given to the impersonator. This reading is strongly suggested by the honorific term huangshi 皇尸 in line six; both Zheng and Kong Yingda read huang 皇 as equivalent to the honorific jun 君, which Kong identifies as a direct form of address. (See the discussion in the analysis below.) Kong adds that “the invocator thereupon invokes the intention of the offering son and announces to the impersonator(s) that the beneficial [rites] are completed” (於是致孝子之意告尸以利成也). Kong's, commentary in Yi li zhushu, 46.246c Google Scholar (“Te sheng kuei shi li”), on the following passage which describes the particu-lar ritual situation is also parallel: “The host leaves and takes his position outside southwest of the door. The invocator faces east and announces the completion of the beneficial [rites]. The impersonator rises, the invocator goes in front, and the host descends [the stairs].” For a parallel passage see also Yi li zhushu, 48.259c (“Shao lao kui shi li”).

137. Zheng Xuan reads ju 具 (= ju 俱) as jie 皆 “all,” which would refer to the whole group of spirits. Chen Huan, Shi Mao shi zhuan shu, 20.30b, relates this to the traditional notion that the king had the prerogative to sacrifice to seven ancestors; see, e.g., Liji zhengyi, 12.107b (“Wang zhi” 王制), and 23.203c (“Li qi” 禮器）. I take zhi 止 as the common particle of perfect aspect, equivalent to yi 矣; see Xuehai, Pei, Gu shu xuci jishi, vol. 2, 778 Google Scholar. In the invocator's announcement, zhi seems to emphasize that the spirits' desired state of appropriate intoxication has been accomplished; as a result, the impersonator(s), who had been dutifully drinking on the behalf of the spirits, could now leave (for the drinking of the impersonator[s], see Li ji zhengyi, 49. 377b [“Ji tong” 祭通] Paper, The Spirits are Drunk, 113-14). According to Baxter, A Handbook of Old Chinese Phonology, 809, zhi 止 in many cases is part of the rhyme pattern, even as a final particle (though not in all of these instances). Takashima, Ken-ichi, “The So-called ‘Third’-Person Possessive Pronoun Jue 早 (=厥) in Classical Chinese,” Journal of the American Oriental Society 119.3 (1999), 418 CrossRefGoogle Scholar, has explicitly ar־ gued that in the Shijing, “one does … find grammatical particles and deictic pro-nouns in the rhyme position; there are numerous examples of 矣 and 之 where this is evident.”

138. The divine protector(s) (the ancestral spirit[s]), return(s) to heaven as the sacrifice is completed and the impersonator(s) is/are led out of the hall; see Kong Yingda's commentary.

139. The removal of the dishes is another fixed element in the process of an ancestral sacrifice. According to the Zhou li and Yi li, the invocator gives the order to remove the dishes once the sacrifice is finished; see Zhou li zhengyi, 49.2024 (“Da zhu” 大祝)； Yi li zhushu, 46.246c, 247b (“Te sheng kui shi li”) 48.259, (״c (“Shao lao kui shi li”). For the women and officials involved, see also Zhou li zhengyi, 7.251 (“Shan fu” 膳夫), 14.538 (“Neiwai chen” 內外臣), 14.554 (“Jiu pin” 九嬪),35.1408 (“Da zong bo” 大宗伯), and 41.1692 (“Wai zong” 外宗).

140. In this line, I understand yan 言as an adverbial marker, attached to bei 備 “complete, altogether”; see Xuehai, Pei, Gu shu xuci jishi, vol. 1, 433 Google Scholar. According to Zheng Xuan, the male participants of the same clan-name (tongxing 同姓) stay for the private banquet after the other guests have left.

141. Stanza 6 implies a shift of location, from the main hall of the ancestral temple to the inner (repose) apartment (qin 寢),still within the temple complex, where the private banquet is held; see Kong Yingda's commentary here and also on “Qiao yan” 巧言 (Mao shi zhengyi, 12/3.186b [Mao 198]). There is no conclusive evidence on the perspectives of the different speeches. The stanza contains four different rhymes and thus is much livelier than the preceding ones. The speech of lines three and four contains the personal pronoun er, which can refer to either the descendant or the spirits. The following speech in lines five through eight contains the direct and hon-orific address form jun 君 “lord.” The final speech is partly parallel to a guci formula in the Yi li issued by the invocator; see Yi li zhushu, 46.247a-c (“Te sheng kui shi li”). Because of this parallel, I assume the voice of the invocator also in the closing lines of Stanza 6. Note that the invocator is present at the banquet, giving announcements and commands. Seen as a whole, the final stanza appears to comprise a sequence of “auspicious words” (guci, ultimately from the spirits) and “congratulatory words” (qingci 慶辭, offered by the relatives).

