The Basket’s Display
Introduction

The Kāraṇḍavyūha is an early Mantrayāna sūtra that is the source of the mantra oṁ maṇipadme hūṁ. The sūtra is thus of particular importance, as this mantra now holds a central role in the Tibetan Buddhist tradition, especially throughout the lay population. This sūtra also records Avalokiteśvara’s transformation into the principal figure of the Buddhist pantheon, greater than all other buddhas, let alone bodhisattvas. In this sūtra, Avalokiteśvara is a resident of Sukhavātī and acts as a messenger and gift bearer for Amitābha, even though he is also described as superior to all buddhas and therefore paradoxically has both a subservient and dominant status.

The appearance in writing of the Kāraṇḍavyūha probably dates to around the fifth century ᴄᴇ. In terms of place, the text indicates familiarity with the cesspits of Vārāṇasī, and assumes the reader’s knowledge of Candradvīpa, the southern part of Bengal where the Ganges Delta is situated. In the Tibetan version, the merchants who wish to sail to Laṅka ask whether the winds are blowing toward the land of the Greeks. This appears to locate their port of departure on the northwest coast of India. In terms of time, the text is located within a culture where the Mahābhārata and the Purāṇas had a dominant place in Indian culture, particularly the Skandha Pūraṇa, probably during the Gupta period of the third to fifth century.

The earliest surviving manuscript is comprised of fragmentary pages from two manuscripts discovered within a Gilgit stūpa in the 1940s. It was written in a hybrid of Middle Indic and Sanskrit, now called Buddhist Hybrid Sanskrit, which was frequently used in sūtras. Adhelheid Mette, who has published these fragments, suggests that it was composed in the fourth or fifth century;1 the script in which it is written had fallen out of use by the early seventh century, and the fragments show variations between the two manuscripts that are the result of the texts having gone through generations of copying. Other existing Sanskrit manuscripts (see below) date from a century or more later than the ninth century Tibetan translation.

According to Lokesh Chandra,2 in 270 ᴄᴇ Dharmarakṣa of Dunhuang translated the Kāraṇḍavyūha into Chinese. Then, between 435 and 443 ᴄᴇ, Guṇabhadra translated it into Chinese again. However, this is a case of misidentification. The sūtra they translated was the Ratna­karaṇḍa­vyūha. The Kāraṇḍavyūha itself was not translated into Chinese until 983 ᴄᴇ, considerably later than the Tibetan translation; the translator was T’ien Hsi-tsai.

The sūtra also exists in a later, longer, and more polished form, entirely in verse and incorporating passages from such texts as Śantideva’s Bodhi­sattva­caryāvatāra, which has great importance within Nepalese Buddhism. Dating to the fifteenth century, it is one of the last Sanskrit Buddhist sūtras. It has not been translated into Tibetan.

Avalokiteśvara is noticeable by his absence in early sūtras where Mañjuśrī figures prominently. In the Sukhāvatīvyūha or The Display of the Pure Land of Sukhāvatī,3 which describes the realm of Amitāyus, the buddha who later became known by the name Amitābha, Avalokiteśvara has yet to appear. He makes his first prominent appearance in the longer Sukhāvatīvyūha4 in which he stands beside Amitāyus as one of his two principal bodhisattva attendants. The other bodhisattva is Mahāsthāmaprāpta, and in a number of subsequent sūtras they are included as a pair in the introductory description of the assembly of those who are listening to the teaching. In one of the Kāraṇḍavyūha’s internal contradictions, both Mahāsthāmaprāpta and Avalokiteśvara are listed as being in the audience awaiting Avalokiteśvara’s appearance.

