In lieu of an abstract, here is a brief excerpt of the content:

Tang Studies 6 (1988) A Note on Nightmare Magic in Ancient and Medieval China DONALD HARPER BOWDOIN COLLEGE Several years ago I made a study of the "Meng fu" ~ JtJB: (Dreamfu) by Wang Yen-shou ::E~. (mid-second century A.D.) and of the relationship of his composition to exorcistic incantations used to alleviate nightmares. Since publishing "Wang Yen-shou's Nightmare Poem" (HJAS 47.1 [1987], 23983 ), I have come across a passage concerning nightmares in the Tun-huang manuscript Pai tse ching kuai t'u 81mm~!iI (White Marsh's Diagrams of Spectral Prodigies) that both sheds additional light on the forms of incantatory magic used against nightmares and contributes to our understanding of the transmission of occult lore from ancient to medieval times. The following note is intended to supplement observations already made in the above article and in an earlier article ("A Chinese Demonography of the Third Century B.C.," HJAS 45.2 [1985], 459-98) on the oldest extant example of Chinese demonographic literature (which occurs on one of two occult manuscripts recovered from the Ch'in tomb at Shui-hu-ti M! fJr. Jm, Hupei). The Tun-huangPai tse ching kuai t'u is an incomplete version of a previously lost work which is first mentioned by the title Pai tse t 'u in the Pao p 'u tzu 1@! l~ T (the Pao p 'u tzu also alludes to the legend of the creature White Marsh whose knowledge of the demon world was recorded by the Yellow God [Huang ti lit *] to create the Pai tse t'u). Two Ch'ing reconstructions of the Pai tse t'u bring together fragments preserved in encyclopedias and other received texts, but the Tun-huang manuscript provides us with an example of the book as it looked in T'ang times, complete with illustrations of some of the demonic entities described in the text.! 1 See Harper, "A Chinese Demonography," 490-98, f~r source citations and bibliographic information relative to the Pai tse t'u (and a discussion of the relationship of the Pai tse t'u to the Shui-hu-ti demonographic text). The reconstructions of the Pa; tse t'u 69 Harper: Nightmare Magic The Pai tse t'u and other demonographic works like it were guides for identifying demons and demonically caused events, and they detailed magical techniques for dealing with the demon world. Comparison of the Tun-huang manuscript with the two reconstructions of the Pai-tse t'u reveals parallels in content, and yet it is also clear that the manuscript does not correspond exactly to the one or several editions of a Pai tse t'u represented by the fragments collected in the reconstructions.2 Deletions, accretions , and other alterations are inevitable in the transmission of are in Ma Kuo-han .~ II ~, Ya han shan fang chi i shu :E m L1JJjj • f*ff; and Hung 1-hsilan m 6 m,Ching tien chi lin ~ A #; #. The main portion of the Tun-huang Pai tse ching kuai t'u is Pelliot 2682. Jao Tsung-i • *6, "Pa Tun-huang pen Pai tse ching kuai t'u liang ts'an chilan" j£ t£ 1!:ao shu" E.~ elt 1&;<$: 85M ff III .& It 1&=+~;lj i£, Tung Wu wenshih hsaeh-pao Jt! ~ Jt ~ !l:lII2 (1977), 97-102 (I am grateful to Dr. Carol Morgan for bringing the latter article to my attention). Beliefs related to the legend of the spirit Pai tse are of interest in themselves. 1 am currently conducting research on the apparent fusion ofthe name Pai tse with the classical name for the giant panda, rna ., which enhanced the image of the panda as a magical creature in Tang times. 2 See Jao, Opecit., and Lin, Opecit., for comparison of the Tun-huang manuscript with the reconstructions. 70 Tang Studies 6 (1988) any book. In the case of the Pai tse t'u, however, the discrepancy between the Tun-huang manuscript and the reconstructions is also a reflection of the nature of occult knowledge and the transmission of occult manuscripts. Since Warring States times numerous manuscripts containing occult knowledge (many not recorded in official bibliographies) circulated in private channels, and their contents were extracted and recombined to form new compilations. It is likely that the first Pai tse t'u was itself a compilation of pre-Han and Han demono-magicallore drawn from manuscripts that were subsequently lost. 3 The central fact ofthePai tset'u during the SixDynasties and T'angwas its nature as a repository for demono-magicallore. Besides the Tun-huang manuscript and the several versions recorded in official bibliographies , there were undoubtedly other demonographic manuscripts that laid claim to the prestige of the ur-Pai tset'u. We are dealing, then, not with the textual history of a single document in medieval China, but with the identification of a sub-class of demonographic literature ascribed to White Marsh. 