Printer Friendly

A Harivamsa hymn in Yijing's Chinese translation of the Sutra of Golden Light.


The Harivamsa (first to third century), well known as an account of the life of the god Krsna, contains also a few hymns to the goddess Durga. These hymns are extremely inter-esting in that they are amongst the earliest written sources on Durga and represent one of the early efforts at establishing the identity, significance, and worship of Durga as the great goddess. (1) One of these hymns appears in Chinese translation in the Buddhist Sutra of Golden Light (Suvarnabhasottama Sutra), in Yijing's [TEXT NOT REPRODUCIBLE IN ASCII] (635-713) rendering of 703, (2) for which the original Sanskrit version does not survive. The presence of the hymn in a Buddhist context is not particularly surprising, given that the Indic Buddhist cosmos is populated with numerous Indic deities of Vedic, Brahmanical, Hindu, as well as folk origin. In the Indic Buddhist context as reflected in Yijing's translation, however, the hymn does not retain the same purpose that it had in the Harivamsa. In fact, it is no longer addressed to Durga, but instead to the goddess Sarasvati. And yet, despite the change of context, purpose, and deity addressed, the hymn remains remarkably effective: a warrior-like nature and countenance adds a fittingly fierce aspect to Sarasvati's role in a sutra for the protection of the state. (3) The effectiveness of the hymn in the sutra context renders its presence justifiable at the level of conventional truth, as the sutra itself suggests, (4) and its inclusion is indeed revealing in terms of the composition and compilation of the Sanskrit sutra.

In the Chinese translation of the sutra, furthermore, the hymn is metamorphosed into a different idiom meaningful in a different cultural context. The goddess is called Biancai tiannu [TEXT NOT REPRODUCIBLE IN ASCII] "Eloquence Goddess," a function held by Sarasvati in the Sanskrit sutra and entirely in line with her depiction from the late Vedic period onwards. We have, then, a hymn invoking Durga, used in the sutra context to invoke Sarasvati, and rendered into Chinese to praise Biancai tiannu, whose name points to Sarasvati's function and whose warrior-like form is borrowed from Durga.

Here we shall examine the above mentioned versions of this hymn from two different contexts and in two different languages: the Sanskrit text as found in the Harivamsa, alongside the Chinese version of the hymn in Yijing's translation of the Sutra of Golden Light. Through an annotated translation of and commentary on both the Chinese and the Sanskrit, we shall find we are in fact dealing with two versions of the same, rather than simply a related, hymn. We shall also ascertain, in large part, the version of the hymn that appeared in Yijing's no longer extant Sanskrit manuscript and point out the ways in which Yijing and his translation team went about rendering the hymn into Chinese verse.


The Harivamsa hymn in question is one of a small number of hymns to Durga appearing in the vulgate Mahabharata epic (mid-second century B.C.E. to the year zero), (5) to which the Harivamsa forms a kind of supplement, and in the Harivamsa itself. Our hymn, like most of these early Durga hymns, is not included in the critical edition, but appears instead in an appendix to it. (6) It consists, in fact, of the latter part of a hymn uttered by the god Visnu in Harivamsa 47:38-57 of the critical edition, "inserted," according to the editor Parashuram Lakshman Vaidya, after 47:52. Our hymn is identified as the Arya stava "praise of she who is noble." The context is Krsna's complex birth story, and the hymn is offered by Visnu when arranging the birth of Krsna. Wicked King Kamsa knows he is to be slain by one of the children of Devaki and Vasudeva (Krsna's parents), and hence confines them and slays their children as each one is born. In the first half of the hymn, Visnu requests the goddess Nidra, "Sleep" (one of the names applied to Durga), to arise in the womb of Yasoda, from whom she will be born at the same time as Krsna from his mother Devaki. The babies will then be exchanged, so that when evil Kamsa sees baby Nidra, who has taken the place of Krsna, and dashes the little girl against a rock, Krsna's life will be spared. Visnu's request ends in a praise of the goddess, which indicates that she is actually Durga, identified with a long series of goddesses. (7) The latter part of Visnu's praise is our hymn, the Arya stava, which found its way into the Sutra of Golden Light.


This sutra was a highly influential text throughout Buddhist Asia, and would have existed in some form already in the first century C.E. The extant Sanskrit text was edited by Johannes Nobel in 1937 and translated into English by Ronald E. Emmerick in 1970, who then revised and corrected his translation in 1990, 1992, and 1996. (8) A new edition of the Sanskrit is being prepared by P. O. Skjaervo, for which he uses a Nepalese manuscript that was not available to Nobel. The translation of the Sutra of Golden Light into numerous ancient languages reveals the importance of this text and the existence of earlier and different versions of it. In addition to Chinese and Tibetan renderings, there are Khotanese, Sogdian, Xi Xia (Tangut), Mongolian, and Old Turkic translations.

The great popularity of the sutra in China is attested by the remarkable number of manuscripts of it from Dunhuang and elsewhere. (9) The Taisho (1912-26) edition of the Buddhist Canon includes three different Chinese versions (10) of the sutra: (a) Dharmaksema's translation of 417, (11) Jinguangming jing [TEXT NOT REPRODUCIBLE IN ASCII] in 4 juan and 18 parivarta (T. vol. 16, no. 663), based on an earlier Sanskrit version of the sutra than the extant version edited by Nobel in 1937; (b) Baogui's [TEXT NOT REPRODUCIBLE IN ASCII] edition of 597, Hebu Jinguangming jing [TEXT NOT REPRODUCIBLE IN ASCII] in 8 juan and 24 parivarta (T. vol. 16, no. 664), including the translations of Dharmaksema from 417, of Paramartha from 552, and of Yasogupta and Jnanagupta from 561-78; (c) Yijing's translation of 703, (12) Jinguangming zuishengwang jing [TEXT NOT REPRODUCIBLE IN ASCII] in 10 juan and 31 parivarta (T. vol. 16, no. 665). The Harivamsa's Durga hymn appears only in Yijing's Chinese translation, in the chapter dedicated to Biancai tiannu (Sarasvati). It is not found in the extant Sanskrit text of the sutra. On the other hand, the hymn does appear in some of the Tibetan translations. (13) The Tibetan versions, however, are beyond my linguistic abilities, and hence I leave their discussion to scholars competent in the field.

The Sarasvati chapter, in the extant Sanskrit, in the edition of Baogui, and in Yijing's rendering, consists of three parts, each of which presents the goddess in a different aspect: in the first part Sarasvati appears as a deity of eloquence; (14) in the second she teaches a ritual herbal bath; and in the third she is invoked by the brahmana Kaundinya. Yijing's Chinese version of the Harivamsa hymn is found in the third part, placed in the mouth of Kaundinya and addressed to Biancai tiannu. (15) It consists of twenty-two stanzas (T. vol. 16, no. 665, p. 437a6-b20), as opposed to the Harivamsa's twenty-nine stanzas (that is, twenty-eight stanzas plus introductory stanza included in 591*, an insertion after Harvamsa 47:54).


The appearance of a Chinese version of a Harivamsa hymn in Yijing's translation of the sutra has been noted previously. As far as I am aware, it is first mentioned in a note found in vol. 16 of the Taisho edition of the Buddhist canon. (16) Its presence was then duly noted in 1929 in the first fascicle of the Hobogirin encyclopedic dictionary of Buddhism. (17) Neither work, however, specified to which hymn of the Harivamsa it referred. Some years later, in 1932, an inaccurate identification of the text from which the hymn derives appeared in the Japanese rendering of Yijing's translation of the sutra in Kokuyaku Daizokyo [TEXT NOT REPRODUCIBLE IN ASCII]: in a very brief note, Watanabe Kaigyoku [TEXT NOT REPRODUCIBLE IN ASCII] stated that Kaundinya's praise was taken from a laud to Durga found in the Mahabharata epic and amplified. (18) Watanabe did not specify which Mahabharata hymn to Durga he had in mind.

Nobel in 1958, in his German translation of Yijing's Chinese rendering of the sutra, located the hymn in the Harivamsa, but quoted only its first ten (of twenty-nine) Sanskrit stanzas in a note. (19) Nobel stated that ten of the eleven stanzas he quoted (20) were almost literally translated in Yijing, but that the rest of the Harivamsa stanzas were certainly not literally rendered, although the Chinese contained numerous allusions to their contents. (21) As my study will attempt as far as possible to demonstrate, Yijing's entire twenty-two-stanza rendering of the hymn is indeed a literal Chinese translation of the hymn in the version that appeared in his Sanskrit manuscript of the Sutra of Golden Light.

More recently, Nagano Sadako [TEXT NOT REPRODUCIBLE IN ASCII], like Watanabe, has claimed that the source for Kaundinya's praise in Yijing is a Mahabharata hymn, but provides a more specific reference: Yudhisthira's hymn to Durga. (22) The small number of Durga hymns found in the Mahabharata and the Harivamsa are indeed closely related, (23) but, as we shall see, Yijing's Chinese version clearly corresponds to the Harivamsa hymn identified by Nobel, and not to a Mahabharata hymn. Watanabe and Nagano, nevertheless, are correct in their understanding of the invoked goddess as Durga.

On the other hand, because Naluoyan [TEXT NOT REPRODUCIBLE IN ASCII] (Narayana) appears in the introductory stanza (591* lines 3-4), (24) the Hobogirin tells us that "Benzaiten [Japanese reading for Biancaitian [TEXT NOT REPRODUCIBLE IN ASCII]] y est identifiee a la deesse Narayani, epouse de Visnu ..." and then lists her manifestations. (25) Nobel, who is familiar with the Hobogirin, likewise considers it to be a hymn to Narayani. The goddess, he explains, is praised under many different names and is identified with other goddesses, such as the consorts of Siva (including Durga) and Indra, and also with Sarasvati, in support of which he refers to passages in the Harivamsa. (26) While different names are frequently applied to different goddesses, (27) in the case of our hymn, the wider mythological context (Krsna's complex birth story) identifies the goddess as Durga. As Thomas Coburn's study of the early hymns to Durga has shown, names and attributes resonant of the Vedic tradition, like Narayani, (28) are bestowed on the wild mountain goddess for the express purpose of integrating her into the Brahmanical mainstream, thereby rendering her "noble" (arya) and hence worthy of worship. In order to establish her identity as mother goddess, furthermore, she is identified with virtually all goddesses, ranging from Vedic, epic, and early Puranic figures to various local, tribal deities. Herein lies the difficulty in identifying the goddess to whom all these names are applied.


The Harivamsa hymn is composed in the sloka meter, consisting of four pada "feet" of eight syllables. Yijing's version of the hymn is also in verse, corresponding metrically, as Edwin Pulleyblank comments, "with normal Chinese seven-word verse of the Tang period, not, of course, with the tonal patterns of so-called 'regulated verse' but in the freer style of seven-word gushi or yuefu, in quatrains with a regular caesura in each line after the fourth word." (29) In Yijing, this hymn proclaimed by the brahmana Kaundinya is introduced as song [TEXT NOT REPRODUCIBLE IN ASCII] (praise verses), one of the Chinese terms for translating the Sanskrit gatha: (30)

 Then, brahmana Kaundinya ... addressed the great assembly as follows:
 "You all, the entire great assembly of humans and gods, should know
 everything in this way and listen wholeheartedly. I now again wish, by
 resorting to the way of conventional truth, to praise that excellent,
 wonderful Eloquence Goddess." He then proclaimed the gatha as
 follows. (31)

The term shidi [TEXT NOT REPRODUCIBLE IN ASCII], referring to the conventional truth (samvrtisatya), is revealing: perhaps it indicates an awareness and an acknowledgment on the part of the sutra compilers that what follows may not be strictly Buddhist material, but that it is nevertheless provisionally admissible in the conventional sphere. (32) This is not to suggest that the sutra compilers lifted the hymn directly from the Harivamsa, for it may well have circulated in different formats and contexts. The hymn could have been considered a part of the common Indic religious realm, but its contents clearly reveal an epic-Puranic mythological universe, and hence the inclusion of the hymn in the sutra may have required, in the minds of the sutra compilers, a word of justification.

In terms of the Chinese translation, a fundamental question is to what extent we may be certain that Yijing and his collaborators interpreted correctly the Sanskrit original on which their Chinese rendering was based. In order to answer this question and to evaluate correctly the quality of Yijing's version of the sutra, including the hymn in question, a few remarks on Yijing's translation team may be helpful. His team consisted of sixteen people, (33) including two Indians: Manicintana (Baosiwei [TEXT NOT REPRODUCIBLE IN ASCII], d. 721 C.E.) (34) from Kashmir, who arrived in Luoyang in 693 and spent twenty-eight years in China, was in charge of verifying the Sanskrit meaning (zheng fanyi [TEXT NOT REPRODUCIBLE IN ASCII]); the Indian Srimata (Limoduo [TEXT NOT REPRODUCIBLE IN ASCII], presence in China documented from 700 to 707) (35) read the Sanskrit text (du fanwen [TEXT NOT REPRODUCIBLE IN ASCII]). Together with Yijing, who had spent over twenty years in India and Srivijaya, Manicintana and Srimata were the members of the translation team who certainly knew Sanskrit. There were probably others on the team with a knowledge of Sanskrit, but this point cannot be ascertained for the time being. Two people, Bolun [TEXT NOT REPRODUCIBLE IN ASCII] and Huibiao [TEXT NOT REPRODUCIBLE IN ASCII], wrote down (bishou [TEXT NOT REPRODUCIBLE IN ASCII]) the translation orally made by Yijing, and all the others, including eminent monks like Faming [TEXT NOT REPRODUCIBLE IN ASCII] and Degan [TEXT NOT REPRODUCIBLE IN ASCII], verified the meaning (zhengyi [TEXT NOT REPRODUCIBLE IN ASCII]) of the Chinese. Based on what the scribes wrote down, Yijing connected the text (zhuiwen [TEXT NOT REPRODUCIBLE IN ASCII]), i.e., revised the composition, and corrected the characters (zhengzi [TEXT NOT REPRODUCIBLE IN ASCII]) he deemed unsuitable. (36) Hence, we have here a qualified translation team, headed by a very cultivated Chinese, Yijing, who, to say the least, had "a rather good practical knowledge of Sanskrit." (37) Possible deficiencies in Yijing's Sanskrit, in any case, would likely have been filled by the two Indians Manicintana and Srimata. I have kept these points in mind throughout my comparative analysis of the Harivamsa hymn with its Chinese translation, especially where it might easily have been assumed that the translation team had misunderstood the Sanskrit.