Each bodhisattva later had a chapter dedicated to him in the White Lotus of Compassion Sūtra,5 but while Avalokiteśvara reached preeminence over all buddhas in the Kāraṇḍa­vyūha, Mahāsthāmaprāpta declined in importance. In the Tibetan tradition, even in the Sukhāvatīvyūha, he has become conflated with Vajrapāṇi. At the time of the composition of the Kāraṇḍavyūha, Vajrapāṇi, who in earlier Buddhism was a powerful yakṣa, appears as one of the gathered bodhisattvas, which is indicative of sūtras that contain mantras. However, this is a recent development, as one of Avalokiteśvara’s qualities given in the sūtra is that he terrifies Vajrapāṇi! Vajrapāṇi would soon join Mañjuśrī and Avalokiteśvara to form the principal trinity of bodhisattvas in the early tantra tradition.

The Kāraṇḍavyūha does not mention Avalokiteśvara’s abode in this world on the Potalaka Mountain, which was a later feature that first appeared in South Indian Buddhism. The origin of the popular four-armed version of Avalokiteśvara appears within the sūtra as the goddess who is the embodiment of the six-syllable mantra, referred to throughout as a vidyā (which is a feminine noun) or often as the queen of mahāvidyās. Many forms of Avalokiteśvara appeared in India, such as the thousand-armed Avalokiteśvara included in fasting practice, and in the eleventh century there appeared the higher tantra form named Jinasāgara, a red, four-armed Avalokiteśvara in union with a consort. This practice was introduced into Tibet in the beginning of the twelfth century.

Eventually Avalokiteśvara practices spread throughout the Buddhist world. There are still ancient Avalokiteśvara statues even in Śrī Laṅka, though the figure is identified as Śiva in Tamil areas and as Maitreya in Buddhist temples. Avalokiteśvara was prominent in China for centuries before the Kāraṇḍavyūha was translated into Chinese. In particular Avalokiteśvara became a dominant figure in Chinese Buddhism as Kuan Yin (or Guanyin in Pinyin), transforming into a female bodhisattva, a process described by Chün-Fang Yü in Kuan-yin: The Chinese Transformation of Avalokiteśvara, as the result of focusing on his incarnation as the Princess Miao-chan.6

The Pillar Testament (Tib. bka’ chems ka khol ma) from the eleventh or twelfth century states that the Kāraṇḍavyūha was one of the texts that descended from the sky in a casket onto the palace roof of the fifth-century ruler of the Yarlung Valley, Lhathothori Nyentsen (Tib. lha tho tho ri gnyan btsan), and that during the reign of his descendant Songtsen Gampo (Tib. srong btsan sgam po), who became the king of most of the Tibetan plateau and introduced Buddhism to Tibet, it was translated by Thönmi Sambhota, the inventor of the Tibetan alphabet.7 In the thirteenth century Nelpa Paṇḍita, rejecting this legend, stated that the casket was brought by a paṇḍita on his way to China.8 However, he only records the maṇi mantra as being within the casket, which happens to be called a za ma tog or “a solid and precious casket” (rinchen za ma tog) and not a reed basket. Nevertheless, this is probably why this sūtra became associated with the legend.

The earliest and only translation of the sūtra appears to be the one presently in the canon. All of the versions of the Kangyur except one have a colophon ascribing the translation of the Kāraṇḍavyūha to Yeshé Dé and the Indian upādhyāyas Dānaśīla and Jinamitra, who collaborated with each other on the majority of their translations. The Narthang Kangyur (snar thang bka’ ’gyur) is alone in attributing the translation to Śākyaprabha and Ratnarakṣita.

Nanam Yeshé Dé (sna nam ye shes sde) was a Tibetan who became the principal translator in the translation program set up under the royal auspices of King Trisong Detsen (khri srong lde btsan, r. 742–798 ᴄᴇ). The translation work took place in a building dedicated to the translation program. It was situated within the circular compound of Samye (bsam yas) Monastery. Yeshé Dé’s name is in the colophon of no fewer than 347 texts in the Kangyur and Tengyur, three of which are his own original works in Tibetan. Jinamitra was invited to Tibet during the reign of Trisong Detsen, and 234 texts name him as Yeshé Dé’s co-translator. Dānaśīla, also known as Mālava, was invited to Tibet from Kashmir during the reign of Ralpachen (ral pa can, r. 815–838 ᴄᴇ) and was involved with the translation of around 165 texts. He was also the author of seven texts, five of which he helped translate. He was still active in Tibet during the reign of King Langdarma (glang dar ma, r. 838–841 ᴄᴇ).

