1. The original inspiration for this article was a manuscript by Jao Tsung-i entitled “New Light on wu.” A revised version of Professor Jao's paper will appear in the next issue of Early China. It will treat in more detail many of the larger paleographie and religiocultural issues concerning wu that are only touched upon in the present work, the primary purpose of which is to to discuss the philological and linguistic implications of the Old Sinitic reconstruction of wu. In the meantime, Professor Jao has published a brief but important article on the West Asian connections of in which he suggests the possibility that such symbols may have influenced the origins of writing in China. See his “Ssu-ch'ou chih lu yin-ch'i te'wen-tzu ch'i-yüan'ti wen-t'i [The Question of ‘The Origins of Writing’ Brought about by the Silk Road]” 絲綢之路引起的「文字起源」問題, Ming-pao yüeh-kan 明報月刊 25.9 (09, 1990), 47–50Google Scholar, an English version of which is forthcoming in Sino-Platonic Papers.

Modern Standard Mandarin transcriptions are in Wade-Giles romanization. For Old Sinitic I follow chiefly the tentative transcriptions of Li Fang-kuei and Axel Schuessler given in the latter's A Dictionary of Early Zhou Chinese (Honolulu: University of Hawaii Press, 1987)Google Scholar, but also take into account those of Bernhard Karlgren, Tung Tung-ho, and Chou Fa-kao given in the latter's A Pronouncing Dictionary of Chinese Characters in Archaic and Ancient Chinese, Mandarin and Cantonese (Hong Kong: The Chinese University of Hong Kong, 1973)Google Scholar. While Li reconstructs a final voiced velar stop for *m^ᵞag, Schuessler omits it. For the significance of this discrepancy, see note 42.

I wish to thank my colleague Donald A. Ringe, a specialist on Indo-European historical linguistics, for helping me with the reconstruction of the word for” wheel[ed vehicle].” Participation in the National Endowment for the Humanities institute, “Perspectives on the Ancient Indo-European World,” held at the University of Texas (Austin) from June 18 to 07 20, 1990Google Scholar, served to sharpen my understanding of many of the issues raised in this article. I am grateful to Edith Porada, Boris Marshack, Irene Winter, Renata Holod, Judith Lerner, G. Azarpay, Parviz Varjavand, and the other authorities on Middle Eastern art and archeology who responded to my requests for assistance in identifying the headgear of the Chou-yüan figures. However, I alone am responsible for all interpretations of the material herein.

2. Sheng-p'ing, Yin 尹盛平, “Hsi-Chou pang-tiao jen-t'ou hsiang chung-tsu t'an-so” 西周蚌雕人頭像族探索 [Investigation of the Racial Affinity of Two Western Chou Human Heads Sculpted of Shell], Wen-wu 文物 [Cultural Relics], 1986.1, 46–49Google Scholar; Ch'üan-fang, Ch'en 陳全方, Chou-yüan yü Chou wen-hua 周原與周文化 [Chou-yüan and Chou Culture] (Shanghai: Shanghai jen-min ch'u-pan-she, 1988), p. 20Google Scholar.

3. The flattened design was presumably intended to enable the pins to lie smoothly along the surface of the user's hair, but is perhaps merely a function of the limitations of the material from which they were made. The shape of the mollusk shell may also help to account for the truncated tops of the heads.

4. The drawings of these pieces supplied by Yin Sheng-p'ing and Ch'en Ch'üan-fang (see note 2) show no trace of ears. The available published photographs of the two pieces are insufficiently distinct to determine with certitude whether ears are present or not.

