With ever new dohā verses, nothing is hidden. —Saraha

The time and place of Saraha are steeped in obscurity despite—or perhaps because of—the abundance of hagiographic material devoted to him. This being the case, it is not surprising that the history of the composition of the Treasury of Dohā Verses is equally opaque. We will probably never know just how the dohās of Saraha were composed, or even if he composed all or any of them. Charlotte Vaudeville states the problem succinctly with regard to Kabir’s poetry: “There is no evidence that Kabīr ever composed a single work or even wrote a single verse.”¹ This applies as well to Saraha and the large corpus of spiritual songs attributed to him; we simply have no direct evidence that Saraha wrote the dohās that are currently available to us. There is no autograph manuscript with Saraha’s signature; the oldest manuscript was scribed perhaps several centuries after his death, and even it, as we shall see, was a compilation of disparate verses written down by a distant member of his tradition. The various recensions of the Treasury of Dohā Verses are so different as to make the creation of any ur-text a feat of the imagination. The very name Treasury of Dohā Verses should lead us to inquire about the circumstance of its creation in the tradition following Saraha, and not in the single-focused intentions of an author presenting a coherent work of literature, or a logically organized work of philosophy. “Dohākoṣa” should not even be translated as The Treasury of Dohā Verses, but rather as A Treasury of Dohā Verses—a compilation of dohā verses. “With ever new dohā verses, nothing remains hidden,” Saraha tells us.² These words ring with irony when (p. 80 ) one reads them with the knowledge that this dohā itself exists only in one of the dozen or so recensions of the Treasury of Dohā Verses. But its absence makes no difference to Saraha’s message, for the verse contains instructions for its own disappearance. The traditions following Saraha took the spirit of this verse to heart, and until one recension of the Treasury of Dohā Verses was canonized in Tibetan translation in the fourteenth century, “ever new dohās” found their way into the Treasuries.

This and the next chapter will present a variety of viewpoints on the development of the Treasury of Dohā Verses and the dohā literature in general. As in the discussion of the tales of Saraha in part I, this chapter constitutes an anthology of sorts, bringing together translations and traditional discussions of dohā songs.³ The next section will provide a brief overview of the place of the dohās within the Tibetan canon, as well a collection of exemplary songs. With the songs fresh in mind, the section following will present what commentators on Saraha’s Treasury of Dohā Verses have said about the creation of such poetic descriptions of tantric practice. The section after that will look closely at another set of writings in which Saraha figures: the anthologies of tantric songs attributed to Phadampa Sangye. Finally, a detailed analysis of the nature of these anthologies, and the means by which they were assembled, will allow us to conjecture about the ways in which the songs of Saraha himself may have been developed by the tradition. Throughout these varied discussions I have sought to ask about the form, literary context, and methods of creation of the dohās.

Dohās in the Tibetan Buddhist Canon

Saraha is the attributed author of twenty-six separate works in the Derge Tanjur. They can be divided into three general groups: The first is a collection of works dealing with the Buddhakapāla Tantra,⁴ including a commentary on difficult points, a sādhana ritual manual, and a longer manual for conducting the maṇḍala ritual associated with the Tantra. The second group is the dohās. The third group consists of four different translations of essentially the same sādhana dedicated to Lokeśvara.⁵ Out of these twenty-six works, eighteen are poetic songs, styled in their Tibetan titles either as do ha (dohā), glu, or some combination of the two terms. A total of eighteen scholars contributed to the translation and transmission of Saraha’s canonical works, and seven of them are named as having worked on the dohās. Although the dohās constitute the larger part of Saraha’s works in terms of titles, they account for only 43 folios out of 123.5 total in the Derge Tanjur. The commentary on the Buddhakapāla Tantra is the longest of the works attributed to him.

The dohās of Saraha are part of a larger corpus of works collected in the Tanjur under the section title, “Commentaries on the General Intention of the Highest Yoga Tantras.”⁶ According to Zhuchen Tsultrim Rinchen, author of the catalog to the Derge Tanjur, there are 287 separate works in this section. These works form the basis of any study of the writings of Saraha and other (p. 81 ) later tantric poet-saints; in three volumes of the Tanjur the greater part of the ecstatic poetry of the Buddhist adepts is collected.⁷ They range in length dramatically, from Saraha’s massive song on Mahāmudra theory in some 700 lines (Sku mdzod ‘chi med r do rje’i glu) to Kāiṇhapāda’s cryptic single line “Diamond-Song,” no doubt one of the shortest works in the Tanjur:

• Homage to glorious Vajrāsana-pāda.
• Said: Kāṇha’s mind not being at all stable, all appears topsy turvy.
• The Diamond-Song by Master Kāṇapāda is complete.⁸

Diamond-songs are perhaps the most elegant of the songs preserved in the Tanjur, as the following translations attempt to show. Dispensing with the often semi-exegetical style of works such as the Treasury of Dohā Verses, they use metaphor, paradox, and suggestion to portray the experience of tantric practice as spontaneous, playful, and transcending logical analysis. The diamond-songs are much shorter than dohās, and they more often employ metaphorical language. Diamond-songs strive for a poetic beauty far exceeding that of the more prosaic dohā materials and feel more like songs, like emotionally charged teachings whose lessons lie in the evocation of a certain feeling rather than in the promotion of a certain doctrine. A few examples of such songs from well-known adepts such as Nāropa, Virūpa, Ḍombi, and Saraha himself will illustrate this feeling. The following anonymous diamond-song expresses in a few words what volumes of Madhyamaka treatises have sought to explicate through reasoned argumentation:

• Homage to the Greatly Compassionate.
• As am I,
• So are you.
• Separation, conception,
• Serpent sprung from an ocean of ambrosia,
• Hissing conception.
• Diamond-song is complete.⁹

Using similar serpent imagery, Nāropa’s diamond-song contains one of the most striking self-reflexive conclusions of any of these songs:

• Homage to Lord Mañjuvajra.
• Cunning serpent of conception,
• Snares us. Our minds:
• As, just as theyconceptualize,
• So, just so are they fettered,
• Feeling much useless suffering.

• Followers of delusion
• Will not be rid of delusion.
• Compassionate hero
• Must meditate on just that.
• (p. 82 ) As, just as he discerns,
• So, just so is he free.

