The lock has an address

The doctrine this chapter is about has an address, and the address will surprise most of the people who repeat it.

They attribute it to the Tantras. At the turn of the Kali age, the story goes, Śiva drove a pin into every mantra on earth and left them all inert, and the Śābara mantras alone he passed over. It sounds like scripture. It gets quoted like scripture.

Its earliest located statement is a couplet in Hindī, written by a commentator on a Hindī poem, and the commentator has a name.

He is Paṇḍit Śivalāl Pāṭhak. His book is the Mānas-mayaṅka, a ṭīkā on the Rāmcaritmānas, and it stands in the standard compendium of Mānas commentary under a siglum rather than a name, as the Mayaṅkakāra, the man who made the Mayaṅka. That compendium is Mānas-pīyūṣa, edited by Añjanīnandan Śaraṇ and printed by the Gita Press at Gorakhpur in 1968. It argues no line of its own. It gathers the named commentaries verse by verse and sets them against each other, so that a reader can watch the tradition disagree with itself in one place.

On the verse this whole treatise is built on, Pāṭhak’s couplet is there, and the gloss under it is the doctrine entire. The Vedic and Tāntric mantras had been made kīlita, pinned. So Śiva composed the śābara mantras for the benefit of creatures. Because the others had been pinned, these alone went on bearing fruit.

When Pāṭhak wrote it, this treatise does not know. The Mānas-mayaṅka has not been opened here and its date has not been fixed, and fixing it is the single most valuable outstanding item in this chapter.

The address, though, is fixed. What the doctrine has not got, so far as this research can see, is a Sanskrit address at all.

The doctrine is easy to look for in the wrong place

It is tempting to conclude that the doctrine has no text behind it at all, and there are two ways to reach that conclusion, both of them wrong, and both of them instructive about how a negative gets manufactured.

The first runs through the popular Hindī compilation the counting chapters run on. Put the locking vocabulary into that book as a search string and four occurrences come back. Four is a small number, and a small number invites the conclusion that the discourse is marginal even among the people supposed to hold it. Open the four and that conclusion collapses. The four occurrences are the doctrine. They state it in full sentences. A word counted with the page never opened is evidence about a search string.

The second runs through the language. The Śābara claim is attached, in the popular tradition, to one verse of the Rāmcaritmānas, and there is a standard Hindī compendium that collates the named commentaries on that poem verse by verse. Anyone searching only the Sanskrit will never reach it. Opened, it carries the doctrine on the page, attributed to a commentator with a name.

So the question this chapter can answer is not whether the doctrine has a text. It is which literature the text belongs to, and that turns out to be a literature almost nobody who repeats the doctrine would name.

The claim, and why it deserves respect first

State it the way its believers state it, without softening and without sneering.

At the turn of the Kali age, the argument runs, Śiva foresaw what would happen to sacred speech in a degraded time. Powerful formulas would fall into unworthy hands. Men without discipline, without a teacher, without the moral formation that ritual competence assumes, would pick up instruments capable of real effect and use them for gain and for harm. So he locked them. He drove a pin into every mantra, a kīlaka, and left them inert. From that moment no mantra would function for anyone who had not first performed utkīlana, the unlocking, under a qualified teacher. That is why, the argument continues, so much practice fails. The instrument is sound. It is bolted shut.

And then the second half. The Śābara mantras were exempt. They were made after the lock, or outside it, or in a speech the lock did not reach. They need no utkīlana, no dīkṣā, no puraścaraṇa. That last is the classical name for the preparatory regimen, and it appears here only as one more category the Śābara corpus is said to dispense with. What it consists of is not this treatise’s business. They work as they are, for anyone, at once.

This is a good argument, and it solves a real problem. The previous chapter set out the apparatus classical Tantric mantra-science assumes: sixty structural defects a formula can carry, of which kīlita, rendered “tied” by the scholar who tabulated the list, is one; ten saṃskāras that repair a defective formula; a sealed transmission from teacher to student. That apparatus is elaborate, Sanskrit, and textual, and the Śābara corpus has almost none of it. The verse at the head of this treatise concedes that the syllables are ill-matched, that they carry no settled sense, and that no recitation is prescribed for them.