Yijing and his collaborators were translators and their main duty was to give the best possible rendering, both in terms of fidelity and readability, for their Chinese audience. Besides court officials, this audience consisted mainly of the cultivated monks of the capitals and of the principal Buddhist centers, as well as educated people who were either Buddhist or who had deep interests in Buddhism. Hence they knew much more about Buddhism than we might imagine. With this audience in mind, therefore, I do not consider, in my own English translation of the Chinese, what people ignorant of Buddhism in Yijing's time or presentday average Chinese readers might understand from his version.

We now turn, then, to the annotated, two-column translation of the hymn, with Yijing's Chinese translation on the left and the Sanskrit Harivamsa hymn on the right. The stanzas appear in Yijing's order, and only the matching Sanskrit stanzas are provided. Yijing's manuscript did not have precisely the same stanza order and wording that appears either in the critical edition of the Harivamsa, which I use here, or in the variants listed in the notes to the critical edition. Hence, it has been necessary, from Yijing's eleventh stanza onwards, to reorganize the order of the Sanskrit stanzas so as to match them up with Yijing's Chinese. Although certain passages in Yijing's version of the hymn remain without an extant Sanskrit correspondent--either in the critical edition or in the variants listed in the notes--it has been possible to find Sanskrit equivalents for most of the Chinese rendering of the Harivamsa hymn. As we shall see and as Pulleyblank likewise notes, (38) the formal requirement of the seven-word verse may have resulted in a certain amount of rearrangement of the material provided by the Sanskrit hymn in Yijing's manuscript.

In order to give a sense of where within the Chinese and Sanskrit versions of the hymn stanzas are found, I have numbered them (1-22 for Yijing; introductory stanza [591* lines 3-4] plus 1-28 for the Sanskrit Harivamsa), in addition to providing line references to the Taisho edition and to the critical edition of the Harivamsa. Translations of the Chinese and the Sanskrit into English are my own. (39) My explanatory notes on the stanzas, furthermore, are presented in abbreviated form due to spatial constraints and will appear in full in a publication provisionally entitled Re-contextualizing the Praises of a Goddess. (40)

 Yijing's Version Harivamsa Hymn to Durga
 Arya Stava

 (p. 437a6-b20) (Crit. ed., app. no. 8 in vol. 2,
 pp. 34-37)

 Stanza 1 (lines a6-a7) Introductory stanza (591* lines
 3-4) (41)
 vaisampayana uvaca
 [TEXT NOT REPRODUCIBLE IN ASCII] aryastavam pravaksyami
 [TEXT NOT REPRODUCIBLE IN ASCII] yathoktam rsibhih pura /
 [TEXT NOT REPRODUCIBLE IN ASCII] narayanim namasyami
 [TEXT NOT REPRODUCIBLE IN ASCII] devim tribhuvanesvarim //

 Vaisampayana (42) said:
 I respectfully honor the goddess I recite the praise to the noble one
 Narayani, (arya stava),
 who has sovereignty in the as it was formerly pronounced by the
 worlds. seers.
 I now sing the praise of that I honor Narayani,
 venerable one,
 as it was formerly pronounced by the goddess, queen of the three
 the hermits. (43) worlds.

Evidently, the reverse order of Yijing appears in the Sanskrit. The first half of Yijing's stanza goes with the third and the fourth pada of the Sanskrit, while the second half of Yijing's stanza goes with the first and the second pada of the Sanskrit. There is no way of knowing whether Yijing and his collaborators translated the contents of this stanza in the precise order in which it was found in the Sanskrit manuscript they were working from, or whether they chose to reverse the order in their Chinese rendering.
 Stanza 2 (lines a8-a9) Stanza 1 (lines 1-2)

 [TEXT NOT REPRODUCIBLE IN ASCII] arya katyayani devi
 [TEXT NOT REPRODUCIBLE IN ASCII] kausiki brahmacarini /
 [TEXT NOT REPRODUCIBLE IN ASCII] janani siddhasenasya
 [TEXT NOT REPRODUCIBLE IN ASCII] durga vira mahatapah //

 You are good fortune, success, Noble goddess Katyayani,
 and peace of mind,
 intelligence, modesty, and glory. Kausiki, practicing the Brahman,
 You are the mother, able to mother of him whose army consists
 generate beings. of successful [warriors]
 Brave fierce (one), you Durga the brave, of great energy.
 constantly practice great

The first half of Yijing's gatha does not match pada one and two of the Sanskrit, where the goddess is invoked as the noble Katyayani and the Brahman-practicing Kausiki, names and epithets recalling the Vedic tradition. (44) No equivalent for these two first pada is to be found anywhere in the Chinese rendering of the hymn. Instead a match for the first half of Yijing's gatha appears in Harivamsa 47:54ab. According to the division of a hymn by Visnu in the critical edition, where the editor accepted the first half into the critical text (Harivamsa 47:38-57) and relegated the second half to an appendix (appendix 1, no. 8, i.e., the Arya Stava discussed here and translated into Chinese in Yijing), Yijing's line 8 corresponds to Harivamsa 47:54ab, one of the last stanzas of the first half of the hymn, and in fact the stanza following which the latter half is to be read. The problem of the match between Yijing and the Sanskrit does not present itself in most manuscript traditions, which include the Arya Stava. (45) The corresponding pada of the Sanskrit hence read, side by side with Yijing:
 [TEXT NOT REPRODUCIBLE IN ASCII] tvam siddhih srir dhrtih kirtir
 [TEXT NOT REPRODUCIBLE IN ASCII] hrir vidya samnatir matih /

 You are good fortune, success, You are success, good fortune,
 and peace of mind, knowledge, fortitude, fame, modesty, knowledge,
 intelligence, reverence, reverence, intellect.
 modesty, and fame. (47:54ab)
 (line a8)

The two characters congming [TEXT NOT REPRODUCIBLE IN ASCII] may be taken together in the sense of "intelligence" or "knowledge," but since each of the characters also has an independent meaning, I translate ming [TEXT NOT REPRODUCIBLE IN ASCII] as "knowledge," corresponding to the Sanskrit vidya, and cong [TEXT NOT REPRODUCIBLE IN ASCII] as "intelligence," corresponding to matih. (46) Likewise, the characters cankui [TEXT NOT REPRODUCIBLE IN ASCII] may be understood together as "shame" or "modesty," (47) but also separately as "reverence" (can [TEXT NOT REPRODUCIBLE IN ASCII]), corresponding to samnatih, and "shame/modesty" (kui [TEXT NOT REPRODUCIBLE IN ASCII]), corresponding to hrih.

The third pada is clearly not a perfect match, but the two versions are indeed related. The substantive janani "mother," she who generates or produces, comes from root jan "to be born or produced." The function of janani, therefore, is found in Yijing's translation: [TEXT NOT REPRODUCIBLE IN ASCII] "you are the mother, able to generate." (48) The problematic word is siddhasena-"one whose army consists of successful [warriors]," i.e., Skanda, son of Siva and Parvati, leader of the army of the gods. Coburn translates siddhasena- as "the one whose army is the Siddhas (Skanda)." (49) The siddha, literally "realized, perfected ones," are a class of semi-divine beings with supernatural powers. However, Skanda's army, as far as I am aware, is not generally considered to consist exclusively of siddha. Skanda, furthermore, has no particular connection with siddha. Therefore I interpret the compound as "the one whose army consists of successful [warriors]."

Yijing, moreover, does not mention Skanda in his role as army general, but simply refers to "beings" or to "the world [of beings]" shijian [TEXT NOT REPRODUCIBLE IN ASCII]. It is likely that siddhasena did not appear in the Sanskrit text Yijing was working from. Variants listed in the critical edition of the Harivamsa include siddhisenasya "of the army of successes/perfections" and siddhasanghasya "of the successful multitude." While the latter comes closer to the corresponding Chinese, Yijing's beings are not characterized as successful, accomplished, or perfected, but perhaps they might be perceived as the successful result of creation. In fact, the "successful" or "accomplished" aspect found in the Sanskrit pada, qualifying the army/multitude, appears in the Chinese in the characterization of the goddess as endowed with the "ability" (neng [TEXT NOT REPRODUCIBLE IN ASCII]) to generate beings, which in turn assures success in the task and an accomplished, perfect result. Nevertheless, the successful generation of beings by the mother is not the precise equivalent of the "mother of the successful multitude."

The fourth pada calls the goddess Durga, qualifies her as brave, and mentions her great tapas. It is quite possible that Yijing had the same reading as the above Sanskrit (durga vira mahatapah) and that he may have rendered an exact Chinese translation: yong meng [TEXT NOT REPRODUCIBLE IN ASCII] "brave-fierce" for durga vira (or rather vira durga) and da jingjin [TEXT NOT REPRODUCIBLE IN ASCII] "great energy," literally "great energy-progress," for mahatapah "of great energy," (50) with the intervening changxing [TEXT NOT REPRODUCIBLE IN ASCII] "constantly practice" to further explain mahatapah and to fill in the required number of characters. (51) Yongmeng [TEXT NOT REPRODUCIBLE IN ASCII] taken together means "brave" and might lead one to think that if "Durga/durga" did in fact appear in Yijing's manuscript, it was left untranslated, perhaps under the constraints of the seven-character verse or because Yijing chose not to include her name in his Chinese rendering. If the characters are read separately, however, yong [TEXT NOT REPRODUCIBLE IN ASCII] as "brave" and meng [TEXT NOT REPRODUCIBLE IN ASCII] as "fierce," the second character meng [TEXT NOT REPRODUCIBLE IN ASCII] might have been intended as a translation of Durga: literally, durga means "difficult to access" and can refer to a place like wilderness, where the fierce goddess Durga dwells, and of which she might be conceived as a personification. Another name of Durga, incidentally, is Candi "fierce one," for which meng would have been a perfect literal rendering.
 Stanza 3 (lines a10-a11) Stanza 2 (lines 3-4)

 [TEXT NOT REPRODUCIBLE IN ASCII] jaya ca vijaya caiva
 [TEXT NOT REPRODUCIBLE IN ASCII] pustis ca tvam ksama daya /
 [TEXT NOT REPRODUCIBLE IN ASCII] jyestha yamasya bhagini
 [TEXT NOT REPRODUCIBLE IN ASCII] nilakauseyavasini //

 In the battlefields, you fight You are Victory (jaya)
 and are always victorious. and Decisive Victory
 You foster and raise [people], you prosperity, forbearance,
 bring [them] under control, and [your] compassion,
 heart is compassionate and forbearing.
 You appear as the eldest sister of Yama. eldest sister of Yama,
 You always wear a blue wild-silk wearing a blue silk garment.

While the Sanskrit personifies the goddess as Victory (jaya) and Decisive Victory (vijaya), the Chinese translation describes how she is always victorious ([TEXT NOT REPRODUCIBLE IN ASCII]) when fighting ([TEXT NOT REPRODUCIBLE IN ASCII]) in battlefields ([TEXT NOT REPRODUCIBLE IN ASCII]). The characters zhangyang [TEXT NOT REPRODUCIBLE IN ASCII] "you foster and raise [people]" are related to pusti "prosperity," but there is no Sanskrit correspondent for tiaofu [TEXT NOT REPRODUCIBLE IN ASCII] "bring [them] under control" in this second pada. Variant readings for pustis ca tvam include tustih pustih "contentment [and] prosperity" and, in reverted order, pustis tustih, neither of which corresponds to tiaofu. There is, however, a correspondent to be found in the first pada of the Sanskrit, where the goddess is identified as Victory and Decisive Victory (jaya ca vijaya caiva). (52) As for the remainder of the second line, ksama daya finds its equivalent, in reverse order, in xin ciren [TEXT NOT REPRODUCIBLE IN ASCII]: ksama "forbearance" is translated as ren [TEXT NOT REPRODUCIBLE IN ASCII], and daya "compassion" as ci [TEXT NOT REPRODUCIBLE IN ASCII], both of which are inherent in the goddess's heart (xin [TEXT NOT REPRODUCIBLE IN ASCII]).

Yama (literally, "twin") and his twin-sister Yami were the first human pair, through whom the race of humans was generated. Yama's sister, therefore, is his consort. Just as the goddess is identified with the female counterparts of various gods, including Narayana (introductory stanza [591* lines 3-4] to the Harivamsa hymn; Yijing, line a6 in stanza 1), she is also identified with Yama's twin-consort.
 Stanza 4 (lines a12-a13) Stanza 3 (lines 5-6)

 [TEXT NOT REPRODUCIBLE IN ASCII] bahurupa virupa ca
 [TEXT NOT REPRODUCIBLE IN ASCII] anekavidharupini /
 [TEXT NOT REPRODUCIBLE IN ASCII] virupaksi visalaksi
 [TEXT NOT REPRODUCIBLE IN ASCII] bhaktanam pariraksani //

 Beautiful and ugly aspects, you are endowed Multiform and
 with them all. varicolored, of many
 kinds of forms [are
 Your eyes can provoke fear in the beholder. Varicolored-eyed,
 [Your] countless superior deeds surpass the
 You save all those who have faith in [you]. [you are] the protection
 of devotees.

