Svarodaya

Width

Ādi Yāmala

The root of the set — a title almost nothing can be recovered about, and the tantric doctrine of one revelation dividing under the ages that makes its position in the list intelligible.

Ādi is first, primordial, original. Where the other six are named for a deity and open a stream, this one is named for priority itself — and that difference is the most interesting thing about it, because almost nothing else about it can be recovered at all.

What cannot be said

It is best to be blunt about the evidence before building anything on it. Narapati names the seven Yāmalas without enumerating them; the Subodhinī supplies the seven names and, at this point in the list, supplies nothing else — no deity, no subject, no chapter, no quotation. No work under the title circulates in a form that can be checked against the name. The rosters of the Yāmala group given in the Śaiva sources — those preserved chiefly through Jayaratha's long extract from the Śrīkaṇṭhīyasaṃhitā in his commentary on the first chapter of the Tantrāloka — vary considerably among themselves and do not settle the matter either.

So this page does not describe a book. It describes the doctrinal position the name occupies, the very well attested tantric habit of thought that makes such a position ordinary, and the bibliographical problem the title exemplifies. Each of those three things is knowable; the text is not.

One revelation, dividing

The natural reading of the name's place in the list is that it stands for the undivided revelation before the streams part — the source of which Rudra, Brahma, Devī, Viṣṇu, Kūrma and Skanda are the sectarian articulations. On that reading the set has a shape rather than a mere length: one root and six branches, and to cite all seven, as Narapati does, is to claim the whole rather than a school. It would explain, too, why the commentator lists it without a deity. Every other member can be glossed by naming who speaks it. This one cannot, because it is prior to the point at which the speaker acquires a sectarian name — and a gloss that supplied one would defeat the title.

The idea is not unusual in this literature; it is close to standard. The clearest case is the Śaiva canon itself, whose five streams are assigned to the five faces of one Sadāśiva — Īśāna, Tatpuruṣa, Aghora, Vāmadeva and Sadyojāta — so that the difference between the mild Siddhānta and the fierce Bhairava tantras is a difference of face and not of source. The scheme is a claim about unity as much as about classification: a single utterer articulating himself into currents of quite different character.

The most detailed early statement of the same logic is in the Niśvāsa, the oldest surviving Śaiva tantra of any kind. Its introductory book, the Niśvāsamukha, sets out a graded account of revelation itself — the ordinary or worldly teaching, the Vedic, the ādhyātmika teaching of the philosophical schools, the Atimārga of the Pāśupata ascetic orders, and finally the Mantramārga, the tantric path of mantras — arranged not as rival systems but as one revelation delivered at ascending levels. The whole Niśvāsatattvasaṃhitā survives essentially in a single Nepalese palm-leaf manuscript generally placed in the ninth century, and the first volume of its critical edition, The Niśvāsatattvasaṃhitā: The Earliest Surviving Śaiva Tantra, was edited by Dominic Goodall in collaboration with Alexis Sanderson and Harunaga Isaacson and published jointly by the Institut Français de Pondichéry and the École française d'Extrême-Orient in 2015; the Niśvāsamukha has since been edited separately by Nirajan Kafle.

The same logic runs through the tradition's account of how scripture descends. A teaching held entire at the summit passes downward through a series of recipients, each of whom receives less than was given, until it reaches human teachers in the abridged form in which it can be used. Loss is built into the model. What survives is by definition a fragment, and the fragment's authority derives from the whole it came out of rather than from itself. This is also the logic of the yuga argument that opens so many tantric works, including the svara literature: an age of shorter lives and weaker capacities requires methods suited to it, and the tantric revelation presents itself as the portion of the original teaching that remains effective under those conditions.

The lost original as a literary form

The pattern is not merely theoretical; the texts are built on it, and they say so in their opening verses.

Round and impossible verse-totals belong to the same habit. Scriptures routinely claim originals of tens or hundreds of thousands of verses, or of a koṭi, ten million, of which what is transmitted is a remnant; the Śrīkaṇṭhīyasaṃhitā's account of the canon works with figures of this kind throughout. The numbers are not estimates of length. They are statements that the extant text is a fragment, which is a doctrinal position and not a bibliographical one.