5. Yin Sheng-p'ing, “Hsi-Chou pang-tiao jen-t'ou.”

6. Minns, Ellis H., Scythians and Greeks: A Survey of Ancient History and Archaeology on the North Coast of the Euxine from the Danube to the Caucusus (Cambridge: Cambridge University Press, 1913), p. xxxvii (fig. O) and passimGoogle Scholar; From the Land of the Scythians: Ancient Treasures from the Museums of the U.S.S.R. 3000 B.C. — 100 B.C. (N.p.: The Metropolitan Museum of Art and The Los Angeles County Museum of Art, distributed by New York Graphic Society, n. d.), color plates 17–18Google Scholar (the celebrated fourth-century B.C.E. golden vase from the tomb at Kul Oba near Kerch in the Crimea), also reproduced in Mallory, J. P., In Search of the Indo-Europeans: Language, Archaeology and Myth (New York: Thames and Hudson, 1989), plate 6Google Scholar, and Flon, Christine, ed., The World Atlas of Archaeology, translated from Le Grand Atlas de l'archeologie (Encyclopaedia Universalis, 1985; rpt. New York: Portland House, 1988), p. 217Google Scholar (a Central Asian Saka delegation from Persepolis). Among the 3,000 human figures depicted with great realism and fine detail on the monumental reliefs of Persepolis (520–c450 B.C.E.), there are dozens of different hair and hat styles, including those of Sogdians, lonians, Ethiopians, Lydians, Babylonians, and Scythians, but none of them resemble the headgear of the Chou-yüan figures. One must assume, therefore, that the latter did not fall within the range of those who paid tribute to the Achaemenid kings of Persia. See Wilber, Donald N., Persepolis: The Archaeology of Parsa, Seat of the Persian Kings (1969; rev. Princeton, N.J.: Darwin, 1989)Google Scholar.

7. All of these are mentioned in Sheng-p'ing, Yin, “Hsi-Chou pang-tiao jen-t'ou”, p. 47Google Scholar and Ch'üan-fang, Ch'en, Chou-yüan, p. 20Google Scholar.

8. Ibid.

9. Hentze, Carl, Funde in Alt-China: Das Welterleben in ältesten China, Sternstunden der Archäologie (Göttingen: Musterschmidt, 1967), Taf. XVGoogle Scholar. The presence of actual Europoid individuals — not just their artistic representations — in China during the Shang period is proven by a couple of skulls (one virtually indistinguishable from that of a modern Englishman born in America) recovered from the cemetery at Hou-chia-chuang 侯家莊. These skulls were apparently deposited in sacrificial pits along with a large number of Mongoloid skulls and a somewhat lesser number of Eskimoid and Oceanic Negroid skulls as well as an unclassifiable fifth group of small crania. The circumstance of their discovery makes it difficult to draw any firm conclusions concerning the racial composition of the Anyang population as a whole, still less of the ruling elite. See Chi, Li, “Notes on the Physical Anthropology of the Yin-Shang Population”, in his Anyang (Seattle: University of Washington Press, 1977), pp. 255–264Google Scholar and Chang, Kwang-chih, Shang Civilization (New Haven: Yale University Press, 1980), pp. 331–335Google Scholar. My thanks to Emma Bunker for these two references. For the present, the safest view is probably that of Chi, Li (Anyang, p. 264Google Scholar): “From the very beginning, the north China plain was a meeting place of many different ethnic stocks and it is partly from the mixture of these groups that the early Chinese population was formed, although we must not forget that the dominant group among these stocks was indubitably the Mongoloid group.” The status of Europoid individuals within Shang society as a whole remains to be determined.

10. The tremendous dissimilarities of these pieces from typical early Chinese representations of human faces are vividly brought home by comparing them with the examples illustrated at Chang, , Civilization, p. 330, fig. 89Google Scholar.

11. Akurgal, Ekrem, Urartäische und altiranische Kunstzentren, Türk Tarih Kurumu Yayinlarindan, series 6, no. 9 (Ankara: Türk Tarih Kurumu Basimevi, 1968), plate 26cGoogle Scholar.

12. See the discussion of the word for chariot near the end of this article and the notes thereto.

13. Surprisingly, the headgear of Celtic warriors depicted on the silver Gundestrup cauldron bear a conspicuous resemblance to those of the Shang and Western Chou figures. The cauldron, however, is much too late (La Tène III, around 80–50 B.C.E.) to take seriously into account in our present investigation. See Olmsted, Garrett S., The Gundestrup Cauldron: its Archaeological Context, the Style and Iconography of its Portrayed Motifs, and Their Narration of a Gaulish Vision of Tain Bó Cúailnge, Collection Latomus, 162 (Brussels: Revues d'Études Latines, 1979), p. 9Google Scholar; Mallory, Indo-Europeans, plate 20. Cf. the warrior's helmet with projecting horns shown in Ross, Anne, Pagan Celtic Britain: Studies in Iconography and Tradition (New York: Columbia University Press, 1967), fig. 88Google Scholar. This bronze figurine from Grevensvaenge, Zealand, which is dated to c. 1250 B.C.E., bears an intriguing resemblance in its facial features to the Western Chou figures.

