The Body Incantatory: Spells and the Ritual Imagination in Medieval Chinese Buddhism

7. I borrow the language of “domains” from the work of Richard Sennett, who—in a stimulating meditation on the “material consciousness” of the craftsman (an idea very helpful for understanding the nature of ritual practice, itself irreducibly material)—speaks of the nature of shifts among cognitive and bodily domains, an idea he traces originally to Claude Lévi-Strauss’s assertion of an intimate connection between, for example, things being “good to eat” and “good to think.” “[Lévi-Strauss] means this literally: cooking food begets the idea of heating for other purposes; people who share parts of a cooked deer begin to think they can share parts of a heated house; the abstraction ‘he is a warm person’ (in the sense of ‘sociable’) then becomes possible to think. These are domain shifts” (Sennett, The Craftsman, 129).

8. Powers, Pattern and Person, 3.

9. Ibid., 268 and 17.

10. Ibid., 4, quoting Keightley, “Archaeology and Mentality,” 93. Just in passing, we might also consider in this light the famous statement of Marshal McLuhan’s that “we shape our tools and afterwards our tools shape us,” as well as more radical conceptions such as Adam Gopnik’s counter that “contraptions don’t change consciousness; contraptions are part of consciousness” (“The Information: How the Internet Gets Inside Us,” New Yorker, Feb. 14, 2011; http://www.newyorker.com/arts/critics/atlarge/2011/02/14/110214crat_atlarge_gopnik, accessed Feb. 12, 2011).

11. See chapter 2 for an example of a scripture apparently revised in reaction to changes in material practice. I borrow the image of “bundling” from Webb Keane, who writes (for example) of the ways that an American flag’s particular bundling of the material and the symbolic causes it to escape doctrinal control—the cotton chosen for the ways it allows a flag to lay smoothly flat also allows it to be burned dramatically. “No one would say that flags are made of cotton in order that they may be burned, but their wholly contingent flammability makes available a potent political symbol. Flammability is bundled together with all the other material characteristics of flags … Bundling gives to material things (including linguistic forms) an inherently open-ended character.” Keane, “On the Materiality of Religion,” 230; see also Keane “The Evidence of the Senses and the Materiality of Religion” for a more in-depth treatment of these ideas.

12. See Waley, A Catalogue of Paintings Recovered from Tun-Huang, xiii, for an early assertion of the importance of “Dhāraṇī Buddhism” in China.

13. The Five Dynasties Period is traditionally said to have stretched from 907 to 960 and the Song Dynasty from 960 to 1267.

14. Kieschnick, The Impact of Buddhism on Chinese Material Culture, 24–82, and throughout.

15. Ibid., 24; see also page 56, where he discusses all such objects as simply “repositories of sacred power.”

16. Strong, Relics of the Buddha.

17. Ryūichi Abé has discussed such visions of the nature of words and bodies in Esoteric Buddhist practice using the evocative and helpful phrase, “the physiology of words” (Abé, “Word,” 305).

18. Qianshou qianyan Guanshiyin pusa guangda yuanman wu’ai dabeixin tuoluoni jing, T no. 1060; vol. 20: 109a. See chapter 3 for a fuller discussion of this passage and its contexts.

19. For example, a dhāraṇī sutra translated by Baosiwei (d. 721), a key figure in the spread of material dhāraṇī practices in the late seventh century, reports that as the Buddha’s worlds-illuminating radiance was returning into his body, it proclaimed itself to be the “Wheel-Turning King One Syllable Mind-Spell of the Wisdom of all Tathāgatas.” Da tuoluoni mofazhong yizi xinzhou jing, T no. 956, 19: 315c.

20. On this work, see Copp, “Anointing Phrases and Narrative Power.”

1. Mantras constitute the oldest class of spells in Indic cultures. They were taken into Buddhist practice, along with much else of traditional Indian religious culture, and often conflated with dhāraṇīs. Parittas are Buddhist words of power found in Southeast Asian traditions. Hṛdaya and vidyā are specialized, and more narrowly contextually based, terms for Buddhist spells in Mahāyāna Buddhism. On mantras, see Abé, The Weaving of Mantra, especially 5–6, 12–13, 262–264; Gonda, History of Ancient Indian Religion, 248–301; Kapstein, “Scholastic Buddhism and the Mantrayāna”; Padoux, “Mantras—What Are They?”; Staal, “Vedic Mantras”; Wayman, “The Significance of Mantras”; among others. On parittas, see Harvey, “The Dynamics of Paritta Chanting in Southern Buddhism”; McDaniel, “Paritta and Rakṣā Texts”; and Skilling, “The Raksā Literature of the Śrāvakayāna.” On hrdaya as a term for spells, see Lopez, Elaborations on Emptiness, 165–186.

2. See Tambiah, Buddhist Saints of the Forest and the Cult of Amulets, for a study of the place of amulets in a modern Buddhist culture.

3. This aversion to the “magical” is not limited to modern scholars of Buddhism—practitioners of what has been called “modernist Buddhism” have been similarly averse to spells and ritual. For surveys of the nature of “modern Buddhism,” or “Buddhist modernism,” see Lopez, Buddhism and Science, and McMahan, The Making of Buddhist Modernism.

4. Müller and Nanjio, The Ancient Palm Leaves, 31–32. John Gager, in his study of curse tablets and binding spells in the ancient world, notes that the tradition of scholarly disrespect for practices such as spells is best seen as stretching back to Lucian’s “Lover of Lies” (Philopsueda, chapter 10). See Gager, Curse Tablets and Binding Spells, 9.

5. Suzuki, The Lankavatara Sutra, 223 n.1. To Müller’s and Suzuki’s characterizations, we might also add that of Charles Luk, who in the preface to his 1966 translation of the Śūralgama Sūtra explains why he omitted the text’s incantation, though it is central to the nature of the text as a whole. “[The] average Western student of Buddhism seems to have little faith in mantras and rituals which should not be published lest they create unnecessary disbelief and confusion and so compromise the beauty of this profound sūtra” (Luk, tr., The Śūralṅgama Sūtra, xx).

6. Starn, “A Historian’s Brief Guide to New Museum Studies,” 84 (cited in Bynum, Christian Materiality, 280). See also Bynum, “Perspectives, Connections, and Objects,” 78–80.

7. Examples of this trend in recent scholarship on East Asian Buddhism are far too numerous to fully mention here, but any listing must feature, for example, Foulk and Sharf, “On the Ritual Use of Ch’an Portraiture in Medieval China”; “Teiser, The Scripture on the Ten Kings and Reinventing the Wheel; Ruppert, The Jewel in the Ashes; Sharf, “Prolegomenon to the Study of Japanese Buddhist Icons”; Kieschnick, The Impact of Buddhism; Benn, Burning for the Buddha; Rambelli, Buddhist Materiality; Mollier, Buddhism and Taoism Face to Face; Robson, Power of Place; and, most broadly, the great tradition of French scholarship on Buddhism, which seems never to have conceived the radical split between doctrinal thought and material culture that has until recently characterized its Anglo-American cousin. See, for example, the works of Demiéville, Pelliot, Soymié, Durt, Drège, Trombert, and Mollier listed in the bibliography. Note that this listing also excludes the work of art historians, who have obviously focused on objects and their practices for much longer than have other scholars.

8. Benn, Burning for the Buddha, 202.

9. Gager notes a similar scholarly neglect of the “irrational” or “magical” among scholars of the classical West. “One reason for this persistent neglect stems surely from the potentially harmful character of these small metal tablets … the potential harm to the entrenched reputation of classical Greece and Rome, not to mention Judaism and Christianity, as bastions of pure philosophy and true religion” (Gager, Curse Tablets and Binding Spells, 3). Similar descriptions of the history of Buddhist Studies might not be far off the mark, as Robson (“Signs of Power”) also notes in his discussion of this issue.

10. See Kieschnick, The Eminent Monk and McBride, “Dhāraṇī and Spells in Medieval Sinitic Buddhism.”

11. See Li Xiaorong, Dunhuang Mijiao wenxian lungao, 295ff. for some vivid examples from the eighth century.

12. See chapter 3 for an exploration of this evidence.

13. Williams and Tribe, Buddhist Thought, 206.

14. Dhāraṇīs are traditionally said to be variations of Sanskrit words, as are the originals of the Buddhist scriptural texts translated into Chinese. Yet, one commonly encounters statements in Indian texts asserting that the spells are incomprehensible—the language of the “barbarians.” Further, as recent research has shown, it is clear that many of the sūtras brought into China were actually in languages other than Sanskrit: probably the language of whatever monk or layman happened to have introduced the text to the Chinese. It seems best to assume, or at least consider the possibility, that the same held true for the dhāraṇīs contained within those texts. The case of spells is more complicated than that of sūtras, however, as the specific sounds of any one spell—originally Sanskrit, at least in inspiration—were said to be of paramount importance and thus, in theory, would likely have been preserved in ways that the merely discursive language of the scriptural narratives were not. The language of origin thus seems an open question: though the theory is clear, prescription and practice are not always as tightly related as the prescribers might hope.

15. See Abé, Weaving of Mantra, 166–167 for a discussion of this shift in dhāraṇīs’ place within sūtras that understands it to be a feature of a more general shift from “exoteric” to “esoteric” styles. See chapter 4 of this book, as well, for further discussion of these issues.

16. Yiqie jing yinyi, T no. 2128, vol. 54: 366c.

17. The characterization of the Chinese transliterations is drawn from Kroll, Dharma Bell and Dhāranī Pillar, 40.

18. “[By] forgetting to examine the implicit presuppositions behind the operation which consists in deciphering, in seeking the meaning of the words, the ‘true’ meaning of the words, philologists expose themselves to the risk of projecting into the words they study the philosophy of words which is implied by the very fact of studying words … You can see that if the philologist were to reflect on what being a philologist means, he would be obliged to wonder whether the use he makes of language coincides with the use made of it by those who produced it; and whether the gap between linguistic usages and interests does not risk introducing into interpretation an essential bias, one that is far more radical than mere anachronism or any other form of ethnocentric interpretation, since it stems from the activity of interpretation itself.” (Bourdieu, In Other Words, 97–98; and, more generally, Bourdieu, Outline of a Theory of Practice.) See also Lamarre, Uncovering Heian Japan, 116–124 (and passim), who notes a related danger in unreflexive textual scholarship that seeks to collapse the various “figural registers” of writing into only the verbal.

19. See Yuyama, “An Uṣṇīṣa-Vijayā-Dhāraṇī Text from Nepal,” 168, for a discussion of one such “corruption.”

20. Padoux, “Mantras—What Are They?,” 300.

21. See Benjamin, Illuminations, 69–82.

22. Malinowski, Coral Gardens and Their Magic, and Tambiah, “The Magical Power of Words,” respectively.

23. Wittgenstein, “Remarks on Frazer’s Golden Bough.”

24. Staal, “Vedic Mantras.”

25. See, for example, Ibid.; Padoux, “Mantras—What Are They?” and “L’énergie de la parole.”

26. Lopez, Elaborations on Emptiness, 167.

27. Austin, How to Do Things with Words. But see also Bourdieu, Language and Symbolic Power, 107–136, for an important critique of Austin for insufficiently emphasizing the role of institutional and other social contexts in his analysis of how words do things.

28. We can also note, again merely in passing, the work of recent philosophers who have taken issue with what they see as the prioritizing of the spoken over the written word in Western cultures, of whom the most famous exemplar is, of course, Derrida, Grammatology.

29. The fraught relationship between these terms in the history of scholarship (to say nothing of that between “magic” and “science”) is familiar. Another side effect of my use of the words spell and incantation is that it might seem to imply that this relationship is at work in the history of Chinese religious spell craft. It is not. The presence invoked by these terms is illusory. As noted before, I use the words incantation and spell because they are good translations of the Chinese word zhou, which is one of the terms most commonly used by Chinese to translate the Sanskrit words dhāraṇī and mantra.

30. For an earlier version of the following discussion, see Copp, “Notes on the Term ‘Dhāraṇī.’” See also Davidson, “Studies in Dhāraṇī Literature I,” and McBride, “Dhāraṇī and Spells,” for detailed discussions of the nature of dhāraṇīs and the term dhāraṇī.

31. Though zhenyan and miyan are often said to have been a translation of mantra, it is clear that the terms referred to dhāraṇīs in medieval China as often as they did mantras.

32. A point that will be especially clear in chapter 4.

33. For Chinese Buddhist traditions of spell craft, see Strickmann, Chinese Magical Medicine, 89–122; Kieschnick, The Eminent Monk, 67–111; and McBride, “Dhāraṇī and Spells.”

34. For an early example of dhāraṇī as memory we can take the discussion of the “four adornments” of the bodhisattva listed in Dharamarakṣa’s (Zhu Fahu; d. 316) translation of the Great Compassion Sūtra (Da ai jing), the fourth of which is called “the adornment of dhāraṇī (zongchi zhuangyan), whereby what is heard is never forgotten” (suowen buwang). T no. 398, 13: 416a. Scholarly explorations of mnemonic conceptions of dhāraṇī include those of Braarvig, “Dhāraṇī and Pratibhāna,” Gyatso, “Letter Magic,” Kapstein, “Scholastic Buddhism and the Mantrayāna,” and Nattier, A Few Good Men,” 291–292, n. 549. For discussions of dhāraṇīs as spells and incantations, see McBride, “Dhāraṇī and Spells,” as well as the following chapters of this book.

35. The word usually used here, chi, itself often simply means “dhāraṇī” (including in these very contexts) as I will discuss later.

36. Bakhtin, The Dialogic Imagination, 64.

37. Nattier, A Few Good Men, 291–292, n. 549.

38. Braarvig, “Dhāraṇī and Pratibhāna,” 19.

39. Strickmann, Chinese Magical Medicine, 103.

40. Kapstein, “Scholastic Buddhism and the Mantrayāna,” 237.

41. For a recent extensive study of related matters, see Overbey, “Memory, Rhetoric, and Education in the Great Lamp of the Dharma Dharani Scripture.”

42. Braarvig, “Dhāraṇī and Pratibhāna,” 18.

43. The place of the Bodhisattvabhūmi in medieval Chinese Buddhism has not yet inspired a high level of interest among scholars, though it was clearly among the most often-cited treatises in Chinese exegetical and doctrinal works, and its fourfold rubric for classifying dhāraṇīs was extremely influential in East Asia. The text exists as one section of the Yogacārabhūmi, though it circulated independently of that larger work and was so translated into Chinese twice in the fifth century, once (Pusa dichi jing; T no. 1581) by Dharmakṣema (Tanwuchen; 385–433), and once (Pusa shanjie jing; T no. 1582) by Guṇavarman (Qiunabamo; 367–431). See McBride, “Dhāraṇī and Spells,” for a concise discussion of Dharmakṣema’s fame as a spell caster, his translation of the Bodhisattvabhūmi, as well as the development of that text’s fourfold dhāraṇī rubric by Jingying Huiyuan (523–592). For a much longer study of Dharmakṣema, including information about the dating of his translations, see Chen, “The Indian Buddhist Missionary Dharmakṣema,” especially p. 258.