4 3 The proliferation of occult literature in the late Warring States and Ch'in-Han periods paralleled the popularity of the fang shih 11 ± (masters of recipes). Occult literature represented the technical or "recipe" (fang) knowledge of such specialists. Very little of this literature survived into the post-Han period intact. However, as can be seen by comparing newly discovered manuscripts like those from the Shui-hu-ti tomb with Six Dynasties and Tang occult literature, there is a direct correlation between their contents. This is also true for the sexual cultivation literature that had been in circulation since late Warring States times (see Harper, "The Sexual Arts of Ancient China as Described in a Manuscript of the Second Century B.C.," HJAS 47.2 [1987], 539-93). For further discussion of Ch'in-Han occult literature in the light of recent manuscript discoveries, see Harper, "Oin-Han Scientific and Occult Manuscripts," forthcoming in Paleographic Sources of Early China, ed. Edward L. Shaughnessy. 4 See lao, op. cit., pp. 539-40, for a summary account of the several editions of the Pai tse I'u that came from the hands of different compilers. On pp. 541-42, parallels are noted between the Tun-huang Pai tse ching kuai I'U and another lost book, the Pai tse Ii ching 131$ ~ Si (White Marsh's Mirror of the Earth). lao thinks the parallels indicate that the Tun-huang manuscript took these passages from the Pai tse Ii ching. It is equally probable that the passages in question were part of a "Pai tse literature" that was disseminated in a number of versions, so that determining the original source becomes a moot issue. 71 Harper: Nightmare Magic Recently discovered Ch'in and Han manuscripts containing demono-magical material (in particular the two occult manuscripts from Shui-hu-ti and several manuscripts from the Former Han tomb at Ma-wang-tui, Hunan) document the antiquity of some of the material contained in medieval manuscripts like the Tun-huang Pai tse ching kuai t'u. If occult books were lost with regularity, these Ch'in-Han manuscripts reveal that ancient occult knowledge enjoyed a surprising continuity of transmission in successive compilations down the centuries. The passage concerning nightmares in the Pai tse ching kuai t'u is a case in point: the closest textual parallel occurs in twoversions in the Shui-hu-ti occult manuscripts. First the Pai tse ching kuai t'u passage:5 When a person has foul dreams at night, rise at dawn, and in the northeast part of the house unbind the hair and chant this incantation: "Po-ch'i, Po-ch'i. He does not drink wine or eat meat, and regularly eats from the land of High Elation. May these foul dreams return home to Po-ch'i. Crushing dreams abate, give rise to great blessings." Chant the incantation like this seven times and there will not be spirit odium. A ~ ~~~\~ Et U iJrIt !~t,;f:i~. Q9u g 1(;)~1E~ ~1.~ ;~i t 1:1.~~~~ ~,~.~ l~ ~~ ~9~ ~: ~~. ~ -K~~ ~9 ~t, -to o7v ~, ~ ~ Next the more complete of the two versions from the Shuihu -ti occult manuscripts:6 5 Pelliot 2682, Tun-huang pao ts'ang, vol. 123, p. 290, bottom. The nightmare passage occurs in a long section of text without illustrations near the end of the manuscript. My transcription retains the orthography of the original manuscript, which does not contain any unattested graphs (e.g., 1\ for ~; see Matsumoto, op. cit., p. 147, no. 61, for a punctuated transcription). 6 YUn-meng Shui-hu-ti Ch'in mu ~.y lit 1%jt!!, ~ Ri (Peking: Wen-wu, 1981), plate 131, slips 883-82, reverse side. This is the version in the first occult manuscript (MS A); the second version is in the second occult manuscript (MS B), YUn-meng Shui-hu-ti Ch'in mu, plate 160, slips 1089-90. See Harper, "A Chinese Demonography," 462-70, for 72 T'ang Studies 6 (1988) When a person has foul dreams, on wakening then unbind the hair, sit facing the northwest, and recite this prayer: "*K'og. I dare to declare you to Ch'in-ch'i. So-and-so has had foul dreams. Flee back home to the place of Ch'in-ch'i. Ch'in-ch'i, drink heartily, eat heartily. Grant so-and-so great broadcloth. If not coins, then cloth. If not cocoons, then silkstuff." Then it will stop. A~ ~.t ~/3 ~f I,1. ~;I~!fJ §t 4~~ L EJ ~ ..