14. Past Worlds: The Times Atlas of Archaeology (Maplewood, N.J.: Hammond, 1988), p. 148Google Scholar.

15. See Eliade, Mircea, Shamanism: Archaic Techniques of Ecstasy, tr. from the French by Trask, Willard R., Bollingen series 76 (Princeton: Princeton University Press, 1964)Google Scholar.

16. For a general anthropological study of the wu (“mage”) in ancient Chinese society, see Chao-t'ao, Liang 梁垒銷, Chung-kuo ku-tai wu-shu — tsung-chiao te ch'i-yüan ho fa-chan 中國古代巫術一宗敎的起源和發展 [Ancient Chinese Magic: The Rise and Development of Chinese Religion] (Canton: Chung-shan University Press, 1989)Google Scholar.

17. In claiming that “mage” is a more accurate translation of *m^ᵞag than “shaman”, I wish not to detract from the vital ongoing debate between David N. Keightley and K. C. Chang regarding the significance of the *m^ᵞag for the alleged theocratic foundations of the Chinese state and the possibility that the king himself may have been the head *m^ᵞag. Chang's, position is presented in his Art, Myth, and Ritual: The Path to Political Authority in Ancient China (Cambridge, Mass.: Harvard University Press, 1983)Google Scholar. Keightley's response may be found in “Royal Shamanism in the Shang: Archaic Vestige or Central Reality?” prepared for the workshop on Chinese divination and portent interpretation, June 20 - 07 1, 1983 (University of California, Berkeley)Google ScholarPubMed and “Shamanism in Guo Yu? A Tale of Xi 覜 and Wu 巫” prepared for the Center for Chinese Studies Regional Seminar, 04 7–8, 1989 (University of California, Berkeley)Google Scholar. I have no view on whether or not there may have been a separate, indigenous tradition of shamanism in China before the introduction of magianism. There is, however, good evidence that the role of king and *m^ᵞag coincide in the crucial matter of the interpretation of the cracks in oracle bone divination. A typical notation would consist of preface (crack-making pu 卜 on [stem]-[branch] day), charge (divination cheng 貞), prognostication (the king, reading the cracks, said wang chart yüeh 王占曰), and verification which records what really transpired later on. See Keightley, David N., Sources of Shang History: The Oracle-Bone Inscriptions of Bronze Age China (Berkeley: University of California Press, 1978), pp. 40–44Google Scholar. Minao, Hayashi 林奈夫, “Chūgoku kodai no shinfu” 中國古代の神巫 (Shamanistic Gods in Ancient China), TōhōGakuhō 東方學報 (Journal of Oriental Studies), 38 (03, 1967), 199–224, p. 211Google Scholar cites examples of early prognostications which begin wu yiieh 巫曰 (“the *m^ᵞag said”). Considering the fact that, like the Shang king himself, the *m^ᵞag was responsible for the very important task of interpreting the divinatory cracks, was sacrificed to after death, and was closely associated with Ti (the high god, also a prefix for deceased royal ancestors), his position at court must have been quite exalted. Or perhaps the king was indeed the chief *m^ᵞag.

18. I reject the old etymology of “shaman” which traces it back through Tocharian ṣamáne to a Prakrit samaṇa or Sanskrit śramaṇa. The ascetic practices of the Buddhist śramaṇa are quite at variance with the ecstatic spiritualism of the shaman.