44. There are several treatments of these passages in the scholarly literature. Among them, Ryūichi Abé’s succinct discussion in his The Weaving of Mantra is especially clear (p. 166). His treatment, however, though in many ways excellent, does not sufficiently distinguish among ideas of dhāraṇīs as spells—such as those found in the Lotus Sūtra—as “mnemonic device,” and as forms of spiritual capacity.

45. Braarvig, “Dhāraṇī and Pratibhāna,” 18.

46. Pusa dichi lun (T no. 1581, 30: 934a). Braarvig, who translates from a Sanskrit version of the text, has a somewhat different rendering (Ibid., 20).

47. P. 2141; T no. 2803; 85: 951c.

48. Braarvig, “Dhāraṇī and Pratibhāna,” 20. We should keep in mind, as Kapstein notes, that in its Sanskrit original this is a difficult passage, and that “not all aspects of its interpretation are entirely secure” (Kapstein, “Scholastic Buddhism and the Mantrayāna,” 238). The Chinese version is, however, rather clear.

49. Pusa dichi lun, T no. 1581, 30: 934a. Xuanzang’s translation of the Yogacarabhūmi, Yuqieshi dilun, in a parallel passage, has zhou zhangju, which amounts to the same thing (T no. 1579, 30: 543a).

50. Pusa dichi lun, T no. 1581, vol. 30: 934a.

51. P. 2141; T no. 2803, 85: 952a.

52. P. 2141; T vol. 85: 952c. The connection between dhāraṇīs and this particular form of kṣānti, the patient acceptance of the emptiness of all phenomena, is found in dhāraṇī sūtras as well, though the relationship is not often so explicitly accounted for. For a convenient example, we may take the Great Vaipulya Dhāraṇī Sūtra (to use Paul Swanson’s translation of the title), the Da fangdeng tuoluoni jing, translated by Fazhong (fl. 401–413) very early in the fifth century. In the beginning of the scripture’s narrative, as a result of hearing the Buddha speak the names of nineteen dhāraṇīs, the bodhisattvas in the audience come to abide in the “patience [of tolerating the knowledge] that dharmas do not arise”—that is, that they are empty. The lesser beings in attendance achieve accordingly lesser spiritual states upon hearing the names (Swanson, “Dandala, Dhāraṇī, and Denarii,” 206–207). Though the bodhisattvas here are not said to engage in the sort of practice described in the Bodhisattvabhūmi (though it is possible that “hearing the names” here might have been understood in something like the contemplative sense described in the Bodhisattvabhūmi), the two texts create the same parallel relationships between textual dhāraṇīs and this most profound form of kṣānti (though, as will become clear in later chapters, the sūtra is closer in spirit to the later understandings of dhāraṇīs as powerful utterances in that simply hearing the words seems to induce the state described).

53. Foshuo wuliangmen weimichi jing, T no. 1011, 19: 681b. The translations of the syllables are highly tentative.

54. Gyatso, “Letter Magic,” 191, also provides an especially clear and concise discussion of this issue.

55. Qing Guanyin jing shu chanyi chao, T no.1801, 39: 998b. The quotation from the Mahā-Prajñāpāramitā-sūtra is from Kumarajiva’s translation (Mohe bore boluomi jing, T no. 223, 8: 421b). For another, roughly contemporary, statement, see Zixuan’s (965–1038) commentary to the Lengyan jing, the so-called “Pseudo-Śūraṃgama” (on which see more below). “The mystic words of the buddhas [that is, dhāraṇīs], occult teachings that only buddhas and buddhas understand—they are not something that other [lesser] saints are able to penetrate” (Shoulengyan yishu zhujing, T no. 1799, 39: 919c).

57. Da piluzhena chengfo jingshu, T no. 1796, 39: 579b.

58. Sūryagarbha vaipulya sūtra (Rizang fen, collected within the larger Da fangdeng da jijing), T no. 397, 13: 254b.

59. Xianjie jing, T no. 425, 14: 4c-5a. Both the Anantamukha and Bhadrakalpika scriptures provide examples of the ways dhāraṇī practices, and the ways dhāraṇīs were construed within them, changed over time. In both texts dhāraṇīs begin as simple runs of syllables to be contemplated in the ways discussed in this chapter. In later versions of the texts, in both Chinese and Tibetan translations, the syllables are replaced by dhāraṇīs of the later incantation variety and the nature of their ritual engagements change correspondingly. See Skilling, “An Arapacana Syllabary in the Bhadrakalpika-Sutra,” for a brief discussion of later Tibetan versions of the Bhadrakalpika Sūtra, where the translated syllables of the Dharmarakṣa translation are replaced by the standard arapacana syllabary. The Anantamukhadhāraṇī Sūtra, which was translated into Chinese several times over the centuries, provides especially stark examples of this transformation.

60. For works that understand dhāraṇīs as codes of various kinds, see, for example, Buswell, Cultivating Original Enlightenment, 271ff; McBride, “Dhāraṇī and Spells,” 97, and Davidson, “Studies in Dhāraṇī Literature I”—though Davidson’s use of the term seems quite different from Buswell’s or McBride’s.

61. See below for a discussion of the Treatise on the Great Perfection of Wisdom.

62. Dazhidu lun, T no. 1509, 25: 408a-b.

63. Mohe bore boluomi jing, T no. 223, 8: 256a; Dazhidu lun, T no. 1509, 25: 407c.

64. Dazhidu lun, T no. 1509, 25: 408b.

65. Dazhidu lun, T no. 1509, 25: 408a. As an aside, it seems likely that we should add this passage to those that Matthew Kapstein has highlighted as having “smoothed the way for and presaged the development of” later Mantrayāna, or even “full-blown Vajrayāna,” conceptions, in that we seem to have here, in these assertions about the ultimate meaningfulness of the syllable “a,” something very close to those later notions of the inconceivable and inconceivably pregnant sounds and sigils that make up mantras and dhāraṇīs. See Kapstein, “Scholastic Buddhism and the Mantrayāna.”

66. Or, as Chou Po-kan has recently argued, by Sengrui, his editor and student (Chou, “The Problem of the Authorship of the Mahāprajñāpāramito-padeśa”).

67. The canonical status of his work has more recently been affirmed in Michel Strickmann’s own deeply learned study, Mantras et mandarins: le bouddhisme tantrique en chine, which relies heavily on Lamotte’s discussion in its own introduction of the nature of dhāraṇīs.

68. Dazhidu lun, T no. 1509, 25: 95c. Cf. Lamotte, tr., Le Traité de la grande vertu du sagesse 1:317. In the translation that follows I have consulted Lamotte’s translation, especially in the case of difficult passages, but the translation is my own.

69. Translation adapted from Lamotte: “Elle est contenue dans un élément (dhātu), une base de la connaisance (āyatana), et un agrégat (skandha).”

70. Dazhidu lun, T no. 1509, 25: 95c.

71. The Song, Yuan, and Ming versions of the text do not have this line, which has tuolinni for “dhāraṇī.”

72. The distinction here seems to be between bodhisattvas and bodhisattva-mahāsattvas, who are characterized by dhāraṇī, samādhi, and kṣānti. The former have not yet attained these powers.

73. I follow Lamotte’s lead on where to end the quotation.

74. Lamotte (321), no doubt wisely, does not translate the names of these dhāraṇī. My translations are highly provisional.

75. Lamotte (321) omits the character “revolutions” (xuan).

76. Dazhidu lun, T no. 1509, 25: 95c-96c.

77. Rṣi, in the Sanskrit.

78. Dazhidu lun, T no. 1509, 25: 97c.

79. On chi as term indicating the “wielding” of texts, see Gimello, “Icon and Incantation,” 225–256.

80. For a succinct discussion of this term, see Teiser, The Scripture on the Ten Kings, 140–141, including note 6.

81. P. 2807, collected in XB vol. 17, p. 11571. For a recent study that at least in part carries forward the work begun here (and makes explicit reference to its earlier version in Copp, “Notes on the Term ‘Dhāraṇī ’”), see Eubanks, Miracles of Book and Body, 50 (and passim).

82. Song Gaoseng zhuan, T no. 2061, 50: 714c.

83. See, for example, the text on the Dunhuang manuscript P. 2094, Chisong Jin’gang jing lingyan gongde ji, which, in praising the power of the popular incantation “a-ra-pa-cana,” proclaims: “when someone chants these real words with a perfect mind it is as if he had chanted all the scriptures stored in the world one time.” T no. 2743, 85: 160a.

84. Boreboluomiduo xinjing, T no. 251, 8: 848.

85. For a discussion of terms such as “one mind” and “mind ground” in Tang Buddhism, see Gregory, Tsung-Mi and the Sinification of Buddhism.

86. T no. 251, 8: 848c.

87. Davidson, “Studies in Dhāraṇī Literature I”—the state-of-the-art study of dhāraṇīs in Indic Buddhist literature—comes to a conclusion about the nature of dhāraṇīs similar to the one I have given here (and in Copp, “Notes on the Term ‘Dhāraṇī’”), though one expressed in the language of “codes,” specifically the kind represented by the nearly infinitely capacious DNA encodings of life. I find this a provocative and in many ways powerfully apt analogy, though simply as a matter of method I prefer to try to account for traditional Chinese-language discussions of dhāraṇīs and their practice as much as possible in their own terms; thus my use of the imagery of “grasp” and “hold” (and to a lesser extent the rhetorical figure of synecdoche), which I take to be a literal translation of chi, the basic linguistic image employed in this literature.

1. Lopez, Elaborations on Emptiness, 167.

2. As well as a range of historical contingencies that are now largely irrecoverable. I do not mean to discount the role of “extra-religious” history in the history of material incantation practice (and the reader will find some attention to it in the following chapters), but this is at its heart a study of ritual logic and I will keep my attention focused there.

3. Their original Indic versions likely date to the sixth century, as Gergely Hidas argues was the case for the early version of the Scripture of the Incantation of Wish Fulfillment, the Mahāpratisarā Mahāvidyārājñī (see Hidas, “Mahāpratisarā-Mahāvidyārājñī, The Great Amulet, Great Queen of Spells,” 17).

4. That is, it seems likely that the full-blown versions of ideas and practices exemplified in such texts reflect a relatively late stage in their development. Ritual improvisations, conversations, lost drafts of scriptures, and the other milieu of their earlier forms would no doubt have existed for some time before a sūtra was composed.

5. Ma Shichang has proposed a different set of rubrics in which to understand this range of practices—recitation, wearing, and devotion (Ma, “Da suiqiu tuoluoni mantuluo tuxiang de chubu kaocha,” 556).

6. As noted in the introduction, the early third-century translation of the Anantamukha dhāraṇī sūtra calls for the writing of syllables as part of a contemplative dhāraṇī practice that, as I see it, bears little relation to what I construe here to be the writing down of spells. I do not count such practices here.

7. Qifo bapusa suoshuo da tuoluoni shenzhou jing, T no. 1332; vol. 21: 536b–561b.

8. A possible exception is the final occurrence in the text, in a moment within a ritual enchanting of water—a singularly strange rite even within dhāraṇī literature—where one is enjoined to “enchant the water seven times and write [the spell?] in the opening [of the water bottle?],” as part of the ritual setup (T 21: 560b).

9. For the origins of this now-long-pervasive notion see Schopen, “The Phrase ‘sa pṛthivīpradeśaś caityabhūto bhavet’ in the Vajracchedikā.”

10. Conze, The Perfection of Wisdom in Eight Thousand Lines and its Verse Summary, 104–106.

11. Qifo bapusa suoshuo da tuoluoni shenzhou jing, T 21: 538b. The “evil paths” are those that lead to birth as an animal, a hungry ghost, or a dweller in a Buddhist hell.

12. See, for example, T 21: 540a.

13. Beihua jing, T no. 157, vol. 3: 167–233.

14. Henghe sha. That is, the Ganges.

15. “Supreme, correct, awakening,” a standard term for the perfect awakening of a buddha, here in its usual Chinese transliteration.

16. Harrison, “Is the Dharma-kāya the Real ‘Phantom Body’ of the Buddha?” 76.

17. Pratyutpanna Sūtra (Banzhou sanmei jing); T no. 418, vol. 13: 911b; translated in Harrison, tr. The Pratyutpanna Samādhi Sutra.

18. A notable exception to this fact, though one probably unique in its particulars, is the 1117 burial of nearly the entire Chinese Buddhist canon inscribed on stone tablets beneath a stupa. Though the immediate reason for this act was not “religious” but simply the fear that the stones would be lost in the chaos attending the imminent fall of the Liao Dynasty, the inclusion of the stupa makes it at least an outlying example of text relic. On this event and its aftermath see Ledderose, “Carving Sutras into Stone Before the Catastrophe.”

19. It is important to note that the texts seem to have been called relics only after they were placed inside the stūpa—the basic architectural marker of the Buddha’s body.

20. Da Tang Xiyu ji (T no. 2087, vol. 51: 920a). See Fahua chuanji (T no. 2068, vol. 51: 92a-b) for another brief account of the use of dharma relics in the “western lands.”

21. Translation from Boucher, “The Pratītyasamutpādagātha,” 11.

23. Skilling, “Traces of the Dharma.”

24. Liebenthal, “Sanskrit Inscriptions from Yunnan I,” 2. Liebenthal describes bricks from the inside of a Yunnan stūpa inscribed (in Brāhmī, a script he states was normally reserved for dhāraṇīs) with both the ye dharmā and dhāraṇīs (Ibid., 31–34, 36).

25. Schopen 1997b, 120–121 (original article published in 1987).

26. Bentor, “On the Indian Origins of the Tibetan Practice of Depositing Relics and Dhāraņīs in Stūpas and Images,” 252ff. This article provides a convenient overview of these issues in Tibetan and South Asian contexts. Curiously, given the popularity of this practice, it seems that spells were only rarely referred to as dharma relics in other ritual contexts. The only example I have found is from Cixian’s (Liao Dynasty, 907–1125) version of the Miaojixiang pingdeng guanmen dajiaowang jing lüe chu humo yi, which states that one should “recite with a focused mind this dharma relic mantra (zhenyan) while burning objects” on the homa pyre (T no. 1194, vol. 20: 935c). The nature of the practice determines the nature of its objects, it would seem.

27. The use of dhāraṇīs did not entirely replace the “Verse on Dependent Origination,” at least within later scriptural accounts known in China. Both the “Sūtra Preached by the Buddha on the Merits of Constructing Stūpas” (Foshuo zaota gongde jing, T no. 699), translated in 680, and the “Sūtra on the Merits of Bathing the Buddha’ (Yufo gongde jing, T no. 698), translated by Yijing early in the eighth century, contain the injunction to employ this verse.

28. T no. 967, vol. 19: 351b.

29. T no 967, vol. 19: 351b. For an extensive and excellent discussion of the ideas and practices of “complete body relics,” as well as of reliquary practices in premodern Chinese and Korean Buddhism in general, see Seunghye Lee, “Arts of Enshrining.”

30. For the new identification of cave 217, see Shimono, Tonko Bakukōkutsu dai niyakujūnana kutsu minamiheki kyōhen no shin kaishaku. Wang Huimin, building on her work, has pushed the identification farther and extended it to other caves, including no. 33.