142. According to Kong Yingda, hou lu 後祿 “subsequent fortune” is the continuous fortune that will result from the successful sacrifice. In his “congratulatory words” (qingci 慶辭, the term used by Zheng Xuan in Yi li zhushu, 48.260a [ “Shao lao kui shi li〃]), the principal guest states: “May the host receive the blessings from the sacrifice. May he with extensive longevity protect and sustain his lineage house.”

143. For the same line, addressed to the host during the feast that follows the sacrifice proper, see “Ji zui” 既醉 (Mao shi zhengyi, 17/2.268a [Mao 247]).

144. For the same line which is, according to Zheng Xuan's commentary, also in the voice of the guests, see “Zhi jing” 執競 (Mao shi zhengyi, 19/2.321 c [Mao 274]).

145. According to Zheng Xuan, these are the “congratulatory words” (qingci) offered by the guests; cf. n. 142 above.

146. The translation takes Zheng Xuan's paraphrase (維君德能盡之) as its point of departure. I read qi 其 here in the function of a second-person possessive pronoun; literally, the line then reads: “is your completing the rites.” Takashima, “The So-called ‘Third’-Person Possessive Pronoun Jue,” 426-27, argues that qi, compared to zhi 之 or jue 厥, is more “distal” (as opposed to “proximal”). This seems to fit well a usage equivalent to an honorific like jun 君 (as Zheng Xuan reads it): honorific expressions, in their basic psychological function, are formal means to generate distance; see also the discussion in the analysis below. I am grateful to Ken-ichi Takashima (personal communication) for clarifying these issues in relation to the present line.

147. The final line is verbatim the closing formula of a guci passage in Yi li zhushu, 48.258c (“Shao lao kui shi li”), issued by the invocator.

148. Grouping the ya and song together has textual support since Eastern Zhou times when the formula ya song (Elegantiae and Eulogia) became a fixed expression. See, e.g., Lun yu zhushu 9.35a (Lun yu 9.15); Wang Xianqian王先謙 (1842-1917), Xunzi jijie 荀子集解 (Zhuzi jicheng 諸子集成 ed.), 14.252.

149. At the same time, one can also imagine how the royal and aristocratic songs, as written texts, might have served as models in the writing of the guo feng. In any case, given the later Han redaction of the received text (see Baxter, “Zhou and Han Phonology in the Shijing”), one may have some doubts about attempts to identify the strata and date the different sections of the Shi jing; the divergences between even the most careful attempts are indicative of how inconclusive the evidence is. Compare, for example, Matsumoto, , Shikyō shohen no seiritsu ni kansuru kenkyū, vol. 2, 471–638 Google Scholar, with the conventional scheme given by Shih-hsiang Ch'en, “The Shih-ching: Its Generic Significance in Chinese Literary History and Poetics,” 16, according to which the Zhou song were first, followed by the ya (together with the Lu song 魯頌 and Shang song), and finally by the guo feng.

150. Note that some of these—notably rhyme, stanza structure, and metrical regu-larity—are much less developed in the Zhou song, probably the oldest layer not only of sacrificial hymns but of the whole anthology. Wang Guowei has interpreted this lack of formalization as an expression of “melodic sostenuto” (sheng huan 聲緩); see Guantang jilin, 2.19a-20a.

151. On the vocabulary, see Dobson, The Language of the Book of Songs; and Xi, Xiang 向熹, Shi jing yuyan yanjiu 詩經語言研究 (Chengdu: Sichuan renmin, 1987), 228–44Google Scholar.

152. However, see below, “Syntactical Patterns of Ritual Speech,” pp. 106-9, for discussion of one distinctive formal pattern of rhythmic speech that is indeed specific to the ritual hymns, compared to the guo feng; and thus serves to qualify any general “rule” of shared ritual speech across the Shi.