Jinamitra and Dānaśīla were also two of the four or five Indian paṇḍitas who played principal roles in the completion of the Mahāvyuttpati, the Sanskrit-Tibetan concordance that was intended to regulate the translation of Sanskrit texts into Tibetan. Work on this dictionary began during the reigns of Trisong Detsen and Senaleg (sad na legs, r. 800–815 ᴄᴇ), but it was completed in the reign of Ralpachen. The catalog for the Tangtong Denkar Palace (pho brang thang stong ldan dkar) collection, which was compiled in 824 ᴄᴇ, lists the Kāraṇḍavyūha.

There is at least one instance in the Kāraṇḍavyūha where the translation does not accord with the Mahāvyuttpati. In describing the twenty peaks of the mountain that is the belief in the existence of an individual self in relation to the skandhas (“aggregates”), the peaks are described as samudgata, which the Mahāvyuttpati translates as “high” (Tib. mtho ba). In the Kāraṇḍavyūha, however, it is translated as “arisen” (Tib. byung ba). Unless the translators changed their minds, this would appear to identify the translation as having taken place before the Mahāvyuttpati was completed. Therefore we can say that the translation was certainly made during the decade between 815 and 824 ᴄᴇ, and presumably in the earlier part of that decade, around 820 ᴄᴇ or earlier. Neither Yeshé Dé nor Jinamitra are specified to have lived beyond the end of Ralpachen’s reign in 824 ᴄᴇ. Yeshé Dé’s remains are said to be interred within a stūpa on Hepori Hill next to Samye Monastery, where he worked on so many translations.

A later translation or revision of the Tibetan version was never made. However, the Kāraṇḍavyūha served as the basis for the eleventh-century Maṇi Kabum (A Hundred Thousand Teachings on the Maṇi Mantra; Tib. ma Ni bka’ ’bum), which was attributed to Songtsen Gampo, although the extracts from the sūtra that it includes are clearly derived from the early ninth-century translation. The Maṇi Kabum was a highly influential work in propagating the practice of Avalokiteśvara, known in Tibetan as Chenrezi (spyan ras gzigs), the repetition of the maṇi mantra, and the identification of Songtsen Gampo as an emanation of Avalokiteśvara; it has had a much greater impact on Tibetan culture than the sūtra upon which it is based.

The title of the sūtra is somewhat ambiguous. A karaṇḍa is usually a basket made of reeds. The karaṇḍa is frequently portrayed in the background of portraits of Indian siddhas as a large pot-bellied basket with a lid, containing collections of scriptures. These siddhas are also portrayed making the hand gesture representing the basket, the karaṇḍamudrā (“basket gesture”). There is even a layperson’s hairstyle named karaṇḍa­makuṭa (“basket crest”), where the hair is arranged on top of the head in the shape of a tall, rounded basket with a lid.

Another word for basket is piṭaka, which forms the basis of the most common metaphor for the Buddha’s teachings, “the three baskets” or tripiṭaka, which contain the Vinaya, Sūtra, and the Abhidharma or its predecessor the Mātṛkā. However, there are many instances in Tibetan literature where za ma tog, the translation of karaṇḍa, means something more solid and smaller than a pot-bellied reed basket, as in the precious casket (rin chen za ma tog) in the legend of the Kāraṇḍavyūha’s appearance to King Lhathothori. The name of the earlier Ratna­karaṇḍa­sūtra could at first seem to mean “precious casket,” but the contents of that sūtra validate the Tibetan translation as The Basket of the [Three] Jewels (dkon mchog gi za ma tog).9 There are also instances in the Sanskrit where the word karaṇḍa is apparently used for something more solid than a reed basket. There is a dhāraṇī in the tantra section of the Kangyur that has in its title the phrase dhātukaraṇḍa (Tib. ring bsrel gyi za ma tog), which means “the casket of relics,” or “reliquary.”10