31. See below for a discussion of the metaphor of seals in dhāraṇī practice.

32. Putichang zhuangyan tuoluoni jing, T no. 1008, vol. 19: 672c.

33. Barrett, “Stūpa, Sūtra, and Sarīra in China,” 51–58, and Chen Jinhua, “Śarīra and Scepter,” 103–116.

34. Barrett, “Stūpa, Sūtra, and Sarīra in China,” 51–58; Antonino Forte, as cited in Chen, “Śarīra and Scepter,” 115.

35. Campany, “Notes on the Devotional Uses and Symbolic Function of Sūtra Texts.”

37. Here again it is important to emphasize the ideal-typical nature of the analyses of practical logics I present in this book. Relics can be (and were and are) sometimes worn as amulets in Buddhist Asia, as well, following modes of practice other than those I explore here. For a vivid example, see Stein Painting no. 247 (figure 2.19) an incantation amulet that names its spell śarīra-gātha (sheli zhi qieta): “reliquary verses.” In addition, the specific material placement of a spell in a stupa must be carefully attended: the work of Seunghye Lee, for example, makes clear that not every dhāraṇī in a stupa or pagoda functioned simply as a relic; incantations were at times sited in positions clearly ancillary to the relics, some of which were also in the forms of spells. (See Lee, “Arts of Enshrining”). The ideal types laid out in the first half of this chapter provide analytical tools, not rigid practical rules.

38. For a later example of the earlier material styles, see for example, the Qianshou qianyan Guanshiyin pusa guangda yuanman wu’ai da beixin tuoluoni jing, T no. 1060, vol. 20:109a. See also next.

39. On fu, see Seidel, “Imperial Treasures and Taoist Sacraments”; Despeux, “Talismans and Diagrams”; Liu Xiaoming, Zhongguo fuzhou wenhua daguan; Mollier, “Talismans”; and Wang Yucheng, Tang Song Daojiao mi zuanwen shili; among many others. On Buddhist fu, see Mollier, Buddhism and Taoism and Robson, “Signs of Power.”

40. Mollier, Buddhism and Taoism.

41. On which see Ibid. and Robson, “Signs of Power.”

42. Brown, The Cult of the Saints, 13–22, especially 17ff. Brown credits David Hume for the original insight. For a discussion of the question of “popular religion” in China, as well as a useful overview of the problematic in the study of European history, see Bell, “Religion and Chinese Culture.” Teiser, “Popular Religion,” offers a critical analysis of both the rubric itself and its recent history in the field of the study of Chinese religions. But see also Lancaster, “Elite and Folk,” especially 88, who defends the model.

43. Sørensen, “Michel Strickmann on Magical Medicine in Medieval China and Elsewhere.” I am grateful to Dr. Sørensen for his comments on this section (and others in the book) and for encouraging me to keep it in the chapter. This section is also intended as a response to an early anonymous reviewer of the book for the press who voiced concerns related to those raised in Sørensen’s essay.

46. As Robert Sharf has argued, “Our identification of a text, doctrine, image, or rite as Indian or Chinese, Buddhist or Taoist, Tantric or Ch’an orients our approach to the material, predisposing us to one set of readings while foreclosing others. It behooves us to reflect on the premises and entailments of such identifications” (Sharf, Coming to Terms with Chinese Buddhism, 21; and see the entire discussion, 21–25).

47. For extensive explorations of the ways Buddhists and Daoists used the same tales, see Mollier, Buddhism and Taoism, and Robson, Power of Place.

48. As we have seen, other much less descriptive (and evocative) terms also appear, such as “to secure” (an) and “to tie” (xi).

50. Shishuo xinyu 2. Translation by Richard Mather, from Liu Yiqing, 30–31.

52. Hanyu dacidian 1:1341, citing lines from the works of the Tang poets Wang Wei and Wei Zhuang.

53. See Seidel, “Imperial Treasures and Taoist Sacraments,” on this subject.

54. Liu Zhaorui, Kaogu faxian yu zaoqi Daojiao yanjiu, 135–138. See also the seminal work of Michel Strickmann on religious seals in China contained in Strickmann, Chinese Magical Medicine.

55. In addition to Liu Zhaorui’s and Michel Strickmann’s works noted earlier, see Wang Yucheng, Daojiao fayin lingpai tan’ao; Li Yuanguo, Daojiao fayin mizang; and various works of Xiao Dengfu, including Daojia Daojiao yingxiangxia de Fojiao jingji. I am, further, preparing a separate study of the pervasive and structuring place of seals in Chinese Buddhism that focuses on archeological and tale evidence.

56. For works that take the amulets in one form or another to have been Chinese additions, see Ma, Da suiqiu tuoluoni mantuluo tuxiang de chubu kaocha, 530; Xiao, Daojia Daojiao yingxiangxia de Fojiao jingji, 773; and Xiao, Daojiao shuyi yu mĳiao dianji, 194–195.

57. Yiqie jing yinyi, T no. 2128, vol. 54: 553b.

58. Zhenyuan xinding shijiao mulu, T no. 2157, vol. 55: 1010a.

59. Yiqie jing yinyi, T no. 2128, vol. 54: 404a.

60. Harper, Early Chinese Medical Literature, 167. Such techniques were not limited to amuletic practices—Harper (Ibid., 166ff) surveys a wide range of other practices. See also Sivin, “Ailment and Cure in Traditional China,” 36ff. and Yu, Shendao renxin, 330–332.

61. Birrell, The Classic of Mountains and Seas, 3.

62. See Campany, To Live as Long as Heaven and Earth, 27, on zhi in Chinese religious writings. Campany makes clear that the term refers not to fungi (as is often assumed), but, “redolent of the numinous,” it is more basically a “generic word for protrusions or emanations from rocks, trees, herbs, fleshy animals, or fungi.”

63. Baopuzi neipian jiaoshi 11: 199.

64. Baopuzi neipian jiaoshi 17: 304.

65. Schafer, “Orpiment and Realgar in Chinese Technology and Tradition,” 80–85, and Obringer, L’Aconit et L’Orpiment, 65–90.

66. For a color plate of the object, see Shōsōin Jimusho, Shōsōin hōmotsu, 16 (see also pp. 104 and 215 of that work).

67. Schafer, “Orpiment and Realgar,” 84.

68. Ishinpō 26: 610b. The text refers to Sun’s text simply as the “Qianjin fang.” On this passage, see Sivin, “Ailment and Cure,” 37.

69. Pregadio, Great Clarity, 129. In a related matter, Yu Xin, in his study of “livelihood religions” (minsheng zongjiao) in Tang-Song Dunhuang, notes evidence for the wearing, or carrying, of sticks of peach wood in almanacs discovered at Mogao, a practice he relates to the ancient belief that arrows of peach wood were effective weapons against certain demons. Yu speculates, in fact, that the carrying of peach wood mentioned in his text (p. 2661) indicates the carrying of such arrows in case they are needed, not the treating of the wood as itself effective (Yu, Shendao renxin, 332). Though Yu’s suggestion here is not fully convincing—for example, the use of demonifugal peach wood idols is found from the early period (Harper, Early Chinese Medical Literature, 169–170)—the caution in reading the texts he urges is well taken.

70. Guanzizai pusa suixin zhou jing, T no. 1103a, vol. 20: 462a. On Zhitong, see chapter four. Given that the phrase suixin is parallel with other phrases used to translate Indic words for amulets or other talismanic objects—ruyi rendering maṇi and suiqiu translating (as we will see in the next chapter) pratisara—it seems very likely that it also translated a word for amulets, a possibility of course made more likely given the contents of the text.

71. Guangda bao louge shanzhu mimi tuoluoni jing, T no. 1006, vol. 19: 641b.

72. Guanshiyin pusa mimizang ruyilun tuoluoni shenzhou jing, T no. 1082, vol. 20: 198c. See, as well, the Jñānagupta, Xuanzang, and Bodhiruci versions of the Amoghapaśa Dhāraṇī Sūtra, which contain similar recommendations (taking the first as representative, see T no. 1093, vol. 20: 401c). In keeping with the caution urged by Yu Xin in his study of the carrying of peach wood effective (Yu, Shendao renxin, 332), it is important to point out here that not every instance of the prescribed carrying of potent substances in Buddhist texts describes an amuletic practice. The Tuoluoni jijing, for example, recommends that such drugs be carried on the person not as periapts, but so that they will be ready to hand if needed (T no. 901, vol. 18: 858a-b). As well, the Dharmaguptaka Vinaya allows that if a monk is sick he is permitted to fit a container (“of bone, ivory, horn, iron, copper, wax, lead, or wood”) filled with smoking herbal preparations into a small sack to be worn from his arm. The Buddha makes clear, however, that such a rig is only allowed if the monk is too weak to hold the container in his hands (Sifen lü, T no. 1428, vol. 20: 877a).

73. Baopuzi neipian jiaoshi 19: 336. Translation Pregadio, Great Clarity, 126–127.

74. Foshuo guanding qiwan erqian shenwang hu biqiu zhoujing; T no. 1331, vol. 21: 501a.

76. As Sivin has noted, “The secret names of gods themselves served as charms, as in the Divine Treasure Canon, Ling-pao ching; so could a list of drugs” (Sivin, “Ailment and Cure,” 31). He cites Kaltenmark 1960 and Schipper 1965. Another apparently related practice was the simple holding of certain words in the palm of one’s hand as talismans. See Yu Xin, Shendao renxin, 335.

77. T 21: 502b. A version of this passage is quoted in Fayuan zhulin; T no. 2122, vol. 53: 925b.

78. Bukong juansuo shenbian zhenyan jing; T no. 1092, vol. 20: 232a. On earlier translations of this scripture and their importance in the development of the use of incantation cords, see above.

79. Guangda bao louge shanzhu mimi tuoluoni jing; T no. 1006, 19: 639a. The earlier, sixth-century translation of the text, the Moli mantuoluo zhou jing, contains one of the earliest injunctions to “hold the spell on your body or your clothes”; see T no. 1007, vol. 19: 658b. As so often with these texts, there is also a later and more elaborate version by Bukong; see T no. 1005.

80. Yizi foding lunwang jing; T no. 951, vol. 19: 226b. See also Bodhiruci’s translation of the Wu foding sanmei tuoluoni jing; T no. 952, vol. 19: 264b.

81. Wugou jin’guang da tuoluoni jing; T no. 1024, vol. 19: 721a.

82. The practice of wearing images is much less well attested. In fact, the only clear account I have found so far is a brief tale of the great Song literatus Su Shi (1037–1101) included in the Ming collection Collected [Accounts] of Departed Births (Wangsheng ji), an anthology of Pure Land devotional tales compiled by Zhuhong (1535–1615). Since the anecdote dates from around the turn of the seventeenth century, it can hardly be taken to reflect Su Shi’s actual practice, to say nothing of those of earlier centuries, but its resemblance to spell wearing makes it at least worth mentioning here. Su Shi, in the days of his exile in the south, painted an image of Maitreya and wore it when he walked about. When people would ask him about it he replied that it was an offering in honor of his deceased mother. He said he’d had her jewelry made into a “foreign-style monk’s staff” (huxi) and painted the image to generate merit for her rebirth in a pure land (yi jian wangsheng). (See Wangsheng ji, T no. 2072, vol. 51: 141a.) No reason is given for why he chose to wear the image, but it seems reasonable to think about the tale within the context of the wearing of spells, seals, and talismans. Other instances mainly come from early collections of “miracle tales,” including tales of people wearing small metallic images of Guanyin in their long hair (dai jingfa zhong) apparently as markers of their Buddhist faith (though in one case the main event of the tale comes when the statuette turns an executioner’s blade). See Guanshiyin yingyan ji, 85, 105, et al. An especially interesting example exists as a fragment of the lost Xuanyan ji quoted in Falin’s (572–640) Bianzheng lun, tells of a man named Fofo, said to have been “a buddha among men,” who wore an image of the Buddha on his back, which forced monks to worship at his back. The man was struck down by lightning while traveling. After he was buried, his coffin was struck by lightning as well and his corpse tumbled out. On his back were the words “vicious and without the Dao” (xiongnüe wudao). T no. 2110, vol. 52: 540a.

83. Contained in Sengyou’s (445–518) Hongming ji, T no. 2102, vol. 52: 48c–49a.

84. Hongming ji, T no. 2102, vol. 52: 48c.

85. For this characterization of the text, see Lai, “The Earliest Folk Buddhist Religion in China.” The foundational study of the scripture remains Makita, Gikyō kenkyū, 148–211, who based his work on the Dunhuang fragments, P. 3732 and S. 2051. For another discussion of this passage and its relevance for an understanding of Buddhist talismans, see Robson, “Signs of Power,” 140, who notes that this particular passage from the sūtra is only evidenced by Zhiyi’s quotation of it. The fact that Zhiyi’s quotations in the two texts do not seem to match each other suggest that he may have exercised an active editorial hand in his use of the work.

86. On the “teachings of men and gods” in the context of Huayan doctrines, see Gregory, “The Teaching of Men and Gods.”

87. Miaofa lianhua jing xuanyi, T no. 1716, vol. 33: 804a.

89. Jin guangming jing wenju, T no. 1785: 50c15

90. Miaofa lianhua jing xuanyi, T no. 1716, vol. 33: 804a.

91. Longshu wuming lun, T no. 1420, vol. 21: 957b-958b. The text is not mentioned in any medieval catalog and is evidenced only by a single manuscript copy held in the Ishiyama-dera in Japan (which was copied into T). Strickmann’s treatment of the text (Strickmann, Chinese Magical Medicine, 170–178) appears in part to have been based on Osabe, Tō Sō Mikkyō shi ronkō, 234–247. For more recent studies, see Robson, “Signs of Power,” 147–149, and Young, “Conceiving the Indian Patriarchs in China,” 287–292. Young suggests convincingly that at least in its present form the text was likely a product of the seventh century (though Strickmann places it in the sixth).

92. On this idea, see also Sivin, “Ailment and Cure,” 31; and Pregadio, Great Clarity, 129.

93. For studies of fu in Chinese Buddhism, see Robson, “Signs of Power,” and Mollier, Buddhism and Taoism.

94. The best study of this manuscript is now Yu Xin, “Personal Fate and the Planets,” but see also Stein, Serindia, 2:1080; Waley, A Catalogue of Paintings Recovered from Tun-Huang, 164; Whitfield and Farrer, Caves of the Thousand Buddhas, 83; Mollier, “Talismans,” 409; and Robson, “Signs of Power,” 155–158.

95. On this point I am in agreement with Mollier (“Talismans,” 409) and Robson (“Signs of Power,” 158).

96. For the identification of the first figure as Mercury, rather than the Pole Star (as has become common in scholarship on the talisman), see Yu Xin, “Personal Fate and the Planets.”

97. For another talisman explicitly drawing on dhāraṇī lore—here one to be inscribed on a seal and worn—see the eighth seal model included among the set on p. 3874, and Wang Yucheng’s transcription and discussion of it (Wang, Daojiao fayin lingpai tan’ao, 46–53).

98. See, for example, Sørensen, “Michel Strickmann on Magical Medicine” (though note that he does not question the “Chineseness” of such practices).