153. See my argument in Die Hymnen der chinesischen Staatsopfer, 112—13,161—68, and The Stele Inscriptions of Ch'in Shih-huang, 126-29, that the Shijing rhymes, even in what we now recognize as their irregularities, are in perfect accord with the rhymes of both the Qin imperial stele inscriptions and the seventeen hymns for the ancestral sacrifices from the early Western Han reign of Han G a ozu 漢髙祖 (r. 202-195 B.C.). Note that the irregularities between categories of rhymes (not individual words) prove the living tradition here. I believe they exclude a mere imitation of the venerated model: no one in Qin or Han times would have been able to conceptualize in abstract categories the old rhyme system to the extent of even taking the deviations into account. The coherence with the Shijing gradually vanishes during Western and espe-daily Eastern Han times, and this raises a problem: if during the relatively short time of two hundred years of the Western Han, under the stabilizing conditions of a relatively centralized state and its imperial court, phonological changes and regional differenees resulted in increasingly diverging rhyming practices, how can we assume that the Shijing rhymes remained perfectly constant from the tenth through the early second century B.C., through eight long centuries of political and cultural diversity? Also, although Zhou bronze inscriptions indicate that the validity of Shi jing rhyme categories can be traced with some coherence as far back as the Western Zhou, it is clear that rhyme in these inscriptions, even in the latter half of the Eastern Zhou, is much less consistent and regular than in the received Shijing. For the rhymes of the bronze inscriptions, see Behr, Wolfgang, “Reimende Bronzeinschriften und die Entstehung der chinesischen Endreimdichtung” (Ph.D. diss., Johann Wolfgang Goethe-Universität Frankfurt, 1996)Google Scholar. Skeptical opinions about the actual age of the Shijing rhymes have been advanced by Serruys, Paul L.-M., The Chinese Dialects of Han Time According to Fang Yen (Berkeley: University of California Press, 1959), 21-22, 53–54 Google Scholar; Barnard, Noel, The Chu Silk Manuscript: Translation and Commentary (Canberra: Australian National University Press, 1973), 213 Google Scholar; and Walker, Galal LeRoy, “Towards a Formal History of the Chuci” (Ph.D. diss., Cornell University, 1982), 412 Google Scholar, passim.

154. One is immediately reminded of Confucius's statement according to which one should “venerate the spirits and keep them at a distance” (敬鬼神而遠之); see Lun yu zhushu, 6.23b (Lun yu 6.22).

155. I borrow this technical term from Falkenhausen's discussion of the structure of Zhou bronze inscriptions; see his “Issues in Western Zhou Studies,” 154.

156. See Falkenhausen, , “Issues in Western Zhou Studies,” 152–56Google Scholar.

157. See the three song of the Zhou song 周頌section: “Si wen” 思文 (Mao shi zhengyi, 19/2.322a [Mao 275]), “Wu” 武 (Mao shi zhengyi, 19/3.329c [Mao 285]), and “Zhuo” 酌 (Mao shi zhengyi, 19/4.336b [Mao 293]).

158. See Hubei sheng wenwu kaogu yanjiu suo, Jiangling fiudian Dong Zhou mu 江陵九店東周墓 (Beijing: Kexue, 1995), 508 Google Scholar; and Harper, Donald, “A Warring States Prayer for Men who Die by Weapons” (unpubl. manuscript), 3 Google Scholar.

159. I am not aware of any commentary that understands er as referring to the spirits.

160. See n. 109 to Stanza 2 above.

161. See Xi, Zhu, Shiji zhuan, 9.13a, 17.9aGoogle Scholar, for “Tian bao” 天保 (Mao shi zhengyi, 9/3. 144a-c [Mao 166]) as responding to the preceding five hymns; and for “Hang Wei” (Mao shi zhengyi, 17/2.266a67־c [Mao 246]) as responding to “Ji zui〃 (Mao shi zhengyi, 17/2.267C-69b [Mao 247]).

162. The change of rhyme between Stanzas 2 and 3 is probably irrelevant for the issue of voices and perspectives of speech since the individual stanzas, by means of their closing guci formulae, are clearly demarcated and relatively self-contained textual units. In my understanding, the issue of rhyme changes becomes significant only within a single stanza.