The Kāraṇḍavyūha is spelled with a long initial a in all existing Sanskrit manuscripts, while every Tibetan edition has a short initial vowel. The long vowel is more likely to be lost than added, as errors generally replace the uncommon with the common. The enhanced vowel is used in Sanskrit to denote affiliation, origin, and ancestry. In the case of kāraṇḍa, the word usually means “ducks”; they live among the river reeds that are used to make baskets. Here kāraṇḍa may be signifying that this sūtra has its origin in the basket that contains the description of Avalokiteśvara’s qualities. A basket or casket is normally spelled without the long vowel: karaṇḍa.

There are also titles in the Tengyur that contain the word ratna­karaṇḍa (without the long vowel) where it means “a casket that is made of a precious material,” even though that meaning is not necessarily evident in Tibetan because of the syntax of the titles in question.11

Therefore, after hesitating between “basket” and “casket” and wishing there was one word for both (or at least a word for a lidded, pot-bellied reed basket), we chose “basket” as the better translation, primarily because of the way karaṇḍa is used in the sūtra itself. This term occurs only within the description of the Avīci hell. The Vaidya edition has visphurad ratnakaraṇḍavat, which means “raging [flame] like a precious casket,” but this appears to be a corruption, with the Cambridge manuscript having visphurantaṃ karaṇḍavat, and the Tibetan not having the equivalent of ratna (“precious”). If karaṇḍa is being used here to describe the shape of the flame, then it is referring to the distinctive shape of the reed basket, wider at its middle. This shape is still associated with za ma tog in contemporary Tibetan, and it is also compared with the shape of an egg.

Vyūha has a wide range of meanings, but is based on the idea of things being set out or displayed, and was therefore translated into Tibetan as bkod pa. The word can also mean “description” or “explanation” and even “chapter.” The sūtra is therefore a display from a basket, or the presentation of its contents.

The later Nepalese version of the sūtra has a longer title, Guṇa­kāraṇḍa­vyūha, which could be translated as A Display from the Basket of Qualities, the “qualities” being those of Avalokiteśvara. Both versions of the sūtra are dedicated primarily to a description of Avalokiteśvara’s qualities, which are stated to be greater than that of any buddha. The use of vyūha in the title is also evocative of the earlier Gaṇḍavyūha, which forms the last chapter of the Avataṃsaka, where gaṇḍa means “supreme” or “best.” The influence of the contents of that chapter is also discernible in this sūtra.

The Kāraṇḍavyūha’s principal content is the introduction of the oṁ maṇipadme hūṁ mantra and the descriptions of its inconceivable benefits. These are also the most quoted sections of the sūtra. However, it contains no instructions on the qualities and benefits of each syllable, of the kind that subsequently became widespread in Tibetan Buddhism. It also gives no explanation of the meaning of the mantra as a whole, a meaning that has been understood in various ways. Donald Lopez has given an account of various interpretations of the mantra in the West in his Prisoners of Shangri-la.12

Alexander Studholme, in his The Origins of Oṁ Maṇipadme Hūṃ, describes how the sūtra was composed within the context of familiarity with, and under the influence of, Purāṇic literature, in particular the Skandapurāṇa. In this sūtra, Avalokiteśvara has taken on various attributes and characteristics of Śiva, to the extent that one passage could be misread as describing Avalokiteśvara to be the creator of the universe. Even so, he is still being described as the creator of its deities, including Śiva and Viṣṇu. In particular, Avalokiteśvara’s mantra is evidence of the influence of Śiva’s five-syllable mantra, oṁ namaḥ śivāya (“Oṁ‍—Homage to Śiva!”), which is found in the Skandapurāṇa together with a description of the benefits of its recitation.