99. Mollier, Buddhism and Taoism, 209.

100. See Sharf, Coming to Terms with Chinese Buddhism, 2.

1. I follow Katherine Tsiang in taking fang to have indicated, most probably, an artisanal workshop rather than a section of the city (Tsiang, “Buddhist Printed Images and Texts of the Eight–Tenth Centuries”).

2. Feng, “Ji Tang yinben tuoluoni jingzhou de faxian,” 50. The last phrase is not legible in the photograph provided in the archeological report; I follow Feng’s description. Existing photographs of the amulet are unclear, but one can consult the example reproduced as plate 7c in Su, Tang Song shiqi de diaoban yin shua.

3. Feng, “Ji Tang yinben tuoluoni,” 50.

4. This xylograph is by far the most studied and published example of a Mahāpratisarā amulet. See, for example, Whitfield, The Art of Central Asia, v. 2, fig. 151; Matsumoto, Tonkōga no kenkyū, 604–609; Ma, “Da suiqiu tuoluoni mantuluo,” 550–552; Fraser, Performing the Visual, 155–158; and Tsiang, “Buddhist Printed Images and Texts,” 218–219.

5. Shouchi. On this term see the introduction.

6. The text paraphrased is the Baosiwei translation, Foshuo suiqiu jide dazizai tuoluoni shenzhou jing, T no. 1154, on which see below. See chapter 1 and below for discussions of the relevant features of amulet culture in China and elsewhere in Asia.

7. Monier-Williams Sanskrit-English Dictionary. See also Hidas, “Mahāpratisarā-Mahāvidyārājñī, The Great Amulet, Great Queen of Spells,” 18ff.

8. Bühnemann, “Buddhist Deities and Mantras in the Hindu Tantras II,” 34.

9. See Hidas, “Mahāpratisarā-Mahāvidyārājñī, The Great Amulet, Great Queen of Spells,” for the translation of Mahāpratisarā as “the Great Amulet.”

10. Note that the drawing of the pelvis seems to indicate that the wearer was male, though since the drawing was made long after the discovery of the tomb, we should be cautious in interpreting its details.

11. Foshuo suiqiu jide dazizai tuoluoni shenzhou jing (Mahāpratisarā-dhāraṇī-sūtra), T no. 1154. Though, following Hidas, “Mahāpratisarā-Mahāvidyārājñī,” and following the basic meaning of “pratisara” itself, I take seriously the fact that “suiqiu” in part referred specifically to amulets and not more generally to “wish fulfillment.” Indeed, we find structurally similar translations—ruyi most prominently—as renderings for amulets or talismans, mani in the case given. However, given that suiqiu was, in the present case, a shortened form of the phrase “suiqiu jide,” “what is wished for is immediately achieved,” I translate it in general as “wish fulfillment” and not as “amulet.”

12. On Manicintana, see Forte, “The Activities in China of the Tantric Master Manicintana.”

13. On the importance of her reign in the history of dhāraṇī practice, see the following chapter, and especially Barrett, The Woman Who Discovered Printing.

14. Dafangguang pusazang jing zhong Wenshushili genben yizi tuoluoni jing, T no. 1181. See Forte, “The Activities in China of the Tantric Master Manicintana,” for a discussion of the translations of these texts. On incantation cords and their relationship to the amulets, see below.

15. In addition to Amoghavajra’s text, Vajrabodhi is said to have made his own translation because he found the Manicintana version lacking. See Song gaoseng zhuan, 712a; Kaiyuan shijiao lu, 571c; and Yiqie jing yinyi, 553a. The latter text glosses two terms in Vajrabodhi’s work. Since neither appears in the received edition of Amoghavajra’s text, Vajrabodhi’s would indeed seem to have been a third version. See Hidas, ““Mahāpratisarā-Mahāvidyārājñī,” for a translation of a Sanskrit version that seems close to Amoghavajra’s. More generally, comparison of the Chinese versions with Hidas’s work makes clear the great extent to which different versions of the scripture differed from one another, and were perhaps local improvisations around a core of amuletic instructions and tales—though this last possibility awaits a broader systematic study of all extant versions.

16. For an overview of the nature of dhāraṇī scriptures, see Copp, “Dhāraṇī Scriptures.”

17. Reflecting the conventions of Esoteric scriptures, the Amoghavajra translation situates its opening scene at “the summit of Great Vajra-Sumeru (da Jin’gang xumilu feng, T no. 1153, vol. 20: 616a), a fact reflected in the later Sanskrit version studied by Hidas, which situates the Buddha at the peak of “Mt. Vajrameru” (Hidas, “Mahāpratisarā-Mahāvidyārājñī,” 184).

19. See Skilling, “The Rakṣā Literature of the Śrāvakayāna,” which makes clear, as well, that protection and reassurance was one of the main concerns of Buddhist iconography from its earliest stage. See also Granoff, “Maitreya’s Jewelled World,” for a discussion of the central place of the “concern for protecting its believers” (p. 181) in Buddhist (and Hindu and Jain) traditions and the ways that demonic predators were transformed into guardians of Buddhism and its adherents, a feature important in the Scripture of the Incantation of Wish Fulfillment and in dhāraṇī traditions in general.

20. In terms of dhāraṇī books and their Daoist counterparts during this period, Michel Strickmann’s work is most helpful. Nearly his entire corpus bears on the subject, but see especially Strickmann, Mantras et mandarins and Chinese Magical Medicine. For Fangshan, see the forthcoming study by Lothar Ledderose, a taste of which is available in Ledderose’s aptly titled “Carving Sutras into Stone before the Catastrophe.” It is important to note, however, that mofa belief in China was not limited to the early medieval period—it made a strong return in the Liao Dynasty (907–1125), for example.

21. For the early Indian history of anti-demonic magic as seen in the Atharva-veda, see N. P. Ahuja, “Changing Gods, Enduring Rituals.” For the early Chinese situation, see Harper, Early Chinese Medical Literature and “Warring States Natural Philosophy and Occult Thought,” and Seidel, “Imperial Treasures and Taoist Sacraments,” as well as the discussion in chapter 1 of this book.

22. Hidas, “Mahāpratisarā-Mahāvidyārājñī,” 18. Pratapaditya Pal, in a discussion of the amulet cords and boxes represented on Gandhāran bodhisattva statues (which he believes would have carried dhāraṇīs) notes that, in addition to the Atharva-Veda, the use of the Sanskrit word pratisarā (or the Prakrit paḍisarā) for amulets is found in the play Pratijñā-Yaugandharāyaṇa, by the probably early fourth-century poet Bhāsa (Pal, “Reflections on the Gandhāra Bodhisattva Images,” 101–102, 113 n. 7).

23. Material traces of the other sorts of amulet practices prescribed in the text have not survived, at least as clearly, a situation due in part simply to the fact that tombs preserve their contents, and provide contexts for them, in ways that little else that remains from the medieval period does. We cannot, thus, assume that non-mortuary practices of the Incantation of Wish Fulfillment were less popular based only on those amulets that have survived.

24. As noted above, in this section I will sample a few of the most important passages in the sutra, with only minimal commentary, in order to present some of the discursive background for the study of the amulets the text prescribes that will follow.

26. On this point, as on others, the Manicintana and Amoghavajra texts diverge. The Amoghavajra text does, in fact, start off with recitation. This feature of the text, like many others in scriptural translations attributed to Amoghavajra, seems an aspect of the increasing absorption of dhāraṇī practice into the conventions of the Esoteric Buddhist programs he propounded, in large part those associated with the Sarvatathāgatatattvasaṃgraha.

28. The word used here, shi, can also refer to the penis. Thus, given that the spell confers the power to create only (or mostly) males, “masculine power” might be a more fitting translation here. But since the passage also emphasizes the powers of persuasion and personal security (for her and for her fetus) conferred by the spell, I give its more usual sense.

29. T 20: 637c-638a.

30. Possibly referring to one of the forms of Maheśvara.

32. This would seem to be a function of the fact that, as I will explore below, the amulet itself is an altar of the iconic variety.

33. For the purposes of this chapter, I will split my presentation of the sutra into two parts, the first outlining its descriptions of how to wear the spell and the other, given later in the chapter, its accounts of how to write it and draw its accompanying imagery.

36. Miaofa lianhua jing, T no. 262, vol. 9: 2a.

39. T 20: 640b. The king’s martial practice finds a striking parallel in the Chinese historical record. As J. J. M. De Groot mentioned in an 1891 article, the emperor Xiaowu (510–535, r. 532–535), the final ruler of the Northern Wei Dynasty, is reported to have had within his retinue the monk Huizhen (d.u.) who, it was said, “bore a seal on his back (fuxi) and wielded a thousand-ox blade (qianniu dao).” See De Groot, “Militant Spirit of the Buddhist Clergy in China,” 128. My translation follows the passage contained in juan 156 of the Zizhi tongjian, which appears to be the earliest extant version of the anecdote. De Groot cites juan 29 of the seventeenth-century Rizhi lu by Gu Yanwu (1613–1682). A thousand-ox blade was a term for a sharp blade wielded with consummate skill. It originated in the Zhuangzi story of Cook Ding, who employed his knife with such skill that though he carved thousands of oxen it never grew dull. It is, short of further information, impossible to determine if the seal noted in the anecdote was simply a display of allegiance to his sovereign—xi typically referring to imperial seals—or a religious practice in line with that of the Incantation of Wish Fulfillment.

40. T 20: 640b. See below for the important implications of the likening of dhāraṇīs to seals.

41. Hidas notes that the later Sanskrit version of the Mahāpratisarā scripture often mentions cognitive engagement with the spell (Hidas, “Mahāpratisarā- Mahāvidyārājñī”).

42. That is, a lay donor to the monastic community.

43. The editors of T identify this as Puskalāvatī, a city in the Peshawar region of Gandhara (in modern Pakistan), a common stopping point for Chinese pilgrims of the medieval period. The Chinese name “Fulfilment City” (manzu cheng) seems only to be attested within translations of this scripture.

45. T no. 2084, vol. 51: 842a-b. The version runs: “Once there was a bhiksu who, [though] his mind embraced pure faith, went against the controlling precepts of the Thus Come One. Without writing out [the proper forms? (Buxie)] he seized for his own uses numerous objects from the common property of the resident samgha. Later, he became severely ill and suffered greatly. At this time, there being no cure for the monk, he cried out in a loud voice. Nearby there was a Brahmin who, hearing his great cry, hurried to the monk’s side. Great compassion arising within him, [the Brahmin] immediately wrote out this Great Incantatory King Dhāraṇī of Wish Fulfillment and fastened it to [the monk’s] neck, whose torments all ceased. The monk’s life thereupon immediately came to an end, and he was born into the Hell of No Interval [that is, Avīci Hell]. Because this dhāraṇī was borne on the corpse of the bhiksu, entombed in its pagoda, as soon as he entered hell all the torments of the sinners there all came to an end. Everyone there, without exception, attained security and joy. All the conflagrations of Avīci Hell, because of the awesome virtuous might of this dhāraṇī, were completely extinguished. …”

46. See appendix 1 for a survey of the 23 Suiqiu amulets considered in this chapter. Su Bai has made clear that others have been discovered but not yet reported or published (Su, Tang Song shiqi de diaoban yinshua, 7–9).

47. Ma, “Da suiqiu tuoluoni mantuluo,” 529–530.

48. At the time of this writing I have not found any photographs of the tubes, boxes, or pendants. Indeed, it is often frustrating that the archeological reports often do not include images of the cases at all.

49. For a photograph of the Jiao Tietou armlet, see Li and Guan, “Xi’an xijiao chutu.”

50. See the previous chapter or a survey of relevant Chinese practices.

51. Zysk, “Religious Healing in the Veda,” 174; Bolling, “The Cantikalpa of the Atharva Veda,” 120; Hidas, “Mahāpratisarā-Mahāvidyārājñī,” 18. See also Bühnemann, “Buddhist Deities and Mantras,” 34.

52. Note that “knotted ‘protection cords’” are still in use in contemporary Tibetan Buddhism, at least in Nepal, as Matthew Kapstein has noted (Kapstein, Tibetan Buddhism: A Very Short Introduction, 2).

53. Ślączka, Temple Consecration Rituals in Ancient India, 68–69, 175.

54. Jingchu suishi ji, comp. Zong Lin (fl. 550s), I am grateful to Ian Chapman for alerting me to the existence of these cords in Chinese amuletic practice. The translation of Zong Lin’s description is his.

55. Wuming luocha ji; T no. 720, vol. 16: 851a-b. For a convenient summary of the contents of this little-studied text, see Busshō kaisetsu dajiten 10: 420. Kings required the protection of such cords not only on the battlefield, it seems. Alexis Sanderson describes an instance where a king was adorned with a “protective wrist thread” (pratisaraḥ) as he prepared for his first foray onto the wedding bed (Sanderson, “Religion and the State,” 250).

56. Rulai fangbian shanqiao zhoujing (perhaps, as the editors of H propose, reading back from a later Tibetan version, the Saptabuddhakasūtra [H, p. 116]); T no. 1334, vol. 21: 565c. It is unclear if (here and in the next accounts) the details that the cords to be used were often “five colored” were additions of the Chinese translations, as seems possible, or were present in earlier Indic versions.

58. The precise identity of these “mountain stakes” is unclear. They would seem to be a special version—perhaps one occurring naturally in the mountains?—of the stakes commonly used to nail down the corners of ritual spaces, sometimes called “vajra stakes” (jin’gang jue; vajrakīla). On the use of similar stakes, at times painted with human images and used in exorcistic rites, see Mollier, Buddhism and Taoism, 85; Whitfield and Farrer, Caves of the Thousand Buddhas, 174–175; and Liu Zhaorui 2007, Kaogu faxian yu zaoqi Daojiao yanjiu, 356–357.

59. Bukong juansuo shenzhou xinjing; T no. 1094, vol. 20: 405.

60. Thus we see, once again, the essential identity of amuletic practices (including the logics of their descriptions) with those known in Chinese as zhaifa, which secure the safety of physical spaces. For an excellent discussion of these techniques in late medieval Dunhuang, see Yu Xin, Shendao renxin.

61. The Susiddhikara (on which see below) and its ancillary ritual manuals typically call for the armlet to be strung with “living children” (huo erzi), a term for the seeds of the “Bodhi Tree” (putishu), under which, legends tell us, Siddhartha Gautama achieved the awakening of a Buddha.

62. For example, see Suxidijieluo jing, T no. 893, vol. 18: 628c.

63. Tuoluoni ji jing, T no. 901, vol. 18: 838a.

64. Putichang suoshuo yizi lunwang jing, T no. 950, vol. 19: 222a.

65. See the helpful brief discussion of samaya in Rolf Giebel’s discussion of the Susiddhikara (Giebel, Two Esoteric Sutras, 11).

66. Powers, Pattern and Person, 61.

67. Shengjing, T no. 154, vol. 3: 101a.

68. In this regard, a (perhaps unique) event in the history of the Tang imperial house takes on added interest: “[The] emperor [Tang Suzong (711–762; r. 756–762)] arranged for a … religious drama to be enacted in the newly completed chapel in the Lin-te Hall of the Ta-ming Palace. His ministers were summoned to render homage to palace attendants who were dressed as Buddhas and bodhisattvas, while army officers portrayed the guardian spirits of Buddhism (jin’gang shenwang)” (Weinstein, Buddhism Under the T’ang, 59). See below for more on this emperor and his involvement with dhāraṇī practice. For a study of some similar portrayals, see Li Yuhang’s study of women’s devotion to Guanyin in the late imperial period, especially her chapter on the Empress Dowager Cixi’s practice of dressing up as the bodhisattva and having her portrait painted or photographed (Li, “Gendered Materialization”).