163. I propose to simplify the rhyme scheme of the present stanza compared to Baxter, A Handbook of Old Chinese Phonology, 680. Baxter includes the “-un” final in the last word of line four (sun 孫) in a “combined rhyme” (he yun 禾口韻) with the “-an” finals of the rhyme words in lines one and two (han 燠, qian 慰).However, “combined rhymes” serve as a last resort only where the overall rhyme pattern requires inclusion of words that usually would not rhyme. By definition, “combined rhymes” are exceptions to and violations of the rule, and are proposed only out of sheer necessity. Yet Stanza 4—and in a curious parallel also Stanza 5—provides enough independent semantic and formal evidence to eliminate the necessity of the “combined rhyme” in line four, leaving lines three and four without any rhyme. Needless to say, removing a troublesome “combined rhyme” only corroborates the validity of Baxter's rigorous system.

164. Yi li zhushu, 48.258c (“Shao lao kui shi li”). The last line rhymes on the penultimate word.

165. For the early (however idealized) coherence of ritual music see Zhou li zhengyi, 43.1781 (“Da si yue” 大司樂) on the musical program of Zhou sacrifices and banquets: “At the great banquets, one does not lead in the sacrificial victim [and does not perform the corresponding music for this ritual step, “Zhao xia” 昭夏]. Everything else is like in the sacrifices.” Again Zhou li zhengyi, 46.1892 (“Zhong shi” 鍾師): “At all sacrifices and feasts, one performs the banquet music.” See also, Falkenhausen, , Suspended Music, 29 Google Scholar. For later imperial times, see my Die Hymnen der chinesischen Staatsopfer, 77-95.

166. For the congratulatory formulae at the capping ceremony, see Yi li zhushu, 3. 13b–c.

167. Reasons for the difference in translations are explained in n. 129 above. As in Stanza 4, Baxter, , A Handbook of Old Chinese Phonology, 680 Google Scholar, proposes a “combined rhyme” that would link line four (“-uk”; gao 告) to lines one and two (“-ək”; bei 備, jie 戒). For reasons given in n. 163 above, I remove this irregularity.

168. Compare also the speech quoted above from the Yi li where the impersonator reverently refers to the “august impersonator” under whose command he is acting.

169. Anyone familiar with the Roman Catholic Mass will recognize the possibility of such a shift of perspective in ritual speech from the central juncture in the Catholic liturgy when the priest performs the speech act of transubstantiation while at the same time commemorating, in a narrative voice, the Last Supper.

170. Perhaps, for example, the Minor Invocator (xiao zhu 小祝), who according to the Zhou li assists the Great Invocator (da zhu 大祝), see n. 101 above.

171. I suspect that this particular quality of Zheng's exegesis—which was eclipsed by Zhu Xi's influential Shiji zhuan throughout late imperial times and has also been widely rejected by modern scholars who propose to free the Shi from their Han political interpretation—is directly related to his intimate knowledge of early Chinese ritual, manifest in his formative and authoritative commentaries on the three ritual canons.

172. See Shang shu zhengyi, 14.97a (“Zi cai” 梓材).

173. Wheelock, , “The Problem of Ritual Language” 50 Google Scholar.

174. Note the two famous Zuo zhuan passages according to which the spirits accept sacrifices only from their own legitimate descendants; see Chunqiu Zuo zhuan zhengyi, 13.99c (Xi 10), 17.130a (Xi 31).

175. See Chunqiu Zuo zhuan zhengyi, 12.93c (Xi 5), quoting the Shu for the statement that ultimately, the spirits do not respond to the sacrificial offerings but to the offering descendant's shining virtuous power. In the received text of the Shu, this sentence occurs in “Jun chen” 君陳 (Shang shu zhengyi, 18.125a), one of the forged “old text” iguwen 古文) chapters.

176. For references, see n. 158 above.

177. See Shuihudi Qin mu zhujian zhengli xiaoza, Shuihudi Qin mu zhujian 睡虎地秦墓竹簡 (Beijing: Wenwu, 1990), 179–255 Google Scholar. There are two types of prayer texts in the Shuihudi burial: one is to be chanted in a ritual for the protection of a horse, the other is a dream incantation for relief from a nightmare; see Sterckx, Roel, “An Ancient Chi-nese Horse Ritual,” Early China 21 (1996), 47–79 CrossRefGoogle Scholar; Harper, Donald, “Wang Yen-shou's Nightmare Poem,” Harvard Journal of Asiatic Studies 47 (1997), 270–71Google Scholar; Harper, , “A Note on Nightmare Magic in Ancient and Medieval China” Tang Studies 6 (1988), 72–75 CrossRefGoogle Scholar.