19. This explanation of Rab Mag was commonly accepted by most biblical commentators until it was disputed by Benveniste, E., Les mages dans l'ancien Iran, Publications de la Société des Études Iraniennes, 15 (Paris: Librairie Orientale et Américaine G.-P. Maisonneuve, 1938), “Appendice”, pp. 28–30Google Scholar. Benveniste would derive it from Babylonian rab mu-gi which is presumably a military title meaning “Master of the Horse” (Magister Equitum). His attempt to account for the phonetic discrepancy by positing an intermediary Aramaic stage strikes me as unnecessarily strained. For the traditional interpretation, see such standard works as Keil, C. F. and Delitzsch, F., Commentary on the Old Testament in Ten Volumes, vol. III, Jeremiah, Lamentations by Keil, C. F., translated from the German by Kennedy, James (Grand Rapids, Michigan: William B. Eerdmans, 1978 rpt.), p. 117Google Scholar; Streane, A. W., The Book of the Prophet Jeremiah together with the Lamentations (Cambridge: Cambridge University Press, 1913, revised version), p. 237Google Scholar; and Dummelow, J. R., ed., A Commentary on the Holy Bible by Various miters (New York: Macmillan, 1936), p. 476Google Scholar.

20. Zaehner, R. C., The Dawn and Twilight of Zoroastrianism (New York: G. Putnam's Sons, 1961), p. 163Google Scholar.

21. Lindtner, Christian, “Buddhist References to Old Iranian Religion,” in A Green Leaf: Papers in Honour of Professor Jes P. Asmussen, Hommages et Opera Minora, 12 (Leiden: E. J. Brill, 1988), pp. 433–444, esp. p. 442Google Scholar.

22. Herodotus, with an English translation by Godley, A. D., 4 vols. (1921–1922; Cambridge: Harvard University Press, 1957), passimGoogle Scholar. It is most interesting that some of the activities of the Magi described by Herodotus correspond very closely with those of the *m^ᵞag in early China. For example, in Book VII. 191 (vol. 3, pp. 508–509 of Godley's edition), we find the following passage: “For the storm lasted for three days; and at last the Magians, by using victims [cut up in pieces and offered to the manes, έντομα] and wizards' spells on the wind, and by sacrificing also to Thetis and the Nereids, did make it to cease on the fourth day …. “Compare note 27 below.

23. Nock, Arthur Darby, Essays on Religion and the Ancient World, selected and edited, with an Introduction, Bibliography of Nock's writings, and indexes, by Zeph Stewart, 2 vols. (Oxford: Clarendon, 1972), p. 309Google Scholar. Nock's essay” Paui and the Magus” (pp. 308–330) is the best study of the word magus in Greek.

24. Messina, Giuseppe, Der Ursprung der Magier und die zarathustriche Religion (Rome: Pontificio Istituto Biblico, 1930), pp. 1, 64, and passimGoogle Scholar. See also Moulton, James Hope, Early Zoroastrianism, Lectures Delivered at Oxford and in London„ 02 to May 1912; The Hibbert Lectures, second series (London: Williams and Norgate, 1913), pp. 118, 197, 323, and 410Google Scholar; Zaehner, , Dawn and Twilight of Zoroastrianism, p. 165Google Scholar. According to Zolar, Zolar's Encyclopedia of Ancient and Forbidden Knowledge (New York: Prentice Hall, 1986), p. 193Google Scholar, Zoroaster was traditionally held to be “the presumable inventor of all magic.”

25. See Burrow, T., “The Proto-Indoaryans”, Journal of the Royal Asiatic Society (1973), ”123-140Google Scholar, esp. p. 139. Gnoli, Gherardo, Zoroaster's Time and Homeland: A Study on the Origins of Mazdeism and Related Problems, Istituto Universitario Orientale, Seminario di Studi Asiatici, Series Minor, 7 (Naples, 1980), pp. 10-11 and 160–161Google Scholar, strongly supports this view and includes the Magi as part of the Zoroastrian complex of religious beliefs, practices, and institutions. Without specifying a specific upper limit, Mary Boyce also opposes the low date of the seventh century B.C.E. for Zarathustra in her A History of Zoroastrianism, 2 vols. (Leiden/Kōln: E. J. Brill, 1975–1982)Google Scholar. The position of Burrow, Gnoli, and Boyce has recently received a measure of archeological confirmation from the recent discovery of a spectacular temple complex at Togolok-21 in the southeastern Karakum desert of Turkmenia, an area known to the ancient Persians as Margush and to the Greeks as Margiana. This ritual center, built circa 1000 B.C.E., preserves evidence of the usage of soma (referred to as haoma in the Avestan hymns) and has structural characteristics which suggest that Margiana was the homeland of Zoroastrianism; see Sarianidi, Victor, “Where Was Zoroaster Born?” Sputnik, 3 (03, 1990), pp. 96–101, esp. p. 98Google Scholar. Even if it cannot be proven that Zoroaster himself was the center of the cult at Togolok-21, current scholarly opinion holds that the site was at least associated with proto-Zoroastrianism; see Parpola, Asko, “The Coming of the Aryans to Iran and India and the Cultural and Ethnic Identity of the Dasas”, Studia Orientaiia, 64 (Helsinki, 1988), 195–302, esp. pp. 236–238Google Scholar and the copious bibliographical references cited there.