69. Tuoluoni ji jing, T no. 901, vol. 18: 796b.

70. Cheng, Luoyang chutu Hou Tang diaoyin jingzhou; and Ma Shichang, Da suiqiu tuoluoni mantuluo. See below for a discussion of the amulet sheet.

71. Han, Shijie zuizao de yinshua pin, 404–410; Ma, Da suiqiu tuoluoni mantuluo, 535. See below for a discussion of this amulet sheet.

72. Kotansky, “Incantations and Prayers for Salvation on Inscribed Greek Amulets,” 114 and passim.

73. For fascinating studies of medieval Arabic block-printed amulets, see Schaefer, Enigmatic Charms; and Elverskog, Buddhism and Islam on the Silk Road, 104–116. Elverskog’s work, which was published after the present book was sent to the press and which I only discovered just before it was to be printed, is especially important for the connections it draws between Buddhist and Arabic printed amulets.

74. Skemer, Binding Words, 26–29 and passim. See also Skemer’s extensive and very helpful bibliography for a list of relevant studies. The study of amulets in the ancient and medieval worlds constitutes, naturally, a vast topic. The connections of this material with Central and East Asian traditions fairly cry out for serious study. A very helpful contribution to this study is Elverskog, Buddhism and Islam on the Silk Road, 104–116 (see previous note and the coda to the present volume).

75. On the ye dharmā and its central place in text relic practices, see the previous chapter.

76. Salomon, Ancient Buddhist Scrolls from Gandhāra, 86, 59–60.

77. Ibid., 86, quoting Bühler, Detailed Report of a Tour in Search of Sanskrit Mss., 29 n. 4.

78. See von Hinüber, Die Palola Sahis, 16–17; von Hinüber, “Namen in Schutzzaubern aus Gilgit,” passim; and Hidas, “Mahāpratisarā-Mahāvidyārājñī, 17, 25–26.

79. Skjærvo, Khotanese Manuscripts, 231, where the text is identified as IOL Khot 28/5 and Kha. i 182a1.

80. Ibid., 193, labeled as IOL Khot 16/4 and Kha. i 53.

81. Ibid., 585, and Emmerick, “Some Khotanese Inscriptions,” 142–143, and pl. II. For the images figured on it see Aurel Stein, Serindia, v. IV: XCI.

82. Emmerick, “Some Khotanese Inscriptions,” 142.

83. Skjærvo, Khotanese Manuscripts, 585.

84. T no. 1154, vol. 20: 641c.

85. I have added the numbers and the paragraphing for ease of reference.

86. See below for an example of an amulet apparently following this prescription.

87. T no. 1154, vol. 20: 641c–642a

88. Strickmann, Chinese Magical Medicine, 184. The connection with bile in medieval China seems to have been especially close. Edward Schafer noted that while traditionally “bezoar” referred to a “concretion found in the fourth stomach of many ruminants, notably the bezoar goat … [the] ‘bezoars’ of medieval China, called ‘ox yellow’ there, did not always match this classic definition. Some, if not most, were biliary calculi, taken from the gall bladders of oxen” (Schafer, The Golden Peaches of Samarkand, 191).

89. It is especially important in the early Tantric Susiddhikara, as well as in earlier texts of dhāraṇī tradition, such as the Tuoluoni jijing and the Amoghapāśa.

90. Strickmann, Chinese Magical Medicine, 188.

91. Schafer, The Golden Peaches of Samarkand, 191.

92. Strickmann, Chinese Magical Medicine, 185.

93. Schafer, The Golden Peaches of Samarkand, 191.

94. See, for example, Yizi foding lunwang jing, T no. 951, 19: 226b; and Wu foding sanmei tuoluoni jing, T no. 952, vol. 19: 264b. See also chapter 1 of this book.

95. Strickmann, Chinese Magical Medicine, 185.

96. Chaves, “The Legacy of Ts’ang Chieh,” 210–211. The power attributed to the ink could, as well, have been due to its physical connection with a Buddhist temple, for reasons that will be become clear in the next chapter.

97. Ma, “Da suiqiu tuoluoni mantuoluo.”

98. T no. 1153, vol. 20: 623c–624a.

99. Taking sanji as an abbreviated form of sanchaji, “trident” (Skt. Triśūla).

100. Thus connecting the amulet’s image with canonical accounts of the realms of buddhas, which nearly always feature such “treasure pools” (baochi). See, for one example among a great many, Prajña’s eighth-century translation of the Ga็ḍavyūha Sūtra (Dafangguang fo huayan jing), T no. 293, vol. 20: 718c.

102. That the diagrams seem to have been used in constructing actual ritual images (or perhaps actual physical spaces) is suggested, as Sarah Fraser has pointed out, by the corrections written on to Stein Painting 172, evidently by the diagram maker’s “supervisor” (Fraser, Performing the Visual, 153). These corrections, for example, “put the deity in the center and move the mudra to the left,” in Fraser’s translation (153), likely are traces of a dialog between the artists and a ritual master.

103. Fraser also draws close connections between diagrams such as Stein Painting 172 and the amulets, particularly Stein Painting 249, and notes the talismanic function of the latter set. She sees a “high/low” distinction in the cultures of the two types, with the handwritten altar diagrams part of highly individualized and sophisticated monastic practices of contemplation, and the amulets, particularly their printed variety, as “popular expressions of prayers” (drawing a link between the printed amulets and other printed icons, on which see below) that were mass-produced and important in the economic life of the monasteries (Fraser, Performing the Visual, 157–158).

104. See, for example, Yizi foding lunwang jing, T 951, vol. 19: 226b; and Wu foding sanmei tuoluoni jing, T 952, vol. 19: 264b.

105. Fraser, Performing the Visual, 153.

106. See, for example, Da fangguang pusazang Wenshuhili genben yigui jing, T no. 1191, vol. 20: 876c–877b; Foshuo shengbaozang shen yigui jing, T no. 1284, vol. 21: 350b and 351c; and Yuqie dajiao wang jing, T no. 890, vol. 18: 560a-b.

107. See Da Song sengshi lüe, T no. 2126, vol. 54:240b-c, and chapter 4 below.

108. Interestingly, two of these amulets seem to date from transitional moments in the history of the production of the amulets. The first, made for “Iron-head” Jiao, was handwritten and painted, yet its central image is an iconic representation of the Bodhisattva Mahāpratisara (see below for the significance of these images). The second was made for a certain “A-luo.” Like late amulets, it is a woodblock print, yet like early amulets, its central image is of the donor receiving divine blessings, and the name of the donor has been written in to the text of the spell. Both amulets are contained in the collection of the Xi’an Forest of Stone Tablets Museum. See Cheng, Xi’an Beilin Bowuguan, 152–153.

109. Ma Shichang, Da suiqiu tuoluoni mantuoluo, 530. For related claims for the Chinese origins of amuletic practices involving dhāraṇīs, see Xiao Dengfu, Daojia Daojiao, 773, and Daojiao shuyi, 194–195, as well as Li, Dunhuang Mijiao wenxian, 295.

110. Strikingly, Hidas, in his study of Indic versions of the Mahāpratisarā dhāraṇī and scripture, notes that though many extant versions of the spell from Gilgit and elsewhere in the Indic world contain names inserted within it, the Sanskrit version of the scripture he studies contains “no definite instruction” to insert names within it. Instructions within the Chinese versions are included as interlinear notes, a fact suggestive of the culture of practice evidenced everywhere surrounding this incantation, including in translations of its scripture.

111. The character/syllable xie (to write, copy, etc.) might initially be taken simply as part of the transliteration of the spell (and, indeed, I took it that way for some time), but three facts strongly suggest that it was taken as a Chinese word: it recurs at each point the name is indicated, regardless of the incantatory words that follow; it bears no relation to the next words in the dhāraṇī; and, finally, the Chen Chouding amulet substitutes “receives and keeps” for “writes,” and makes clear that the syllable xie was taken as a Chinese word (at least in that one instance).

112. Dazuigu darani kanchū, T no. 2242, vol. 61: 747a-b.

113. See the next chapter for further discussion of this phenomenon. For an extended study of such glosses and what they reveal about dhāraṇī practice in medieval China and Japan, see Copp, “Anointing Phrases and Narrative Power.”

114. Bibliothèque Nationale, Catalogue des Manuscrits Chinois, 165.

115. See the next chapter for a discussion of this account.

116. Bibliothèque Nationale, Catalogue des Manuscrits Chinois, 454.

117. An and Feng, “Xi’an Fengxi chutu de Tang yinben,” 89–90. An and Feng date the sheet largely based on three stylistic judgments: they note that the vajradhara figure resembles mid-eighth century forms at Dunhuang, that the headscarf of the kneeling figure matches “High Tang” images, and that the shortness of the clouds’ “tails” marks them as either early or High Tang styles.

119. See chapter 1 for a discussion of the relationship between seal practices, the Baopuzi and related works, and inscribed Buddhist incantations.

120. Ma, “Da suiqiu tuoluoni mantuoluo,” 539.

121. See Drège, “Les Premières Impressions”; Su, Tang Song shiqi de diaoban yin shua; Barrett, “Stūpa, Sūtra, and Sarīra”; and Strickmann, Chinese Magical Medicine, 123–193.

122. Zwalf, Buddhism: Art and Faith, 70 (fig. 82), and the British Museum online collection notes.

123. See Schopen, Figments and Fragments of Mahāyāna Buddhism, 338, for the identification and 314–344 on this dhāraṇī in broader contexts. The stamp also bears the text: “whosoever constructs a caitya after having written this dhāraṇī and thrown it inside, will gain the merit of having constructed 100,000 caityas.”

124. Daizong chaozeng sikong dabian zheng guangzhi sanzang heshang biaozhi ji, T no. 2120, vol. 52: 829b. For more on this passage see chapter 3. The text notes, as well, that Amoghavajra used the Incantation in 761 to cure the emperor when he had fallen ill.

125. The word I have translated as “text” here, ben, was in fact often employed, in the compound zhouben, or “spell text,” in naming single sheets of incantations, often those belonging to individual practitioners, including at least one example of a Mahāpratisarā amulet—the Chengdu xylograph with which this chapter opened (and to which I will return below). The term was also used as a general term for the texts of spells, particularly in catalogs. Most known examples of zhouben were found within the Dunhuang cache and consist not of elaborate pairings of spell and image but simpler manuscripts that more often than not took the forms of single sheets inscribed with the text of a transliterated dhāraṇī. The manuscript known as S. 165, “Changxin’s spell text,” apparently the personal spell sheet of a monk named Changxin (otherwise unknown), is a typical example.

126. Foshuo suiqiu jide dazizai tuoluoni shenzhou jing, T no. 1154, vol. 20: 637c.

127. For the image, see Su, Tang Song shiqi de diaoban yinshua, 4, 127; and Ma, “Da suiqiu tuoluoni mantuoluo,” 536–537, 569. The printed edition of the spell on this sheet, incidentally, may be the earliest version of the transliterated spell yet known. As Ma notes, its Chinese transliterations differ from those of transmitted editions of the Baosiwei spell, whose title this sheet bears—a trait shared by the version of the spell found on S. 4690, as discussed below.

128. An and Feng, “Xi’an Fengxi chutu de Tang yinben,” 88.

130. See T no. 1154, vol. 20: 642a.

131. Two versions of these spells are found at T no. 1154, vol. 20: 640a (the version found on the Koryŏ edition) and 644b (the one dating to the Ming edition). Close examination of the text of the spells inscribed on S. 4690 sheds some light on the discrepancies between the two versions. In general, it is closer to the Koryŏ version, but at several points it matches the Ming spell quite precisely. The two versions included in T seem to have been but two of many versions used in the history of Buddhist practice in China, and not necessarily “canonical” in anything but a rather contingent sense (i.e., they were enshrined in printed canons). Taking only the two received versions as evidence, one would be tempted to conclude that the much later Ming edition was in fact a text developed after the earlier Korean printing. But such a conclusion is complicated by the existence of the version on this amulet model, which very likely predated both of them. The four spells on S. 4690, as best as I can transcribe them, are given in appendix 2.

132. The amulet can be viewed on the International Dunhuang Project website (http://idp.bl.uk/) under the search value “EO 1182.”

133. With, again, the exception of the “A-luo” sheet, which blurs my sharp ideal types, and which I take (provisionally) as the product of a transitional or perhaps backward-looking practice.

134. See the following chapter for a discussion of these objects.

135. Mair’s analysis proceeds in terms of the distinction between “bianxiang” and “maṇḍala” (Mair, “Records of Transformation Tableaux,” 3–4), while Teiser uses “narrative” and “icon” (Teiser, Scripture on the Ten Kings, 42–43). Teiser’s analysis focuses on the hanging silk paintings that I will discuss below. Though I have tried in this section to apply the rubrics with care, and to adjust them to the material I consider here, Teiser’s warning that “we must remember that the distinction is modern and arbitrary, and at times impedes an understanding of medieval culture” is well taken (Ibid., 43).

136. Wu Hung, The Wu Liang Shrine, 133.

137. Teiser, Scripture on the Ten Kings, 42–43.

138. Wu Hung, The Wu Liang Shrine, 133.

139. See chapter 1 for an exploration of this ritual logic.

140. Stein Painting no. 247.

141. Bukong juansuo shenbian zhenyan jing, T no. 1092, vol. 20: 299c.

142. But see the Ruiguangsi prints discussed below.

143. On the Dunhuang Mahāpratisarā invocations, see Chen Huaiyu, “Dunhuang P. 2058v wenshu.” Note that, as mentioned earlier, the monk Xingsi’s amulet (P. # 3982), which the compilers of the Pelliot Catalog date to the tenth century—contemporary with the manuscript versions of the ritual invocations—also contains an invocation (though a much simpler one) within the text of its spell.

144. See Suzhou bowuguan, Suzhou bowuguan cang Huqiu Yunyansi ta, Seunghye Lee, “Arts of Enshrining,” and Eugene Wang, “Ritual Practice Without a Practitioner?”

145. At least none have yet been made public. The slow pace of the publication of archeological finds in China imposes the need for caution here. There is, however, later textual evidence suggesting that the Mahāpratisarā was used for amulets in later periods in Japan; Jōnen’s (fl. 1154) collection of ritual modules, the Gyōrinshō, contains an account of how to make amulets that clearly drew on knowledge of Chinese examples (T no. 2409, 76:304b).

146. See chapter 4 for a discussion of this shift in regards to the material studied in this book.

147. The sole exception being the holes, and the possible traces of thread they still bear, within the Chen Chouding amulet.

148. See, now, the excellent discussion of this small pillar and its place in the Ruiguang si pagoda in Lee, “Arts of Enshrining.”