She who is "multiform and varicolored" (bahurupa virupa ca) in the Sanskrit is described as having both "beautiful, good" (hao [TEXT NOT REPRODUCIBLE IN ASCII]) as well as "ugly" (chou [TEXT NOT REPRODUCIBLE IN ASCII]) aspects (rongyi [TEXT NOT REPRODUCIBLE IN ASCII]) in Yijing. The original sense of virupa- is color-oriented (53) and therefore translated here as "varicolored." The goddess has not only many forms, but also various colors. Hence she is endowed with many kinds of forms (anekavidharupini). Yijing conveys this sense by describing the goddess as having all aspects, beautiful as well as ugly.

With time, however, virupa- acquires also the negative sense of "deformed, ugly," in which case bahurupa virupa ca might be rendered as "multiform and misformed." This pair, however, does not fit together, neither in a complementary nor in a contrastive sense--unlike the well-matched "multiform and varicolored." If we were to assume that the Chinese had interpreted virupa as "misformed" and translated it as "ugly aspects" (chou rongyi [TEXT NOT REPRODUCIBLE IN ASCII]), an explanation would be required for how "multiform" (bahurupa) had become "beautiful aspects" (hao rongyi [TEXT NOT REPRODUCIBLE IN ASCII]). At best one could imagine that bahu, meaning "much, many, plentiful, great, large, abounding in," might have been understood as characterizing a form of abundant proportions, hence a "beautiful/good aspect," like that of Chinese beauty of the time.

It is in fact possible to understand Yijing's first line as both an interpretive rendering based on the more original meaning of virupa-, as well as a literal translation based on the later sense of the word. It is even conceivable that the twofold sense was aimed at in the Chinese.

This same virupa- reappears in the third pada of the Sanskrit, where the goddess is described as virupaksi: in the more original sense of virupa-, she is "varicolored-eyed," meaning that her eyes are of different colors; in the later sense of virupa-, she is "deformed-eyed," perhaps signifying cross-eyed or squint-eyed. Either way, an unnatural, and perhaps even frightening condition to behold, as we read in the corresponding Chinese: "your eyes can provoke fear in the beholder." The characterization of the goddess as visalaksi "wide-eyed" appears to have been rendered into Chinese together with the preceding virupaksi, as her wide eyes, like her deformed eyes, may well have been interpreted as something unnatural and hence provoking fear in the beholder.

As for Yijing's third line, which describes the goddess's countless superior/victorious attitudes/deeds dominating the world(s), it appears to be related to the second pada of the Sanskrit, "of many kinds of forms" (anekavidharupini), in that wuliang [TEXT NOT REPRODUCIBLE IN ASCII] "countless" corresponds to aneka "many." If xing [TEXT NOT REPRODUCIBLE IN ASCII] is interpreted as "attitudes" rather than "deeds," the meaning of the character comes closer to the "aspects" (rongyi [TEXT NOT REPRODUCIBLE IN ASCII]) of the goddess mentioned in the first line and the "forms" (rupa) assumed by her in the Sanskrit. The remaining characters, i.e., sheng [TEXT NOT REPRODUCIBLE IN ASCII] "superior" and yue shijian [TEXT NOT REPRODUCIBLE IN ASCII] "dominate the world(s)," may be fillers for the required seven-character line. Yijing's third line, therefore, might possibly be a translation of anekavidharupini.

The final pada of the stanza finds a far more evident match in the corresponding Chinese rendering: she who in the Sanskrit is the "protection of devotees" (bhaktanam pariraksani) is the one who saves all who have faith in her.
 Stanza 5 (lines a14-a15) Stanza 4 (lines 7-8)

 [TEXT NOT REPRODUCIBLE IN ASCII] parvatagresu ghoresu
 [TEXT NOT REPRODUCIBLE IN ASCII] nadisu ca guhasu ca /
 [TEXT NOT REPRODUCIBLE IN ASCII] vasas tava mahadevi
 [TEXT NOT REPRODUCIBLE IN ASCII] vanesupavanesu ca //

 You dwell in deep and dangerous places On wild mountain peaks,
 on mountain cliffs,
 in caves and on river banks. on rivers and in caves,
 You dwell among large trees and in your dwelling, great goddess,
 [O] goddess, you mostly dwell there. is in forests and in groves.

This stanza describes the natural dwellings of the goddess. The original meaning of ghora is "wild," but it also conveys the sense of "horrifying," i.e., the subjective reaction of the perceiver to the wild environment. Yijing's manuscript conveys precisely the same sense by using the expression shenxian [TEXT NOT REPRODUCIBLE IN ASCII] "deep and dangerous."
 Stanza 6 (lines a16-a17) Stanza 5 (lines 9-10)

 [TEXT NOT REPRODUCIBLE IN ASCII] sabarair barbarais caiva
 [TEXT NOT REPRODUCIBLE IN ASCII] pulindais ca supujita /
 [TEXT NOT REPRODUCIBLE IN ASCII] mayurapaksadhvajini
 [TEXT NOT REPRODUCIBLE IN ASCII] lokan kramasi sarvasah //

 Should folk from mountains, forests, By Sabaras, Barbaras,
 and wilds
 constantly honor [you, O] goddess, and Pulindas, well-honored,
 And using peacock's feathers make with a standard of peacock
 standards, (54) feathers,
 you will, at all times, constantly you stride the worlds in all
 govern [their] worlds. their parts.

The Sabaras were a wild tribe of the Deccan; the Barbaras were barbarians of the lowest origin, and the Pulindas were likewise a wild tribe. The Sabaras and the Pulindas dwelt in mountains and forests, and their names eventually became interchangeable, representing wild tribes in general. (55) Clearly for metrical reasons, instead of rendering their three names phonetically, Yijing and his collaborators referred to them as peoples dwelling in three places: mountains, forests, and wilds (i.e., wide open spaces, like fields, but uncultivated). (56)

When the goddess is described as striding the worlds in all their parts (lokan kramasi sarvasah), the sense is that she covers the worlds with her steps. This depiction intentionally recalls the three steps taken by Visnu, referred to already in the Rg Veda (e.g., 1:22:17-18), and thereafter by his incarnation as Vamana ("dwarf"). Through their strides, they produce and preserve the worlds. (57) This pada is rendered in the Chinese as: "you will, at all times, constantly govern [their] worlds." While yu yiqie shi [TEXT NOT REPRODUCIBLE IN ASCII] may be a translation of nityasas "constantly, always," a variant reading for sarvasas "completely, in all its parts," the remainder of the Sanskrit pada, if Yijing's manuscript in fact had the same reading (lokan kramasi), is conveyed interpretatively rather than literally. The goddess does not stride; she governs, supervises, and thereby protects the worlds (hu shi [TEXT NOT REPRODUCIBLE IN ASCII]). Among the principal meanings of the character hu [TEXT NOT REPRODUCIBLE IN ASCII] is jianshi [TEXT NOT REPRODUCIBLE IN ASCII] "watch" or "supervise," "control," "govern." The expression hu shi became a Buddhist term generally understood as "protecting the world(s)," but it is clearly a "protection" derived from "watching" or "control." Evidently this term was carefully chosen for its connotations and represents, if Yijing had the reading kramasi, his interpretation of the goddess's striding of the worlds (lokan kramasi).

Nobel's understanding of hu shi, on the other hand, is that the goddess simply protects the world ("... so beschutzt sie ... die Welt."), for he claims that kramasi in the Sanskrit is doubtless to be read as trayasi "you protect," to which, he continues, it is palaeographically close. (58) Although his interpretation would clearly simplify matters in terms of the Chinese, it is not possible to justify it for the Sanskrit: amongst the thirty-seven (59) manuscripts in eight different scripts and the four printed editions consulted for the critical edition of the Harivamsa, trayasi is not listed in the notes of the critical edition as a variant reading for kramasi. The reading trayasi, furthermore, would greatly reduce the richness of connotation and intended suggestiveness inherent in kramasi, which was perhaps overlooked by Nobel. The author(s) of the Harivamsa hymn, as we have already seen, very carefully, purposefully, and effectively chose his (their) words of praise, invoking Durga with epithets and descriptions resonant of the Vedic tradition so as to render her respectable and venerable in the orthodox circles of the twice-born (arya).
 Stanza 7 (lines a18-a19) Stanza 6 (lines 11-12)

 [TEXT NOT REPRODUCIBLE IN ASCII] kukkutais chagalair mesaih
 [TEXT NOT REPRODUCIBLE IN ASCII] simhair vyaghraih samakula /
 [TEXT NOT REPRODUCIBLE IN ASCII] ghantaninadabahula
 [TEXT NOT REPRODUCIBLE IN ASCII] visruta vindhyavasini //

 Lions, tigers, and wolves always Thronged by cocks, goats, sheep,
 surround [you];
 cows, goats/sheep, and cocks, lions, and tigers,
 besides, obey you.
 Shaking large bells and emitting dense with the sounds of bells,
 the multitudes of the Vindhya widely heard, dwelling in the
 mountains all hear the echo. Vindhya [mountains],

As a goddess of the wilds, she is surrounded with throngs of wild animals. The order of the list of animals is not the same in the Chinese as it is in the Sanskrit, and the Chinese list is in fact longer, including also wolves and cows. The character yang [TEXT NOT REPRODUCIBLE IN ASCII] means sheep as well as goats, both of which are included in the Sanskrit (chagala; mesa).

The Sanskrit term ghanta refers here to a suspended bell with an interior clapper. (60) It is an essential feature of Hindu temples, hung at the gateway of small shrines and in the foyer of larger temples, to be rung by incoming worshippers. The stanza, however, may not describe a goddess temple, but rather the goddess in her wild habitat in the Vindhya mountains, thronged by animals and accompanied by the dense sounds of ghanta. This bell, furthermore, may be suspended or hand-held, as parallel textual evidence indicates: in another early hymn to Durga, also in the Harivamsa, the goddess is described as holding the ghanta in her hand(s), and also as wearing garlands of [suspended] ghanta. (61)

The Chinese rendered ghanta as da lingduo [TEXT NOT REPRODUCIBLE IN ASCII] "large bells." The late Neolithic ling [TEXT NOT REPRODUCIBLE IN ASCII] was a small bell with an internal clapper, intended for suspension by means of the loop in its crown. It would be attached, for instance, to a shaman's belt, and later to horse-and-chariot gear and to canopy curtains. Subsequent types of ling, however, appear to have been handheld. The duo [TEXT NOT REPRODUCIBLE IN ASCII], dating possibly from the Shang dynasty (ca. 1700-1040 B.C.E.), was a hand-held clapper-bell with a cylindrical shank. (62)

We are told in Yijing that these bells are shaken (zhen [TEXT NOT REPRODUCIBLE IN ASCII]), and that the echo of the sound emitted by their shaking is heard by the inhabitants of the Vindhya mountains. This description suggests, then, that the bells are not sounded by the devotees of the goddess, but rather that the bells, if they are suspended, are attached to the clothing, ornamentation, or body of the goddess, producing sound as she moves. It is unlikely that the undomesticated, wild animals of the mountains should wear bells. This depiction would point, therefore, to the ancient type of ling and to the garland of ghanta mentioned above. As for the hand-held duo, it would correspond to the hand-held ghanta. The da lingduo should probably be understood as "large bells." (63) But if we were dealing with both ling and duo, we would have, most interestingly, a very nice match with the above-mentioned passage from another Harivamsa hymn to Durga, where the goddess holds both ghanta in her hand(s), as well as wears them suspended on (a) garland(s).

In the final pada of the Sanskrit the goddess is described as visruta "widely heard, famous," a well-chosen epithet literally linked with the immediately preceding ghantaninadabahula "dense with the sound of bells," and as vindhyavasini "dwelling in the Vindhya [mountains]." Yijing, in his Chinese rendering, clearly brings out the sense of visruta, conveying the thickness of the sound echoing through the goddess's mountainous habitat: as she shakes the bells, the sounds they emit carry their echo to the multitudes dwelling in the Vindhya (Pintuo [TEXT NOT REPRODUCIBLE IN ASCII]) mountains, who hear them.

Vindhyavasini was a local goddess worshipped in the geographical area of the Vindhya mountains by its tribal inhabitants, enumerated in the previous stanza. (64) The center of her worship today is a temple in the bustling pilgrimage town of Vindhyachal, midway between Allahabad and Varanasi, where the Vindhya mountain range reaches the river Ganges. (65)
 Stanza 8 (lines a20-a21) Stanza 7 (lines 13-14)

 [TEXT NOT REPRODUCIBLE IN ASCII] trisulapattisadhara
 [TEXT NOT REPRODUCIBLE IN ASCII] suryacandrapatakini /
 [TEXT NOT REPRODUCIBLE IN ASCII] navami krsnapaksasya
 [TEXT NOT REPRODUCIBLE IN ASCII] suklasyaikadasi priya //

 Carrying a trident, a round topknot on the Carrying a trident and a
 head, spear,
 on the right and on the left always holding sun-and-moon-bannered,
 sun and moon banners;
 the dark [fortnight] of the moon, on the as the ninth [lunar day],
 ninth and eleventh [lunar] days, [you are] dear to the
 dark fortnight,
 during this time, one should honor [you]. as the eleventh, [you
 are] dear to the bright

As if ready for battle, the goddess carries trident and spear, and has sun and moon banners as her emblem--rather than the previously mentioned peacock-feather standard (Yijing's line a17 in stanza 6 and the Sanskrit line 10 of stanza 5). While the Sanskrit reads "trident" (trisula) and "sharp-edged spear" (pattisa) (66) for the weapons she wields, the Chinese has only one weapon, the trident (sanji [TEXT NOT REPRODUCIBLE IN ASCII]), followed by a description of her hairstyle, consisting of "a round topknot on her head" (touyuanji [TEXT NOT REPRODUCIBLE IN ASCII]). (67) The odd combination of weapon and hairstyle in Yijing suggests either a different wording in his Sanskrit manuscript or some kind of misreading. While a pattisa or pattisa is a "sharp-edged spear," a patta, on the other hand, may mean a "fillet or cloth worn round the head, a turban." Although one could imagine the Chinese might have read pattisa as patta, the leap from a turban to a topknot is rather considerable.