A reader who knows both bodies of material is left holding a genuine question. If mantra requires all of that, why is this material supposed to work without any of it?

The lock answers in one move. The apparatus is a response to a historical event; the event has an exception; the Śābara mantras are the exception. Economical, dignified, and it makes a puzzling corpus intelligible in a sentence.

That is why it spread. The question this chapter can now actually answer is where it spread from.

The doctrine has a commentary, and the commentary has a name

Mānas-pīyūṣa is the standard Hindī compendium of Rāmcaritmānas commentary, edited by Añjanīnandan Śaraṇ and published by the Gita Press at Gorakhpur in 1968. It does not argue a line of its own. It collects the named ṭīkā tradition, verse by verse, and prints the commentators against each other. Its treatment of the verse this treatise is built on carries the answer in the first note.

There it reports the Mayaṅkakāra, the author of the Mānas-mayaṅka commentary, and gives his couplet, which runs सर्पादिक विषहरण कलि, साँबर रचे तुरंत। सो उमेश कलि अघ दहन मानस यश बिरचन्त॥

The commentator’s gloss on it, paraphrased, is the doctrine entire: the Vedic and Tāntric mantras had been made kīlita, so Śiva composed the śābara mantras for the benefit of creatures; because the others had been pinned, these alone went on bearing fruit.

There it is. Not a content farm. A Gita Press compendium, reporting a named commentary, on the exact verse the whole claim hangs from.

Per the Mānas-pīyūṣa siglum table, the Mānas-mayaṅka is by Paṇḍit Śivalāl Pāṭhak, with a ṭīkā by Bābū Indradevanārāyaṇ Siṃh. The Mānas-mayaṅka itself has not been opened for this treatise, and its date has not been fixed. What is established is that a named Hindī commentator on Tulsīdās held this doctrine, and that the standard 1968 compendium reports him holding it. When the commentary was written is the next thing to find out, and it is the single most valuable outstanding item in this chapter.

So the honest statement is this. The exemption doctrine has a commentarial pedigree inside the Hindī Mānas tradition, attached to this verse, reported in a scholarly compendium. That is where the popular tradition got it.

Four lines that were counted and not read

The second attestation sits in a book that is easy to dismiss on sight, and it should not be dismissed.

The compilation is nameable, and this chapter will name it. Durlabh Śābar Mantra, presented by Tāntrik Bahal, published by Rājā Pocket Books of Burāṛī, Delhi, new edition 2016, read here as a scanned full text of 323,692 characters. A Delhi paperback. That is the object every count in chapters 10 to 13 runs on.

In it, this, printed here in running prose rather than as a framed quotation because it is modern Hindī and not scripture: इसका एक कारण यह भी है कि अन्य सभी मंत्रों को ‘कीलित’ किया … है, अत: वे ‘उत्कीलन’ किए जाने पर ही अपना प्रभाव दिखाते हैं, परंतु [शाबर] मंत्रों को कीलित नहीं किया गया है

All other mantras have been made kīlita, and therefore show their effect only once utkīlana has been performed on them; these have not been made kīlita. The bracketed word is lost in the scan and is supplied from the sentence around it, which is about the śābara mantras throughout. A second passage says of the same corpus that no utkīlana of any kind is required.

That is the doctrine, in the book, twice. The report stops there, and deliberately. What these passages assert is that no operation is needed; there is no operation in them to describe, and if there were, it would not be printed here.

Then a third statement, in a different frame and with a different reason, and it turns on a word no search for the locking vocabulary would return. Dattātreya, addressing Matsyendranātha on Mount Girnār, says he will give his own workings in the Hindī language, so that even a barely lettered person can use them easily. Since the Kali age came, he says, the brahmins have become Veda-less, angry, greedy and lawless, and the power of the pure mantras has departed; so let the teacher speak the mantras and tantras without a kīlaka, and let the people have no difficulty. He then says he will declare those mantras which in the Kali age can be perfected without a kīlaka, and adds, in the same breath, that this must be kept from the unrighteous.