• Insight hydra
• Devours selfhood.
• Insight hydra
• Forever nourishes contemplation with milk,
• Slaying the mongoose of conception.

• The hundred-headed insight serpent
• Forever devours Nāropa.
• Who knows this?
• Nāropa himself knows.
• Nāropa asks this of himself.
• Great yogin Nāropa’s diamond-song is complete.¹⁰

Ḍombi’s song employs imagery from the hunt. The forests become the illusory nature of the world seen from an unenlightened state, the journey into the forest becomes the spiritual quest. The kill, repeated in the refrain of the song, hails the adept’s triumph:

• Homage to the Buddha.
• Into the many forests of illusion,
• To slay beasts Minister Ḍombi goes.
• Backward travels, where beasts travel, I see.
• A sharpened arrow Ḍombi readies: it plunges in.

• Five arrows has he, five let loose straight on.
• Ḍombi takes up the readied bow, and there they go.
• Backward travels, where beasts travel, I see.
• A sharpened arrow Ḍombi readies: it plunges in.

• No bow, no string, no reed, no tip has he.
• Ḍombi, doubtless, slays with certainty.
• Backward travels, where beasts travel, I see.
• A sharpened arrow Ḍombi readies: it plunges in.

• Setting the five Victors upon your diadem,
• Slay the beasts again and again.
• Backward travels, where beasts travel, I see.
• A sharpened arrow Ḍombi readies: it plunges in.
• Ḍombi’s song is complete.¹¹

Virūpa’s diamond-song again uses hunting imagery to evoke a vivid sense of the immediacy of realization; the reader is fooled, following the arrow to its target, expecting something solid to strike and—nothing:

• Homage to the Buddha.
• Taking up the arrow of the diamond yoginī,
• (p. 83 ) Espying objects at which to aim,
• I shoot the arrow of co-emergence:
• No target, not one, nothing struck.

• Hear the message of co-emergence:
• An arrow which connects with all things.
• I shoot the arrow of co-emergence:
• No target, not one, nothing struck.

• No sensation exists beyond these.
• “Of this,” “This is,” have passed into nonduality.
• I shoot the arrow of co-emergence:
• No target, not one, nothing struck.

• Not even a price of twenty shells;
• Such a slave makes Virūpa.
• I shoot the arrow of co-emergence:
• No target, not one, nothing struck.
• The diamond-song of Virupa is complete.¹²

The songs of Saraha do not begin and end with the three most popular dohās—the King, Queen, and People Dohās. Fifteen more songs attest to the popularity of this form of spiritual instruction. Another trilogy—or perhaps quartet—is attributed to him, as well as an Alphabet Dohā in which the first word of each verse begins with a consecutive letter of the Apabhraṃśa alphabet, as well as two songs simply titled “song.” The first of these songs uses an impossible vision—milking the sky—to express the freedom from convention heralded by the adepts. This untitled poem consisting of a simple verse and refrain, is surely one of Saraha’s most evocative:

• Intertwined are the natures of emptiness and compassion
• Indivisible, unceasing, emptiness exists.
• I see the empty ḍākinī,
• Milking, milking, and drinking the sky.

• She churns the sky in sky unseen,
• Upon the earth, bound by samsara she does not dwell.
• I see the empty ḍākinī
• Milking, milking, and drinking the sky.

• A ḍākinī such as she wanders from home, from the root;
• Stainless is the magnificent taste of compassion.
• I see the empty ḍākinī,
• Milking, milking, and drinking the sky.

• Why should anyone else do what Saraha says?
• Day and night he walks to drink the sky.
• I see the empty ḍākinī,
• Milking, milking, and drinking the sky.
• Saraha’s Song is complete.¹³

(p. 84 ) Saraha’s second diamond-song speaks of joyous feelings that accompany the fruits of spiritual insight:

• Homage to the Buddha.
• “Profound, Profound!” says all the world,
• Yet, within the unborn there is something joyous.
• Lo, the depths of mind are difficult to know, so,
• When co-emergence is dissected, it is not there.

• Mind unstationed settles unperceived,
• The yogic fires burns all things.
• Lo, the depths of mind are difficult to know, so,
• When co-emergence is dissected, it is not there.

• Desire for yoga is nothing taught by the master,
• Upon the blessings of the master nothing is gained.
• Lo, the depths of mind are difficult to know, so, When co-emergence is dissected, it is not there.

• If you desire to seek out glorious, magnificent bliss,
• Desiring to seek the center, you fall blindly down the well.
• Lo, the depths of mind are difficult to know, so,
• When co-emergence is dissected, it is not there.
• The song of Sarahapa is complete.¹⁴

With some idea of the place of Saraha’s songs in the Tibetan Buddhist canon, and a healthy sampling of the diamond-songs of the adepts, we may turn now to what the tradition has to say about the creation of these works.

Saraha the Singer

The classic hagiographic account of Saraha’s singing—of the beginnings of the Treasury of Dohā Verses as an oral song—is found in Tales of the Eighty-Four Adepts. Here as in all other accounts, it is stressed, that Saraha sang his songs not in order to argue some particular doctrine or preserve his poetry for posterity, but rather to bring his disciples to enlightenment. As we have seen, this scene occurs at the close of Saraha’s tale and represents the culmination of his spiritual life (on earth, that is). As the Buddha himself gave teachings following his enlightenment, in this hagiography the singing of the Dohā Trilogy embodies Saraha’s activity as an enlightened being. The act of singing is nothing less than an enlightened act, geared toward liberating his disciples. As the Tales of the Eighty-Four Adepts tells us: “Then all the brahmins and the king paid homage to him and requested spiritual instructions, so he sang songs to the king, the queen, and the people, and these became known as the Dohā Trilogy. The brahmins renounced their own teachings and entered into the teachings of the Buddha. The king and his court attained spiritual boons.”¹⁵

Karma Trinlaypa develops a richer account of the initial singing of the (p. 85 ) Dohā Trilogy. Here one gets the strong impression that this tale was developed precisely to explain the presence of three distinct Dohās, each bearing the name of the group to which Saraha sang: King, Queen, and People. He writes:

[Saraha], on behalf of the common people, put into song one-hundred and sixty dohā verses and led them onto the correct path. There, on behalf of the king’s queens, who had also so beseeched him, he put into song eighty dohā verses, and by these he introduced even them to the purport of how things really are.