149. Reading wu neng sheng.

150. I have used the transcription of the text found in Ma, “Da suiqiu tuoluoni mantuoluo,” 546.

151. The image can be viewed on the International Dunhuang Project website (http://idp.bl.uk/) under the search value “EO 1232.” For the ritual prescriptions, see, for example, the description of a rite in the Usnīsavijaya tradition that calls for, among other things, vajras at the frame of an icon and wheels at the corners: the manual attributed to Śubhākarasimha known as the Zunsheng foding yuqie fa yigui (T no. 973, vol. 19:375c); see also the later Japanese summary in the Gyōrinshō (T no. 2409, vol. 76: 78c).

152. It should be noted that the Tang Dynasty had fallen in 907.

153. For a fuller description and analysis of this painting, see Whitfield, The Art of Central Asia: Vol. 2, Plate 7. The image and Whitfield’s analysis are also included in the website of the International Dunhuang Project (www.idp.bl.uk), under the search value “1919,0101,0.14.” See also Waley, A Catalogue of Paintings: 26.

154. Teiser, Scripture on the Ten Kings, 42.

155. See Tsiang, “Buddhist Printed Images and Texts,” for an overview of early Buddhist printed images.

156. It is crucial to emphasize the very different natures of icons such as Stein Painting 237 and dhāraṇī amulets such as those that are the subject of this chapter. Cf. Fraser, Performing the Visual, 155–158.

157. For a study of this print in the context of concepts of the “True Visage” and the cult of Mañjuśrī on Wutaishan, see Choi, “Quest For the True Visage.”

158. Jiao’s amulet, found on Fenghao Rd in Xi’an in the early 1980s (see Ma, “Da suiqiu tuoluoni mantuoluo,” 528–529), confounds any simple picture of historical progression in the design and production of the amulets. It features an image of the eight-armed deity at its center, yet it is a handwritten and painted manuscript that includes Jiao’s name within the text of the spell. Ma Shichang, because of its visual similarity with Madame Wei’s amulet, deemed it to be roughly contemporary with that painting, dating it to the middle of the eighth century. Yet the presence of the multiarmed deity makes this unlikely. I think, instead, that it is most likely either an “archaized” product of the following century, created by someone who knew the older amulet traditions and sought to merge them with the new iconography, or a truly transitional piece whose makers adapted the new iconography at a time (or place) where the new styles of xylograph icons had not yet taken hold.

159. See Liu Yongzeng, “Mogaoku di 148 ku nanbei kan tianjing tuxiang jieshuo,” 524–526, for a discussion of this painting. See also Li Ling, “Dasuiqiu tuoluoni zhoujing de liuxing yu tuxiang,” who claims to be able to read the entire inscription as “Suiqiu pusa.”

160. Wong, “Divergent Paths.”

161. Waley, “A Catalogue of Paintings.”

162. As Liu Yongzeng notes, a nearly identical tableaux is found in the slightly later Cave 156.

163. Tsiang, “Buddhist Printed Images and Texts,” 246–247.

164. Abé, Weaving of Mantra, 124–125. See 487–488 n. 60 for Abé’s discussion of the nature of this work.

165. That is, literally, an “Indic case” (fanqie) in which a collection of inscribed leaves were kept.

166. Hizōki 2:12. See also Ishida, Mandara no kenkyū, 44–45, for a discussion of this form of the Zuigu bosatsu within the context of the Womb Mandala.

167. Yōson dōjōkan, T no. 2468, vol. 78: 50c.

168. Betsugyō, T no. 2476, vol. 78: 132c.

169. Yuqie dajiao wang jing, T no. 890, vol. 18: 568a-b. Though he does not mention this text, Mevissen has tracked these images to the eleventh- and twelfth-century iconographic manuals known as the Sādhanamālā and the Nispannayogāvalī, sources that, if Mevissen is correct about their dates, significantly postdate the Chinese translations. See Mevissen, “Studies in Pañcarakṣā Manuscript Painting,” 362–363, and Mevissen, “Images of Mahapratisara in Bengal,” 99, where he notes that the earliest actual statues exhibiting this form date only to the eleventh century at the earliest.

170. Mevissen, “Images of Mahapratisara,” 99–108.

172. Some of the Javanese examples seem to include objects, such as in several cases a jewel in place of a “jeweled banner” (Ibid., 116), not found elsewhere among these images.

173. The two similar pieces are numbers three and six in his account. Of the two, the most completely readable sculpture is the latter, a bronze image from Java now held in Paris’s Musée Guimet (Ibid., 102–103). Its possible match is a metal sculpture held in the Bangladesh National Museum in Dhaka (Ibid., 100, 103). Neither piece is precisely datable.

176. See Chou, “Tantrism in China,” 290 n. 29, on the identification of this land as Java.

177. Da Tang gu dade sikong dabianzheng guangzhi Bukong sanzang xingzhuan, T no. 2056, vol. 50: 292c. A slightly different version is collected in Song gaoseng zhuan, T no. 2061, vol. 50: 712b. See also Chou, “Tantrism in China,” 290, for a translation of the latter work, which I sometimes followed.

178. This is, indeed, how Mevissen seems to have taken it (Mevissen, “Images of Mahapratisara,” 117). It should be noted, however, that since no full Western language translation of the Mahāpratisarā dhāraṇī sūtra existed at the time of his research on the iconography, he may not have had access to the scriptural tales I discuss here.

179. This is either the timil, a large fish mentioned in the Vedas, or (as the editors of T have it), the timilgila, the “eater of the timil,” an even greater beast. The authors of the Fan fanyu take timilgila to mean “fish of the gods” (tianyu); T no. 2130, vol. 54: 1024b.

180. Reading xiao for suo.

181. T no. 1153, vol. 20: 621b–622a.

1. Rulai fangbian shanqiao zhoujing (perhaps, as the editors of H propose, reading back from a later Tibetan translation, the Saptabuddhakasūtra [H, p. 116]); T no. 1334, vol. 21: 565c.

2. Bukong juansuo shenzhou xinjing; T no. 1094, vol. 20: 405c.

3. Foding zunsheng tuoluoni jing; T no. 967, vol. 19: 351c.

4. I have followed the edition of this text in Li Xiaorong, Dunhuang Mijiao wenxian lungao, 59–60, which was based on the versions found on the manuscripts Sh.M. 48.6; S. 5598; and S. 5560.

5. Shimono, Tonko Bakukōkutsu dai niyakujūnana kutsu. Though I think Shimono is correct, her identification of this painting remains controversial. See, for example, Wang, Shaping the Lotus Sutra, for an alternative reading of some of the paintings at issue here.

6. Li Xiaorong, Dunhuang Mijiao wenxian, 64–65. In Japan, especially, another spell was most closely associated with these mortuary practices—the Kōmyō shingon, or Mantra of Light. For an in-depth discussion of this mantra and its practices—which bear a close resemblance to the much earlier Chinese techniques of the Incantation of Glory—especially as seen through the work of its earliest proponent, the thirteenth-century monk Myōe, see Unno, Shingon Refractions, and Tanabe, Myōe the Dreamkeeper, 137–152.

7. Zongshi tuoluoni yi zan, T no. 902, vol. 18: 898b.

8. “Dhāraṇī pillars” is the most common name for the structures in English. A better translation of their most common name in Chinese—jingchuang—is “scripture pillar.” They were also labeled “treasure pillars” (baochuang) and “shadow pillars” (yingchuang), among other things (see Liu Shufen, Miezui yu duwang, 52).

9. Note that the text does not specify a stone banner or pillar.

10. The Buddha is addressing the god Indra.

11. The class of beings, sometimes called “titans” or “demigods” in English, which constitutes one of the possible “paths” of rebirth.

12. T no. 967, vol. 19: 351b. That is, unsurpassed and perfect awakening. The text gives the usual Chinese transliteration of the Sanskrit term given here.

13. Longhua si ni Wei Qiyi zunsheng chuang ji, in BQ 47: 12. See Liu, Miezui yu duwang, 10–11 and 134–5 for discussions of this inscription. For a discussion of the term mizang, see chapter 4.

14. Davidson, “Atiśa’s A Lamp for the Path of Awakening,” 290, discusses this “economy of basic forms” not only as a core feature of institutions but more narrowly as “one of the keys to understanding Buddhism during its final phases on Indian soil.”

15. For more information on this pillar, see Liu Shufen, Miezui yu duwang, 62.

16. See, for example, Ibid., 46.

17. In such places the pillars often became new versions of older practices: in tombs they were like the tomb-securing (zhenmu) objects and texts of ancient Chinese practice; in homes and within cities they became forms of the “methods of the abode” (zhaifa) used to make them safe from perils, demonic or otherwise. For an exploration of such practices, see Yu, Shendao renxin.

18. Liu Shufen, Miezui yu duwang.

19. On this association, see Zhiru, The Making of a Savior Bodhisattva, 147–150, et al.

20. Howard, “The Dhāraṇī Pillar of Kunming.”

21. Schopen, Figments and Fragments of Mahāyāna Buddhism, 350–370.

22. Liu, Miezui yu duwang, 131–139.

23. On this coffin, see Shen, “Praying for Eternity.”

24. Frankfurter, “The Magic of Writing and the Writing of Magic.” More recently, I encountered the work of Don C. Skemer, whose characterization of these modes of efficacy as “material magic” is in some ways closer to my own. In his discussion of Anglo-Saxon charms he notes that “there are a number of apotropaic texts that were supposed to be copied on writing supports that could then be applied directly to the body or used together with other remedies. … Material magic could take the form of inscribing a few sacred names, Christian symbols, and cryptic words on stones, wafers, plates and other objects, which could then be worn from the neck or eaten or written directly on the body” (Skemer, Binding Words, 78–79).

25. These notions were not limited to the Chinese—there is much in them that is applicable to wider Buddhist practices and understandings. My relatively narrow focus on medieval China is not meant to deny or obscure this fact.

26. Frankfurter, “The Writing of Magic,” 191. He notes that these distinctions were first put forth in the Cornell lectures of Marcel Détienne.

27. Ibid., 192. According to Frankfurter, the reasons why ancient Egyptians saw written spells as powerful have much to do with the nature of their hieroglyphic writing system—notably, its pictographic nature. The Chinese script, of course, is not primarily a pictographic system, save in a few, albeit prominent, cases. It is largely a phonetic script (there are several different kinds of characters, but the phonetic is the largest group). Frankfurter stresses the pictograph’s ability to “reify their subjects as well as the things expressed,” which he contrasts with Greek writing’s close relationship with speech, as a partial explanation for the perceived potency of written Egyptian. The images of the script echoed the images of religious art and architecture, so that writing about a god captured something of that god. In addition, he describes the fact that the Egyptian script was the “indispensable and dynamic center of the Egyptian cultic-priestly world,” and that it “was maintained by and for a priesthood with the intention of encoding or fixing ritual and cosmology in a timeless and ideal reality.” He notes how the ritual and social status of the script lent instant authority, in the form of an “archaistic timelessness,” to even the most ad hoc improvisations of priests. This was partly achieved by the fact of the script’s “discontinuity with popular spoken Egyptian”—a feature, we may note, shared to some degree with literary Chinese (Ibid., 191–193). This is not the place for a full-scale comparison between Egyptian and Chinese talismanic writing, though such a project would doubtless be fascinating—not to mention provide balance to what to my mind is Sinology’s overemphasis on Chinese-Greek parallels.

29. For a convenient collection of these texts, see Huang and Wu, Dunhuang yuanwen ji.

30. Frankfurter, “The Writing of Magic,” 196.

31. As opposed to the sense of the term contagion as employed by the mythographer James Frazer, whose writings yet remain interesting to a study of material incantations (Frazer, The Golden Bough, 43 et al.)

32. This image introduces further distinctions into the traditional Buddhist picture of how spiritual potency (of whatever vintage) is imbued into an object. For example, one might argue that the nearly physicalist account of incantatory potency I advance here merely describes an epiphenomenon of the way the power of words is imagined in many forms of Buddhist practice, that the “charge” certain dhāraṇīs were thought to lend to the material they touched was simply the kind of merit generation common to many sorts of Buddhist ritual. This is certainly part of it; and as we saw before, the two modes were at times conflated. But such an objection might ignore the fact that, especially in the case of incantatory anointment, the substances so empowered were then themselves capable of empowering—that the dust and shadows described in the sūtra then have the same qualities that the scriptural words have. Seeing the empowered material as secondary to the buddhavācana itself, in these ritual contexts, misses something very important in the world the sūtra creates in its descriptions. It is also, perhaps, and particularly when we are interpreting materialized incantations such as amulets and pillars, a sign that scholars’ love of words might blind them to the value of normally mute stuff. One might also counter that, in the case of the Incantation of Glory, I am inflating out of all proportion the importance of what is in the text a minor point, and one that the only Chinese commentary on its sūtra passes over almost totally in silence. (Foding zunsheng tuoluoni jing jiaoji yiji, Fachong (fl. eighth century), T no. 1803). Yet the pillars prescribed in that small portion of the scripture filled the landscape of medieval China, and images of their potent dust and shadows were key factors in this spread and key features of the Buddhism of the age.

33. See BQ 46: 6 for a pillar called a “shadow pillar.” The imagination of shadows (or of the term ying, which can also mean “reflected image”) in Buddhist writings deserves a study of its own. See, for example, the Zanyang shengde duoluo pusa yibaiba ming jing, translated by Tianxizai (*Devaśanti?; d. 1000), which states that a person’s shadow can be made impervious to demonic harm (T no. 1106, vol. 20: 476a). See also the Da weide tuoluoni jing, translated by Jñanagupta (523–600), which names shadows as one possible vector of poison, in a list that includes the roots, stalk, branches, flowers, and fruit of a plant (T no. 1341, vol. 21: 834b-c).

34. Tanabe, Myōe the Dreamkeeper, 9. For a meditation on the nature of fantasy in Buddhism, see pages 1–10 of Tanabe’s book.

35. Sontag, Illness as Metaphor, 3.

36. The pillar is dated to 939. See Longyou jinshi lu, Tang section, 70–71. I am grateful to Kuo Liying for this information.

37. For the listing of copies of the Scripture of the Incantation of Glory, see Dunhuang yanjiu yuan, Dunhuang yishu zongmu suoyin xinbian, pp. 7 and 55 of the index section; for the list of the Huayan jing copies, see pp. 5–6; for those of the Lengyan jing, see pp. 7 and 107.

38. Forte, “The Preface to the So-Called Buddhapālita Chinese Version of the Buddhoṣṇīṇa Vijaya Dhāraṇī Sūtra,” and Chen Jinhua, “Śarīra and Scepter,” are exemplars here.

39. See Barrett, “Stūpa, Sūtra, and Sarīra.”

40. Li Xiaorong, Dunhuang Mijiao wenxian, 42–73, and Lü, Mijiao lunkao, 77–108.

41. Liu Shufen, Miezui yu duwang.

42. Forte, “The Preface,” 6.

43. The issue of how to render the name of this monk is a vexed one, due to conflicting reports of how his name was translated in Chinese: Juehu and Jue’ai. I follow Forte, who renders it Buddhapāli[ta], based on Zhisheng’s translation of Juehu. Forte notes that he does so not out of any clear sense that Zhisheng was correct but “to avoid confusion among readers” (see Forte [n.d.], 23–24, n. 115 for a discussion of these matters). I follow him in this.