In classical Sanskrit grammar, padme would be the locative case, which has led to the interpretation of oṁ maṇipadme hūṁ as “jewel in the lotus.” However, mantras are typically given in the vocative or dative case, usually with the name of a deity being invoked. Padme is in fact the vocative for padma, this being Buddhist Hybrid Sanskrit. In classical Sanskrit, the e-ending vocative form is only used for feminine nouns.13 P.C. Verhagen has translated one of the few native Tibetan texts to be found in the Tengyur, a grammar text that uses this very mantra to explain the e-ending vocative form for masculine nouns.14 This vocative form of masculine nouns is a characteristic of the Magadhi, or northeastern Middle Indic, dialect. However, this form appears to have been much more widespread, extending as far as Sanskrit loan words in the Tocharian language of Central Asia.15 Maṇipadma is therefore a compound and is a name for Avalokiteśvara meaning “Jewel Lotus.”

The sūtra itself is rarely read in Tibet, other than in the annual ritual chanting of the Kangyur, and as mentioned above it has been eclipsed by the eleventh-century Maṇi Kabum. There is no evidence of it having had any significant impact on religious life in Tibet in the preceding centuries. In spite of the eventual importance of the oṁ maṇipadme hūṁ mantra, the sūtra is still primarily known only through select quotations. One reason for this is that very little of the teaching and meditation practice of the Maṇi Kabum is to be found in the sūtra.

Another reason is the difficulty involved in reading the sūtra due to its structure of narratives within narratives. After a buddha is initially introduced, he is subsequently only referred to as “Bhagavat,” and it is easy for readers to lose track of which level of the narrative they are reading. Although the speakers’ names were not repeated in the original, we have added them in here for clarity. We have not marked these insertions with square brackets, again for the sake of readability.

Another problem with the sūtra is that although it is a compilation of narratives, the sūtra does not always use its source material in a skillful manner. The Sanskrit original itself does not compare well with the clarity and style of writing found in other sūtras. There are abrupt transitions, inconsistency in the use of pronouns, and the contents of one part of the narrative appear to be in contradiction with those of another. For example, the Buddha tells the tale of the merchants being rescued from the land of the rākṣasīs in the first person, but there are sporadic lapses into what must have been the original third person of the narrative. The asura king Bali’s account of his downfall likewise transitions from a first- to a third-person account. In common with many other Mahāyāna sūtras but perhaps more frequently than most of them, the Kāraṇḍavyūha refers to itself within its own narrative as a sūtra that is being taught, requested, or longed for, but appears to describe itself as being comprised of verses, almost as if the Kāraṇḍavyūha is a different sūtra that is simply being referred to in this sūtra.

The sūtra assumes that the reader is familiar with the Mahābhārata and the Rāmāyaṇa, the two great epics of Indian literature, and the story of Viṣṇu’s avatar as a dwarf deceiving Bali, the lord of the asuras. Tibetan readers, however, would be unfamiliar with personages referred to in passing in the text, such as Śukra, who is both the deity of the planet Venus and counselor for the king of the asuras. Viṣṇu is usually referred to as Nārāyaṇa in the sūtra, but in the passage where he rescues the Pāṇḍavas and other kṣatriyas of Mahābhārata fame, he is referred to as Daśarathaputra (“son of Daśaratha”), which is actually the name of Rāma, another of Viṣṇu’s avatars. This may be because the story of the dwarf avatar also appears in the Rāmāyaṇa when it is told to Rāma, that is, Daśarathaputra.

The sūtra also includes a variation of a well-known jātaka tale in which the Buddha as a horse saves merchants from the island of the rākṣasīs,16 which has been retold with variations many times in Buddhist literature. Here it is retold with Avalokiteśvara as the horse and the Buddha as the head merchant who is being rescued. However, this too implies an unexplained internal contradiction: the sūtra had earlier narrated how Avalokiteśvara, in the form of a handsome man, had converted all the rākṣasīs from their cannibalistic ways to become devotees of Buddhism.