Then, because King Māhāpala himself came to request that [Saraha] resume his former demeanor, on his behalf [Saraha] put into song forty dohā verses and led even the great king upon the path of reality. Among other things [Saraha] sang many diamond-songs, and he accomplished immeasurable benefit for living beings.¹⁶

We are told very few details from Indian sources regarding Saraha’s singing, and even less about the process of putting the songs into writing. We can, however, look to the Tibetan traditions concerning their creation for hints regarding the general circumstances of Saraha’s creative efforts and—in proper deference to the fact that these sources are Tibetan, not Indian—for some notion of how Tibetans conceived the origins of these works. More often than not, these traditions speak of Saraha as a singer of songs, but not as a writer. The copying down of the Great Brahmin’s inspired aphorisms were seen as the work of his disciples, grand-disciples, and spiritual descendants. In this the work of Saraha bears yet another similarity to that of Kabīr. As Vaudeville writes: “Kabīr’s followers…do not assert that the Prophet himself wrote down the numerous compositions attributed to him. They hold that he composed them orally and that they were subsequently written by his immediate disciples.”¹⁷ This is certainly the view of Karma Trinlaypa, who presents two opinions regarding the creation of the Dohā Trilogy: “Some say that while the Dohā Trilogy is that which was put into song by the illustrious Saraha, because they were merely the successes of his spiritual experience given voice, they were without division into larger and smaller verses. Because what Saraha spoke was at a later time put into writing by the master Nāgārjuna, the three distinct larger and smaller works came about on the basis of how they were spoken.” Others, by contrast, “allege that these works were not even compiled by Nāgārjuna, but that Śahara, upon being liberated by these instructive teachings, composed the three texts for the benefit of Maitripa. The former explanation is authoritative.”¹⁸

If the dohās are not generally considered by Tibetan historians to have been put into writing or redacted by Saraha, then the question remains as to how and why they were recorded. Here Karma Trinlaypa presents us with two scenarios: one in which Saraha’s pupil wrote down his songs, and one in which Saraha’s grand-pupil writes them. In either case Saraha himself is not characterized as a writer, nor are the three separate songs of his Dohā Trilogy understood by Karma Trinlaypa to be the product of Saraha’s compositional skills. (p. 86 ) There were no divisions between the three Treasury of Dohā Verses at the time when Saraha sang them; for Karma Trinlaypa, in fact, the Dohā Trilogy did not exist until Saraha’s disciple created it. On the contrary, Saraha’s mastery lay not in his skill as an author of philosophy, but precisely in his ability to bring his vision to his disciples in an inspiring way through song; the work of writing, editing, and redaction was left to his students. The anonymous Tales of dohā Lineages favors the second alternative claim, although it would have Śabareśvara be the direct disciple of Saraha: “The people reviled and slandered them, so [Saraha] put the Dohā Trilogy into song for the sake of the king, the queen, and the common folk. Master Balpo says that at that time Śahara…heard Saraha put the Dohā Trilogy into song, and set them down in writing.”¹⁹ Pawo Tsuklak Trengwa paints a similar picture of the singing of the dohās, embellishing it with the notion that the Dohā Trilogy was in fact a presentation of the three bodies of the Buddha. He then notes a very interesting bit of hearsay: “The people beseeched him not to act [contrary to society], so he sang the People Dohā, which teaches principally the enlightened body of emanation. The king’s queens beseeched him, and he sang the Queen Dohā, which teaches principally the enlightened body of enjoyment. The king himself beseeched him, and he sang the King Dohā, which teaches principally the enlightened body of dharma. So singing, everyone was liberated, and the kingdom became empty.” As an aside, Tsuklak Trengwa also makes reference to the material conditions of the early dohās: “It is said that later the three texts were written on palm leaves and spread after falling into the hands of two brother scribes.”²⁰

A very different notion of Saraha’s role in the creation of the songs emerges in the picture developed by Chomden Raldri in the opening passage of his Ornamental Flower for the Dohās: “The Great Brahmin of Southern India, Saraha, heard [the teaching] under the arcane lord Vajrapāṇi, the student of the Buddha Vajradhāra, and under the ḍākinī of primordial awareness, Sukhasiddhī, and composed this text of the Dohā…[Saraha] taught it to Master Padmavajra, Noble Nāgārjuna, and Śabareśvara. This text is known to have been composed when Master Padmavajra made a request. Even the vocative words such as ‘boy’ are actually [present] in the [text].”²¹ Here Saraha is said to have composed the text of the Treasury of Dohā Verses for the benefit of his student, Padmavajra. The use of the vocative address is given by Chomden Raldri as proof in support of this claim. For Chomden Raldri, Saraha was indeed the writer of the Treasury of Dohā Verses; but this does not imply that Saraha was the sole creator of the work or its teachings, for though Raldri does not use such language, it is clear from the following account that for him Saraha was the recipient of a revelation, and thus the dohās were for him a kind of revealed literature. We thus find here a vision of Saraha as a writer, but also of Saraha as a mediator between the realm of enlightened beings and the realm of Buddhist aspirants.

Kongtrul’s origin tale of the Tales of the Eighty-Four Adepts and the songs of the adepts also provides an imaginative vision of the transfer of the dohās from voice to the written word. As was seen in an earlier chapter, in Kongtrul’s (p. 87 ) tale a certain king beseeches the adepts to come and make blessings for his deceased mother. In a cantankerous manner, the magic yogins give a glimpse of their message of enlightenment in song and then quickly depart. According to Kongtrul: “Each adept sang a dohā diamond-song and went back alone to where they had been. Then the king created statues of each adept as representations of them and wrote the dohās of each on the front of them. He then made offerings to [these statues].”²² Again we see that it was not Saraha and the rest of the adepts who put their songs in writing. This time it is the king who makes the change, who takes the fleeting words of the yogins and immortalizes them in stone. The early formation of the songs did not end there, however, for according to Kongtrul’s tale it was in fact no less than the ḍākinīs who taught the songs that were eventually to spread to Tibet in a written format:

Two ḍākinīs appeared in person to [Vīryaprabha] and taught the dohās, the Ratnamāla, and the tales that went with them. He took their meaning into his experience and thus acquired a distinctive realization and became a master among adepts.