Read that last clause twice. The passage that abolishes the gate reinstates it in the next sentence.

The normalisation, stated

Two forms out of a family prove nothing on their own. Here is the whole family, in the same file, with the rule that produced it.

Kīlita stands at two. Utkīlana at two. Kīlana returns four, of which two are inside utkīlana and are not separate occurrences, and two stand free in mantra lines that will not be reproduced here. Kīlaka stands at two, and both are the passage above. Kīla returns twenty-five, almost all of them other words. And there is a fifth occurrence of the locking vocabulary that no string search could ever have found, because the scan renders it उत्तीलन, with the k read as a t. It sits in the sentence that most plainly states the exemption.

One further count runs the other way, and it inflates rather than deflates. A substring search for ताला, “lock,” returns fifteen in this file. Three of those fifteen are the noun, all three inside mantra lines. The rest are तालाब, a pond, वार्तालाप, a conversation, तालियाँ, clapping, and scanner noise. A substring count in Devanāgarī without a word boundary will do that every time, in either direction.

Which is the whole lesson twice over. A count is only as good as its normalisation, the normalisation has to be printed, and a frequency figure with nobody reading the sentences underneath it is not evidence about a doctrine. It is evidence about a search string.

One rule for this book

Four chapters of counting run on a book of this kind, so the rule governing its use has to be stated, and it is stated here once for chapters 9 to 13 together.

This compilation is admissible, named, dated and imprinted, as evidence about what the popular Śābara literature contains. It is admissible in the same way when it states a doctrine as when it yields a number. It is not evidence about any critically edited Śābara Tantra, and it never will be. What remains uncitable is the anonymous web tier, and the objection there has nothing to do with the material being popular. It is that those items carry no author, no date and no imprint, so nothing about them can be checked at all.

What the Sanskrit says instead

Now go back to the Sanskrit, because the doctrine’s absence from it is the finding that survives.

There is a half-line in the Nāradapurāṇa that most people who repeat the claim have never seen.

शिवेन कीलिता चेयं तदुत्कीलनमुच्यते ।

śivena kīlitā ceyaṃ tadutkīlanam ucyate And this [vidyā] has been locked by Śiva; that [rite] is called its utkīlana.

Nārada Purāṇa, Pūrvabhāga 1.87.39 · also numbered 1.3.87.39 where the tṛtīya-pāda is spelled out; first hemistich only #

Phrase by phrase. Śivena, by Śiva, instrumental, the agent. Kīlitā, pinned, a feminine singular participle. Ca iyam, and this one, feminine singular again. Tat, that. Utkīlanam ucyate, is called its utkīlana, its unpinning. “And this one has been pinned by Śiva; that is called its utkīlana.”

Look at the word carrying the load. Iyam. This one. A demonstrative, pointing at a particular vidyā, the technical term for a formula belonging to a female deity, which the passage has been discussing for some time.

The next verse names her. At 1.87.40 the vidyā is Chinnamastā’s, and the section closes there: 1.87.41 opens by announcing another manifestation of the Goddess, which fixes the boundary and fixes what iyam was pointing at. The verse also attaches a Kali-age efficacy claim to that one formula, reported here in transliteration as evidence and not as recommendation: uditā chinnamasteyaṃ kalau śīghram abhīṣṭadā.

Hold that beside the popular claim for a second, because the valence has flipped. In the Purāṇa the Kali age is why this formula is worth having now. In the popular claim the Kali age is why nothing works now. Same age, opposite direction, one verse apart. Whether the modern claim descends from these verses by misreading, or arrived some other way and attached itself to them afterwards, we do not know. They sit close enough to make the first possibility hard to ignore, and that is inference, flagged as inference.

Earlier in the same text, at 1.86.41 to 47, it handles śāpoddhāra, the lifting of a curse laid on a formula. That stretch has been read verse by verse, because a single generalising verse in it would change this chapter. It does not generalise. It gets narrower. The curse averted is the goddess’s own. The participle utkīlitāṃ is feminine singular. The vidyā freed of the curse is a particular vidyā. At 1.86.46 the curse removed is named as Bālā’s. Every operation in the passage is attached to a named or demonstratively marked formula.