26. Jao Tsung-i's paper (see note 1) cites dozens of occurrences of *m^ᵞag in the oracle bone inscriptions. His analysis shows that they were involved with animal (chiefly canine but also ovicaprid) sacrifice and libationary ritual and that they were associated with earth deities of the four directions. Hayashi Minao (see note 17) cites oracle bone inscriptions where *m^ᵞag are associated with porcine sacrifice, praying for rain, and other tasks that were central to the functioning of the Shang court.

27. Chin-hsiung, Hsü 許進雄, Chung-kuo ku-tui she-hui: Wen-tzu yü jen-lei-hsü eh te t'ou-shih 中國古代社會一文字與人類學的透視 [Ancient Chinese Society: From the Perspective of Writing and Anthropology] (Taipei: Taiwan Commercial, 1988), pp. 387, 440–441Google Scholar states that the *m^ᵞag were thought to be capable of calming the winds and that this is often reflected on the oracle bones. He also holds that the *m^ᵞag were healers, but this view is based on post-Shang sources and supposition. Hsü interprets as a pictograph of an implement used by the *m^ᵞag in ritual ceremonies. There is, however, no paleographical or archeological evidence to support this interpretation. A more convincing suggestion is that of Meng-chia, Ch'en 陳夢家, Yin-hsü pu-tz'u tsung-shu 殷虛辭線述 [A General Account of Oracle Bone Inscriptions from the Wastes of Shang], K'ao-ku-hsü eh chuan-k'an chia-chung 考古學專刊甲種 [Archeological Monographs, Series A], No. 2 (Peking: K'e-hsüeh, 1956), p. 579Google Scholar who believes that symbolizes the four directions. This is a reasonable interpretation because of the clear associations of the *m^ᵞag with the four directions in the oracle bone inscriptions. For other religious and ritual practices of the *m^ᵞag during the Shang period, see pp. 575–578, 590, of the same work. Ch'en cites one sacrifice which required nine canine victims.

28. Gershevitch, Ilya, “Zoroaster's Own Contribution”, Journal of Near Eastern Studies, 23 (01-October, 1964), 12–38 (on p. 25)CrossRefGoogle Scholar.

29. Cited in Jao Tsung-i, “New Light on wu.”

30. Bouisson, Maurice, Magic: Its Rites and History, translated from the French by Almayrac, G. (London: Rider, 1960), p. 17Google Scholar.