The goddess, in the Sanskrit, is identified with two lunar days, and the pada may be read in two ways. Either, as translated above, that as the ninth, she is dear (priya) to the dark fortnight, and as the eleventh, she is dear to the bright fortnight. Or, taking priya separately, (68) as the Chinese have interpreted it: "As the ninth [lunar day] of the dark fortnight and the eleventh of the bright [fortnight,] you are dear [to us]." Variant readings include switches among bright and dark fortnights, as some manuscripts read navami suklapaksasya "as the ninth [lunar day] of the bright fortnight" for the third pada and others krsnasyaikadasi tatha "so also [are you] the eleventh [lunar day] of the dark [fortnight]" for the fourth pada. (69) The first of these variants (navamai suklapaksasya) would actually accord with the great Durga Puja (Navaratra "nine nights") performed during the first nine lunar days of the bright fortnight in the month of Asvin (September-October) and culminating on the ninth. (70) Yijing's manuscript seems to have had the second variant (krsnasyaikadasi tatha), (71) which places both the ninth and the eleventh in the dark fortnight. It is on these days that the goddess is to be worshipped, the Chinese tells us.
 Stanza 9 (lines a22-a23) Stanza 8 (lines 15-16)

 [TEXT NOT REPRODUCIBLE IN ASCII] bhagini vasudevasya
 [TEXT NOT REPRODUCIBLE IN ASCII] rajani kalahapriya /
 [TEXT NOT REPRODUCIBLE IN ASCII] avasah sarvabhutanam
 [TEXT NOT REPRODUCIBLE IN ASCII] nistha ca parama gatih //

 Manifesting as the younger sister of the Sister of Vasudeva,
 great god Vasu (Vasudeva),
 seeing battle, [your] heart is always Night, fond of quarrel,
 Among those who look at all sentient beings Abode of all beings,
 the goddess is the most victorious and [their] apex and supreme
 unsurpassable one. goal.

Yijing reads "great god Vasu" (Posu da tian [TEXT NOT REPRODUCIBLE IN ASCII]), which indicates that the reference is not to the person known as Vasudeva, the father of Krsna, who is not a god; but to his son, the great god Vasudeva Krsna. The basic text adopted by the Taisho edition has the character nu [TEXT NOT REPRODUCIBLE IN ASCII] following "great god Vasu." In this case, the meaning would be "great goddess Vasu" or "daughter of the great god Vasu." However, note 7 (p. 437) indicates that the editions of the Song, Yuan, and Ming dynasties, the old Song edition (1104-48) belonging to the library of the imperial household, and the Saidaiji manuscript have, more specifically, mei [TEXT NOT REPRODUCIBLE IN ASCII], younger sister. As we see in the Sanskrit hymn, the goddess is in fact Vasudeva Krsna's sister (bhagini vasudevasya), and hence I think that the character mei must have been the correct reading. This line refers to the goddess's role in the birth story of Krsna. (72)

The second pada in the Sanskrit describes the dark goddess as night personified (rajani) and as fond of quarrel (kalahapriya). In Yijing, on the other hand, like a compassionate goddess, her heart is grieved to see battle. The second and the third pada, however, are a better match, as she who, in the Sanskrit, is the abode, the apex, and the supreme goal of all beings, in the Chinese is, among those who look to all sentient beings, (73) the most superior/victorious and unsurpassable one. What is interesting is that Yijing's version ("the goddess is the most victorious and unsurpassable one") suggests a warrior goddess, like the one fond of quarrel (kalahapriya) described in the second pada of the Sanskrit, rather than the one whose heart is grieved to see battle. Her compassionate grief, in context, seems a little amiss. This victorious, unsurpassable goddess, furthermore, already appeared both in Yasogupta/Jnanagupta's translation in the Baogui edition of the Sutra of Golden Light (p. 387c20 [TEXT NOT REPRODUCIBLE IN ASCII]), as well as in Yijing (p. 437b26 [TEXT NOT REPRODUCIBLE IN ASCII]).
 Stanza 10 (lines a24-a25) Stanza (lines 17-18)

 [TEXT NOT REPRODUCIBLE IN ASCII] nandagopasuta caiva
 [TEXT NOT REPRODUCIBLE IN ASCII] devanam vijayavaha /
 [TEXT NOT REPRODUCIBLE IN ASCII] ciravasah suvasas ca
 [TEXT NOT REPRODUCIBLE IN ASCII] raudri sandhya tvam eva ca //

 Manifesting as the daughter of the Daughter of the cowherd Nada,
 cowherd Nanda, (74)
 when fighting together with the gods, bringing about the decisive
 you always obtain victory [for them]. victory of the gods,
 Able to live long and in peace in the bark-clad and well-clad,
 you are harmony and endurance; you are precarious twilight are you.
 even violence and evil.

Nobel considers muniu [TEXT NOT REPRODUCIBLE IN ASCII] to refer to the goddess, but clearly understands that huanxi nu [TEXT NOT REPRODUCIBLE IN ASCII] is the daughter of Nanda. (75) As we see in the Sanskrit, the cowherd in question is in fact male (gopa) and he is called Nanda ("joy"). Yijing renders the cowherd's (muniu) name literally as huanxi [TEXT NOT REPRODUCIBLE IN ASCII] ("joy"). The final nu [TEXT NOT REPRODUCIBLE IN ASCII] of the Chinese line is a translation of suta, both meaning "daughter."

The first two pada refer to the goddess's role in the birth story of Krsna. (76) Nanda is the spouse of Yasoda, Krsna's foster mother. The victory of the gods, including Krsna, is brought about through the goddess, who is born of Yasoda and then changes place with Krsna, thus sparing Krsna's life at the hands of evil king Kamsa, deluding him (mohayitva ca tam kamsam ...) (77) to his eventual death at the hands of Krsna.

For the third pada, it would seem that Yijing's manuscript read cira "long [time]" instead of cira "bark," (78) leading to the divergence between the Harivamsa's "clothed in bark" (ciravasah) and Yijing's "living for long" (ciravasa; jiuzhu [TEXT NOT REPRODUCIBLE IN ASCII]). This discrepancy in turn led to a differing interpretation of suvasah "well-clothed" as suvasa "living well," rendered into Chinese as "living safely" (anzhu [TEXT NOT REPRODUCIBLE IN ASCII]). (79)

The final pada identifies the goddess with the dangerous, mysterious transition period between the day and the night: precarious twilight (raudri sandhya). Yijing seems to have interpreted raudri in the sense of "violent," for his translation reads bao'e [TEXT NOT REPRODUCIBLE IN ASCII] "violence," or, if the characters are taken separately, "violence and evil." Instead of twilight (sandhya), Yijing has heren [TEXT NOT REPRODUCIBLE IN ASCII] "harmony and patience," which likely corresponds to sandhi "peace." This discrepancy may have resulted from a different reading (possibly a scribal error) in his manuscript, or as a misreading of sandhya. (80)
 Stanza 11 (lines a26-a27) (81)


 The teachings of the four Vedas of the great brahmanas,
 the magical incantations, etc., you have penetrated (them) all
 Among gods and hermits (82) you have attained sovereignty.
 You are able to make the seeds and the great earth.

Yijing's stanza 11 is particularly difficult to match up with Sanskrit stanzas in the Harivamsa hymn, and I think that it is one of two (together with Yijing's stanza 19) stanzas in the Chinese for which there is no corresponding Sanskrit. (83)
 Stanza 12 (lines a28-a29) Stanza 12 (lines 23-24)

 [TEXT NOT REPRODUCIBLE IN ASCII] siddhih samyatrikanam ca
 [TEXT NOT REPRODUCIBLE IN ASCII] vela sagarayayinam /
 [TEXT NOT REPRODUCIBLE IN ASCII] yaksanam prathama yaksi
 [TEXT NOT REPRODUCIBLE IN ASCII] naganam suraseti ca //

 At the time of the assembling of the [You are] the success of
 goddesses, travelers
 like the tide of the great sea you and the seashore of sea-farers,
 will certainly come in response.
 Among dragon deities (naga) and the first yaksi among the yaksa,
 you are the head, who can subjugate and "Surasa" among the naga,

Although we are dealing, at least certainly in the second half, with one and the same stanza, the Sanskrit manuscript Yijing was working from did not have the same wording as what appears in the critical edition of the Harivamsa. The Chinese says that at the time of the assembling of the goddesses, the goddess will certainly come in response, like the tide of the great sea, either in its regularity or in its majesty. The Harivamsa verse, on the other hand, identifies the goddess as the success of travelers (84) and as the seashore of sea-farers, to which they inevitably return. In other words, here too she responds to the needs of her worshippers, whether in travel or in safe return from sea-faring. However, the sea imagery is used quite differently: while in the Chinese the goddesses are like the tide of the ocean coming to shore, in the Sanskrit the goddess is not the sea, but the shore of the sea-farers.

Whereas the first half of the stanza in our two versions is related through the dependability of the goddess and through the (different) use of sea imagery, the second half is far more closely connected. In the Sanskrit the goddess is the prominent female amongst both yaksa and naga: she is the number one yaksi amongst yaksa--perhaps, one might conjecture, the wife of Kubera, god of wealth and chief of the yaksa; she is also Surasa, the demoness (raksasi) mother of the naga. While neither the foremost yaksi nor Surasa are mentioned in Yijing, the goddess is identified as the head of the naga ("dragon deities" longshen [TEXT NOT REPRODUCIBLE IN ASCII]) and the yaksa (yaocha [TEXT NOT REPRODUCIBLE IN ASCII]), capable of subjugating them.
 Stanza 13 (lines b1-b2) Stanzas 13-14 (lines 25-27)

 [TEXT NOT REPRODUCIBLE IN ASCII] brahmavadiny atho diksa
 [TEXT NOT REPRODUCIBLE IN ASCII] sobha ca parama tatha /
 [TEXT NOT REPRODUCIBLE IN ASCII] jyotisam tvam prabha devi
 [TEXT NOT REPRODUCIBLE IN ASCII] naksatranam ca rohini //
 rajadvaresu tirthesu
 nadinam samgamesu ca /

 Among women you are the supreme Brahman A speaker of Brahman, and
 practitioner. also initiation,
 When you utter words, you are like the likewise the supreme beauty,
 ruler of the world.
 you are the light of the
 luminaries, O goddess, and
 Rohini among the
 constellations. (85)
 In the places where kings dwell, you are At the gates of kings, at
 like a lotus. (86) fords,
 If you are in a river ford, you can be
 compared to a bridge or a raft. and at the confluences of

The first half of the stanza is, again, only loosely related. It would seem that Yijing had, in his Sanskrit manuscript, brahmacarini (= fanxing [TEXT NOT REPRODUCIBLE IN ASCII]) "Brahman practitioner," i.e., a celibate woman, instead of brahmavadini "Brahman speaker." In fact, there is a manuscript of the Harivamsa (S1) that has the alternate reading brahmacarini, but the rest of the pada remains the same as above and hence does not agree with Yijing's version. The diksa, with which the goddess is also identified in the first pada of this Harivamsa stanza, is an initiation ceremony for the study of the Vedas. Yijing's version, on the other hand, identifies her as the supreme brahmacarini among women. The brahmacarin as the individual in the first, i.e., student, stage of life, incidentally, is the one required to receive the diksa in order to begin his studies in the Vedas.

The second pada of the Sanskrit then identifies the goddess as supreme loveliness (sobha ca parama tatha), whereas Yijing turns to the power of her words: when she speaks, she is like the ruler of the world (Shijianzhu [TEXT NOT REPRODUCIBLE IN ASCII]), who is Prajapati. (87) As the Brahmanical creator Prajapati produces the world by pronouncing words, the goddess's speech is likewise most potent and powerful, and hence renders her comparable to him. (88) In addressing her speech, Yijing's version comes closer to the first pada's brahmavadini "Brahman speaker." The Sanskrit's "supreme loveliness," on the other hand, calls to mind Yijing's characterization of the goddess as the "supreme brahmacarini among women." From the point of view of contents, reversing the order of the two first pada would render a slightly better, but still imperfect by far, match. Meter, however, does not allow for such a switch.

There are, therefore, plenty of connections between the Chinese and the Sanskrit, but Yijing's translation represents a different version of the Sanskrit than what we find in the critical edition of the Harivamsa. The last two pada of the Sanskrit do not appear in the Chinese.