178. For the two earliest received versions of the text, allegedly recarved accord-ing to the original stones that were still extant in Song times, see Rong Geng 容庚, Gu shike iingshi 古石刻零拾(Peiping: privately published, 1934); Moruo, Guo 郭沫若, “Zu Chu wen kaoshi” 言且楚文考釋, in his Tiandi xuanhuang 天地玄黃 (Shanghai: Dafu, 1947), 606–25Google Scholar; Chavannes, Edouard, “Les inscriptions des Ts'in ” Journal Asiatique, Neuvième Série 1 (1893), 475–82Google Scholar; and Chavannes, , Les mémoires historiques de Se-ma Ts'ien (Paris: Ernest Leroux, 1895–1905), vol. 2, 544–49Google Scholar.

179. If we accept these as incantations to individual spirits, following David Hawkes and others before him. See his The Songs of the South (Harmondsworth: Penguin, 1985), 95–101 Google Scholar, passim; and Hawkes, , “The Quest of the Goddess,” in Studies in Chinese Literary Genres, ed. Birch, Cyril (Berkeley: University of California Press, 1974), 42–68 CrossRefGoogle Scholar.

180. As noted above (p. 61), these hymns also record responses from cosmic spirits. In contrast to the Zhou ancestral hymns, these responses are not given as speeches but take the form of auspicious atmospherical phenomena.

181. Falkenhausen has proposed that for the Zhou bronze inscriptions the guci were appended to existing archival texts at the time when these were “transferred from its wooden or bamboo slip version onto bronze” ( Falkenhausen, , “Issues in Western Zhou Studies,” 166 Google Scholar).

182. Xu, , “Jinwen guci shili,” 43 Google Scholar. Longevity and permanence are two different albeit related issues: longevity is invoked for the offering descendant, permanence for the survival and ritual continuity of his lineage. In “Chu ci,” the “ten thousand years” of Stanzas 2 and 3 refer to the former, while the final exhortation refers to the latter.

183. See my The Stele Inscriptions of Ch'in Shih-huang, ch. 4, and my Die Hymnen der chinesischen Staatsopfer, 151-52.

184. The same idea of continuity is expressed in the famous Li ji passage on the function of tripod inscriptions (Li ji zhengyi, 49.378c-379a [“]i tong” 祭統]),employing the pun ming 銘 /ming 名: by commemorating and inscribing (銘) his ancestor's achievments the descendant creates a name (名) for himself. Through the inscription, future generations will therefore recognize and praise both him and his ancestor.

185. As observed above (n. 118), the last line of Stanza 2 is the only shared line that occurs in the guo feng. Note, however, that this line occurs not only in numerous bronze inscriptions but also in two other Shi jing hymns. It is clear that in this case the guo feng song borrows a stock phrase from the language of the ancestral cult, not vice versa.

186. See Dobson, W.A.C.H., The Language of the Book of Songs (Toronto: University of Toronto Press, 1968), 247–55CrossRefGoogle Scholar; Kanru, Lu 陸侃如 and Feng Yuanjun 瑪沉君, Zhongguo shishi 中國詩史 (Beijing: Zu ojia, 1956), vol. 1, 45–46 Google Scholar; Xi, Xiang, Shijing yuyan yanjiu, 228–44Google Scholar.

187. Karlgren's translation of Stanza 4 illustrates the problem:

We are very respectful, our rules and rites have no error, the officiating invoker makes the announcement, he goes and presents it to the pious descendant; fragrant is the pious sacrifice, the spirits enjoy the wine and food; they predict for you a hundred blessings; according to the (how much =) proper quantities, according to the (proper) rules, you have brought sacrificial grain, you have brought millet, you have brought baskets, you have arranged them; forever they will give you the utmost (blessings); those will be in myriads, in myriads of myriads.

See Karlgren, Bernhard, “The Book of Odes, Kuo feng and Siao ya,” Bulletin of the Museum of Far Eastern Antiquities 16 (1944), 246–47Google Scholar. The main defect in Karlgren's translation, which completely obscures the structure of the original, is obvious: since the two direct speeches are not identified, the relation between “we” and “you” remains unclear.

188. It is significant that my textual analysis provides independent support for Baxter's reconstruction of the Shi jing rhyme system, thus underscoring the importance or historical phonology for the study of Zhou texts.