31. It has long been asserted that there is some connection between the word *m_yag (“mage”) and *m^ᵞag (“make postures to musical accompaniment, dance”). For example, L. C. Hopkins attempted to show that the graphs for *m^ᵞag (”mage”), the negative *m^ᵞag 唤 (“not have, not, no, don't”), and *m^ᵞag (“dance”) “can all be traced back to one primitive figure of a man displaying by the gestures of his arms and legs the thaumaturgie powers of his inspired personality”; “The Shaman or Chinese Wu: His Inspired Dancing and Versatile Character”, Journal of the Royal Asiatic Society (1945), 3-16, p. 5Google ScholarPubMed. See, too, Hopkins', earlier article “The Shaman or Wu 巫: A Study in Graphic Camouflage”, The New China Review, 2.5 (10, 1920), 423–439Google Scholar plus 1 plate, which follows Schindler, Bruno, Das Priestertum im alten China, 1. teil: Königtum und Priestertum, Einleitung und Quellen (Leipzig: Spamer, 1919), pp. 14–29Google Scholar. There are several objections to this type of graphemic analysis. The first is that it is based on the erroneous identification of the various oracle bone graphic forms for *m^ᵞag (“not”) and *m^ᵞag (“dance”) –which are indeed graphically related –as being equivalent to that for *m^ᵞag (”mage”). The second is that it confuses word with script In the study of Sinitic etymologies, phonology should always take precedence over graphology. This is clear from the fact that graphs are frequently borrowed to write homophonous or near-homophonous words that are semantically completely unrelated (e.g. *mlag [“wheat, i.e. Triticum aestivum”] as the written form of *mlag [”come”]). Furthermore, the origin of Sinitic languages antedates the rise of the Chinese script by many centuries. Hence the latter must be regarded as secondary and the former as primary, an observation so painfully obvious that it is often forgotten.

32. Knights of the religious military order known as the Templars, founded circa 1118 at Jerusalem by the Crusaders, had a large Cross Potent woven into the right breast of their robes. For an illustration, taken from a twelfth-century manuscript in the Universitä tsbibliothek at Heidelberg, of one such Crusader, see The American Heritage Dictionary of the English Language (1981), p. 1325Google Scholar.

33. The most influential of the grimoires is the Key of Solomon, published under the title Clavicula Salomonis c. 1456 but probably composed during the fourteenth or fifteenth century; see Encyclopedia of Occultism & Parapsychology, 3 vols., ed. Shepard, Leslie (Detroit: Gale Research, 1984), vol. 2, p. 718bGoogle Scholar. A Greek version of the Key in the British Museum is held to date from the twelfth or thirteenth century. “Solomon enjoyed a great legendary reputation as a powerful magician and a worshipper of strange gods, and as far back as the 1st century AD the Jewish historian Josephus referred to a book of incantations for conjuring up demons, supposedly written by Solomon. The Testament of Solomon (not the same as the Key) dating from c 100 to 400 AD, lists the names and powers of demons which Solomon had subdued with his magic ring. In the 13th century Roger Bacon, himself a reputed magician, knew of magical works attributed to Solomon and c 1350 Le Livre de Salomon, containing methods of evoking demons, was burned by order oi Pope Innocent VI.” Man, Myth, & Magic: The Illustrated Encyclopedia of Mythology, Religion and the Unknown, 11 vols, plus index, ed. Cavendish, Richard (New York, London, and Toronto: Marshall Cavendish, 1983), vol. 5, p. 1181Google Scholar.

The antiquity of the essential contents of the grimoires is further evidenced by the strong influence they received from the Hebrew Kabbalah (Jewish tradition of mystical speculation passed on primarily by word of mouth). The most important collection of Kabbalistic knowledge is the massive Zohar (Sejer ha-Zohar [Book of Splendor]), written for the most part during the late thirteenth century in Spain, in all likelihood by Moses b. Shem Tov de Léon of Guadalajara, a small town northeast of Madrid. A notable precursor is the Sefer Yetsirah [Book of Creation] probably written sometime between the third and sixth centuries C.E.

Another ancient source for the grimoires was Babylonian astrological lore, especially as elaborated by the Chaldeans who ruled for nearly a century until they were overthrown by the Persians in 89 B.C.E. Finally, the magical papyri stemming from the large numbers of Greeks who settled in Egypt starting around the fourth century B.C.E. represent perhaps the most significant early textual bases for the grimoires. They were also important sources for the Kabbalah; see Scholem, Gershom, Kabbalah (1974; New York: Dorset, 1987), p. 185Google Scholar. The more substantial of the Greek magical papyri date from the kte third through fifth centuries C.E. but some of the smaller texts derive from as early as the second century. The classical scholar Arthur Darby Nock was of the opinion that the longer texts were actually handbooks used by practicing magicians while the shorter ones were extracts designed for sale to nonprofessionals who desired them for specific needs and occasions; “Greek Magical Papyri,” The Journal of Egyptian Archaeology, 15 (1929), 219-235, esp. p. 220Google Scholar. Prominent among the signs on the Greek magical papyri was the Cross Potent (especially in its Pomee variety with knobs at the ends of the arms instead of bars — as it frequently occurs in the grimoires — often tilted). For examples, see Preisendanz, Karl, ed. and tr., Papyri Graecae Magicae: Die Griechischen Zauberpapyri, Sammlung Wissenschaftlichen Commentare, 2 vols. (1928–41; rev. ed. Stuttgart: B. G. Teubner, 1973–1974), vol. 2, pp. 140, 170–171Google Scholar. For an excellent brief introduction, see the article on “Magical Papyri” by Rees, B. R. in Cavendish, , ed., Man, Myth, & Magic, vol. 6, pp. 1689–1691Google Scholar.