The Tibetan translation occasionally transliterates the Sanskrit rather than attempting to find a Tibetan equivalent, particularly when it comes to fauna and flora‍—even the Sanskrit word for “wolf” is simply transliterated as tarakṣa. There are also instances of obscure translations of words that do not agree with the Mahāvyuttpati.

In some passages, we relied more on the Sanskrit than we had originally anticipated because there is evidence that the manuscript from which the Tibetan translation was made had suffered from scribal corruption, as revealed by the surviving Sanskrit and confirmed by the English translation of the Chinese. For example, when describing the maṇḍala as adṛṣṭa (“not seen”), this was corrupted to aṣṭa (“eight”); a mountain made of padmarāga (“ruby”) was corrupted to padmarakta, which was translated as “red lotuses” (pad ma dmar po); and in the middle of the Buddha’s describing Avalokiteśvara’s qualities, ayaṃ (“this”) was corrupted to ahaṃ (“me”) so that the Buddha seems to be describing himself.

There are also omissions of sentences in the Tibetan (whether as the result of omission in the original Sanskrit manuscript or later copies of the Tibetan) that affect the narrative or meaning. The omissions are particularly evident when there are lists of qualities or meditations that are more easily left out in the process of copying manuscripts. On the other hand, there are also instances of members of lists that are preserved in the Tibetan but omitted in the available Sanskrit texts.

The most egregious flaw in both the Tibetan and Chinese translations, and one which has already attracted scholarly attention, occurred on rendering the obscure term ratikara, which literally means “that which creates joy,” and is also the name of one of the apsarases that are in the audience for this sūtra. The later Nepalese version used instead dvīpa, the common word for “lamp,” but both the Chinese and Tibetan translators, even with the assistance of Sanskrit scholars, were understandably stumped by this odd word, particularly as the ratikara laughs and speaks. Both Yeshé Dé and T’ien Hsi-tsai chose to make it refer to the rākṣasī wife speaking in her sleep, as she is the only other person in the room and is the merchant’s paramour. This entailed interpolating the word “sleeping” into the translation. However, the result makes little narrative sense, whereas the unlikely meaning of lamp, which we therefore preferred (see 2.­7), does make narrative sense.

Our aim was to make the most readable, accurate, and coherent version of the sūtra as it is preserved in the Tibetan translation. The Degé edition and the version in the critical edition of the Kangyur were therefore our principal sources.

Sanskrit manuscripts do not necessarily reflect the original form of a text, even though they are in the original language, because they may have their own accretion of omissions and additions that have occurred in the centuries following the time a Tibetan or Chinese translation was made. There has not yet been a critical edition from all available Sanskrit manuscripts, but we consulted three Sanskrit editions, the most important being a palm-leaf manuscript from the Cambridge University Library, which was written in the beginning of the second millennium before the development of the Devanāgarī script. It is notable for being closer to the Tibetan. Of easier access but less representative of the original text are the Sāmaśrami edition of 1872 and the 1962 Vaidya edition that is based closely on Sāmaśrami. The Sāmaśrami is available on the Online Sanskrit Texts Project of the Theosophical Network, and the Vaidya is openly available on the internet. To complete the translation of some difficult passages, we also referred to the Gilgit manuscript fragments, though they were not readily accessible. Silfung Chen’s online English translation from the Chinese proved interesting in its correspondences with these editions.

Nevertheless, as noted above, there were a number of points where we relied on the Sanskrit to fill in missing elements, words, members of a list, and sometimes whole sentences, although it is possible that some of the latter may have been later additions to improve the flow and clarity of the sūtra’s sometimes clumsy narrative. Where our translation favors the Sanskrit over the Tibetan, annotations indicate that this is the case.

An important objective was readability, so the syntax does not necessarily reflect that of the Tibetan or Sanskrit versions. For example, an active construction may be used instead of a passive construction found in the original. The inconsistencies of first and third person have been resolved, and, as noted above, names are repeated when otherwise the reader might lose track of who is speaking or to whom the text is referring. Hopefully this will make reading the sūtra in English far less challenging than attempting to do so in Tibetan or Sanskrit. Readers will find the variant readings in Tibetan and Sanskrit in the notes if they wish.