The second half of the stanza in Yijing corresponds to the two first pada of the following stanza of the Harivamsa. We have here a list of places where the goddess dwells: at the royal gates (rajadvaresu) or, in Yijing, in the dwellings of kings ([TEXT NOT REPRODUCIBLE IN ASCII]), in river fords (tirthesu; [TEXT NOT REPRODUCIBLE IN ASCII]), and in the Sanskrit also at the confluence of rivers (nadinam samgamesu ca). While in the Sanskrit the first two pada are incomplete in and of themselves, Yijing's version explains what the goddess may be likened to in these places: in the dwellings of kings, she is like a lotus; and in river fords she is like a bridge or a raft to assist people to the other shore.
 Stanza 14 (lines b3-b4) Stanzas 14-15 (lines 28-29)

 [TEXT NOT REPRODUCIBLE IN ASCII] purne ca purnimacandre
 [TEXT NOT REPRODUCIBLE IN ASCII] krttivasa iti smrta //
 [TEXT NOT REPRODUCIBLE IN ASCII] sarasvati ca valmikeh
 [TEXT NOT REPRODUCIBLE IN ASCII] smrtir dvaipayane tatha /

 Your face is like the full moon. And when the moon of the full
 moon night is full,
 The learning you are endowed with [you are] remembered as
 constitutes the abode on which one "Krttivasas" ('[antelope-]
 relies. hide-clad'),
 [As] eloquence you excel like High and the eloquent speech
 Peak (Valmiki), (sarasvati) of Valmiki,
 [and as] memory you are just like likewise the memory in
 Island (Dvaipayana). Dvaipayana (Vyasa),

Krttivasas is Siva, who, while engaged in meditation, wears or sits on the hide of an antelope, and hence is called "hide-clad." His spouse is therefore known as the feminine Krttivasas. Siva's spouse Parvati is also an ascetic practitioner, who engages in austerities to win her husband. (89)

If we compare these pada with Yijing's version, we do find a full moon in his first line, but the connection ends here. In the Chinese, the goddess's face is compared to the full moon, and her learning is the abode people rely on. (90) Clearly, Yijing had a different reading. While what follows in the two versions may at first sight appear unrelated, we do in fact have a match. In the Sanskrit, the abilities of the traditionally accepted authors of the two epics, the Ramayana and the Mahabharata, are attributed to the goddess: she is the sarasvati of Valmiki and the memory of Krsna Dvaipayana Vyasa. She is identified with the goddess of speech/knowledge Sarasvati, who is the speech that eloquently expresses knowledge, and the knowledge eloquently expressed through speech. Yijing translates sarasvati as biancai [TEXT NOT REPRODUCIBLE IN ASCII] "eloquence," which is necessarily dependent on knowledge. She is also identified with the outstanding memory of Dvaipayana, whose feats of compilation are believed to include the Vedas, the Mahabharata, and the Puranas. In the Chinese, the names of Valmiki and Dvaipayana do not appear in phonetic transcription, but in literal translation: valmiki "of a termite hill (valmika)" (91) finds its correspondent in gaofeng [TEXT NOT REPRODUCIBLE IN ASCII] "high peak," a considerably larger hill, it would seem; and dvaipayana "of an island (dvipa)" (92) in the entirely literal rendering zhouzhu [TEXT NOT REPRODUCIBLE IN ASCII] "island." (93)

A termite hill measures, on average, two to three meters in height, (94) enough to cover a seated human body, but hardly what one might call a "high peak." Nevertheless, in ancient cosmogony it is commonly described as a "high place" in relation to the waters from which it emerges and from which it is believed to swell. (95) In order to maintain a very high level of humidity within the mound, some species of termites will penetrate more than forty meters below the surface of the earth to reach water and bring it up. (96) This being said, we do not know how well versed Yijing and his collaborators might have been, either on termite-mound cosmogony or in the lifestyle of termites, but the absence of the word "termite" in the Chinese translation might suggest that their manuscript could have had a slightly different Sanskrit reading. It is, however, also possible that they simply aimed to express the idea of a mound--albeit in exaggerated terms--rather than the literal meaning.
 Stanza 15 (lines b5-b6) Stanzas 15-16 (lines 30-31)

 [TEXT NOT REPRODUCIBLE IN ASCII] suradevi ca bhutesu
 [TEXT NOT REPRODUCIBLE IN ASCII] stuyase tvam svakarmabhih //
 [TEXT NOT REPRODUCIBLE IN ASCII] indrasya carudrstis tvam
 [TEXT NOT REPRODUCIBLE IN ASCII] sahasranayaneti ca /

 The asura and the multitude of And [as] the goddess of liquor
 gods (deva) (sura) among the spirits,
 in unison praise these merits you are praised by your own deeds.
 [of yours]
 up to [your] thousand-eyes, and You are a lovely sight to
 Sovereign Shi (Sakra), head Indra, (97)
 [of the gods],
 looks [at you] with earnest and and [you are called]
 deep heart. "thousand-eyed."

Instead of suradevi ca "and [as] the goddess of liquor," Yijing's Sanskrit manuscript probably read sura-adeva "gods and non-gods" in the locative plural suradevesu, a variant reading noted in the critical edition of the Harivamsa. (98) This reading would account for the asura (phonetically rendered as asuluo [TEXT NOT REPRODUCIBLE IN ASCII]) and deva (literally translated as tian [TEXT NOT REPRODUCIBLE IN ASCII]) in Yijing's first line. (99) The goddess is praised, in the Sanskrit, by her own deeds (stuyase tvam svakarmabhih), perhaps originally "because of her good/virtuous deeds (su-karma-)," a reading that comes closer to the second line of the Chinese, where her merits (gongde [TEXT NOT REPRODUCIBLE IN ASCII]) are praised.

In the third pada of the Sanskrit, the goddess is, to Sakra (Indra), a lovely, pleasing, even beloved, sight (carudrsti). She is called, after Indra, "thousand-eyed" (sahasranayana). The implication is that she is identified as Indrani, Indra's beloved consort, for which we find confirmation on line 44 of this same hymn, where she is in fact called Indrani. In Yijing's version, her thousand eyes are the topmost ("up to" naizhi [TEXT NOT REPRODUCIBLE IN ASCII]) of her merits, mentioned in the second line. Indra looks upon the goddess "with earnest and deep heart": gauncha [TEXT NOT REPRODUCIBLE IN ASCII] "look" corresponds to drsti "sight," and [TEXT NOT REPRODUCIBLE IN ASCII] can be traced to caru in the sense of "beloved, dear."
 Stanza 16 (lines b7-b8) Stanza 19 (line 37) and
 stanza 18 (line 36)

 [TEXT NOT REPRODUCIBLE IN ASCII] asa tvam manusanam tu
 [TEXT NOT REPRODUCIBLE IN ASCII] tustis ca krtakarmanam /
 (line 37)
 [TEXT NOT REPRODUCIBLE IN ASCII] svaha tustir dhrtir medha
 [TEXT NOT REPRODUCIBLE IN ASCII] vasunam tvam vasumati //
 (line 36)

 When beings hope and wish for something, You are the hope of men
 you have full power to allow them to and the contentment of those
 realize it swiftly. who have done their duty.
 (line 37)
 Also you grant quick understanding and Svaha, contentment,
 eloquence, and you provide hearing and steadfastness, wisdom;
 keeping (dharani).
 You are the first in the great earth. among the Vasus you are the
 wealthy one. (line 36)

In the first pada, the hope (asa, xiqiu [TEXT NOT REPRODUCIBLE IN ASCII]) of people (manusa "men"; zhongsheng [TEXT NOT REPRODUCIBLE IN ASCII] "beings") unites the Chinese translation with the Sanskrit. In the second line, the two versions stray further apart, but there remains a certain sense of completion (krtakarma) or fulfillment vaguely relating them. In the third line, we have a list of qualities, with which the goddess is identified in the Sanskrit and which she grants in the Chinese. The exclamation svaha is uttered when offering oblations into the sacred fire. In personified form, Svaha is the spouse of the fire god Agni (Mahabharata 3:220:1-7). In Chinese it is usually rendered phonetically as Suohe [TEXT NOT REPRODUCIBLE IN ASCII] (e.g., p. 435a17). Neither svaha nor delight (tusti) appear in the corresponding Chinese. The noun dhrti "taking hold of, maintaining, firmness, resolution, steadfastness," from the root dhr "to hold, bear, maintain," corresponds to wenchi [TEXT NOT REPRODUCIBLE IN ASCII] "hearing and keeping," that is, dharani, likewise from the root dhr. The term wenchi is also related in meaning to medha in the sense of "retentiveness," (100) but medha, in addition to "wisdom," also means "intellect, intelligence," which corresponds to cong [TEXT NOT REPRODUCIBLE IN ASCII] "sharpness of mind." However, "eloquence" (bian [TEXT NOT REPRODUCIBLE IN ASCII]) is not to be found in the Harivamsa stanza. (101) While svaha, as already noted, is left untranslated in the Chinese, it is in fact related to dharani (wenchi), in that it generally appears within dharani.

The fourth pada in the Sanskrit calls the goddess the wealthy one among the Vasus, a class of deities. The substantive vasu means "wealth, riches," and vasumati (in our text, vasumati, as meter requires) "she who has wealth" is the earth. (102) Some manuscripts of the Harivamsa read, instead, vasumdhara, likewise meaning "earth." The pada may then also mean: "you are the earth among the Vasus" (vasunam tvam vasumati). The earth (vasumati, vasumdhara) is what connects the Sanskrit with the Chinese, which reads "you are the first in the great earth." Yijing's version, however, is so different in meaning that their manuscript must have had a variant reading, but perhaps one which preserved vasumati or vasumdhara?
 Stanza 17 (lines b9-b10) Stanza 19 (line 38)

 [TEXT NOT REPRODUCIBLE IN ASCII] disas ca vidisas caiva
 [TEXT NOT REPRODUCIBLE IN ASCII] tatha hy agnisikha prabha //

 In these worlds of the ten directions, [You are] the [four]
 directions and the
 intermediate directions,
 You are like a great light always so also fire-crested Prabha
 and everywhere reflecting (light, lustre).
 up to the gods and demons, (103)
 birds and beasts;
 all attain their heart's desire.

If the intermediate directions (vidisa-) are considered to be four, then, together with the four directions (disa-) and the zenith and the nadir, we arrive at the ten directions mentioned in the Chinese. The directions are said to be ten already in Satapatha Brahmana 6:2:2:34 and 8:4:2:13: dasa disah.(104) There is no match in the Harivamsa for the second half of Yijing's stanza.
 Stanza 18 (lines b11-b12) Stanzas 21-22 (lines 42-43)

 [TEXT NOT REPRODUCIBLE IN ASCII] narinam parvati ca tvam
 [TEXT NOT REPRODUCIBLE IN ASCII] pauranim rsayo viduh //
 [TEXT NOT REPRODUCIBLE IN ASCII] arundhaty ekabhartrnam
 [TEXT NOT REPRODUCIBLE IN ASCII] prajapativaco yatha /

 Among women, you are like Mountain Among women you are Parvati.
 Peak (Parvati).
 Like ancient hermits, you stay Seers know [you] as the ancient
 long in the world. one. (105)
 As a young goddess, you are Among monogamous women, [you are]
 always desireless. Arundhati,
 The real words are just like according to the words of
 [those] of the great lord of Prajapati.
 the world (Maha Prajapati).

Parvati is literally "she who is of the mountain (parvata)," in the sense that she is the daughter of Himalaya, the highest of mountains. The Chinese translate Parvati's name as shanfeng [TEXT NOT REPRODUCIBLE IN ASCII] "mountain peak," or simply "mountain," since feng means both "peak" and "mountain." Although shan would have been sufficient to translate parvata, from which the name Parvati is derived, we have seen that there is, in this Chinese translation, a clear preference for binominal expressions in the rendering of names: Gaofeng [TEXT NOT REPRODUCIBLE IN ASCII] for Valmiki, Zhouzhu [TEXT NOT REPRODUCIBLE IN ASCII] for Dvaipayana, and Biancai [TEXT NOT REPRODUCIBLE IN ASCII] for Sarasvati.

The reference to the goddess as ancient (paurani) in the Sanskrit also makes allusion to the Puranas. (106) In the Chinese she becomes one who "stays long in the world." The seers (rsi) of the Harivamsa, furthermore, are rendered as the "ancient hermits" in Yijing. The Chinese could have been working from a manuscript with a variant Sanskrit reading, and, if so, I would suggest that paurana may have appeared as an attribute of the seers, rather than of the goddess. They might also have produced an interpretive translation because a greater number of characters was required for the stanza: the seers, who are known to be ancient, knew the goddess to be likewise, i.e., that she lives long in the world.

According to a note in the Taisho edition (p. 437, n. 14), shao nutian [TEXT NOT REPRODUCIBLE IN ASCII], literally "young goddess," is "Kaumari" (sic). Both the consort of Skanda Kumara (Karttikeya) and Durga are known as Kaumari. (107) Durga is also called kanya kumari "girl-virgin," and this is probably what the editors of the Taisho edition had in mind. (108) Yijing's shao nutian is most likely a rendering of kumari (or kumari devi), as the editors of the Taisho edition (despite the diacritical mark under the "m" and the double derivation) suggest.

Arundhati, wife of sage Vasistha, is the model of conjugal excellence and wifely devotion. (109) Therefore, among women with one husband, Arundhati is the greatest, and thus the goddess is identified with her in her excellence. In Yijing, on the other hand, there is no sign of Arundhati. Instead of being likened to a virtuous, devoted married woman, the goddess is compared to an unmarried "young goddess" for her desirelessness. An entirely different line must have appeared in Yijing's Sanskrit manuscript.

In the final pada of this stanza, there is in fact agreement between the Harivamsa and Yijing. What precedes this pada, we are told--and it is not clear if this means only the previous pada or more--is on the authority of, or in the Chinese "just like," the words of Prajapati "lord (pati) of the progeny/creatures (praja)." (110) In Yijing he is called the "great lord of this world" (Da Shizhu [TEXT NOT REPRODUCIBLE IN ASCII]), that is, the great (maha) Prajapati, as noted in the Taisho edition (p. 437, n. 15). A more usual appellation of the great Prajapati is Da Shengzhu [TEXT NOT REPRODUCIBLE IN ASCII] "great lord of living beings," (111) which accords more closely with the Sanskrit "lord of progeny" (praja-pati).
 Stanza 19 (lines b13-14)


 Looking everywhere at the different classes of beings in the world
 up to the celestial palaces in the world of desire,
 there is only the goddess who alone is called venerable.
 One fails to see beings capable of surpassing her.