44. My translation draws on Forte’s (in “The Preface”) and Lamotte’s (in “Mañjuśrī”), and differs from theirs largely on stylistic grounds, except as noted.

45. The phrase here is qianling, literally “submerged their spirits” [or perhaps “souls,” or “powers”].

46. The conditions that prevent one from meeting a buddha: life as a hell dweller, a hungry ghost, an animal, life in the heavens, in Uttaru-kuru (where all is pleasurable), being deaf, blind, or dumb, being a non-Buddhist philosopher, and living in an age between appearances of a buddha (see, for example, the Sifen lü, T no. 1428, vol. 22:567a).

47. My reading follows Forte. But note that the text can also simply be read, “coming from out of the mountains.” However, as we will see, that the bodhisattva (and later Buddhapālita himself) resides inside a mountain is a key feature of these legends.

48. I follow Forte’s suggestion that this date is not the year of Buddhapālita’s return to China but of the completion of the version of the scripture attributed to him and Shunzhen. “This procedure was very familiar to the compilers of historical records when relating different events: the date of the central and final event following a series of related events is stated at the beginning” (Forte, “The Preface,” 8 n. 39).

49. The text reads “dadi.” As Forte points out, this is an abbreviation of the Gaozong Emperor’s posthumous title, “tianhuang dadi,” “Heavenly August Grand Emperor.” The title was conferred soon after his death. See Forte, “The Preface,” n. 41.

50. Foding zunsheng tuoluoni jing xu, T no. 967, vol. 19: 349b. For a translation and analysis of this preface and its historical context, see Forte (n.d.).

51. Chen, “Śarīra and Scepter,” 107.

52. Kaiyuan shijiao lu, T no. 2154.

53. Chen, “Śarīra and Scepter,” 111.

54. Ibid., 109. The matter rests on the portrayal of a monk named Buddhapālita found in the Xiuchan yaojue, XZJ 110: 834a. For a discussion of this text as a source for Tang history, see Forte, “The Preface,” 26–28.

55. Barrett, “Stūpa, Sūtra, and Sarīra,” 28. Though, as Barrett himself has noted, her engagement with Daoist practice, and her attempts to use it as another ground for her imperium, were significant.

56. On the connections between Wu, Wenshui, and Wutaishan, see Ibid., 18 (especially n. 38, where Barrett credits the observation to the scholar Du Doucheng), and Chen, “Śarīra and Scepter,” 109.

57. Barrett, “Stūpa, Sūtra, and Sarīra,” 18, citing Fayuan zhulin, T no. 2122, vol. 52: 393a and 596a.

58. Barrett “Stūpa, Sūtra, and Sarīra,” 20–22. Barrett has done much to demonstrate the tight connections between the Empress’s legitimation projects, permutations of the relic cult, and the rise of block printing. See Barrett, “The Rise and Spread of Printing,” and The Woman Who Discovered Printing, for very helpful summaries of this work. Chen, “Śarīra and Scepter,” also speaks to these matters.

59. Birnbaum, “The Manifestations of a Monastery,” 120. For the alternate characterization of Fazhao’s wuhui nianfo, see Weinstein, Buddhism Under the T’ang, 73, 175 n. 28. For further discussions of Fazhao, see Sasaki, Jōen Hōshō no jiseki ni tsuite, Tsukamoto, Nangaku Shōen den, and Robson, Power of Place, 301–302.

60. Birnbaum, “The Manifestations of a Monastery,” 120. For further discussions of the caves of Wutaishan, see Birnbaum, “Secret Halls of the Mountain Lords.”

61. Song gaoseng zhuan, T no. 2061, vol. 50: 717c.

62. Weinstein, Buddhism Under the T’ang, 81.

63. Daizong chaozeng sikong da bianzheng guangzhi sanzang heshang biaozhi ji. Yuanzhao (d. 778). T no. 2120, vol. 52: 837c–838a. Amoghavajra also requested that an image of the Bodhisattva Samantabhadra (Puxian pusa) be erected, and that the people be exempted from both their corvee and tax duties.

64. See chapter 4 for a discussion of Zanning’s account.

65. Weinstein, Buddhism Under the T’ang, 82.

66. Forte, “The Preface,” 1, quoting Chou, Tantrism in China, 322. The memorial is located in Daizong chaozeng sikong da bianzheng guangzhi sanzang heshang biaozhi ji, T no. 2120, vol. 52: 852c–853a.

67. Jiaju lingyan foding zunsheng tuoluoni jing, T no. 974c.

68. Chen Jinhua notes that Wu Che was a “fourth generation grandson of Wu Shirang, one of Empress Wu’s uncles” (Chen, “Śarīra and Scepter,” 110 n. 199).

69. Jiaju lingyan foding zunsheng tuoluoni jing, T no. 974c, vol. 19: 386a.

70. Foding zunsheng tuoluoni jing, T no. 967, vol. 19: 350a.

71. Anathapindika, whose name means “almsgiver to those without protection.” His given name was Sudatta (see the Sudatta Sutta and the Cullavaga 6.). The Jeta Grove was the monastery provided by him for the Buddha.

72. Foding zunsheng tuoluoni jing, T no. 967, vol. 19: 349c–352a.

73. Forte, “The Preface.”

74. Foding zunsheng tuoluoni jing, T no. 967, vol. 19: 350a. All citations of the sūtra, regardless of the specific translation, refer to the versions of the text found in volume 19 of T.

75. Snellgrove, Indo-Tibetan Buddhism, 52.

76. For this picture, see Sawada, Jigokuhen, and Teiser, The Scripture on the Ten Kings and Reinventing the Wheel.

77. This text, undated though probably from the ninth or tenth century, seems to be the archival script of a lecture given to the participants and audience of a confession and precepts ritual, though as with most works of its kind we have precious little contextual information with which to make a judgment. Modern scholars have entitled the text “A Piece on the Preaching of the Three Refuges and the Five Precepts” (Shuo sangui wujie wen); it is on the manuscript known today as S. 6551. I have also consulted the edition of the text found in Zhou, Dunhuang bianwen, 2: 1010–1032.

78. Zhou, Dunhuang bianwen, 1013, amends this to “is made to ache,” (quan suan). But this seems unnecessary.

79. That is, a Buddhist teacher. The term is derived from the Sanskrit kalyānamitra.

80. Here the text interpolates “call out the [names of ?] the Buddha’s sons” (cheng fozi). Perhaps at this point the names of those about to confess were ritually spoken. Such interpolations (and interpellations) are also found in the texts of dhāraṇīs, including all versions of the incantations of Glory and Wish Fulfillment (see chapter 2 and below).

81. That is, an instant. Emending ban to na, following Zhou, Dunhuang bianwen.

82. The original is in verse, with lines of seven syllables. I have not tried to reproduce any of its prosodic features in my translation.

83. Zhou, Dunhuang bianwen, 2: 1015.

84. Kroll, Dharma Bell and Dhāranī Pillar, 46 n. 41.

85. Ibid., 46 n.42, quoting Shiji jing section of the Chang ahan jing 20, T no. 1, vol. 1: 133c. We might also note that according to the Chinese translation of the Lokasthāna (Daloutan jing), produced by Fali in the third century, males and females of this heaven “perform the matters of yin and yang by employing winds” (T no. 23, vol. 1: 297b). See also Baochang’s Jinglü yixiang 1, T vol. 53: 1, for a similar account, drawn from the Sanfa dujing, a text that is possibly Chinese in its provenance. Note that this source also states that “in enacting their desires they are like humans” (xing yu ru ren), a statement that would seem to weaken the grounds for any assertion that pleasure and intimacy at a distance are so very unlike their more physical versions, at least as understood by Chinese audiences.

86. Wu, “On Rubbings,” 34.

87. Lamarre, Uncovering Heian Japan, 116. See the introduction for further discussion of these issues.

88. Liu Shufen, Miezui yu duwang, 46.

89. Zeitlin, “Disappearing Verses,” 76.

90. XB 760, v. 13: 8981. A somewhat different version of this text, not attributed to Lü Shou, is found in JSCB 66: 45, where it is called Tian Pu deng jingchuang.

91. We should be clear that they do not always employ such precise imagery. At times they simply state that the shadow or the dust touches one.

92. According to the Hanyu da cidian, zhan (1) has the wider range in modern usage, including the negative usages of “contagiousness” and “taint.”

94. Wenxuan 28. Ruan Ji (210–263) quotes this image in the Wenxuan version of the fifth of his “Song of My Cares” (Yonghuai), where he writes that “Clear dew covers the poolside lilacs; chill frost condenses on the wild grass” (ningshuang zhan yecao; Wenxuan 23). We might note in passing the juxtaposition here of “covering” and “infusing” (here rendered as “condensing”). This is the most common basic pairing in inscriptions related to the Incantation of Glory, signaling, perhaps, a deeper rhetorical or tropical structure in Chinese writings in general.

95. T no. 968, vol. 19: 354b.

97. See also the previous chapter, where I argue that the latest translation of the Mahāpratisarā dhāraṇī sūtra also reflected then-current practices and conceptions.

98. Fo mimi xin. I take xin here in the sense of “heart/essence,” rather than “mind,” because of the former’s close association with spells, as in the term hṛdaya.

99. Tihuang wanhui. The meaning of this phrase is not entirely clear to me. Tihuang means, literally, “stairs and boats,” and often is an image of journeying, over land and sea. I read it as indicating that the words of the spell offer access to all regions, along both vertical (stairs) and horizontal (boats) axes.

100. XB v. 20: 13587. Like Lü Shou’s composition, this lyrical evocation of the beauty and power of stone engraved with the “phrases of the Buddha’s Crown” is not a simple example of the picture I have presented so far. The complications it offers are salutary, however; attention to them helps to show the almost ironic position these accounts of the magic of the embodied spell had within other more-established patterns of Buddhist discourse. Composers of texts such as the one at hand were keenly aware of these ironies and the rhetorical conflicts they engendered are detectable in phrases such as those translated above. The practice and theory of dhāraṇīs fit for the most part within a larger discourse founded in rather strict ideas of purity and impurity. In many cases, and certainly in the scriptures of the spell, this discourse employs the figures of earthly filth versus the light of the Buddhas and their pure lands. “Dust” is one of the chief images of delusion and the impurities that hinder one’s advancement among the paths of rebirth. For the Chinese context, one need only invoke the famous dust that occludes the mind’s otherwise bright mirror in Shenxiu’s verse in the Platform Sūtra (leaving aside Huineng’s equally famous dust-denying reply, which would take us into discursive realms quite distant from both dhāraṇī practice and traditional karma talk). More broadly in the Chinese language as well, and in a precise echo of our terminology, the term dust drenched (zhanchen) can denote the state of being stained, or defiled. Yet here on the pillars dust is one of the two principal vectors of purification itself. Hence, perhaps, its ambivalent position in our inscription: whoever is lucky enough to be infused by the dust is thereby lightened of his load of dust—or, not quite—this isn’t homeopathy, after all. In the line in which the latter image occurs, it is not the dust that infuses but the wind that (elsewhere) bears it. The anonymous writer of these lines was careful to keep his discourses neat. We see the same tactic in other inscriptions as well, such as one from the Tang that proclaims that “the mightiest of dhāraṇīs is the Buddha’s Crown. Inscribe it on a pillar: “The flying dust [of delusion that causes] karmic-actions to amass: when the shadow turns, [these] calamities vanish” [XB v. 20: 13535]. Such small moments in the accounts of the dhāraṇī suggest that the tropes of the Incantation of Glory did not fit cleanly into the places made for them in the literate Buddhism of the Tang. In this the spell was not alone; the growth of a religious tradition does not only require rearrangements on the grand scale.

101. Reading 附近 for 俯近.

102. Baqiong 81: 24.

104. Za ahan jing, T no. 99, vol. 2: 304b.

105. Dasheng bensheng xindi guanjing, T no. 159, vol. 3: 304.

106. From Jingde chuandeng lu, The Jingde Record of the Transmission of the Flame, T no. 2076, vol. 51: 208.

107. From Fayuan zhulin, T no. 2122, vol. 53: 569b.

108. XB 19: 13532. During the Kaiyuan period (713–741) of the Tang. Also XB 19: 13546–13547, from 869 (Xiantong 10).

109. XB 19: 13533. (n.d.)

110. XB 19: 13531–13532. This pillar was originally constructed in 731, then rebuilt in 855, then rebuilt again in the Qianyou period (951–954) of the short-lived Northern Han Dynasty (951–979), whence this phrase.

111. Jiaju lingyan foding zunsheng tuoluoni ji, T no. 974c, vol. 19: 386.

112. Dasheng yujia jin’gang xinghai Manshushili qianbei qian, T no. 1177a.

113. That is, monasteries.

114. “Forest,” and by extension a secluded, quiet, abode for monastics.

115. Dasheng yujia jin’gang xinghai Manshushili qianbei qian, T no. 1177a, vol. 20: 726c.

116. Qianshou qianyan Guanshiyin pusa guangda yuanman wu’ai dabeixin tuoluoni jing, T no. 1060.

117. That is, three of the four possible forms of birth described in Buddhist cosmology.

118. Qianshou qianyan Guanshiyin pusa guangda yuanman wu’ai dabeixin tuoluoni jing, T no. 1060, vol. 20: 109a.

119. A fact that provides further evidence that this text was steeped in the Buddhist tale literature of its age.

120. Da fangguang fo huayan jing ganying zhuan, T no. 2074, vol. 51: 175a.

121. The latter work claims the tale is taken from an otherwise unknown text called the Jing tianji youji (Sanbao ganying yaolüe lu, T no. 2084, vol. 51: 837b), the former simply from a “record of the Western Regions” (Da fangguang fo huayan jing ganying zhuan, T no. 2074, vol. 51: 175a). Fazang’s work simply states that the tale was “transmitted from western countries” (xiguo xiangchuan).

122. Huayan jing chuanji, T no. 2073, vol. 51: 169c.

123. It should be noted that this is true even though the tale here functions mainly as the setup for a claim for the true way to enact the spell as prescribed in this text: verbally.

124. The term here is chi. See the introduction for a discussion of its meanings and centrality in the discourse of incantation practice.

125. The terms used here closely parallel those used in the tales. For example, this text reads (yi shui guan zhang), while Fazang’s and Feizhuo’s texts both read (yishui guan zhang). The words for “wash” here are homophones. Similarly, the texts all use the term zhansa to describe the drenching of the ant, a binome that occurs rather rarely in canonical Buddhist texts (eleven times, according to an electronic search). While this is certainly faint evidence for a connection, the possibility that the Divākara text was influenced by Fazang’s seems worth thinking about, and is made more likely by the fact that Divākara and Fazang worked together as translators. See, for example, Fazang’s Huayan jing tanxuan ji, T no. 1733, vol. 35: 111c.