Buddha Śākyamuni is at Jetavana Monastery with many disciples. Lights shine upon the monastery and miraculously transform it. The bodhisattva Sarva­nīvaraṇa­viṣkambhin asks the Buddha where the lights came from. The Buddha explains that they came from Avalokiteśvara, who had just visited the Avīci hell and the city of the pretas, and then describes those visits.

Then Buddha Śākyamuni recounts being a merchant at the time of Buddha Vipaśyin and how he heard him describe how various deities, including Śiva and Viṣṇu, were created from Avalokiteśvara’s body.

Buddha Śākyamuni then recounts being Bodhisattva Dānaśūra at the time of Buddha Śikhin and how light rays shone from Buddha Śikhin. In response to questioning by Bodhisattva Ratnapāṇi, Śikhin says that the lights and other omens are a sign of the approach of Avalokiteśvara, who then arrives from Sukhāvatī with an offering of lotuses from Buddha Amitābha.

After Avalokiteśvara’s departure, Śikhin describes to Ratnapāṇi how Avalokiteśvara’s accumulation of merit is inconceivable by using a series of analogies. Then he describes how Avalokiteśvara teaches this very sūtra to the asuras in the form of an asura.

Buddha Śākyamuni then states that he was a rishi (ṛṣi) at the time of Buddha Viśvabhū. Before repeating what Viśvabhū taught, Śākyamuni relates how Avalokiteśvara taught upside-down beings in the realm of gold and four-legged beings in the land of silver. There then follows a long description of Avalokiteśvara’s visit to the asuras in the land of iron. Avalokiteśvara teaches the asuras the inconceivable merit that comes from making offerings to a buddha. Bali, the king of the asuras, tells Avalokiteśvara that he had in the past made an offering to the wrong recipient. He had imprisoned all the kṣatriyas, but Viṣṇu secretly freed them and came to him in the form of a dwarf asking for two steps of land. Bali offered him three, but Viṣṇu took on his divine form and covered the whole world in two steps. He then banished Bali to the underworld where he now dwells for having failed to fulfill his promise.

Avalokiteśvara then describes to him the suffering in hells that awaits those who have not made offerings to the Buddha.

Avalokiteśvara then radiates light rays to where Viśvabhū and his pupils are residing in Jetavana Monastery. Bodhisattva Gaganagañja asks Viśvabhū where the lights came from. Viśvabhū states that the lights are a sign that Avalokiteśvara is coming. However, Avalokiteśvara first goes to a land of darkness to teach the yakṣas and rākṣasas about the merit that comes from this sūtra.

Avalokiteśvara then goes to the Śuddhāvāsa realms, where in the form of a brahmin he begs from a poor deva. The deva goes into his empty palace to give him whatever he has, but finds it full of jewels and food that he then offers to the brahmin. Avalokiteśvara in the form of the brahmin tells the deva that he is a bodhisattva from Jetavana Monastery.

Avalokiteśvara then descends to Siṃhala Island, the land of the rākṣasīs, in the form of a handsome man. He agrees to be their husband if they follow his instructions, which they do, giving up killing.

Avalokiteśvara then travels to Vārāṇasī, where in the form of a bee he buzzes the prayer of homage to the Three Jewels to the insects in a large cesspit, liberating them.

Avalokiteśvara then goes to Magadhā, where starving beings have been eating each other for twenty years, and he causes a rain of food to fall. One of the people, a man who is hundreds of thousands of years old, realizes that only Avalokiteśvara could have caused this miracle, and tells the others of the benefits of making offerings to him.

Avalokiteśvara then goes to Buddha Viśvabhū. Bodhisattva Gaganagañja meets him, Viśvabhū teaches the six perfections, and the audience disperses. This is the end of part one.

Part two begins with Sarva­nīvaraṇa­viṣkambhin asking for teachings from Buddha Śākyamuni, who lists the samādhis that Avalokiteśvara possesses.