What kīlita actually is

And then the datum that settles the register the word belongs to.

The Nāradapurāṇa’s defect list, the forty-nine that chapter 8 counted verse by verse, closes at 1.64.20 by announcing that the marks of those defects will now be given. They follow, through 1.64.57. One of them is this:

Nāradapurāṇa, Pūrvabhāga, 1.64.28a: māprā namāmi ca padaṃ nāsti yasminsa kīlitaḥ

That mantra in which the specified element is absent is kīlita. Which element, this treatise will not unpack, and the surrounding verses that specify the others are left where they are. The structure of the sentence is the finding, and the structure is a general classificatory rule.

Read what that does. The Purāṇa defines kīlita as a property of how a formula’s syllables are arranged. Something is missing from the string, and a formula with that gap is called pinned. It is a description of a defect, in the same register as blind, deaf, cut and burnt, which are the neighbouring items on the same list.

It is a diagnosis. It is not the residue of a cosmic act.

Nobody locked anything. On the tradition’s own most technical statement, a mantra is kīlita the way a machined part is out of tolerance.

It also rules out an absolute that the earlier passages invite, that the demonstrative is always there. At 1.64.28 there is no demonstrative and no particular goddess, because the referent is a class. The accurate statement is stronger and survives contact with the corpus: in every passage reached, kīlita is either attached to a named or demonstratively marked vidyā, or defined as a structural property of a formula’s syllables. In neither case is it a historical act performed on mantra as such.

The other Sanskrit occurrence outside the Nāradapurāṇa points the same way and breaks the same absolute. The Agnipurāṇa, at 125.6cd, has stambhanaṃ kīlanādyañca karotyeva hi sādhakaḥ: it is the practitioner who performs the immobilising and the pinning and the rest. No demonstrative there either, and the agency runs from the operator onto a target. Whatever kīlana is in that literature, it is something people do, and Śiva is not doing it.

Both those Sanskrit readings come out of electronic texts, and the two files do not stand in the same condition. GRETIL’s Agnipurāṇa file names its printed source in its own header: the edition of Rājendralāl Mitra, Asiatic Society of Bengal, Calcutta, 1870 to 1879, three volumes, Bibliotheca Indica 65,1 to 3, with the data entry credited to Jun Takashima. The volume covering adhyāya 115 to 268, where 125.6cd stands, is an open download on the Internet Archive. So the edition behind 125.6cd has a name, and the page can be found by anyone who wants to check the line. Nothing here has been collated against that printed page.

The Nāradapurāṇa is in a weaker position, and the difference is worth stating separately. GRETIL’s Nāradapurāṇa file names no edition at all. Its header records only that the text was input by members of the Sansknet project, and it warns that the text still wants proof-reading. Both Nāradapurāṇa passages above were read in that file, in the Pūrvabhāga numbering GRETIL uses, and Purāṇic verse numbers move between editions, so the numbering here should be expected to differ from Bühnemann’s bare “64.14–19” by the Pūrvabhāga prefix alone. An edition still has to be found for those two passages, and until it is, the verse numbers in this chapter point at an electronic text and at nothing printed.

Two silences, one of which was a method artifact

Two silences in the modern scholarship look, at first, like they carry the case. One of them does. The other is an artifact of the instrument used to find it.

Gudrun Bühnemann’s study of how mantras are selected and perfected, published in 1991, is the standard scholarly treatment of exactly this ground. It walks the defect lists. It walks the ten refinements. It is a technical map of everything a practitioner is supposed to do to a formula before use.

It never mentions Śiva locking anything, never mentions an unlocking rite, and never mentions the Kali Yuga. That check has been run across every page of the article and every footnote in it, with one page read as an image because its text layer came back empty. Those three negatives hold, notes included.

A fourth negative is easy to add beside them and it is false. It is not true that the only kīla- stem anywhere in the article is kīlita, sitting at number sixteen in a table of defects. Bühnemann lists kīlaka, the wedge, among the items the Tantric texts specify for a mantra, and a footnote on the same page takes the term up again. Two further occurrences of the stem, one in the body and one in a note.