Another ancient source for the grimoires was Babylonian astrological lore, especially as elaborated by the Chaldeans who ruled for nearly a century until they were overthrown by the Persians in 89 B.C.E. Finally, the magical papyri stemming from the large numbers of Greeks who settled in Egypt starting around the fourth century B.C.E. represent perhaps the most significant early textual bases for the grimoires. They were also important sources for the Kabbalah; see Scholem, Gershom, Kabbalah (1974; New York: Dorset, 1987), p. 185Google Scholar. The more substantial of the Greek magical papyri date from the kte third through fifth centuries C.E. but some of the smaller texts derive from as early as the second century. The classical scholar Arthur Darby Nock was of the opinion that the longer texts were actually handbooks used by practicing magicians while the shorter ones were extracts designed for sale to nonprofessionals who desired them for specific needs and occasions; “Greek Magical Papyri,” The Journal of Egyptian Archaeology, 15 (1929), 219-235, esp. p. 220Google Scholar. Prominent among the signs on the Greek magical papyri was the Cross Potent (especially in its Pomee variety with knobs at the ends of the arms instead of bars — as it frequently occurs in the grimoires — often tilted). For examples, see Preisendanz, Karl, ed. and tr., Papyri Graecae Magicae: Die Griechischen Zauberpapyri, Sammlung Wissenschaftlichen Commentare, 2 vols. (1928–41; rev. ed. Stuttgart: B. G. Teubner, 1973–1974), vol. 2, pp. 140, 170–171Google Scholar. For an excellent brief introduction, see the article on “Magical Papyri” by Rees, B. R. in Cavendish, , ed., Man, Myth, & Magic, vol. 6, pp. 1689–1691Google Scholar.

34. Dee's seal is preserved in the British Museum as Sloane MS 3188, fol. 30. See also the mandala-like Sigillum Dei Aemeth illustrated in King, Francis, Magic: The Western Tradition (London: Thames and Hudson, 1975), fig. 15Google Scholar. Complicated sixteenth-and seventeenth-century magical diagrams feature the Cross Potent as a protective device for those who employ the esoteric practices described in the grimoires; see King, Magic, fig. 37. It is also very closely linked with names for God and other symbols of power; King, Magic, figs. 6, 9–13.

35. King, Magic, fig. 36.

36. The classic statement of this view is by Ho, Ping-ti, The Cradle of the East: An Inquiry into the Indigenous Origin of the Techniques and Ideas of Neolithic and Early Historic China (Hong Kong: Chinese University of Hong Kong and University of Chicago Press, 1975)Google Scholar.

37. Compare Shafer, Robert, “Eurasial,” Orbis, 12 (1963), 19–44Google Scholar, also “The Eurasial Linguistic Superfamily”, Anthropos, 60 (1965), 445–468Google Scholar, and Pulleyblank, Edwin G., “Prehistoric East-West Contacts Across Eurasia”, Pacific Affairs, 47.4 (Winter, 1974–1975), 500–508CrossRefGoogle Scholar. The most complete presentation of Indo-European influence on the formation of Chinese civilization to date is that of Tsung-tung, Chang, “Indo-European Vocabulary in Old Chinese: A New Thesis on the Emergence of Chinese Language and Civilization in the Late Neolithic Age”, Sino-Piatonic Papers, 7 (01, 1988), 56 pagesGoogle Scholar. Chang is currently working on a book that will greatly expand his treatment of the subject.