~~ ~b ~ ~1-!~ jt. tk- i~ ~~:t-ttt..- fir ~1i% g~ i~ s£ ~ jl; ~j', .p~ ='~if p, ~ ~~~ JZJ ~~ ~'j ~ ~ a -i!.={t b ~ =\~ Obviously the two passages are not identical. Yet both stipulate the same type of ritual action preceding the nightmare incantation (only differing over a northeast orientation versus the northwest); and the Pai tse ching kuai t'u incantation is unmistakably a later version of the same incantation in the Shui-hu-ti manuscript? In terms of content, the most significant connection information on the two bamboo-slip occult manuscripts; and Harper, "Wang Yen-shou's Nightmare Poem," 270-71, for further details concerning the nightmare passage and the language of the incantation. In transcribing the text I have glossed only those graphs that are not attested in the received literature with the meaning they have in the Shui-hu-ti text (see my comments on transcribing Ch'in-Han manuscript~ in"Oin-Han Scientificand Occult Manuscripts"). 7 Thematically, both incantations conjure the dream controller, instruct the nightmare demons to return home (!alei M) to their overlord-and-jailor (Po-ch'i/Ch'in-ch'i), and request a reversal of the illfortune associated with the nightmare to create good fortune in its place. I am not certain of the significance of the "land of High Elation" in the first incantation. Perhaps it refers to an ultimate land of bliss in popular belief. Apparently Po-ch'i's divine power derives from his association with this land, and the dream controller does not require food sacrifices. In the Shui-hu-ti incantation, the "prayer" (tao tl1) is specificallya type of prayer in which offerings are promised to a spirit if it performs the requested deed (after which the offerings must be delivered in sacrifice); hence the exhortation to Ch'in-ch'i to "drink heartily, eat heartily." And the request for good fortune takes a material form in coins, cloth, and silk. 73 Harper: Niglrl1nare Magic between the two incantations is the spirit conjured to control the nightmare demons: Po-ch'i {a ~ in the Tun-huang manuscript and Ch'in-ch'i ~~f in the Shui-hu-ti manuscript. Prior to reading the Tun-huang manuscript the only occurrence of a spirit Po-ch'i known to me was in the curse recorded in the Hou Han shu ~ 71. account of the No it exorcism performed at court at New Year, in which Po-ch'i is ordered to "eat dreams." Even though Po-ch'i is conjured there in a list of twelve fierce spirits, each one charged with the elimination of a specific spectral evil, the No curse alone would allow one to speculate on the existence of an independent belief in a "dream controller" to whom any individual might have appealed in time of need; and in "Wang Yen-shou's Nightmare Poem" I proposed identifying Poch 'i with the Shui-hu-ti Ch'in-ch'i. The Shui-hu-ti incantation thus represented solid evidence for a popular belief in a spirit whose specialty was neutralizin~ nightmare demons in the ancient magico-religious tradition. To encounter a medieval nightmare incantation of the same type as the Shui-hu-ti incantation, but in which the spirit is actually identified as Po-ch'i I think proves the supposition. I should emphasize the significance of the existence of two incantation texts, one pre-Han and the other T'ang, that derive from popular magico-religious tradition. A genetic link between them is established as much by the style and thematic structure of the two incantations as by the occurrence of the name Po-ch'i. Before 8 Harper, "Wang Yen-shou's Nightmare Poem," 271 (the second version of the incantation writes YOan-chi ~ ~ instead of Ch'in-ch'i). Jao Tsung-i and Tseng Hsient 'ung US ~ Ji!i, Yun-meng Ch'in chien jih shu yen-chiu ~ .1ifliiE .liJf 1l (Hong Kong: Chinese Univ. Press, 1982), p. 28, argue for the same identification. There is another Po-ch'i in popular lore, the son of the Chou minister Yin Chifu P a 1ft • Po-ch'i was wrongfully executed because of his stepmother's slander, and metamorphosed into a bird (the bull-headed shrike) which thenceforth bore the name Po-lao fa * ("Po's tribulation"). The story and the popular beliefs associated with the bull-headed shrike are presented in an essay by Ts'ao Chih .