Then Buddha Śākyamuni recounts being a head merchant who became stranded on Siṃhala Island with other merchants. Each of them goes to live with a rākṣasī. One night, a talking lamp warns the head merchant that the women are all rākṣasīs. As proof, the lamp directs him to an iron fortress where other merchants are being kept prisoner and then eaten. Then the lamp tells him of Bālāha,17 a miraculous horse on which the merchants can escape. As they flee upon the horse, all the other merchants look back, fall off the horse, and are eaten by the rākṣasīs, while the head merchant reaches home safely. Buddha Śākyamuni states that Avalokiteśvara was the horse.

Buddha Śākyamuni then begins a description of two pores on Avalokiteśvara’s body and their inhabitants.

Sarva­nīvaraṇa­viṣkambhin, to the Buddha’s approval, describes the benefits that come from this sūtra.

Buddha Śākyamuni describes another pore and explains to Sarva­nīvaraṇa­viṣkambhin that the pores are immaterial and cannot be seen even by buddhas.

Buddha Śākyamuni describes two more pores, saying that those who remember Avalokiteśvara’s name, meaning the six-syllable mahāvidyā, will be reborn in them, but that no one, not even the buddhas, know this mantra.

After Buddha Śākyamuni describes more benefits that come from the mantra, Sarva­nīvaraṇa­viṣkambhin declares his intention to obtain it.

Buddha Śākyamuni recounts his own fruitless search for it until, after meeting trillions of buddhas, he finally met Buddha Ratnottama who directed him to Buddha Padmottama. Padmottama describes the incalculable benefits that come from saying the mantra once and then describes his own long fruitless search for the mantra until he came to Buddha Amitābha, who instructed Avalokiteśvara to give the mantra to Padmottama. Avalokiteśvara does so through a maṇḍala made of precious stones and gives the instructions on how to make the maṇḍala.

Buddha Śākyamuni follows this narrative with a description of how incalculable the benefits are from even one syllable of the mantra.

He then tells Sarva­nīvaraṇa­viṣkambhin that he can only obtain it from an unnamed dharmabhāṇaka who has lost his monastic vows and lives in Vārāṇasī. Sarva­nīvaraṇa­viṣkambhin goes to him in a huge procession of people and offerings.

The dharmabhāṇaka describes the benefits of the mantra and, at the urging of Avalokiteśvara, who appears in the sky, gives the mantra to Sarva­nīvaraṇa­viṣkambhin, who returns to Buddha Śākyamuni. Seventy million buddhas recite the mantra of the goddess known as both Cundi and Cundā.

Buddha Śākyamuni then describes five more of Avalokiteśvara’s pores.

Buddha Śākyamuni then describes the oceans that come from Avalokiteśvara’s big toe, and says there are no more pores but those ten. Then omens of Avalokiteśvara’s arrival appear. He leaves Sukhāvatī and comes to Buddha Śākyamuni and offers him lotuses from Buddha Amitābha.

Buddha Śākyamuni then directs Maheśvara and Umādevī to receive the prophecies of their future buddhahood from Avalokiteśvara.

Buddha Śākyamuni then gives a teaching on the incalculability of Avalokiteśvara’s merit and listing the samādhis he has.

Then Buddha Śākyamuni recounts when he was with Buddha Krakucchanda and saw Samantabhadra and Avalokiteśvara both practicing various samādhis. Krakucchanda declares that not even the buddhas have Avalokiteśvara’s samādhis.

Buddha Śākyamuni then describes the benefits that come from this sūtra, and Avalokiteśvara departs.

Then Ānanda requests teachings on monastic conduct. Buddha Śākyamuni prophesies how there will be monks who do not maintain their conduct in the future and who should be expelled. He describes the tortures in hell and other rebirths that await laypeople who misuse the property of the saṅgha.

Ānanda departs and the sūtra concludes.

This outline is intended as a guide to the complicated narrative levels of the sūtra.