Yijing's stanza 19 finds no corresponding pada in the Sanskrit.
 Stanza 20 (lines b15-b16) Stanza 23 (lines 46-47)

 [TEXT NOT REPRODUCIBLE IN ASCII] samgramesu ca sarvesu
 [TEXT NOT REPRODUCIBLE IN ASCII] agniprajvalitesu ca /
 [TEXT NOT REPRODUCIBLE IN ASCII] naditiresu ghoresu
 [TEXT NOT REPRODUCIBLE IN ASCII] kantaresu bhayesu ca //

 If in fearful places of battling In all confrontations [of armies]
 or fallen into a fire pit, and conflagrations of fire,
 in river fords that are dangerous on frightful river-banks,
 and difficult, when there are in forests and dangers,
 bandits and thieves,
 in all situations you can make it
 so that one removes [one's own]

The addition of "fearful places" (kongbu chu [TEXT NOT REPRODUCIBLE IN ASCII]) in the first line of Yijing may have occurred because more characters were required for the verse. The third and the fourth pada of the Sanskrit are squeezed into the third line of Yijing's stanza. There are numerous variant readings in the Harivamsa manuscripts for these two last pada. (112) Instead of ghoresu "terrible," we find, most notably, cauresu "amidst thieves" and core va "or when there is a thief," corresponding to the Chinese zeidao shi [TEXT NOT REPRODUCIBLE IN ASCII]. Variants for the final pada include vanesupavanesu ca "in forests and groves"; vanesu ca guhasu ca "in forests and caves"; bhayesu ca ripusv api "in dangers and also amidst enemies"; guhasu ca vanesu ca "in caves and forests." The forests, however, do not appear in the Chinese. Instead, Yijing's fourth line describes the goddess's ability to make the devotee remove his own fear, a line which we do not find in the Sanskrit and which does not correspond to any of the variant readings listed in the notes to the critical edition of the Harivamsa.
 Stanza 21 (lines b17-b18) Stanza 24 (lines 48-49)

 [TEXT NOT REPRODUCIBLE IN ASCII] pravase rajabandhe ca
 [TEXT NOT REPRODUCIBLE IN ASCII] satrunam ca pramardane /
 [TEXT NOT REPRODUCIBLE IN ASCII] pranatyayesu sarvesu
 tvam hi raksa na samsayah //

 Put in a pillory due to the king's Abroad and in the king's prison,
 or murdered by enemies, and in the crushing by
 enemies, (113)
 if one can single-pointedly
 concentrate one's mind [on you]
 without displacing it,
 one will certainly be freed from in all risks to life,
 all distresses.
 you are protection, no doubt.

In the first pada, rajabandhe "in bondage to the king" corresponds to Yijing's "put in a pillory due to the king's law." There is no equivalent for pravase "abroad" in the Chinese, but perhaps Yijing's manuscript might have read pravada-, in the sense of "proclamation," instead of pravasa-, which might then be understood as the proclamation of the king's law. The second pada, satrunam ca pramardane "and in the crushing by enemies" is rendered in Chinese as "or murdered by enemies." (114) There is no equivalent in the Sanskrit of Yijing's third line, "if he can single-pointedly concentrate his mind [on you] without displacing it." A Sanskrit passage closely resembling this Chinese line appears instead in line 50 (stanza 25) of the Harivamsa: tvayi me hrdayam devi tvayi buddhir manas tvayi / "My heart is [absorbed] in you, O goddess, [my] intellect in you, [my] mind in you." Whereas the Sanskrit emphasizes the protection bestowed on the devotee by the goddess in these dangerous situations, in Yijing the focus is on the concentration efforts of the individual, which will free him from distress. (115)
 Stanza 22 (lines b19-b20) Stanza 25 (lines 51, 50)

 [TEXT NOT REPRODUCIBLE IN ASCII] raksa mam sarvapapebhyah
 [TEXT NOT REPRODUCIBLE IN ASCII] prasadam ca karisyasi //
 (line 51)
 [TEXT NOT REPRODUCIBLE IN ASCII] tvayi me hrdayam devi
 [TEXT NOT REPRODUCIBLE IN ASCII] tvayi buddhir manas tvayi /
 (line 50)

 Good and bad people, you protect Protect me from all sins,
 them all
 compassionate and merciful, you always and [I pray] you will show
 appear before [their eyes]. [me] mercy. (line 51)
 Therefore, with utmost sincerity, My heart is [absorbed] in you,
 O goddess,
 I bow to and take refuge in you, [my] intellect in you, [my]
 great goddess. mind in you. (line 50)

While in the Sanskrit the reciter invokes the goddess for his own protection from sins, in Yijing Kaundinya tells of the goddess's protection extending to all people, good and bad. (116) The second half of Yijing's stanza communicates the brahmana's devotion to her, expressed also in the last two pada of the Sanskrit.


Comparative analysis of Yijing's translation with the Harivamsa hymn reveals, as we have seen, numerous different readings that must have appeared in Yijing's Sanskrit manuscript, some entirely unknown lines, and a differing verse order for more than half of the hymn. By modifying the order of some of the stanzas and pada of the Sanskrit to match Yijing's sequence of stanzas, it has been possible to ascertain, in large part, the version of the hymn that appeared in Yijing's no longer extant Sanskrit manuscript. On the whole, we can say that Yijing's Sanskrit original was remarkably close to the Harivamsa hymn and must not have been based on a radically different edition, but rather on one of the various editions of the hymn being circulated.

Yijing's version of the hymn represents a good example of early eighth-century Chinese translation techniques when faced with verse restrictions. Where Yijing's manuscript had the same reading as the Harivamsa hymn, the faithful Chinese rendering produced by the translation team attests to their high level of competence. A comparison of the Sanskrit version of the Harivamsa hymn with Yijing's Chinese rendering, furthermore, leads to a better understanding of both the Sanskrit and the Chinese. In terms of the Sanskrit, a comparison with Yijing's hymn requires a closer reading and hence a better understanding of the literal meaning of the Sanskrit. In the case of the Chinese, it allows for a greater comprehension of the literal meaning that the characters are intended to convey and for a recognition of the Indian names rendered literally in characters. Moreover, by understanding the original context of the hymn as an invocation of Durga in relation to the birth story of Krsna, it is possible to make sense of the contents of the Chinese version of the hymn, despite its entirely re-contextualized placement in the "Biancai tiannu" chapter of Yijing's rendering of the Sutra of Golden Light.



I thank Profs. Antonino Forte of the Universita degli Studi di Napoli "L'Orientale" and Werner Knobl of Kyoto University for their detailed comments and stimulating discussions on this article. I also wish to express my gratitude to two anonymous readers of the journal for their helpful comments. Just before the proofs of this article arrived, Prof. Yuko Yokochi of Kyoto University kindly provided comments, only a few of which I have been able to incorporate here. I am likewise most grateful to her for sending me her recently completed Ph.D. dissertation, "The Rise of the Warrior Goddess in Ancient India: A Study of the Myth Cycle of Kausiki-Vindhyavasini in the Skandapurana" (Univ. of Groningen, 2004), where she also addresses the female divinities of the Harivamsa. According to Yokochi, the goddess invoked in the Harivamsa hymn discussed here is Nidra-Vindhyavasini, who evolves into the warrior goddess popularly called Durga.

1. See Thomas B. Coburn, Devi-Mahatmya: The Crystallization of the Goddess Tradition (Delhi: Motilal Banarsidass, 1984). For the dating of the Harivamsa, see Daniel H. H. Ingalls, "The Harivamsa as a Mahakavya" in Melanges d'Indianisme a la memoire de Louis Renou (Paris: De Boccard, 1968), 394, and Andre Couture, L'enfance de Krishna (Paris: Cerf, 1991), 77.

2. Jinguangming zuishengwang jing [TEXT NOT REPRODUCIBLE IN ASCII], 10 juan, T. vol. 16, no. 665. German translation by Johannes Nobel, Suvarnaprabhasottamasutra: Das Goldglanz-Sutra, ein Sanskrittext des Mahayana-Buddhismus: I-tsing's chinesische Version und ihre Ubersetzung, i: I-tsing's chinesische Version (Leiden: E. J. Brill, 1958).

T. = Taisho shinshu Daizokyo [TEXT NOT REPRODUCIBLE IN ASCII], compiled under the direction of Takakusu Junjiro [TEXT NOT REPRODUCIBLE IN ASCII] and Watanabe Kaigyoku [TEXT NOT REPRODUCIBLE IN ASCII], vols. 1-85 (Tokyo: Issaikyo kankokai, 1924-32).

3. See Catherine Ludvik, "La Benzaiten a huit bras: Durga deesse guerriere sous I'apparence de Sarasvati," Cahiers d'Extreme-Asie 11 (1999-2000): 292-338, and "From Sarasvati to Benzaiten" (Ph.D. diss., University of Toronto, 2001), 248-78.

4. See below, p. 712.

5. For the dating of the Mahabharata, I follow Alf Hiltebeitel, Rethinking the Mahabharata: A Reader's Guide to the Education of the Dharma King (Chicago: Univ. of Chicago Press, 2001), 18.

6. The Harivamsa, ed. Parashuram Lakshman Vaidya (Poona: Bhandarkar Oriental Research Institute, 1971), 2: 34-37, app. I, no. 8. Partial French translation (chapters 30-78) by Andre Couture, L'enfance de Krishna (henceforth cited as Krishna). The methods and decisions of the editors of the critical editions, including that of the Mahabharata and that of the Harivamsa, have been called into question. See, e.g., Madeleine Biardeau, "Mythe epique et hindouisme d'aujourd'hui," Indologica Taurinensia 5 (1977): 43-53, and "L'arbre sami et le buffle sacrificiel," in Autour de la Deesse hindoue, ed. Biardeau (Paris: Ecole des Hautes Etudes en Sciences Sociales, 1981), 215-43, for discussions of two of the early Durga hymns from the Mahabharata.

7. Further confirmation of the goddess's identity as Durga appears later in Harivamsa 65:49-57.

8. Suvarnabhasottamasutra: Das Goldglanz-Sutra, ein Sanskrittext des Mahayana-Buddhismus, ed. Johannes Nobel (Leipzig: Otto Harrassowitz, 1937). All references to R. E. Emmerick's translation herein are to his third (revised) edition, The Sutra of Golden Light: Being a Translation of the Suvarnabhasottamasutra (Oxford: Pali Text Society, 1996).

9. See, e.g., Lionel Giles, Descriptive Catalogue of the Chinese Manuscripts from Tunhuang in the British Museum (London: The British Museum, 1957), 53-62.

10. It should be pointed out, however, that until 730 there were five canonical versions published over a period of almost three centuries (early fifth to early eighth). Suffice to say that the history of these versions, which I hope to address in a future publication, is extremely complex.

11. Sengyou [TEXT NOT REPRODUCIBLE IN ASCII] (445-518), Chu sanzang ji ji [TEXT NOT REPRODUCIBLE IN ASCII], 15 juan, T. vol. 55, no. 2145, 2.11b17. See n. 12 of the same page for the date, where we read that the Song, Yuan, and Ming editions of the Buddhist Canon specify that Dharmaksema's translation appeared in the fifth month of the sixth year Xuanshi [TEXT NOT REPRODUCIBLE IN ASCII], that is between June 1-29, 417. Chen Jinhua, "The Indian Buddhist Missionary Dharmaksema (385-433): A New Dating of His Arrival in Guzang and His Translation," T'oung Pao 90 (2004): 215-63, discusses the issue of Dharmaksema's arrival in the Northern Liang (r. 397-439) capital of Guzang [TEXT NOT REPRODUCIBLE IN ASCII] (present-day Wuwei [TEXT NOT REPRODUCIBLE IN ASCII], Gansu), and in this connection calls into question the date of 417 for Dharmaksema's translation of the Sutra of Golden Light.

12. See the colophon of the sutra discovered in Dunhuang reproduced in Antonino Forte, Political Propaganda and Ideology in China at the End of the Seventh Century, 2nd ed. (Naples: Universita degli Studi di Napoli "L'Orientale," and Kyoto: Italian School of East Asian Studies, forthcoming), pl. 33.

13. Nobel, in his notes to his German translation of Yijing's Chinese (henceforth abbreviated as I-tsing), provides, where he considers relevant, some of the wording used in two Tibetan translations: Tib. II is a ninth-century translation from Sanskrit, and Tib. III, an early-ninth-century translation from Yijing's Chinese. On these Tibetan versions, see Nobel's Suvarnaprabhasottamasutra: Das Goldglanz-Sutra, ein Sanskrittext des Mahayana-Buddhismus: Die Tibetischen Ubersetzungen mit einem Worterbuch, i: Die Tibetischen Ubersetzungen (Leiden: E. J. Brill, 1944) and Suvarnaprabhasottamasutra: Das Goldglanz-Sutra, ein Sanskrittext des Mahayana-Buddhismus: I-tsing's chinesische Version und ihre Ubersetzung, ii: Die Tibetische Ubersetzung (Leiden: E. J. Brill, 1958). See also Claus Oetke, Die aus dem chinesischen ubersetzten tibetischen Versionen des Suvarnaprabhasasutra (Wiesbaden: Steiner, 1977) for a study of the Tibetan translation from Yijing's Chinese, where Oetke discusses further Tibetan renderings of the sutra. There are, moreover, Sogdian, Old Uighur, Tangut, Mongolian, and Old Turkic translations of Yijing, which are likewise beyond the scope of this study. For recent scholarship on the Sutra of Golden Light, see, e.g., Kanaoka Shuyu [TEXT NOT REPRODUCIBLE IN ASCII], Konkomyokyo no kenkyu [TEXT NOT REPRODUCIBLE IN ASCII] (Tokyo: Daito shuppansha, 1980); Natalie Dawn Gummer, "Articulating Potency: A Study of the Suvarna- (pra)bhasottamasutra" (Ph.D. diss., Harvard University, 2000); and Seishi Karashima and Jiang Zhongxin, "Sanskrit Fragments of the Sutra of Golden Light from the Lushun Museum Collection," Hualin [TEXT NOT REPRODUCIBLE IN ASCII] 3 (2003): 331-81.

14. Note that in Dharmaksema's translation the chapter consists only of this first part.

15. For a study of the Sarasvati chapter as found in the extant Sanskrit and in the Chinese translations, see Ludvik, "From Sarasvati to Benzaiten," 227-78.