126. Zuisheng foding tuoluoni jingchu yezhang zhoujing, T no. 970, vol. 19: 360b.

127. Comparison with worlds far removed in time and place from the one under examination is probably rarely a helpful move in a strictly historical study, but here one might be forgiven for wanting to remind the reader of a modern work, if only for the pleasure it gives. William Butler Yeats’s “Sailing to Byzantium” might echo something here: “ …Consume my heart away; sick with desire/And fastened to a dying animal/It knows not what it is; and gather me/Into the artifice of eternity.//Once out of nature I shall never take/My bodily form from any natural thing,/But such a form as Grecian goldsmiths make/Of hammered gold and gold enameling. …” In considering any parallels here, however, one would do well to keep in mind that Yeats himself reportedly sought enhanced life not through perfect artificial substance but through the grafting of a monkey’s thyroid gland onto his body.

128. T no. 970, vol. 10: 361a.

130. For an extended meditation on this and related themes, see Cole, Mothers and Sons in Chinese Buddhism.

131. See the discussion of this text in the introduction.

132. Zhou, Dunhuang bianwen 1: 273–333.

133. For an illuminating discussion of this exegetical tactic in the context of the Chaṇdogya Upaniṣad, see Lincoln, “How to Read a Religious Text,” 130ff.

134. As will soon be apparent, these quotations from the text often differ from the received edition of the Da zhidu lun. The differences seem sometimes to be simply scribal errors, while at others they reflect substantive variations.

135. The corresponding passage in the received edition of the Treatise reads, “Contemplate the five impure marks of the body/person,” T 25: 198c.

136. The text presents the five impurities in an order different from that of the received text, in which the first impurity is “the impurity of the birth location.” “The impurity of the seed” is the second in that text.

137. Or, karma and kleśa.

138. The line in the corresponding received text reads, perhaps, “it is not a remnant of a marvelous treasure.” T vol. 25: 199a.

139. The received text reads here, “it is not from pure whiteness born, but issues from the urinary tract” (p. 199a). We might also note in passing that the fact that, according to legend, the Buddha was born out of his mother’s side when he “descended to birth.” See S. 2440 (Zhou, Dunhuang bianwen, 1056).

140. The received version reads “birth location” (p. 198c).

141. The received version reads, “This body is reeking and filthy, it was not born amid blossoms” (p. 198c).

142. The Yiqie jing yinyi describes the Campaka tree as follows: “it is tall with large and exceedingly fragrant flowers whose aroma travels far on the wind” (T no. 2128, 54: 363b). There are many different baoshans spoken of in Buddhist texts; it is unclear if one specific mountain (or range) is intended here. The Taisho edition reads, “Indeed not from Zhampu, nor from a Jewel Mountain” (p. 199a).

143. Morris and Hardy, eds., The Alguttara Nikāya, 4: 377; discussed in Wilson, Charming Cadavers, 50–51.

144. T vol. 19: 353b.

145. T vol. 19: 350a.

146. Miaofa lianhua jing, T no. 262, vol. 9: 53b.

147. Benn, Burning for the Buddha.

148. Quoted (with permission) from the draft version of his paper “Self-cultivation and self-immolation: preparing the body for auto-cremation in Chinese Buddhism.”

149. The text is unclear on this point. But while the prose sections are simply commentary on the Lotus Sūtra text, the portions in verse seem to be directed toward the audience in more personal ways.

150. That is, chang.

151. Zhou, Dunhuang bianwen, 231–232.

152. Most notably in the Song gaoseng zhuan (p. 871c) and the Da fangguang fo Huayanjing ganying zhuan (p. 176c), on which see below.

153. Huayan jing chuanji, T no. 2073, vol. 51: 167a.

154. Mrozick says that the only extant Sanskrit manuscript of the text, however, is dated to the thirteenth or fourteenth century. According to Hōbōgirin, the Chinese translation (Dasheng ji pusa xuelun; T no. 1636) dates to the early eleventh century.

155. Mrozick, “The Relationship Between Morality and the Body,” 179. I have removed her parenthetical notations of the Sanskrit for the terms of this sentence, as well as the one quoted below.

156. Ibid., 179–180. The Compendium is here quoting the Tathāgataguhya sūtra.

158. Campany, To Live as Long as Heaven and Earth, 21–22.

159. The difficulties of translating this term are well known and I think at this point moot, given its wide acceptance in contemporary usage. Campany, though he seems to mostly agree with this position, follows Bokenkamp’s usage in mainly translating it as “pnuema,” a term that echoes qi’s primary sense of “breath.” His capsule description of the term is helpful: “for religious thought and practice, the significance of qi is that all things are made of it, exist in it, and share it; heaven and earth, gods, humans, lesser spirits, animals, plants, minerals—all are consubstantial, despite the great range of qualities exhibited by qi in these various forms, and this consubstantiality provides a kind of ladder connecting all levels of being, a ladder that could be climbed by systematically working on and transforming the qi constituting oneself” (Ibid., 18–19).

160. Ibid., 21–22.

162. T no. 2074, vol. 51: 174b-c. In another version, the boy takes the dirt from under his fingernails and gives it to the monk to take as medicine (T no. 2073, vol. 51: 165c).

163. T no. 967, vol. 19: 351a-b.

164. Owen, “A Monologue of the Senses,” 245.

165. T no. 1803, vol. 39: 1028a–1033c. It appears in Shūei’s (in China, 862–866) catalog, Shinshosha shōraihōmon tō mokuroku, Newly Copied Buddhist Texts Requested and Brought [From China]; T no 2174a, vol. 55: 1110b), where it is called simply Foding zunsheng tuoluoni jing shu, Commentary on the Scripture of the Glorious Dhāraṇī of the Buddha’s Crown, an alternate title that appears in the body of the received version of Fachong’s text. The oldest extant manuscript of Shūei’s work dates from the Kamakura period (1185–1333), but elements from it were cited in texts discovered at Dunhuang. For a fuller exploration of this text, see Copp, “Anointing Phrases and Narrative Power: A Tang Buddhist Poetics of Efficacy.”

166. At least three glossaries of the words that make up the Incantation of Glory remain among our sources. Two were apparently produced within Amoghavajra’s (Bukong 705–774) Esoteric Buddhist circle in the eighth century; an anonymously produced third is contained in the Beijing collection of Dunhuang manuscripts (B. 7323 [Shuang 13], in DHBZ 105: 466a–471b). The two glossaries from Amoghavajra’s circle appear in works whose existence in late medieval China is evidenced mainly by their presence in ninth-century Japanese catalogs of works brought back to Japan from China: Annen’s (fl. 884) Sho ajari shingon mikkyō burui sōroku, Comprehensive Catalog of the Shingon [i.e., “real word”] Esoteric Teachings of All the Ācāryas, T no. 2176, vol. 55., and Shūei’s (in China, 862–866) Shinshosha shōraihōmon tō mokuroku, Newly Copied Buddhist Texts Requested and Brought [From China]. The first set of glosses, which circulated independently, was called Foding zunsheng tuoluoni zhuyi, Commentary on the Meaning of the Glorious Dhāraṇī of the Buddha’s Crown; T no. 974d. Its authorship is attributed to Amoghavajra. The second, which I will focus on here, is much larger and offers more details for analysis. Other glosses on other spells also exist, two of which I will discuss below.

167. On the status of the Qianfusi among its peers, see Xiong, Sui-Tang Chang’an, 265.

168. On the Amoghavajra version of the Renwang jing (T no. 246) see Daizong chaozeng sikong da bianzheng guangzhi sanzang heshang biaozhi ji, T no. 2120, vol. 52: 831b; Weinstein, Buddhism Under the T’ang, 78; and Orzech, Politics and Transcendent Wisdom. On his version of the Da xukong pusa suowen jing (T no. 404) see Zhenyuan xinding shijiao mulu, T no. 2157, vol. 55: 888a.

169. Daizong chaozeng sikong da bianzheng guangzhi sanzang heshang biaozhi ji, T no. 2120, vol. 52: 834c.

170. It is important to note that, though the spell is treated the way it is in this text at least in part because it is a dhāraṇī in a sūtra, a part of a narrative, the glosses were appended to versions of the spell that circulated free of the sutra as well. The meaningfulness of the spell was thus not simply a result of its enclosure in a narrative.

171. Most of these categories seem to be of Fachong’s own coinage, though a few of them occur elsewhere. “Clearing away the bad trajectories,” for example, is an explanatory rubric in Bodhiruci’s translation of the “Assembly of Bodhisattvas of Inexhaustible Wisdom,” collected within the Ratnakūṭa sūtra, though it is unclear what, if any, connections might obtain between these two texts (Da baoji jing, T no. 310, vol. 11: 648c). It occurs as the fifth in a series of ten elements that constitute the bodhisattva’s perfection of ethics, or śila (jie).

172. I have inserted the numbers of the explanatory headings for reference and treated the text within each heading as if it were a sentence. At times Fachong provides alternate glosses on a term; I have included them in brackets. In addition, I have included the Sanskrit of those translated terms, such as oṃ, or abhiṣeka, I deem especially common in Buddhological usage.

173. It should be emphasized that, though, as I have presented it here, this text might appear to be a performance script, a prayer or invocation, this appearance is an illusion. It was not intended for performance; indeed, the shape I have given it is artificial. The “text” is the product of Fachong’s interpretive analysis of the dhāraṇī—the actual performance piece. Monks did not chant this translation, in other words, but would have read it (or heard it expounded) for the understanding it provided of the dhāraṇī and its practices. This is a crucial, if obvious, distinction to keep in mind, and one that Fachong himself is at pains to make in his text.

174. See Tanabe, Myōe the Dreamkeeper, 138ff, for a discussion of the thirteenth-century Japanese monk Myōe’s use of this imagery in his explanation of the power of the kōmyō shingon, a spell whose practice in medieval Japan (as noted earlier) bore striking resemblances to the much earlier enactments of the Incantation of Glory. Myōe described the workings of the mantra’s power through analogy with a scene in the Huayan jing where “all the buddhas … spoke into their right hands and rubbed Samantabhadra’s [the figure who, empowered in this way, speaks for the Buddha in the text] head, thus transferring to him their virtue through their hand-held words” (138).

175. See, for example, Kapstein, “Scholastic Buddhism and the Mantrayāna.”

176. T vol. 39: 1029b.

177. According to Yixing in his commentary on the Mahāvairocana sūtra, xiang in this phrase does not denote cognitive activity but “that which is distinguished [from something else],” i.e., in a process homologous to the productions of thought. He goes on to explain the term by noting that all the different mantras (or real words) derive from one single syllable (a), which is to make a distinction between the undifferentiated (the one “one-syllable mantra”) and the differentiated (the countless multisyllabic spells). Da piluzhena chengfo jingshu, T no. 1796, vol. 39: 775b–776a. The phrase seems to derive from the Mahāvairocana sūtra.

178. It is unclear if “held” (zhi) simply means “made by the hands,” or if it refers to ritual stamps that one holds in ones hands. The verb is employed to describe both activities in Buddhist texts.

179. T vol. 39: 1030a. The passage goes on to detail practices involving the four directions.

180. That is, the affective and cognitive hindrances to awakening; on which see canonical discussions in the Yogacārabhūmmi (T no. 1579, vol. 30: 327a and 345b) and the Cheng weishi lun (T no. 1585, vol. 31: 1a).

181. T vol. 39: 1029c. Note the close association here between the empowering luminosity of the spell and the practice of anointing the dead with enchanted dirt, discussed early in this chapter.

182. Shaiji also at times renders the Sanskrit “sukha,” or “joy.”

183. My translation of this line is tentative. It reads zhiqian jiekong. I take the first two characters to be shorthand for phrases such as “pozhi qianmi,” as found (for example) in the Mohe zhiguan (T no. 1911, vol. 46: 68b).

184. References to “sweet” or “refreshing” rains have an ancient lineage in Chinese literature, stretching at least back to the Shijing. The poem “Futian” from the Xiao Ya section reads “in offering to the ancestor of the fields, in praying for refreshing rains.” The Chunqiu zhengyi clarifies the meaning of “refreshing rains” in this poem: “The Shijing states: ‘In praying for refreshing rains.’ Here [the text] mentions ‘distressing [literally, ‘bitter’] rains.’ There is only one kind of rain, and there is no difference between them in terms of tasting ‘bitter’ or ‘sweet.’ If it nurtures [the myriad things], it is ‘refreshing’; if it harms [the myriad things], it is ‘distressing.’” (Chunqiu Zuo zhuan zheng yi 42.1379). My thanks to Alexei Ditter for providing these references. The translations here are his (Ditter, “Genre and the Transformation of Writing in Tang Dynasty China,” 168 n51).

185. T vol. 39: 1030b.

186. That is, the “wisdom that is like a great spherical mirror” (da yuanjing zhi), the “wisdom [that sees] all things as equal in nature” (pingdeng [xingzhi]), the “wisdom of subtle observation” (miao guan [cha zhi]), and the “wisdom of the completion of deeds” (cheng suozuo zhi). See Cheng weishi lun, T no. 1585, vol. 31: 56a.

187. T vol. 39: 1033a.

188. See chapter 1 for an exploration of the ritual relationships seen to obtain between seals and spells.

189. T vol. 39: 1030c.

190. As I explore in a separate article on these glosses (“Anointing Phrases and Narrative Power”), the complications it lends what seems on the surface a rather simple text are not limited to narrative structure. Fachong presents a split-screen picture of the action of the spell, describing on the one hand the ways that any speaking of it universally transforms (almost) all beings of the cosmos and, on the other, reiterating a particular practice of the spell. This turns out to be characteristic of his method. His text is simultaneously a narrative of the spell’s unfolding, a guide to the ritual implementation and practices of the spell, and a study book of its philosophical nuances.

191. Or (see below), “momo.” Both appear to be transliterations of the Sanskrit enclitic me, which acts as a first person pronoun.

192. Each of the canonical versions of the sūtra contains a version of this instruction. In the Buddhapālita version it is contained in both versions of the spell, the first is located at T no. 967, vol. 19: 350c, where it reads simply “call out a name” (cheng ming); the second is at T no. 967, vol. 19: 352b. It reads “the one who receives and keeps should himself chant his name at this point.” The instructions occur twice in the Du Xingyi version. In both instances the text reads “meme is to say: so and so, the one who receives and keeps [the spell], should at this point chant out his name” (T no. 968, vol. 19: 353c). The first Divākara version reads simply “call out your own name” (T no. 969, vol. 19: 356b), while the second seems to be the locus classicus for the many manuscripts and inscriptions that have momo instead of meme. Its explanatory note reads, “call out a name instead of these characters [that is, momo]” (T no. 970, vol. 19: 359b). The Yijing version, finally, reads, “Chant your name yourself: ‘I so and so’” (T no. 971, vol. 19: 362c). Similar instructions are fairly common in dhāraṇīs from at least the seventh century, though not all spells have this component. See, for example, the eleventh “great spell of the horse-headed bodhisattva Guanshiyin” (Matou Guanshiyin pusa dazhou), collected in the Foshuo tuoluoni jing (T no. 901, vol. 18: 835b), compiled by Atikūṭa (var. Atigupta; Adijuduo) (fl. 653–4); and the second “shouted seal” (huanyin, a genre of spells that seems to have been limited to Bodhiruci’s translations), the “Yiqie dinglunwang tongqing huanyin,” contained in the Yizi foding lunwang jing (T no. 951, vol. 19: 257a), compiled by Bodhiruci (Putiliuzhi), active in China, 693–727; d. 727).