A string search will not find either, because the scan renders the word as ki/aka and kifaka. A search on the stem cannot return a string the stem is not in. One search string, one silence, and the silence read as a fact about the literature: that is the mechanism, and it is worth naming every time it appears, because it is how most manufactured negatives get made.

The finding survives, and kīlaka strengthens it rather than costing it anything. A wedge named as a component of a formula is one more piece of evidence that kīla- in this literature is architecture. The accurate statement is narrower and it is this: the only kīla- stem in the article attached to a defect is kīlita, and nothing anywhere in the article attaches any kīla- stem to Śiva, to an unlocking rite, or to the Kali age. The absolute goes. The negative it was propping up did not need it.

The second silence does not survive at all.

A string count of Goudriaan and Gupta’s Hindu Tantric and Śākta Literature, the survey volume in Gonda’s History of Indian Literature, returns zero instances of kīlana, utkīlana or śāpoddhāra. The count is accurate. It is also close to worthless, because Goudriaan and Gupta write in English and translate the operations they describe. Just before their Śābara Tantra entry they quote, in translation, a hostile rite from the Uḍḍīśatantra in which a copper pin is driven into an effigy. That is a kīlana, quoted and discussed, in a volume reported as containing none. The rite is not reproduced here.

A survey written in English will not contain the Sanskrit word for an operation it renders as a pin. The three occurrences of “lock” in the entire volume are locks of hair.

So the negative has to be stated as what it is, and the accurate version is narrower and defensible. Goudriaan and Gupta nowhere describe a general pinning of mantras in the Kali age. They nowhere connect the Śābara texts to such a doctrine. Their Śābara Tantra entry describes those texts by subject matter and by language and says nothing about locking at all. The zero count itself carries none of that weight.

The motive nobody wrote, and the motive somebody did

Now the part that is a real absence.

Nothing in the Sanskrit gives a reason. Nāradapurāṇa 1.87.39 says the vidyā has been pinned by Śiva and then moves straight to procedure. There is no clause of purpose in it. The curse-lifting passage gives none. The defect list gives none, because a defect list does not need one. The prevention of misuse by the unworthy in a degenerate age, which is the motive every popular telling supplies, has not been found in any Sanskrit source reached by this research.

What the vernacular sources supply instead is not that motive either, and the difference is worth sitting with.

The Mānas-mayaṅka gives no reason for the pinning. It reports the pinning as a state of affairs and makes Śiva’s composition of the śābara mantras the remedy for it.

The 2016 compilation gives a reason, and it is not moral protection of the instrument. It is institutional collapse. The brahmins went bad, so the pure mantras stopped working, so a teacher should give the formulas without a kīlaka in a language ordinary people read. The blame runs at the custodians. Whoever composed that passage was not defending the gate. He was writing an indictment of the people holding it.

And there is a much older form of the same move, which Goudriaan and Gupta report from the Mahānirvāṇatantra at 1.18 and following: a second revelation was needed because in the deteriorating course of time people had become unable to understand what had been given to their forebears. Degeneration, a fresh dispensation, a lowered threshold. That is the genre the lock belongs to, and the genre is old, respectable and pan-Tantric.

The lock is one member of it. The oldest member it is not.

Older warrants, and none of them pins anything

The corpus has other charters, and they were in print in India long before any of this was searched for in English.

Hazārīprasād Dvivedī, in Nāthasampradāya in 1950, reports from the Gorakṣasiddhāntasaṃgraha that the first of the twelve Kāpālika teachers named in the Śābaratantra is Ādinātha, and that several of the twelve disciples are principal teachers of the Nāth path. Shashibhusan Dasgupta gives the same lists in English four years earlier, in Obscure Religious Cults in 1946, footnoted to the Śābara-tantra as quoted in the Gorakṣasiddhāntasaṃgraha. Chapter 3 works that charter properly. What matters here is negative and simple. It is a full origin story for the corpus, with named figures and a named source, and there is no pin in it anywhere.