38. Information from Wailes, Bernard and Past Worlds, p. 190Google Scholar.

39. Kohl, Philip L., Central Asia: Palaeolithic Beginnings to the hon Age (Paris: Éditions Recherche sur les Civilisations, 1984), p. 155Google Scholar. Kohl notes (p. 157) that millet was also found at Sapalli-tepe and suggests that this too indicates a Bronze Age link to the east. My thanks to Barbara Stephen for this reference.

40. Campbell, Colin writing in the New York Times (02 17, 1986), 10Google Scholar. I am grateful to Edward L. Shaughnessy for this reference. See also the recent report by Hsi-kuang, Li 李希光 in Jen-min jih-pao 人民曰報 [People's Daily], overseas edition (08 22, 1990)Google Scholar.

41. See Shaughnessy, Edward, “Historical Perspectives on the Introduction of the Chariot into China”, Harvard Journal of Asiatic Studies, 48 (06, 1988), 189–237CrossRefGoogle Scholar, and “Western Cultural Innovations in China, 1200 B.C.”, Sino-Platonic Papers, 11 (07, 1989), 8 pagesGoogle Scholar.

42. The initial *k in this transcription might conceivably be aspirated and the *l might be an *r. Also, as in *m^ᵞag, the presence of the final velar might be disputed. None of these adjustments would require a rejection of the relationship between Indo-Iranian and Old-Sinitic with regard to the words for “magician” and” wheel[ed vehicle]” that I am proposing. So long as the correspondences between Indo-Iranian and Old Sinitic are regular, the differences between theitl may be attributable to their phonotactic constraints. Nonetheless, it is frustrating that the reconstructions for Old Sinitic are still so crude. This makes linkages with alphabetical languages whose archaic phonology is more accurately worked out all the more urgent. Naturally, because historical linguistics is inductive (in principle), it is essential to continue to refine the entire system of Sinitic reconstructions so that they are both increasingly accurate and internally consistent.

43. Tomas Gamkrelidze and Vyachislav Ivanov have noted that Indo-European *k^wék^wlo- resembles words for “vehicle” in Sumerian (gigir), Semitic (*galgal-), and Kartvelian (i.e., Transcaucasian *grgar), indicating that the Proto-Indo-Europeans may have been in contact with these other language groups during the fourth millennium B.C.E. when wheeled vehicles were invented; see Mallory, , Indo-Europeans, p. 163Google Scholar.

44. Lane, George Sherman, “The Indo-European Labiovelars in Tocharian”, in Indogertnanica (Festschrift für Wolfgang Krause), eds. Hartmann, Hans and Neumann, Hans (Heidelberg: Carl Winter, 1960) pp. 72–79, esp. pp. 72 and 75Google Scholar; Normier, Rudolf,” “Tocharisch n̄kät/n̄akte ‘Gott,” Zeitschrift für Vergleichende Sprachforschung, 94 (1980), 251–281, esp. p. 263Google Scholar.

45. Miller, Wsewolod, Die Sprache der Osseten: Grundriss der iranischen Philologie, eds. Geiger, Wilhelm and Kuhn, Ernst, Band, Anhang zum ersten (Strassburg: Karl J. Trübner, 1903), p. 26Google Scholar.

46. For pre-Ossetic loanwords in Tocharian, see Winter, Werner, “Baktrische Lehnwōrter in Tocharischen”, Donum Indogermanicum: Festgabe für Anton Scherer, ed. Schmitt-Brandt, Robert (Heidelberg: Carl Winter Universitätsverlag, 1971), pp. 217–223Google Scholar, and for the same in Eastern European languages, see Abaev, V. I., Istoriko-Etimologicheskii Slovar' Osetinskogo Yaz'ika, 3 vols. (Moscow and Leningrad: Akademii Nauk SSSR, 1958–1973)Google Scholar.

47. Pokorny, Julius, Indogermanisches etymologisches Wōrterbuch, 2 vols. (Bern and Munich: Francke, 1959–1969), vol. 1, p. 695Google Scholar.