[Vīryaprabha] also created a book in which the various dohās were anthologized. He taught it to a scholar named Kamala. [Kamala] taught it to the adept Śahara Jamaripa, who taught it to the scholar form Maghada, Abhayadattaśrī. He composed life stories and a commentary on these dohās, and with the translator of Minyag, Mondrup Sherap in Tibet, translated all of these [teachings], redacted them through explanation and listening, and caused them to spread.²³

In this and the previous passages, Kongtrul develops an extremely complicated notion of the seminal transmission of the songs of the adepts. First of all, the songs are twice removed from their singers before they are put into writing, for it is the ḍākinīs who teach the songs to Vīryaprabha, and they in turn are presumed to have heard them from the adepts. The manner of their hearing is not spelled out by Kongtrul, however--an omission that makes this early moment in the life of the songs all the more fantastical in this tale. Vīryaprabha, who was not even present when the adepts sang their songs, is here heralded as the anthologizer, the keeper of the book. It is he who allows the songs to be transmitted in a coherent fashion through several Indian scholars until they finally reach the hands of Abhayadattaśrī. And yet Kongtrul does not stop there in his tale of the development and transformation of the songs and life stories of the adepts, for in the process of translating these from their former Indic language in Tibetan, Abhayadattaśrī and his assistant from Minyag, Mondrup Sherap, redacted (gtan la bob pa) the songs yet again through the interplay of explanation (‘chad) and listening (nyan).

In Kongtrul’s tale, then, the rich life of the songs in both oral and written formats runs through four moments: the initial singing, the teaching of the songs orally, the anthologizing and writing, and finally the redaction and translation. (p. 88 ) I believe we can get no closer to a plausible picture of the general circumstances of the creation of the dohās than that developed in this tale from nineteenth-century eastern Tibet.

Saraha and the Anthologies of Phadampa Sangye

I will now expand the discussion of the development of dohā compendia by looking at another tradition in which Saraha figures: the anthologies of tantric songs held to be transmitted to Tibet by Phadampa Sangye. Descriptions within these composite works provide further examples of Tibetan notions of the process of literary formation—from oral composition to written compilation—and a more in-depth view of the late Indic and early Tibetan literary context in which the works of Saraha were brought to Tibet.

The verse anthology is one of the more intriguing literary forms in which the spiritual teachings of early-medieval Indian tantric adepts reached Tibet, not only by virtue of the often beautiful poetry arranged within its various frameworks, but also because of the supporting architecture itself: the rubrications, settings, the plethora of names and epithets, all of which structure readers’ engagement with the verse. The larger part of such works entered Tibet thanks to the South Indian teacher Phadampa Sangye. These anthologies present many questions regarding their composition, compilation, translation, and transmission. By introducing a selection of them with a particular focus on their organizational features, passages within the works that describe their own creation, as well as similar passages found in the colophons of the works, this section will explore these questions, with particular attention to the figure of Saraha in the anthologies.

The anthology as such is not unknown in Indian literature, and there are numerous examples of works structurally similar to the collections of tantric songs under discussion here. Perhaps the Pāli language Theragāthā and Therigāthā (ca. 500–100 bce), which contain single and multiple verse songs attributed, respectively, to individual monks and nuns, can be considered their distant literary relatives. Numerous anthologies of single verses devoted to secular themes are extant in Prākrit, mostly from Jaina authors, and dated anywhere from the eighth to the fourteenth century.²⁴ The earliest dated anthology of Sanskrit verse, the Subhāṣitaratnakoṣa, was most likely compiled by the Buddhist monk Vidyākara at Jaggadala Monastery in Bengal sometime during the end of the eleventh or the beginning of the twelfth century—the very era in which Phadampa Sangye was active in Tibet.²⁵

The anthology most similar to ours that is still extant in any Indic language is no doubt the Caryāgītikoṣa, a collection of fifty songs attributed to many of the same names of later tantric Buddhism as are in Phadampa Sangye’s anthologies, dating anywhere from the eighth to the twelfth century,²⁶ and thus potentially contemporaneous with our collections. Despite their structural similarity, however, no works designated as caryāgīti (spyod pa’i glu) are to be found in the anthologies of Phadampa Sangye, nor are any of the songs generally (p. 89 ) referred to by the Tibetan term glu, mgur (alternatively ‘ghur in early manuscripts) being the common term for “song.” The fact that none of the seventeen anthologies connected with Phadampa Sangye has any Indic witnesses whatsoever makes it impossible to determine whether these works existed in India in the forms in which we find them in the Tibetan Canon.

In terms of their transmission history, the anthologies of verses from tantric adepts found in the Tibetan canonical collections can be divided into two broad groups: those associated with Abhayadattaśrī’s Tales of the Eighty-Four Adepts, and those tracing their introduction to Tibet back to Phadampa Sangye, who, according to Go Lotsawa, traveled to Tibet in 1097 and remained there for some twenty years until his death in about 1117.²⁷ Two anthologies connected to the collection of stories attributed the Abhayadatta are preserved in the Peking edition of the Tibetan Canon, namely the Heart of Realizations of the Eighty-Four Adepts attributed to Pawo Osal/*Vīraprabhāsvara,²⁸ and the Jewel Garland attributed to *Dhamadhuma.²⁹ Though they share structural and topical features with the anthologies of Phadampa Sangye, one crucial difference is that the songs in the anthologies of the Eighty-Four Adepts’ were connected by later Tibetan writers to tales of their lives; in the case of Phadampa Sangye’s anthologies we find no such hagiographic elaboration.

The majority of tantric song anthologies found in the Tanjur are connected with Phadampa Sangye. A number of these compedia are also found in the first volume of the collected works of Phadampa Sangye and his Tibetan disciples, a five-volume collection entitled The Profound Oral Lineage Descending from Pacification, the Heart of the Holy Teachings.³⁰ Of seventeen anthologies said to be brought from India by Phadampa Sangye, ten are shared between the canon and his collected works,³¹ one is found only in the canon, and six are found only in the Profound Oral Lineage. Several also found their way into separate collections of songs, hagiographies, and iconographic manuals.