Dasgupta has something sharper still in the same book, from the Vimalaprabhā, the commentary on the Laghukālacakratantra, which he read in a manuscript in the Royal Asiatic Society of Bengal. The doctrines of the Pāramitā way, he reports, are set down wholly in Sanskrit; those of the Mantra way are given in Sanskrit, in Prakrit, in Apabhraṃśa, and in non-Sanskritic speech such as that of the Savaras and others.

Take it at his report and hold it against the lock. Here is a Buddhist Tantric commentary of the first rank saying that mantra-teaching is deliberately delivered in non-Sanskritic speech, and naming the Śabaras as the example. That is a warrant for vernacular mantra which needs no cosmic event, no exemption and no pin. It says the vernacular was always part of the plan.

Somebody who had that answer available did not need the lock. Which tells you something about when the lock was reached for.

What is left standing

Three things.

The doctrine exists and it has an address. It is stated in a named Hindī Mānas commentary reported by Mānas-pīyūṣa, and in a named 2016 Delhi paperback at least three times, in two different frames, with two different reasons. It arrived with an author and a text.

The doctrine is not in the Sanskrit. Across the Nāradapurāṇa, the Agnipurāṇa, the Tantrāloka, the Kubjikāmatatantra, the Mālinīvijayottara, the Mātṛkābheda, the Sātvata, the Svacchanda, the Toḍala and the Uḍḍāmareśvara as electronically searched, no kīla- stem and no word for the Kali age stand in the same verse. The window is one verse, and it is stated here because a chapter whose whole lesson is that a count is only as good as its printed normalisation has no business publishing an unscoped absolute. Widen the window and the negative changes, and this chapter has already printed the case that changes it: kīlitā at Nāradapurāṇa 1.87.39 and the Kali age at 1.87.40, one verse apart, pulling in opposite directions. Two other apparent hits in the sweep are the particle kila and a compound of pralaya, and neither is the stem. And the standard modern study of mantra-perfection does not raise the subject. In that literature kīlita is a defect in an arrangement of syllables, and kīlana is something a practitioner does to somebody else. That is absence across the sources reached in this research, at the window stated, and no more. Nobody has read all Sanskrit literature.

And the shape of the two, put together, is the actual finding. A doctrine that everyone who repeats it attributes to the Tantras lives, so far as this research can see it, in the vernacular commentarial and popular literature instead. It is a Hindī doctrine about Sanskrit, told in Sanskrit’s vocabulary, and it does the work of explaining, to people who know the classical apparatus exists, why the material in their hands does not have any.

One count this chapter does not publish. There is a coercive register in the compilation, conditional curse lines among it, and no figure for it is given here, because no criterion for identifying such a line has been stated that a reader could apply. A number without a printed rule behind it is the thing this chapter has just spent two sections warning against. The coercive register is chapter 10’s to describe, on whatever count it can publish a rule for.

Why somebody reached for it

Return to the question the lock was built to answer, because the question was real.

The Śābara corpus arrives with no saṃskāra apparatus, no initiatory chain, no purification sequence and no prescribed count. Set beside the machine of the previous chapter, it looks like a tradition with the rules missing. Anyone who knows both is owed an explanation.

The exemption supplies one, and it supplies it in the most respectable available currency: a decree of Śiva. It does not say the Śābara mantras escaped the system. It says Śiva exempted them. The absence of apparatus stops being a deficiency and becomes a licence.

That is a rationalisation in the precise sense: a reason produced afterwards for something already being done, phrased in the vocabulary of the authority it is working around. Traditions manufacture warrants constantly, usually in good faith, usually because a practice they inherited needs a defence nobody ever gave them. Tulsīdās, in the verse this treatise opened with, did something structurally similar for his own language, in the open, in his own voice.

The difference is that Tulsīdās signed his.

This one was signed too, as it turns out, by a commentator on Tulsīdās whose book almost nobody has read and whose date this treatise cannot yet fix.

The next chapter leaves the Sanskrit apparatus behind and goes into the corpus itself, to the word that turns out to be its signature, and to what that word does: duhāī, an appeal made in someone’s name. Chapter 10 has the counts, and it has the normalisation rules this chapter demands of any count.

There was a lock, and there was a key, and both of them are in Hindī.

We were looking in the wrong language the whole time.