16. T. vol. 16, no. 665 (Yijing's translation of the sutra), p. 437, n. 1.

17. Hobogirin, Hobogirin [TEXT NOT REPRODUCIBLE IN ASCII]: Dictionnaire encyclopedique du bouddhisme d'apres les sources chinoises et japonaises, fasc. I (Paris: Adrien-Maisonneuve and Tokyo: Maison Franco-Japonaise, 1929), 64.

18. Watanabe Kaigyoku, "Kokuyaku Konkomyosaishookyo" [TEXT NOT REPRODUCIBLE IN ASCII], in Kokuyaku Daizokyo [TEXT NOT REPRODUCIBLE IN ASCII], vol. 13 (Tokyo: Kokumin bunko kankokai, 1932), 145, note. Although the Harivamsa forms a supplement to the epic, it remains an entirely separate, independent text. Hence it is not referred to as the Mahabharata, but always as the Harivamsa. One might assume, therefore, that Watanabe meant the Mahabharata and only the Mahabharata.

19. Nobel, I-tsing, 249-50, n. 3 (which begins on p. 248). The edition of the Harivamsa used by Nobel is the one edited by Nimaichandra Siromani, Ramagovinda, and Ramahari Nyaya Panchanan, published in Calcutta in 1839.

20. These stanzas correspond, in the critical edition of the Harivamsa and its supplements, to the introductory stanza for the hymn (vol. 1, 327, note 591* lines 3-4), 47:54 (vol. 1, 327), followed by lines 3-18 (vol. 2, 34-35, app. I, no. 8). The stanza that according to Nobel is not translated in Yijing appears in lines 1-2 (vol. 2, 34, app. I, no. 8).

21. Nobel, I-tsing, 248-49, n. 3.

22. Nagano Sadako, "Konkomyokyo ni okeru 'Benzaiten' no seikaku" [TEXT NOT REPRODUCIBLE IN ASCII], Indogaku bukkyogaku kenkyu [TEXT NOT REPRODUCIBLE IN ASCII] 36.2 (1988): 239 (720). Yudhisthira's hymn to Durga appears in an appendix to the critical edition of the Mahabharata: The Mahabharata, ed. V. S. Sukthankar, S. K. Belvalkar, and P.L. Vaidya et al. (Poona: Bhandarkar Oriental Research Institute, 1933-66), vol. 5 (Virata Parvan, around 4:5), app. I, no. 4, 300-305. Takayasu Suzuki, "Rites and Buddhism--A Perspective from the Sarasvatiparivarta in the Suvarnabhasa," Indogaku bukkyogaku kenkyu 52.2 (2004): 13 (942) follows Nagano on this point.

23. See Ludvik, "La Benzaiten a huit bras," 304-13; and "From Sarasvati to Benzaiten," 252-62.

24. See p. 714 below.

25. Hobogirin, I: 64-65. "Narayani" on p. 64 should read "Narayani" (lengthening of the second syllable). Also on p. 64, "Elle se montre aussi sous la forme de Vasudeva ..." is an error: she does not appear as Vasudeva (Krsna) himself, but as the sister of Vasudeva. See pp. 722-23 below.

26. Nobel, I-tsing, 249, n. 3.

27. Ibid. also draws attention to a hymn from the island of Bali, in which the goddess Indrani is invoked under a plurality of names.

28. See p. 714 below. In the context of this Harivamsa hymn, I cannot agree with Nobel's suggestion that the term Narayani has a folk element to it (I-tsing, 249, n. 3). I would say that it is used here with the precisely opposite effect in mind.

29. Personal communication of 4 December 2002. I would like to express my gratitude to Prof. Pulleyblank for his comments. Baogui's edition of the Sutra of Golden Light, in Yasogupta/Jnanagupta's translation of the Da Biantianshen [TEXT NOT REPRODUCIBLE IN ASCII] chapter (T. vol. 16, no. 664, 388a4), explicitly states that a gatha consists of "seven words/characters" ([TEXT NOT REPRODUCIBLE IN ASCII]), as we see below in the Durga hymn.

30. See Oda Tokuno [TEXT NOT REPRODUCIBLE IN ASCII], Bukkyo dai jiten [TEXT NOT REPRODUCIBLE IN ASCII] (1917; rev. rpt. Tokyo: Daizo shuppan, 1995), 193b, 980c.

31. Yijing's translation of the sutra, T. vol. 16, no. 665, 437a2-5.

32. I would like to thank the anonymous reader who brought this important point to my attention.

33. A certain number of manuscript colophons contain the list of the translators. They present some discrepancies as to the number and names of the translators. The colophon of S. 523 (Giles, Catalogue, 58, no. 2102) is reproduced in Forte, Political Propaganda, 2nd ed., pl. 33 (private communication). According to Forte, the last two monks of the list of eighteen (including Yijing) mentioned in S. 523 may have been added in 705 at the time when an imperial preface was written.

34. On Manicintana, see Forte, "The Activities in China of the Tantric Master Manicintana (Pao-ssu-wei [TEXT NOT REPRODUCIBLE IN ASCII]: ?-721 A.D.) from Kashmir and of his Northern Indian Collaborators," East and West n.s. 34.1-3 (1984): 301-47.

35. On Srimata, see Forte, "Manicintana," 327-29.

36. On the translation process in China, see, for instance, Kenneth Ch'en, "Some Problems in the Translation of the Chinese Buddhist Canon," Qinghua xuebao [TEXT NOT REPRODUCIBLE IN ASCII] [Tsing Hua Journal of Chinese Studies], n.s. 2.1 (1960): 178-88; Jian Tang, "Historical Stages of the Buddhist Religious Movement before the Seventh Century China and the Translation Processes of the Mahayana Sutras," in Proceedings of the Thirteenth International Symposium on Asian Studies, 1991 (Hong Kong: Asian Research Service, 1992), 159-72; Daniel Boucher, "Gandhari and the Early Chinese Buddhist Translations Reconsidered: The Case of the Saddharmapundarikasutra," JAOS 118 (1998): 471-506.

37. August Barth, CEuvres de Auguste Barth, IV: Comptes rendus et notices (1887-1898) (Paris: Ernest Leroux, 1918), 453, expressed this opinion as early as 1898: "I-tsing [Yijing] parait avoir eu une assez bonne connaissance pratique du sanscrit."

38. Private communication.

39. Yijing's Chinese was previously translated by Nobel into German (I-tsing, 250-56), and the Harivamsa hymn was rendered into English by Coburn (Devi-Mahatmya, 279-81) and into French by Andre Couture (Krishna, 353-55). Where warranted, references to these translations appear in my notes to the hymn.

40. In preparation.

41. Appears in Harivamsa, crit. ed., 1: 327, note.

42. Vaisampayana is the narrator.

43. The character xian [TEXT NOT REPRODUCIBLE IN ASCII] (or the combination xianren [TEXT NOT REPRODUCIBLE IN ASCII]) is the most common way of rendering the Sanskrit rsi. In Yijing's line, it is also possible to understand wangxi xianren [TEXT NOT REPRODUCIBLE IN ASCII] together, as Nobel (I-tsing, 250) has done: "die fruheren Seher" ("the ancient seers").

44. On the remarkably rich suggestiveness of these carefully chosen names and epithets, see my Recontextualizing the Praises of a Goddess.

45. The Arya Stava appeared in all manuscripts used by Parashuram Lakshman Vaidya for his critical edition of the Harivamsa, except for three Malayalam palm-leaf manuscripts ([M.sub.1-3]) representing the southernmost manuscript tradition of Kerala (see crit. ed., 1: 327, note; 2: xxx and 34).

46. Nobel (I-tsing, 250) translates congming [TEXT NOT REPRODUCIBLE IN ASCII] as "die schnelle Auffassungsgabe" ("quick understanding") and claims there is no Sanskrit correspondent, suggesting that it is doubtful whether samnati is correct--presumably as a match for congming ("Es ist zweifelhaft, ob samnati richtig ist."). His rendering of the Chinese hymn into German, furthermore, has the goddess in the third person throughout. In accordance with the Sanskrit, I have translated the entire hymn as addressed to the goddess in the second person.

47. As Nobel (I-tsing, 250) translates them: "Schamhaftigkeit."

48. Nobel, ibid., translates as "Als Mutter bringt sie die Welt hervor" ("as mother she brings forth the world").

49. Coburn, Devi-Mahatmya, 279. Couture (Krishna, 353) does not translate Siddhasena.

50. The noun tapas, from the root tap "to heat," is, at its most literal, "heating," and hence energy or inner vitality.

51. My interpretation of this line differs from Nobel's (I-tsing, 250). He understands it as "(Durch ihren) Heldemut vollbringt sie immerdar (Taten der) Energie" ("Through her valor, she always accomplishes [deeds of] energy").

52. Nobel, (I-tsing, 251) translates zhangyang [TEXT NOT REPRODUCIBLE IN ASCII] as "das volle Gedeihen" ("full prosperity") and hence equates it with pusti; tiaofu [TEXT NOT REPRODUCIBLE IN ASCII] as "die vollstandig Bezahmung" ("complete restraint") and suggests it corresponds to vijaya.

53. It occurs several times in the Rg Veda (e.g., 1:73:7c), often in connection with the night, which is black, and dawn, which is pink. Rig-Veda-Samhita: The Sacred Hymns of the Brahmans together with the Commentary of Sayanacharya, 2nd ed., 4 vols., ed. F. Max Muller (1890-92; rpt. Varanasi: Chowkhamba Sanskrit Series Office, 1966).

54. While the character fan [TEXT NOT REPRODUCIBLE IN ASCII] "flag, standard" (in [TEXT NOT REPRODUCIBLE IN ASCII]) is adopted in the Taisho edition, a note (437, n. 4) indicates that the editions of the Song, Yuan, and Ming dynasties, the old Song edition (1104-48) belonging to the library of the imperial household, and what I assume to be the Nara-period (eighth-century) manuscript at Saidaiji in Nara (abbreviated [TEXT NOT REPRODUCIBLE IN ASCII]) have the alternate character chuang [TEXT NOT REPRODUCIBLE IN ASCII], also meaning "flag, banner, standard." On the banner, see "Ban [TEXT NOT REPRODUCIBLE IN ASCII]" in Hobogirin, I: 49-50. Regarding the edition marked as [TEXT NOT REPRODUCIBLE IN ASCII] in the Taisho edition, it is unfortunately omitted from the list of abbreviations at the end of vol. 16. In the Showa hobo somokuroku [TEXT NOT REPRODUCIBLE IN ASCII] (Taisho shinshu Daizokyo: Bekkan [TEXT NOT REPRODUCIBLE IN ASCII], Tokyo: Taisho Issaikyo kankokai and Daizo shuppan, 1929-34), 1: 285c, we read that, among the texts used for the collation of Yijing's version of the Sutra of Golden Light by the editors of T. no. 665, there was a manuscript from Saifukuji [TEXT NOT REPRODUCIBLE IN ASCII] and another one from Saidaiji [TEXT NOT REPRODUCIBLE IN ASCII]. We are left wondering, then, which of the two monasteries is indicated by the abbreviation [TEXT NOT REPRODUCIBLE IN ASCII]. However, since in some of the volumes of the Taisho edition the abbreviation for Saifukuji is given as [TEXT NOT REPRODUCIBLE IN ASCII], it is to be surmised that [TEXT NOT REPRODUCIBLE IN ASCII] is the abbreviation for Saidaiji. On the text of Yijing's version of the Sutra of Golden Light at Saidaiji, see Showa hobo somokuroku, 1: 1072b5.

55. Otto Bohtlingk and Rudolph Roth, Sanskrit-Worterbuch, 7 vols. (1855-75; rpt. Delhi: Motilal Banarsidass, 1990).

56. Nobel's translation of this passage (I-tsing, 252) reads: "Falls Stamme von wilden Menschen aus den Berg-waldern" ("Should tribes of wild people from mountain forests"). It does not convey the sense of the three tribes dwelling, in Yijing, in three locations.

57. For a fuller discussion of the significance of Visnu's and Vamana's strides and of the suggestiveness of the use of lokan kramasi sarvasah in relation to the goddess in the Harivamsa hymn, see my Re-contextualizing the Praises of a Goddess. On Visnu in Vedic mythology, see Arthur Anthony Macdonell, Vedic Mythology (1897; rpt. Delhi: Motilal Banarsidass, 1965), 37-42. On the Vamana incarnation, see Vettam Mani, Puranic Encyclopaedia: A Comprehensive Dictionary with Special Reference to the Epic and Puranic Literature (Delhi: Motilal Banarsidass, 1975), 823-25.

58. Nobel, I-tsing, 252, n. 3.

59. By my count the list of manuscripts used for the critical edition of the Harivamsa totals thirty-seven (see crit. ed., vol. 1, x-xii). The editor, however, says there are thirty-six (vol. 1, x).

60. On the different types of ghanta and the time periods to which they belong, see Claudie Marcel-Dubois, Les Instruments de musique de l'Inde ancienne (Paris: Presses Universitaires de France, 1941), 32-36; Stanley Sadie, ed., The New Grove Dictionary of Music and Musicians, 2nd ed. (London: Macmillan Press, 2001), 9: 799-800, s.v. "Ghanta."

61. ghantahastam namasyami ghantamalakulam tatha (Pradyumna's hymn to Durga, Harivamsa, crit. ed., vol. 2, 378, app. I, no. 30, line 370).

62. Lothar von Falkenhausen, Suspended Music: Chime-Bells in the Culture of Bronze Age China (Berkeley and Los Angeles: Univ. of California Press, 1993), 68, 122, 133-36; Yuan-Yuan Lee and Sin-Yuan Shen, Chinese Musical Instruments (Chicago: Chinese Music Society of North America, 1999), 46; Sadie, New Grove Dictionary, 2: 658, 27: 808. Pairs of small, hand-held bells attached with a cord running through the loop in their crown and struck together are described in Tang-period literature, where they are called pengling, shuanling, etc., and they are depicted in fifth- to sixth-century cave reliefs (Sadie, 2: 658).