•.tit (192-232), "Ling ch'in o niao lun" ~ *";'ffi.~ », Ts'ao chi ch'uan p'ing .•.~ ~ Wf (Peking: Wen-hsOeh ku-chi k'an-hsing she, 1957).This folktradition isnot connected to the dream controller Po-ch'i. 74 T'ang Studies 6 (1988) the discovery of the Shui-hu-ti manuscripts it might have been surmised that the Tun-huang incantation was an example of borrowing the name of a spirit from a classical precedent, in this case from the Hou Han shu No curse which was widely known in medieval times. But the Shui-hu-ti incantation documents a popular tradition, pre-existing the No curse, as the source of the later incantation. The occurrence of the name Po-ch'i in the Hou Han shu No curse probably reflects the fact that the popular deity was already known by that form of the name during the Han. I would argue further than the Tun-huang incantation bears witness to a version of the nightmare incantation that was recorded in Han occult literature, and was subsequently included in various editions of the Pai tse t'u during the SixDynasties. 9 The relationship between the Shui-hu-ti occult manuscripts and the Tun-huangPai tse ching kuai t'u isnot limited to this single example. With the discovery of a number of occult manuscripts in Ch'in and Han tombs it is now possible to delineate the early development of magico-religious practices and their transmission into medieval times using original sources - contemporary manuscripts of occult writings - rather than the fragmentary and often hostile testimony that has survived in the received literature . In carrying out this scholarly investigation it is important to focus attention on the Tun-huang and Ch'in-Han occult manuscripts as a distinctive body of literature in its own right. It is certainly worthwhile to use these manuscripts to trace the sources of certain ideas in books like the Huai nan tzu 7ft Wi T for the Han period or the Pao p 'u tzu for the Six Dynasties, yet by treating the occult manuscripts as adjuncts to the "higher" intellectual tradition as it has been handed down in the received literature one risks overlooking their unique value for constructing a broad picture of the contemporary spiritual and intellectual scene. The manuscripts can reaffirm the received record, but they 9 For a similar argument with respect to the Han origins of certain SixDynasties Taoist nightmare incantations, see "Wang Yen-shou's Nightmare Poem," 272-76. 75 Harper: Nightmare Magic can also bring it into question with evidence that runs counter to conventional truths based on received sources. In this respect the Tun-huang manuscript Pai tse ching kuai t'u is also distinct fromPai tse t'u reconstructions based on extant fragments in the received literature. The composers of medieval encyclopedias or commentaries had a variety of reasons for selecting quotations from the Pai tse t'u that had little to do with a concern for the careful preservation of the Pai tse t'u itself. Since one aspect of the Pai tse t'u is its nature as a demon onomasticon, it is typically cited as authority for demon proper names, with the result that a large part of the magical lore recorded in the medieval Pai tse t'u, among which the Po-ch'i nightmare incantation is an example, was lost. A reconstruction of the Pai tse t'u based on the extant fragments cannot get around this tacit editing of the original text. The Tun-huang manuscript, while incomplete , is a genuine example of medieval demonographic literature . The seemingly loose organization of demonological material in the manuscript, and the variety of apparently everyday events that are treated, are both evident in the contents of Ch'inHan occult manuscripts as well. If the received fragments of demonographic literature might be construed as sometimes extravagant fantasies that need not be seriously considered as part of the fabric of spiritual and intellectual life, the Ch'in-Han and Tun-huang manuscripts indicate otherwise; which is all the more reason for acknowledging the special significance of these occult manuscripts in studying ancient and medieval Chinese culture. 76 ...


Additional Information

Print ISSN
pp. 69-76
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.