Much of this material is now editable because of a long and unglamorous recovery. The Nepalese-German Manuscript Preservation Project, which from 1970 filmed the manuscript collections of Nepal on an enormous scale, is the single reason a great many tantric texts previously known only as titles in other people's lists can now be read; the critical editions of the Trika scriptures rest on it — Somadeva Vasudeva's edition and study of the yoga chapters of the Mālinīvijayottara (Pondicherry, 2004) and Judit Törzsök's edition of chapters of the Siddhayogeśvarīmata among them. Some titles were recovered that way. Others were not, and there is no reason to expect that they all will be.

Titles that exist only in lists

That last point is the general problem, and it has a name in the field: the tantric canons contain a substantial population of works attested nowhere except in enumerations of works. The list of sixty-four Bhairava tantras contains titles never quoted by any author and never found in any collection. The Pāñcarātra reckons its Saṃhitās at a conventional hundred and eight and knows over two hundred names, most of them occurring only inside other Saṃhitās' lists. Enumerations of the Yāmalas differ from one another in three or four of their eight members.

The standard study of how these lists behave is Mark Dyczkowski's The Canon of the Śaivāgama and the Kubjikā Tantras of the Western Kaula Tradition (1988), which collates the scriptural rosters against each other and shows how routinely a title names a stream of transmission rather than an object anyone possessed. Teun Goudriaan and Sanjukta Gupta's survey of this literature (1981) reaches the same conclusion from the other end, working title by title through what can and cannot be traced; Sanderson's classification of the Śaiva canon supplies the framework against which both operate. Their collective finding is not that the medieval lists are fraudulent. It is that a title in this literature does a different job from a title in a modern bibliography — it asserts a lineage, and the assertion can be perfectly sincere and still correspond to no volume.

The recovery cuts both ways, which is why the caution is not a counsel of despair. The Kaulajñānanirṇaya, ascribed to Matsyendranātha and among the most important witnesses to the Kaula turn, was a name in citations until P. C. Bagchi edited it from an early Nepalese manuscript in the 1930s. Titles do come back. The point is only that one cannot assume in advance which will.

Six and one

There is a formal point as well as a doctrinal one. Lists in this literature are rarely flat. A set of seven very often turns out to be six plus one, with the odd member standing in a different relation to the whole than the others do — not another item but their ground. The same shape appears in the five faces of Sadāśiva, where the face turned upward is not a direction among the four but their source; and in the śāntyatīta level of the Śaiva cosmic hierarchy, which is not a further world above the worlds but what lies beyond them all.

If the Ādi Yāmala is read that way, the list is not a bibliography of seven equivalent books at all. It is six sources and an account of why they may be used together — which would make it, for a compiler who intended to draw on Śaiva, Śākta, Vaiṣṇava and Kaumāra material in a single treatise, a rather convenient thing to be able to cite.

The other reading

There is a plainer possibility, and it should be stated. Ādi is a common enough element in Sanskrit titles that an Ādiyāmala may simply have been a particular work, no more cosmically freighted than an Ādipurāṇa or the Ādiparvan of the epic — first in a sequence, or first composed, or attributed to a first teacher. On that reading the seventh member of the list is an ordinary text that happens not to have survived, and the interpretation above is a construction placed on an accident of naming.

The two readings are not as far apart as they look. A work called Ādipurāṇa is first in its sequence because it is held to be closest to the source, and the ordinal sense and the cosmological sense of ādi shade into each other throughout Sanskrit. A tradition that names a book "the First" is in either case making a claim about proximity to an origin; the question is only whether the claim was attached to an actual manuscript or to a position in a scheme.

Nothing decides between the two, and no argument from the surrounding literature is likely to. Both readings are set out here because that is the state of the evidence, and because a page that pretended otherwise would be doing exactly what the scholarship on these canons exists to prevent.