Both verse and prose can be found in the anthologies, though verse is by far the more common form of expression. The four prose works, consisting of the Orb trilogy³² and the Miraculous Lineage,³³ are themselves made up of formulaic aphorisms which read almost like verse with an occasional hypermetric line. Of the thirteen verse collections, seven contain “symbolic” songs (brda mgur, or in the archaic orthography of the manuscript of the Profound Oral Lineage, brda’ ‘ghur), three contain verses called “expressions of realization” (rtogs brjod), two contain “diamond-songs” (rdo rje’i ‘ghur), and finally one contains verses merely referred to as “songs” (’ghur) (See appendix 1). Symbolic songs are most often made up of a series of poetic one-liners, riddles whose punch line is a spiritual experience alluded to but never named. When ten or twenty lines of these are strung together, the effect is dizzying, with one—seemingly—unrelated “symbol” or metaphorical expression following upon another, as in the case of this passage from Precious Symbolic Instructions on the Great Seal:

• Space and the three thousand [worlds] are inseparable.
• A one-winged bird cannot sail through the sky.
• (p. 90 ) The head and the limbs are part of the same body.
• The essence of ice, with stonelike form, is water.
• The mirror and the reflection have no difference.³⁴

In Songs of the Glorious Diamond Ḍākinīs, however, the symbolic songs are more integrated poems in which a single metaphor is developed for several lines. Thus, in the ḍākinīs Tamala’s song we find not only the forest elephant used as a symbol, but its binding straps, food, ornaments, captivity, and finally escape all employed to symbolize the process of spiritual realization.³⁵

“Expressions of realization” are more philosophical in character, employing the terminology common to Great Seal works, as well as more straightforward admonitions against incorrect spiritual practices and views of the ultimate, such as Saraha’s plea to aspirants not to search anywhere for truth but in one’s own mind,³⁶ or Āryadeva’s maligning of the external forms of the four empowerments.³⁷ While both symbolic songs and expressions of realization are operating under the same apophatic strictures, in which the ultimate realization cannot by definition be named, symbolic songs implicitly put this into practice by talking riddles around the intended subject (“An iron boulder is difficult to roll uphill!”³⁸), whereas expressions of realization come right out and exclaim, “you can’t call it this, you can’t call it that!”

Finally, the “diamond-songs” contained in the Garland of Golden Droplets and Diamond-Songs of the Adepts stand somewhere between symbolic songs and expressions of realization, in that similes, philosophical language, and admonitory warnings are combined in single verses. Nāgārjuna’s verse illustrates this technique well:

• The unrealized should not undertake the practices of the realized;
• If they do, they are like the commoners doing harm to the king’s law.
• The realized should not undertake the practices of the unrealized;
• If they do, they are like elephants sinking in the mud.³⁹

Eleven of the seventeen anthologies share a simple common pattern of organization that consists of an alternating sequence of one-sentence rubrications followed by single- or multiple-verse songs, or prose aphorisms. The rubrics invariably include the name of the performer, be it deity, ḍākinī, adept, and so on, followed by a verbal action, more often than not the phrase “put into song” (‘ghur/mgur du bzhengs). In cases where only one verse per performer is given, the names make up a significant part of the work, almost overshadowing the songs themselves. With the relatively small amount of philosophic content conveyed in each verse, one is led to wonder what type of connection the reader made between any one verse and a name. A more elaborate system is found in the Miraculous Lineage, where six sets of six performers are arranged, each with the same order: king, queens, seers, brahmins, great personages, and ministers. Here the names have an artificial feeling, all composed of the same elements rearranged (such as ye shes, zla ba, nyi ma, ‘od zer, etc.). It is possible that the reading of such names was akin to reading a devotional (p. 91 ) work, such as the liturgical works developed by Tāranātha—and later Kongtrul—around the names of the Eighty-Four Adepts.

Five of the anthologies employ larger sections, referred to variously as dum bu, brul tsho, le’u, or left unspecified (see appendix 1). Precious Symbolic Instructions (no. 4) contains six separately titled sections containing smaller sets of single verses from ḍākinīs, with each section in turn said to be compiled by a different ḍākinī. Similarly, Diamond-Songs of Adepts contains nine chapters with beautiful titles such as Eggshell of Unknowing Cracked or Splendor of Primordial Awareness Sparkling, each with forty-one to forty-four single name-and-verse pairs. Finally Symbolic Songs of the Diamond Ḍākinīs and Core Instructions are subdivided along topical lines, both employing the technical terminology of Great Seal teachings.

The anthologies of Phadampa Sangye mention several locations where the songs are said to have initially been sung, some of them well-known in Buddhist history and mythical geography such as Vajrāsana, Gṛdhrakūṭa, others more general locations such as caves, bamboo groves, or famous cities such as Taxila. The most popular setting for the songs, however, is the cremation ground, from which hail single verses in Symbolic Lineage of the Great Sea and Precious Symbolic Instructions on the Great Seal, and in which the entirety of Songs of the Glorious Diamond Ḍākinīs, Realization Expressions of the Thirty-Five Ḍākinīs, and Diamond-Songs of the Adepts are said to have been sung. Phadampa Sangye gives us a brief characterization of the cremation grounds and the lands surrounding them in the introductory passage to Diamond-Songs of the Adepts, calling it “a region difficult for humans to travel, a village where ghouls and zombies wander, a place where the inhuman of the earth wander, the place of the action ḍākinīs, [where] Death’s hair horripilates and the demons are dread-filled upon sight of it.”⁴⁰ In literary terms, the cremation ground was also taken up as a poetic topic in its own right in twelfth-century India, as can be seen by the macabre section devoted to it in Vidyākara’s own verse anthology, the Jewel Treasury of Elegant Sayings (Subhāṣitaratnakoṣa), where ghouls, corpse-eating birds,⁴¹ and zombies rule the night.⁴²

Two of these anthologies, Songs of the Glorious Diamond Ḍākinīs and Realization Expressions of the Thirty-Five Ḍākinīs, mention the well-known gathering of tantric singers and aspirants, the gaṇacakra, at which, in these two cases, ḍākinīs are each requested to sing a verse for the group. Here the initial moment of composition occurs in a group setting in which, from what little is said in this regard, the purpose of singing songs is to inspire one’s peers and instill faith in the teachings. In the first of these anthologies, two or possibly three types of ḍākinīs are present: ḍākinīs of action (las kyi mkha’ ‘gro) prepare for the gathering by killing a young prince for the feast, and ḍākinīs of primordial awareness (ye shes kyi mkha’ ‘gro) are then invited, and from among the latter ḍākinīs reality’s noblewomen (dbyings kyi gtso mo) are invited to sing. Several other anthologies state more generally that the songs contained therein were sung, if not composed, when reality’s noblewomen convened for discussion, once again characterizing the anthology as a group exercise in spirituality.

(p. 92 ) It is not until the final anthology, the Compared Mind-Experiences of Twenty-One Male and Female Yogic Practitioners, set in a gathering of Phadampa Sangye’s Tibetan disciples, which was no doubt based upon the vision of the Indic ganacakra portrayed in the previous collections, that we find a more concrete explanation of the motives for adepts singing to each other: simply because the songs of many aspirants collected together in one place and time are a cause for great joy.⁴³ Though ḍākinīs are by far the most common singers in these collections, other figures are present as well, including yogic practitioners both male and female in the Orb trilogy, deities in the Symbolic Lineage, adepts and masters in Arcane Songs of the Mind, and kings, queens, minister, and brahmins in the Miraculous Lineage. It thus appears that few social groups (real, mythic, or both) were excluded from having their names counted among the singers of spiritual verse (see appendix 1).

According to the anthologies themselves, the compilation of the songs fell to various hands. In both Arcane Songs of the Mind and Diamond-Songs of the Adepts it was the ḍākinīs who are said to have compiled the songs of adepts, masters, and male and female yogic practitioners. The songs of kings, queens, and other social groups of the Miraculous Lineage are also said to have been compiled by ḍākinīs. Curiously, however, none of the songs composed by the ḍākinīs were then compiled by them. Phadampa Sangye is said to be the compiler of the Symbolic Lineage, the singers of which Go Lotsawa later characterizes as deities whom Phadampa Sangye had encountered in visions.⁴⁴ He is also held to be the compiler of the Orb trilogy, two of which he passed directly to his disciple Kunga.

In no fewer than seven cases, the process of compilation is associated not exclusively with oral performance, but with writing as well. The Symbolic Lineage and the Miraculous Lineage both contain interlinear notes stating that Phadampa Sangye wrote them down on white paper and subsequently brought them to Tibet. This notion raises many questions about the material life of this literature, for according to Diringer the oldest datable Indic paper manuscript is dated 1231,⁴⁵ over a hundred years after Phadampa Sangye’s death, and paper is not known to have been in use before the eleventh century in India. This fact would put Phadampa Sangye’s works at the beginning of this trend in writing technology. Certainly the more common writing material was the palm leaf (tāla), a word that Tibetans merely transliterated when referring to this type of book, and were it not for the consistent use of the word “paper” (shog) in all of the anthologies that mention writing material, palm leaf would be the assumed material for a manuscript of this time. Still more curious is the mention of a paper scroll (shog dril) at the close of Symbolic Songs of the Diamond Ḍākinīs, for this form of book was by all accounts unknown in India.⁴⁶ One possible answer to this puzzle is that Phadampa Sangye came across these book technologies in China,⁴⁷ where he is said to have traveled in between his stays at Dingri.⁴⁸ Alternatively, we might suspect that the terms “paper” and “paper scroll” are used either figuratively or inaccurately in the notices to the anthologies.

The introductory note to the Silver Orb, the first work in the Orb trilogy, (p. 93 ) tells us that the work was contained in a book on whose cover was written Collected Explanations of the Stainless Crystal Orb, thus suggesting that these three works existed as an independent written work before their inclusion in the Profound Oral Lineage. Finally, at least two of the anthologies tell us in introductory passages that the ḍākinīs themselves had a hand in the writing down; Arcane Songs of the Mind states that the ḍākinīs wrote down songs of the eighty male and female adepts, and Diamond-Songs of the Adepts (no.14) says that the ḍākinīs wrote the songs of the yogins down on white paper with “unforgettable formulas” (mi bsnyel ba’i gzungs). It is clear that in this tradition the written word was important in its own right and was considered an important aid for people in religious pursuits. That both reading and writing were valued is made especially clear in Diamond-Songs of the Adepts, in which the stated purpose for the ḍākinīs writing the songs down is the benefit of later generations, and at the close of which Phadampa Sangye prays that doubts will be eased merely by the reading of these songs.⁴⁹

Finally, seven of the anthologies (nos. 3, 6–10, and 12) are said to come from a larger entity referred to only as the Arcane Treasury (Gsang mdzod), alternatively known as the Arcane Treasury of the Ḍākinīs, a collection that never seems to be mentioned on its own terms, but rather as that out of which the smaller collections originate. It is thus never clear whether the Arcane Treasury is a material collection, a more abstract categorization of teachings, a poetic epithet of the songs of the ḍākinīs, adepts, and masters, or perhaps some combination of all three. In the Miraculous Lineage we are told that thirty-two of thirty-six songs were initially contained within something called a red notice (dmar byang)⁵⁰ within the Arcane Treasuryito which four songs were later added. Here the notion of treasury has a decidedly physical feeling. The paper scroll on which Symbolic Songs of the Ḍākinīs is said to have been written is also contained in the Arcane Treasury, as is the written work Arcane Songs of the Mind. Whatever the case, the Arcane Treasury can be loosely conceived of as the body of teachings out of which (however metaphorically this might be read) reality’s noblewomen drew the teachings they sought fit to bestow upon Phadampa Sangye.⁵¹

Zhama Lotsawa Tonpa Senge Gyalpo is mentioned as translator at the close of eight out of the seventeen anthologies (nos.1, 4–8, 10, and 12), and it appears that he conducted all of his work at Phadampa Sangye’s Tibetan residence of Dingri Langkhor. It is unfortunate that we as yet have no record of the language spoken by Phadampa Sangye or the languages in which the anthologies were written down. In a way we thus owe any sense we might make out of this sometimes nonsensical verse to the efforts of Zhama Lotsawa. Phadampa Sangye himself is the stated translator of Diamond-Songs of the Adepts, curiously also the only anthology found exclusively in the Tibetan canon and not in the Profound Oral Lineage.

Saraha first appears in the Orb series of anthologies. The Pure Silver Orb, located only in the Profound Oral Lineage, is an anthology of brief prose aphorisms from fifty-four male and female adepts. It forms a trilogy with the Pure Golden Orb and the Pure Crystal Orb, both also containing prose aphorisms (p. 94 ) and verse of fifty-four male and female adepts, although each set of fifty-four is slightly different (as the comparative table in appendix 2 shows). From the note at the beginning of the Silver Orb we can gather that this trilogy existed before its inclusion in the Profound Oral Lineage, as a separate volume entitled the Collected Explanations of the Stainless Crystal Orb. The introductions to each work relate something regarding their compilation: The Silver Orb states that these sayings were compiled by Phadampa Sangye, the Golden Orb tells us that “the great lord of yogis, glorious Mipham Gonpo [Phadampa Sangye] was in harmony with the realizations of the fifty-four male and female yogic practitioners. He spoke of the increase of [his] experience in simple words, and put it down in writing.”⁵²

The Crystal Orb takes pains to claim that the work is an exact duplicate of Phadampa Sangye’s words, free from any interpolation on the part of the translator, Zhama Lotsawa. The teachings contained are passed on “like the continuity from one butter lamp to another,” as the introduction to the collection tells us. The songs of Saraha and the other fifty-three adepts it contains “were spoken from the warm harvest of Padampa Sangye himself through the heat of appropriate means and auspicious coincidence, with no intervening interpolation in the words of the translator, just like they were copied from one piece of paper to another.”⁵³

Arcane Songs of the Mind, another anthology said to come from the Arcane Treasury, contains single four-line verses from eighty male and female adepts. It was, according to its colophon, “put down in writing by the ḍākinīs, and from the Arcane Treasury was granted to Padampa Sangye as the noblewomen of reality convened.” In contrast to symbolic verse, these pithy sayings use the more straightforward technical terminology of the Great Seal teachings and are quite similar to the dohā songs attributed to Abhayadatta’s eighty-four adepts. This, and the fact that many of the adepts’ names are shared in both collections, are no doubt the reasons that it was later included in the Teaching Cycles of the Eighty-Four Adepts. The work begins with a verse from the Great Brahmin, Saraha:

• Hey, the root of samsara and nirvana is the mind’s nature,
• Realizing this, [you] must, without meditation, settle [yourself] fully, without artifice.
• Settled in oneself, Oh what a mistake it is to search elsewhere.
• [This is] the natural state, without [talk like] “This is it, this isn’t.”

If Saraha’s verses are cited individually in the anthologies of Phadampa Sangye, he is also the only adept to whose teachings an entire anthology is dedicated. Saraha’s Core Instruction consists of seven symbolic teachings (brda’ bstan pa)⁵⁴ arranged in what appears to be an expanded variant on the fivefold scheme found in Symbolic Songs of the Diamond Ḍākinīs. The first verse from the work runs as follows:

• Homage to glorious Heruka.
• The ḍākinī’s blessing is a symbol of realization.
• (p. 95 ) Like a cure of spells pronounced,
• The peacock’s food is not [that] of others.
• The sesame is illuminated by the lamp.
• Sandal is the scent of the deer’s musk-sack.
• Waves of water are the ocean itself.
• Then nature of clarity is like white cotton.
• [Though it] cannot produce a mountain,
• The snowy mountain is nothing but water.
• The odor and the garlic are no different.
• The quality of [both] sun and moon is clarity.
• Do not search for the footprint of a bird.
• The directions of the maṇḍala are equal.
• The sky is no topic for the sophist.
• The zombie is like a jewel in the land of activity.
• Incomparable is the miraculous crystal jewel.
• A symbolic instruction by the great hunter Saraha to Kamalaśīla.⁵⁵

Given that it is the only work in the corpus dedicated to a single figure, Core Instruction is included here as an anthology with some hesitation. Yet in terms of how the compilers have arranged the work, it bears close similarity to the others. It is organized into individual poems, which are in turn divided into separate chapters that close with Saraha’s name, all lending the work a composite feel. Core Instruction closes with the intriguing statement that these teachings were given directly to Phadampa Sangye (alias Kamalaśīla), a statement that puts the modern reader in a historiographic quandary. Of course, all of the anthologies make some claim of a direct link between Phadampa Sangye and the adepts. He appears to have met them all. Yet to “meet” the masters is in a sense to have shared in their tradition.

Diamond-Songs of the Adepts: The Shining Suchness of All Yogins is the only collection attributed to Phadampa Sangye found in the Tibetan canon and not found in the Profound Oral Lineage. Containing nine chapters, each containing forty-one to forty-four verses, it is by far the longest of the anthologies. Within each beautifully titled chapter⁵⁶ each verse is associated with one male or female adept whose name is always accompanied by an epithet, such as “Huluka, Yoginī of Self-liberated Thought,” “Matila the River Yogin,” or “Earnest Junutari Steeped in Primordial Awareness.” The introductory passage of Diamond-Songs of the Adepts contains the fullest description of the mythical landscape of phantasmagoria in which these songs were situated, a land of zombies and ḍākinīs:

The wheel of teachings is the ultimate in the accumulation of merit, [teachings by which] sentient beings, the cause, are without regression in the two accumulations, the path on which buddhahood, the result, is obtained. The place where it is turned is a region difficult for humans to travel, a village where ghouls and zombies wander, a place where the inhuman of the earth wander, the place of the action ḍākinīs, [where] Death’s hair horripilates and the demons are (p. 96 ) dread-filled upon sight of it. At the eight great cremation grounds, the yogins of perfected realization and the yogins whose minds are purified in unborn reality assembled, and then the certain knowledge of the the ultimate truth that had been born in the minds of each one, whatever introductions to the co-emergence of the natural meaning, the primordial awareness existing in each, and the realizations dawning from anywhere were put into songs of experience as the offering ceremony was blessed. With unforgettable formulas the ḍakinīs at that place such as *Sukhaśrībhadrī (su kha ta skal ba bzang mo) set them down in writing on white paper for the benefit of later generations.⁵⁷

Here, in a most specific statement of purpose, we are also told that in this case it was the ḍākinīs who wrote down the teachings of the yogins, using unforgettable formulas (mi bsnyel ba’i gzungs) to preserve them for the future faithful. This picture stands in contrast to the more common scene, in which Phadampa Sangye or some other master writes down the songs of the ḍākinīs. Finally, in a verse following the colophon, we are told that these teachings were rolled up (‘gril), a turn of phrase reminiscent of the scroll (shog dril) in which Symbolic Songs of the Diamond Ḍākinīs was said to be contained.

From Saraha to Tradition

As much as anything, Saraha is heralded by his Tibetan hagiographers as a singer, a poet of enlightenment drinking in the sky. To kings, queens, and common people he sang of his spiritual experience—a bard from the other side, from the realm of the ḍākinīs. His songs represent the culmination of his spiritual career and are the ultimate means by which he expressed his enlightenment and brought his disciples along the path. And yet the power of the act of singing is not merely something over which Saraha and the adepts of old held sway. Songs were not just to be sung by masters teaching disciples, by Saraha preaching to kings and queens. They were in fact means to teach oneself, to sing to oneself about one’s true nature; the self-exhortations of Nāropa in his Diamond-Song could in turn be internalized by any member of the tradition. Advayavajra exhorts the readers of his commentary on the Treasury of dohā Verses to sing the songs he appends to that of Saraha—precisely in order to reach realization:

• Nondual diamond-songs,
• Showing the primordial awareness of suchness—
• With these nondual words of ultimate meaning,
• Sing, people, songs for yourselves.⁵⁸

Despite the importance laid upon song in the Tales of the Eighty-Four Adepts and other hagiographies, very few of the dohās and diamond-songs themselves attributed to Saraha are reflective of the song as a medium of spiritual teaching.

(p. 97 ) Nevertheless, the Treasury of dohā Verses itself does have something to say on the topic of singing:

• Fools, know all these words spoken by Saraha.
• The nature of the innate cannot be told with words,
• Yet the instructions of the master may be seen with the eye.
• Delighting in both dharma and nondharma, partake;
• In these there is not a speck of evil.
• When the innate mind has been purified,
• The enlightened qualities of the master will enter your heart.
• Realizing so, Saraha sings this song,
• Though he has not seen a single mantra, a single tantra.⁵⁹

The first two lines of this passage embody the contradiction inherent in any attempt to speak of a spiritual experience which is by definition inexpressible. The words spoken by Saraha entail their own undoing, for as is seen time and time again, the predominantly apophatic rhetoric of Buddhism cannot help but turn to kataphatic declarations describing either the ultimate enlightened state directly or, more obliquely, the power of language about the enlightened state. The reader must know the words of Saraha despite the fact that his subject is ineffable. In an ironic twist, the power of Saraha’s words is precisely their message of ineffability. This seems ultimately to debase the power of the word, and yet the final lines suggest something more; it is not the written word of the tantras that holds the power to express the inexpressible, but song itself. Much as the tales of his life tell us, the realization of the enlightened state encourages Saraha not to write another treatise, another commentary, but to inspire others through the medium of song, which stands above the ordinary language of treatises and tantras.

It is perhaps this claim that gave the commentators on the Treasury of dohā Verses license to write according to the meaning of the dohās as taught by the masters and not according to the letter. The rhetoric of creative orality that both the songs and tales proffer is strongly stated by the commentators on Saraha’s Treasury of Dohā Verses. As Advaya Avadhūti makes very clear, since the enlightened state—described in dohā commentarial literature as the essential meaning (snying po’i don, sārārtha)—is by nature ineffable, then the words of the dohās are merely indicators of the truth, not the truth itself. He writes:

• The essential meaning is not an object of thought, so
• Even though someone like me speaks in words, they are far from the meaning.
• Words have no connection to the meaning. Nonetheless
• Through this commentary on meaning, which relies on the lamp
• of spiritual memory,
• May [you] not rely on the texts of non-Buddhists and others.⁶⁰

And again:

• (p. 98 ) Others explain by commenting in accordance with the root [verses] of the text.
• My tradition writes the root in accordance with the explanation.
• Though we may not remember the scriptural words of the baskets,
• Mantras and tantras are not completed by writing the words of the scriptures.
• Relying only upon the mental-spiritual inspiration of Glorious Śabarapāda,
• I shall write this mnemonic of a drop of the ambrosia of [Śahara’s] speech,
• For the benefit of myself and the faithful like me,
• Summarizing only the instructions on the meaning.⁶¹

This passage provides a strong defense for the primacy of the master’s oral instructions over the written word, in conformity with the general tantric rhetoric of orality. And yet here Advaya Avadhūti stretches the limits of his derogatory remarks on written teachings by claiming that since the essential meaning lies not in the words themselves, then even Saraha’s dohās themselves, the source-text of his commentarial efforts, can be rewritten in conformity with the oral teachings of his immediate master, Śahara. The implications of this idea will be detailed in the following Chapter 7, but for now it is sufficient to emphasize that this gives Advaya Avadhūti himself, and other commentators after him, the license to change, rearrange, and transform Saraha’s words himself. In short, Advaya Avadhūti gives himself permission to “author” the words of Saraha by claiming that the real message of Saraha is not any text of the Treasury of Dohā Verses but rather in the meaning that lives in the hearts of the masters who have realized the message of the dohā.

This chapter has surveyed a series of related bodies of literature. The variety displayed above is in fact indicative of the variety to be met with in the three volumes of the Tanjur that make up the Commentaries on the General Intentions of the Highest Yoga Tantras, for almost all of the poetic songs translated here come from this section. I have not endeavored here to look at the songs as philosophy, but as religious literature with a historical development and identifiable generic features. The anthologies funneled through Phadampa Sangye should in the future be an integral part of any discussion of the dohās and diamond-songs: In a single anthology are contained more diamond-songs than are separately listed in the Tanjur, and there are seventeen such anthologies!

The central point to understand here is that these songs were almost exclusively held by the creators of tradition to have been orally composed, and only later written down. Saraha, the adepts, the ḍākinīs, kings, and queens—a whole social universe of enlightened beings—all sang diamond-songs as an expression of realization. But this is not to belittle the power of writing: The ḍākinīs, Phadampa Sangye, Karma Trinlaypa, and others all allow the importance (p. 99 ) of writing for the perpetuation of the dohās, for the benefit of later spiritual seekers. Even Advaya Avadhūti makes his critique of the written word not in another song, but in a written commentary. Orality is praised as the medium of the adepts, but writing and rewriting are the acknowledged lifeblood of tradition. It is to issues of writing and redaction that we now turn.