63. Da Tang xiyuji [TEXT NOT REPRODUCIBLE IN ASCII] (by Xuanzang [TEXT NOT REPRODUCIBLE IN ASCII] [d. 664] and Bianji [TEXT NOT REPRODUCIBLE IN ASCII], T. vol. 51, no. 2087), 2: 879b20 and 8: 913b2, e.g., mentions the compound lingduo and the impression is that it indicates just the bells that hang from the corners of stupas. Li Rongxi, The Great Tang Dynasty Record of the Western Regions (Berkeley: Numata Center for Buddhist Translation and Research, 1996), 70, appropriately translates lingduo as "bells," but curiously has "bells and chimes" (235) as the translation of the second instance above.

64. On Vindhyavasini, see Cynthia Ann Humes, "Vindhyavasini: Local Goddess yet Great Goddess," in Devi: Goddesses of India, ed. John Stratton Hawley and Donna Marie Wulff (Berkeley and Los Angeles: Univ. of California Press, 1996), 49-76, and Yuko Yokochi, "The Goddess in the Krsna Legend: Reconsidered," Studies in the History of Indian Thought (Indo shisoshi kenkyu [TEXT NOT REPRODUCIBLE IN ASCII]) 13 (2001): 42-46. Vindhyavasini in Somadeva's Kathasaritsagara of the eleventh century is discussed in Fabrizia Baldissera, "Candika/Candi, Vindhyavasini and Other Terrific Goddesses in the Kathasaritsagara," in Wild Goddesses in India and Nepal: Proceedings of an International Symposium, Berne and Zurich, November 1994, ed. Axel Michaels, Cornelia Vogelsanger, and Annette Wilke (Bern: Peter Lang, 1996), 71-103.

65. Humes, "Vindhyavasini," provides a colorful description of the pilgrimage to the Vindhyachal temple.

66. Couture (Krishna, 354) understands the pattisa to be "le foudre" ("the thunderbolt").

67. Nobel (I-tsing, 252) translates touyuanji [TEXT NOT REPRODUCIBLE IN ASCII] as "einen runden Haarschopf" ("a round tuft of hair"). In Chinese yuanji by itself designates a round topknot or chignon. The specification tou "on the head" is redundant but is here needed to fill in the suitable number of characters in the verse.

68. As Coburn (Devi-Mahatmya, 279), and Couture (Krishna, 354) have done.

69. Manuscript [G.sub.1] (Baroda, Oriental Institute Library) has both of these variant readings: navami suklapaksasya krsnasyaikadasi tatha.

70. On the Durga Puja, see, e.g., David Kinsley, Hindu Goddesses: Visions of the Divine Feminine in the Hindu Religious Tradition (Berkeley and Los Angeles: Univ. of California Press, 1986), 106-15.

71. It is also conceivable that a simple copyist error may have occurred in the Chinese: [TEXT NOT REPRODUCIBLE IN ASCII] might have been written instead of [TEXT NOT REPRODUCIBLE IN ASCII], under the influence of [TEXT NOT REPRODUCIBLE IN ASCII] of [TEXT NOT REPRODUCIBLE IN ASCII].

72. See p. 708 above.

73. Nobel (I-tsing, 253) translates Yijing's third line as "Sie schaut hin auf alle Wesen" ("She looks towards all beings").

74. The editor in the Taisho edition (p. 437, n. 8) suggests [TEXT NOT REPRODUCIBLE IN ASCII] is "Gopinanda."

75. Nobel I-tsing, 253: "die Kuhhirtin, die Tochter des Nanda" ("the cowherd [fem.], daughter of Nanda").

76. See p. 708 above.

77. Harivamsa 47:56.

78. As noted also by Nobel (I-tsing, 253, n. 9).

79. Nobel (I-tsing, 252) translates [TEXT NOT REPRODUCIBLE IN ASCII] as "Lange Zeit verweilt sie ..." ("for a long time she dwells ...").

80. Nobel's translation (I-tsing, 253) of Yijing's last line reads: "... zeigt Friedfertigkeit und Geduld, (aber) auch Wut und Hass" ("she displays peace and forbearance, [but] also enmity and rage"). For the corresponding Sanskrit, he was convinced that samdhya (misprinted in his n. 9 on p. 253 as samdhya] was a corruption, and suggested instead saumya "lovely," or, less likely, santa "peaceful." He then connected ciravasa "bark/rag-clad" with raudri and suvasa "well-clad" with saumya. The reading saumya, however, does not appear as a variant for samdhya in the notes of the critical edition of the Harivamsa.

81. At this point Yijing's version begins to deviate from the order of the stanzas in the Sanskrit. In fact, the two Sanskrit stanzas that follow (lines 19-22, stanzas 10-11) are left out altogether in the Chinese version. As for the Sanskrit match to Yijing's stanza 11, the nearest approximation is a combination of stanzas.

82. Nobel (I-tsing, 254) understands tianxian [TEXT NOT REPRODUCIBLE IN ASCII] in Yijing's third line together as "die himmlischen Sehern" ("the heavenly seers"). Oda (Bukkyo dai jiten, 1251c), however, separates tian and xian.

83. For a discussion of the pada in the Harivamsa that are related to Yijing's stanza, see my Re-contextualizing the Praises of a Goddess.

84. Couture (Krishna, 354) translates the first pada as "On dit que tu es la bonne etoile des navigateurs" ("It is said that you are the good star of sailors").

85. Rohini is the fourth lunar mansion and the favorite consort of the moon. See John Dowson, A Classical Dictionary of Hindu Mythology and Religion, Geography, History, and Literature (1879; rpt. New Delhi: Oriental Books Reprint Corporation, 1973), 269.

86. While the character zhu [TEXT NOT REPRODUCIBLE IN ASCII] "dwell, stay" (in Yijing's third line) is adopted in the Taisho edition and is certainly the preferable one, it might be noted that the editions of the Song, Yuan, and Ming dynasties, and the old Song edition belonging to the library of the imperial household have the alternative character wei [TEXT NOT REPRODUCIBLE IN ASCII], meaning "position, rank, throne." See n. 11 on p. 437.

87. Nakamura Hajime [TEXT NOT REPRODUCIBLE IN ASCII], Bukkyogo daijiten [TEXT NOT REPRODUCIBLE IN ASCII] (Tokyo: Tokyo shoseki, 1981), 817b. Nobel (I-tsing, 254, n. 5) considers Shijianzhu to be Brahman, and refers also to Yijing's line b12 of stanza 18, where he identifies Da shizhu [TEXT NOT REPRODUCIBLE IN ASCII] likewise as Brahman (255, n. 7). Both Shijianzhu and Da shizhu (see below), however, refer to Prajapati. Perhaps Nobel had the Puranic creator god Brahma, a later derivation of the Brahmanical Prajapati, in mind.

88. On Prajapati and Speech, see Ludvik, "From Sarasvati to Benzaiten," 70-72.

89. On Parvati, see Kingsley, Hindu Goddesses, 34-54, esp. 41-46. Specifically on the asceticism-practicing Parvati, see N. P. Joshi, Tapasvini Parvati: Iconographic Study of Parvati in Penance (New Delhi: New Age International, 1996).

90. Nobel (I-tsing, 254) translates Yijing's second line as: "Mit vielem Ruhm ist sie ausgestattet und macht ihn zu ihrer Stutze und ihrer Statte" ("With many glories is she endowed and she makes them her support and resort").

91. On valmika as "termite hill," rather than ant-hill, see John C. Irwin, "The Sacred Anthill and the Cult of the Primordial Mound," History of Religions 21 (1982): 339, and especially Ditte Konig's extensive study on the mythology and worship of termite hills in India, Das Tor zur Unterwelt: Mythologie und Kult des Termitenhugels in der Schriftlichen und Mundlichen Tradition Indiens (Wiesbaden: Franz Steiner, 1984), an ancient cult which survives to the present day. As the story goes, following a career in thievery the celebrated sage Valmiki sat for years without moving, repeating the word mara ("Rama" inverted), until he was covered with a termite hill. When he emerged from the termite hill, he was called "Valmiki." On Valmiki, see Vaman Shivaram Apte, The Practical Sanskrit-English Dictionary, rev. edn., ed. P. K. Gode and C. G. Karve (1957; 4th rpt. Kyoto: Rinsen Book Co., 1998), 1418-19; Irwin, "Sacred Anthill," 356; Mani, Puranic Encyclopaedia, 822-23; and Konig, Das Tor, 198-200. On the practice of asceticism in termite hills, see Irwin, "Sacred Anthill," 356 and Konig, Das Tor, 198-209.

92. He was called Dvaipayana because he was born on an island in the Yamuna river. On Dvaipayana, see Mahabharata 1:54:2-6; 1:57:55-75; 1:99:6-15; Hiltebeitel, Rethinking the Mahabharata, esp. 32-91; and Bruce M. Sullivan, Krsna Dvaipayana Vyasa and the Mahabharata: A New Interpretation (Leiden: E. J. Brill, 1990).

93. Nobel (I-tsing, 254), on the other hand, translates Yijing's third and fourth lines as "Das ausgezeichnete Erscheinen ihrer Beredsamkeit ist wie ein hoher Berggipfel und macht sie zu einer Insel fur alle, die (ihrer) gedenken" ("The excellent appearance of her eloquence is like a high mountain peak and makes her an island for all who remember [her]").

94. Konig, Das Tor, 8.

95. Irwin, "Sacred Anthill," 344-47.

96. Ibid., 348. The air-conditioned mound makes it a favorite for cobras, who come and go through the ventilation shafts. Mineral-rich, translocated soil is also brought up by the termites--hence the legends of buried treasures in termite mounds guarded by snakes (Irwin, 352-53; Konig, Das Tor, 210-35).

97. Coburn (Devi-Mahatmya, 280) inaccurately translates this pada as: "the fair appearance of Indra," rather than to Indra. Couture (Krishna, 354) takes carudrsti as a name, which he renders as "la Femme aux beaux yeux" ("the woman with beautiful eyes").

98. See the critical edition, 2:36, notes.

99. Nobel (I-tsing, 254), on the other hand, translates the line as "die Gotterscharen der Asuras usw." ("the godly multitude of the Asuras, etc.").

100. Nobel (I-tsing, 255) translates ju wenchi [TEXT NOT REPRODUCIBLE IN ASCII] as "ein gutes Gedachtnis" ("good memory").

101. Nobel, ibid., translates congbian [TEXT NOT REPRODUCIBLE IN ASCII] as "eine schnelle Auffassungsgabe" ("quick comprehension").

102. Cf. Couture (Krishna, 355), who renders Vasumati as "la Terre" ("the Earth").

103. Nobel (I-tsing, 255) translates shengui [TEXT NOT REPRODUCIBLE IN ASCII], in the third line of Yijing's stanza, as "Damonen" ("demons").

104. The Satapatha Brahmana in the Madhyandina-Sakha with Extracts from the Commentaries of Sayana, Harisvamin, and Dvivedaganga, ed. Albrecht Weber (1855; rpt. Varanasi: Chowkhamba Sanskrit Series Office, 1964).

105. Coburn's translation of line 42 is inaccurate: "Among women (mentioned) in the Puranas, seers know you to be Parvati" (Devi-Mahatmya, 281).

106. Couture (Krishna, 355) renders Paurani as "Celle dont parlent les Purana" ("She of whom the Puranas speak").

107. On Durga as Kaumari, see Coburn, Devi-Mahatmya, 143-44, 313-30.

108. On Kaumari see Paul Thieme, "'Jungfrauengatte.' Sanskrit kaumarah patih ...," Zeitschrift fur Vergleichende Sprachforschung 78.3-4 (1963): 161-248; rpt. in Paul Thieme: Kleine Schriften (Wiesbaden: Franz Steiner, 1971), 2: 426-513.

109. See Apte, Sanskrit-English Dictionary, 219; Mani, Puranic Encyclopaedia, 56.

110. Nobel (I-tsing, 255) translates "Im Sprechen der Wahrheit ist sie wie der grosse Weltenherr" ("in speaking the truth she is like the great lord of the world"), and interprets Da Shizhu to be Brahman (255, n. 7). Cf. my n. 87 above.

111. Oda, Bukkyo dai jiten, 1138b; Nakamura, Bukkyogo daijiten, 1277d [for Maha-prajapati].

112. See the critical edition, 2:37, notes.

113. Coburn translates this pada as "trampling down of enemies" (Devi-Mahatmya, 281). In the context of the stanza, however, where situations precarious to the devotee are listed, it would seem more appropriate to understand satrunam ca pramardane as a difficulty in which he finds himself, i.e., where he is crushed by enemies, rather than engaged in crushing them, as Couture (Krishna, 355) understood the pada.

114. Nobel (I-tsing, 255-56): "oder wenn sie aus Hass getotet werden sollen" ("or when they are to be killed out of hate").

115. Line 50 of the Sanskrit, however, also comes close to Yijing's line b20 (see stanza 22), for which there is no alternative Sanskrit correspondent. Although the Harivamsa's line 50 accords more closely with Yijing's line b18 than with b20, in view of maintaining, as far as possible, the four-pada pattern in the matching of Sanskrit and Chinese stanzas, I have paired the Harivamsa's line 50 with Yijing's line b20.

116. Nobel (I-tsing, 256) translates Yijing's second line as "... steht ihnen (daher) aus Mitgefuhl und barmherzigem Gedenken immerdar vor Augen" ("she always stands before their eyes out of sympathy and compassionate remembrance").
COPYRIGHT 2004 American Oriental Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Ludvik, Catherine
Publication:The Journal of the American Oriental Society
Geographic Code:9INDI
Date:Oct 1, 2004
Previous Article:Did Chinggis Khan have a Jewish teacher? An examination of an early fourteenth-century Arabic text.
Next Article:On a five-colored cloud: the Songs of Mount Wutai.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters